Anda di halaman 1dari 709

Handbooks of Sociology and Social Research

Michael J. Shanahan
Jeylan T. Mortimer
Monica Kirkpatrick Johnson Editors

Handbook of
the Life Course
Volume II
Handbooks of Sociology and Social
Research

Series Editor
John DeLamater, University of Wisconsin-Madison MADISON,
Wisconsin, USA
Each of these Handbooks survey the field in a critical manner, evaluating
theoretical models in light of the best available empirical evidence.
Distinctively sociological approaches are highlighted by means of explicit
comparison to perspectives characterizing related disciplines such as
psychology, psychiatry, and anthropology. These seminal works seek to record
where the field has been, to identify its current location, and to plot its course
for the future. If you are interested in submitting a proposal for this series,
please contact the series editor, John DeLamater: delamate@ssc.wisc.edu.

More information about this series at http://www.springer.com/series/6055


Michael J. Shanahan
Jeylan T. Mortimer
Monica Kirkpatrick Johnson
Editors

Handbook of the Life


Course
Volume II
Editors
Michael J. Shanahan Jeylan T. Mortimer
Department of Sociology Life Course Center and Department
University of North Carolina of Sociology
Chapel Hill, NC, USA University of Minnesota
Minneapolis, MN, USA
Monica Kirkpatrick Johnson
Department of Sociology
Washington State University
Pullman, WA, USA

ISSN 1389-6903
Handbooks of Sociology and Social Research
ISBN 978-3-319-20879-4 ISBN 978-3-319-20880-0 (eBook)
DOI 10.1007/978-3-319-20880-0

Library of Congress Control Number: 2002042769

Springer Cham Heidelberg New York Dordrecht London


Springer International Publishing Switzerland 2016
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or
part of the material is concerned, specically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microlms or in any other physical way,
and transmission or information storage and retrieval, electronic adaptation, computer software,
or by similar or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specic statement, that such names are
exempt from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in
this book are believed to be true and accurate at the date of publication. Neither the publisher nor
the authors or the editors give a warranty, express or implied, with respect to the material
contained herein or for any errors or omissions that may have been made.

Printed on acid-free paper

Springer International Publishing AG Switzerland is part of Springer Science+Business Media


(www.springer.com)
Contents

Introduction: Life Course Studies Trends, Challenges,


and Future Directions ........................................................................... 1
Michael J. Shanahan, Jeylan T. Mortimer,
and Monica Kirkpatrick Johnson

Part I Foundations of Life Course Studies and Future Research


Institutionalization of Life Course Studies ......................................... 27
John Bynner
Age, Cohorts, and the Life Course ...................................................... 59
Glen H. Elder, Jr. and Linda K. George
Opening the Social: Sociological Imagination
in Life Course Studies........................................................................... 87
Dale Dannefer, Jessica Kelley-Moore, and Wenxuan Huang
The Changing Social Construction of Age
and the Life Course: Precarious Identity
and Enactment of Early and Encore Stages of Adulthood ........ 111
Jeylan T. Mortimer and Phyllis Moen
Structuration of the Life Course: Some Neglected Aspects .............. 131
Gunhild O. Hagestad and Pearl A. Dykstra

Part II Changing Social Contexts and Life Course Patterns


Family Heterogeneity Over the Life Course....................................... 161
Sandra Hofferth and Frances Goldscheider
Educational Pathways .......................................................................... 179
Robert Crosnoe and Aprile D. Benner
College for All: New Institutional Conflicts
in the Transition to Adulthood ............................................................. 201
Claudia Zapata-Gietl, James E. Rosenbaum,
Caitlin Ahearn, and Kelly Iwanaga Becker

v
vi Contents

Changes in Educational Inequality


in Cross-National Perspective .............................................................. 223
Pia N. Blossfeld, Gwendolin J. Blossfeld,
and Hans-Peter Blossfeld
Work Over the Gendered Life Course ................................................ 249
Phyllis Moen
Military Service in Lives: Where Do We Go From Here?................. 277
Andrew S. London and Janet M. Wilmoth
Criminal Justice and the Life Course ................................................. 301
Sara Wakeeld and Robert Apel
Disaster and Life Course Processes ..................................................... 321
Jack DeWaard

Part III Health and Development Through the Life Course


Early Childhood Poverty: Short and Long-Run
Consequences Over the Lifespan ......................................................... 341
Ariel Kalil, Greg J. Duncan, and Kathleen M. Ziol-Guest
Does the Body Forget? Adult Health,
Life Course Dynamics, and Social Change ......................................... 355
Mark D. Hayward and Connor M. Sheehan
Living Healthier and Longer: A Life Course
Perspective on Education and Health ................................................. 369
Monica Kirkpatrick Johnson, Jeremy Staff,
John E. Schulenberg, and Megan E. Patrick
Life Course Lens on Aging and Health ............................................... 389
Kenneth F. Ferraro
Mental Health ........................................................................................ 407
William R. Avison
Agency Across the Life Course ............................................................ 431
Steven Hitlin and Hye Won Kwon
Cognitive Development and the Life Course:
Growth, Stability and Decline .............................................................. 451
Duane F. Alwin, Jason R. Thomas, and Linda A. Wray

Part IV Life Course Research Methodologies


Longitudinal Qualitative Research...................................................... 491
Joseph C. Hermanowicz
Causality in Life Course Studies ......................................................... 515
Ravaris Moore and Jennie E. Brand
Contents vii

The Logic and Practice of Growth Curve Analysis:


Modeling Strategies for Life Course Dynamics ................................. 541
Ross Macmillan and Frank Furstenberg
Three Generation Studies: Methodological
Challenges and Promise........................................................................ 571
Terence P. Thornberry
Neighborhood, Place, and the Life Course ......................................... 597
Christopher R. Browning, Kathleen A. Cagney,
and Bethany Boettner

Part V The Life Course and Policy: Building the Nexus


Life Course Research and the Shaping of Public Policy ................... 623
John H. Laub
Epidemiological Perspectives on the Life Course .............................. 639
Michael E.J. Wadsworth and Diana Kuh
The Influence of Social Welfare Policies
on Health Disparities Across the Life Course..................................... 661
Pamela Herd
Life Course Risks and Welfare States Risk Management ................ 677
Martin Diewald
Longitudinal Studies and Policy for Children
in Low- and Middle-Income Countries:
Evidence from Young Lives .................................................................. 689
Paul Dornan
Lags and Leaps: The Dynamics of Demography,
Economy and Policy and Their Implications
for Life Course Research ...................................................................... 705
Angela M. ORand and Amie Bostic
About the Editors

Michael J. Shanahan is Professor of Sociology and Faculty Fellow at the


Carolina Population Center and at the Center for Developmental Science,
University of North Carolina at Chapel Hill. He is interested in individual
differences in the life course, a theme that he has studied with reference to
genetics, health, and personality. His recent articles appear in Social Forces,
Developmental Psychology, Journal of Health and Social Behavior, and
Social Science and Medicine. His current research examines associations
between social circumstances and gene transcription both early and later in
the life course.

Jeylan T. Mortimer is Professor of Sociology and Founding Director of the


Life Course Center, University of Minnesota, and Principal Investigator of
the longitudinal, three-generation Youth Development Study. Her interests
span the areas of work and socioeconomic attainment, the life course, and
youth and transition to adulthood. Her recent articles appear in Social
Psychology Quarterly, Social Forces, the Journal of Research on Adolescence,
the Journal of Marriage and Family, Developmental Psychology, the Journal
of Vocational Behavior, Longitudinal and Life Course Studies, and
Demography. Her current research examines intergenerational change and
transmission of achievement orientations, the impacts of prior parental expe-
riences and trajectories on their adolescent children, and the sources of resil-
ience in the transition to adulthood. Professor Mortimer is Chair of the
Section on Aging and the Life Course in the American Sociological
Association. She previously served as Chair of the ASA Sections on Children
and Youth and Social Psychology.

Monica Kirkpatrick Johnson is Professor of Sociology and Honors


College Distinguished Professor at Washington State University. Her research
interests are in the areas of work, family, and education across the life course,
with particular focus on well-being and achievement in adolescence and the
transition to adulthood. Her recent research has appeared in the American
Journal of Sociology, Longitudinal and Life Course Studies, Social Science
Research, Journal of Marriage and Family, Social Forces, and Social
Psychology Quarterly. Her current research focuses on families and young

ix
x About the Editors

peoples adaptations to the changing nature of the transition to adulthood,


including familial nancial support in young adulthood and its implications
for attainment and well-being; change in educational and work ambitions tied
to changing personal and historical circumstances; and the interplay of strati-
cation and the development of social psychological resources in promoting
well-being and attainment.
Introduction: Life Course
Studies Trends, Challenges,
and Future Directions

Michael J. Shanahan, Jeylan T. Mortimer,


and Monica Kirkpatrick Johnson

The Handbook of the Life Course was published second volume (i.e., with a distinct charge): the
in 2003 (Handbook I) and aspired to provide identication of new and emerging problems,
an overview of key theoretical perspectives, concepts, methods, research questions, and ana-
concepts, and methodological approaches that, lytic strategies. Nevertheless, although the hand-
while applied to diverse phenomena, are united books have unique purviews, they are not wholly
in their general approach to the study of lives distinct. Handbook I provides an overview of life
across age phases (Mortimer and Shanahan course studies but also identies many themes for
2003). In revisiting Handbook I, we were heart- future research. Indeed, Section VII is entitled
ened to discover its continued usefulness and The Future of the Life Course and includes
relevance. A cursory examination of citation pat- essays by leading scholars of life course sociol-
terns reveals that it has been cited frequently as a ogy. Handbook II is forward-looking, charting
corporate work, and many chapters were also directions for future research, yet the chapters in
well-received. Further, most of its chapters the second volume must necessarily provide
remain relevant, inspiring, and creative contribu- foundations on which to build. Thus, many of the
tions more than a decade after they were written. present chapters provide concise, selective histo-
Indeed, with little effort, any of its essays could ries of their topics. In the nal analysis, both vol-
be updated. umes offer overviews of subelds and directions
Given the sustained usefulness of Handbook I, for future research; the difference is one of
we agreed to edit a new Handbook II not as a emphasis, with the present volume highlighting
second edition (i.e., an update) but rather as a the latter.
In describing the contents of this volume, we
note connections between the two handbooks,
M.J. Shanahan (*) and unique features of each. First, however, we
Department of Sociology, University revisit an observation made in Handbook I: that
of North Carolina, Chapel Hill, NC, USA
e-mail: mjshan@email.unc.edu life course studies is growing by proverbial
leaps and bounds, in large part because of its
J.T. Mortimer
Life Course Center and Department of Sociology, growing use in elds beyond sociology. This dif-
University of Minnesota, Minneapolis, MN, USA fusion creates exciting opportunities for interdis-
e-mail: morti002@umn.edu ciplinary work, but it also challenges the eld to
M. Kirkpatrick Johnson maintain a coherent, paradigmatic core.
Department of Sociology, Washington State
University, Pullman, WA, USA
e-mail: monicakj@wsu.edu

Springer International Publishing Switzerland 2016 1


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_1
2 M.J. Shanahan et al.

1000
800
600
400
200
0

1960 1970 1980 1990 2000 2010


Year
Total Publications Biomedical / Epidemiology
Sociology Psychology

Fig. 1 Number of life course publications per year

1 Growth, Diffusion, would change appreciably if these limitations


Opportunities could be addressed.
and Challenges Figure 1 shows the number of life course
journal publications per year overall and for
Although gauging the growth of life course stud- sociological, psychological, and biomedical/epi-
ies is difcult, we gain some descriptive traction demiological Web of Science categories. The
by way of a simple citation analysis (Shanahan results reveal several interesting trends. First,
and Freeman 2012). Drawing on the Thomson/ there was very little publication activity prior to
Reuters Web of Science, we searched for papers 1990. Indeed, as late as 1998, there were only 84
in which the topic, title, or theme included life sociological journal articles featuring the life
course. The resulting record count (number of course. (There were, however, citations to books
journal publications) can be broken down into and chapters prior to 1998 that are not included
year and disciplines (Web of Science in the gure). Second, 1990 was a tipping point,
Categories).1 This strategy is subject to several after which the number of life course publications
important qualications (principally, papers can began to increase appreciably. In fact, regression
be classied into more than one Web of Science analyses support the hypothesis that 1990 is the
category, and books and chapters are often not likely deection point for total journal publica-
included), but we doubt the overall conclusions tion count. Third, also conrmed by regression
modeling, the growth rate in publications follows
1
a quadratic pattern. This pattern of rapid growth
We thank Autumn McClelland for assistance with these
analyses. Results available on request from Michael characterizes the total and discipline-specic
Shanahan. journal publication count. Finally, annual growth
Introduction: Life Course Studies Trends, Challenges, and Future Directions 3

rates in other disciplines reveal that life course course research? Or, has life course become an
publications in sociology have been at or slightly ambiguous rubric? On the other hand, this same
below the total growth rate (for all publications). diffusion creates opportunities for intellectual
Over a 10-year period (from 2000 to 2010), cross-pollination. A great deal of scientic prog-
Biomedical/Epidemiology publications increased ress depends on the concepts and methods from
by 302 %, whereas the total rate of growth for all one eld of study informing the science of
life course publications was 267 % (for another eld. Moreover, many branches of sci-
Sociology, 249 %; for Psychology, 263 %). encenatural, behavioral, and socialare
Given limitations to these analyses, we should adopting a systems view that encourages viewing
not be overly concerned with exact estimates. the full complexity of research questions, thereby
Viewed in broad terms, however, the results are traversing traditional disciplinary boundaries.
consistent with our impressions: life course stud- Thus, the challenge facing life course sociol-
ies are clearly characterized by vibrant growth, ogythe coherent core versus interdisciplinary
and much of this growth, especially in the past diffusionis a particular instance of a problem
5- to 10 years, is driven by publications beyond faced by many successful elds of study. We
the discipline of sociology. invite readers to consider this tension between
This interdisciplinary diffusion is a testament cohesiveness and cross-pollination as they peruse
to the value of life course studies, but it also the two handbooks.
raises the issue of intellectual cohesiveness. The
issue of cohesiveness is yet more vexing because
of the status of the life course as a paradigm. 2 Organization of Handbook II
Traditionally, the concept of the life course refers
to the age-graded roles that structure (or create We begin Handbook II with essays on the histori-
patterns in) biography. The concept in turn gave cal emergence of life course studies, explicating
rise to a set of related ideas, hypotheses, and and critically evaluating key foundational con-
techniques that collectively create a paradigm, or cepts and ideas that animate this eld of inquiry.
imaginative framework. Compared to a theory The historical perspective offered by Bynner
a set of formally interrelated propositions that emphasizes the importance of longitudinal data to
organizes observations and generates predic- the emergence of life course research and its insti-
tionsthe content, boundaries, and emphases of tutionalization; he urges heightened efforts to
a paradigm are less easy to identify with a high connect basic science with policy. Fundamental
degree of consensus. Thus, even within sociol- concepts covered in Section I include the birth
ogy, the life course has a diversity of meanings. cohort, which locates people in history, and age
This point is vividly illustrated by several chap- phases, the basic division of the biography into a
ters in this volume that focus on health, yet with sequence of age-graded segments. Section I closes
little substantive overlap among them. with three essays urging researchers to focus on
The diffusion of the life course and its status societal forces that create biographic patterns,
as a paradigm create both a challenge and an focusing on the increasing differentiation of age
opportunity for life course studies. On the one grading and its subjective consequences; on the
hand, as the life course paradigm diffuses from frequent misuse of models positing a free-stand-
sociology into other elds of study, the challenge ing individual in life course explanations of indi-
of intellectual cohesiveness arises: is there an vidual differences; and on the central role of
intellectual core that informs life course studies intergenerational families in bridging macro and
of diverse phenomena such as occupational micro perspectives. These essays provide a much-
careers, criminal careers, cardiovascular disease, needed corrective as life course scholars continue
cognition, and functional limitations? Or, much to study individual-level phenomena such as
as James House (1977) did for social psychology, health and genetics, and as they engage in inter-
can we identify distinctly different faces of life disciplinary research.
4 M.J. Shanahan et al.

Section II applies the life course perspective to closes with essays by leading scholars who are
the changing institutions and organizations that building bridges between basic life course
powerfully shape biographical patterns within research and policy.
and across cohorts and age phases. Traditionally, We turn now from this organizational portrait
these social forces have included the family, edu- to commentary on the chapters in this volume.
cation, and work, all of which are covered in the Our purpose is not to summarize each chapter,
present volume. Particularly in the American but rather to highlight select themes and future
context, military service in theaters of armed directions for research and, in so doing, encour-
conict and time spent in the criminal justice sys- age close readings of chapters. Table 1 presents a
tem shape the lives of a substantial proportion of highly select summary of broadly-stated themes
the population, and there is now growing aware- for future research that are to be found across
ness across nations of the potentially profound essays in this volume. Again, our purpose with
long-term effects of both natural and man-made Table 1 is not to comprehensively list all future
disasters. Excepting disasters and military ser- directions for life course studies, but rather to
vice, these topics are also covered in Handbook I, highlight select, major themes and to encourage
and the present contributions seek to highlight closer reading of the volume.
recent developments and future directions.
Section III presents a series of essays on health
and development. As shown in Fig. 1, health has 3 Section I: Foundations of Life
emerged as a major theme in life course research. Course Research
In the American context especially, life course
studies include a micro interest in individual dif- John Bynners essay on the institutionalization of
ferences, and this theme is explored in several life course studies is an appropriate starting point
chapters that offer developmentally-informed life because he discusses the historical emergence of
course perspectives on physical and mental health, life course studies and their expansion and con-
poverty, cognition, and agency. Handbook I also solidation to the present, including the recent
included essays on agency and health, as well as founding of the Society for Longitudinal and Life
on connections between childhood and adult- Course Studies. Bynners thesis is that the history
hood, substance use, personality, and biology. of life course studies is best told in terms of the
Having covered fundamental concepts and history and expansion of longitudinal data collec-
applications of the life course perspective, tion. He provides a superb overview of the prolif-
Handbook II turns to research methods in Section eration of longitudinal data in Europe and the
IV. Handbook I examines models that are well- United States. Indeed, he notes that longitudinal
suited to common types of data in the life course data collection is now industrial in scope, with
tradition: age-period-cohort, event-history, panel, consortia emerging to harmonize efforts and to
and latent pathways. There is also a chapter on develop a clear understanding of the types of data
the qualitative study of social change and peo- that are being collected with respect to broad
ples narrative autobiographies. Handbook II themes such as child health and household
covers new topics that are at the forefront of life dynamics. Handbook I also began with a histori-
course research: longitudinal qualitative research, cal perspective (Elder, Crosnoe, and Johnson),
causal analysis, growth curve models, three- but one told with different emphasis: the emer-
generation studies, and spatial analysis. gence of life course principles based on early
Finally, Section V returns to a major theme of empirical work. With their differing historiogra-
John Bynners rst chapter: the increasing neces- phiesthe central role of data and the emergence
sity to connect life course research with policy. of theorythese chapters jointly provide a con-
Presently, few life course scholars have been cise, rich story of the emergence of a new, vibrant,
trained in policy analysis and few policy-makers multidisciplinary, and international intellectual
draw on a life course framework. Handbook II community.
Introduction: Life Course Studies Trends, Challenges, and Future Directions 5

Table 1 Select recommendations for the future of life course studiesa


I. Increasing the international scope of life course studies
A. Promotion of cross-national data harmonization through consortia (Bynner)
B. Development of international ethical standards governing data collection, storage, and anonymity/condentiality
(Bynner)
C. Extension of data collection to middle- and low-income countries (Bynner, Dornan, Wadsworth)
II. Major substantive themes for future research
A. The role of early childhood and adolescent experiences in the subsequent life course
1. Evaluate the contemporary view that early childhood is a critical or even sensitive period (e.g., via embedding
mechanisms) (Dannefer, Hayward, Kalil)
2. Identify the salient features of early environments that predict the later life course (including social context as
constituted by forms of adversity, Ferraro; policy, Herd; family, Hofferth)
3. Identify early individual differences that most differentiate people (for later health, Avison, Hayward; for school
performance, Crosnoe; for targeting policy to promote education, income, and health, Diewald, Dornan)
4. Identify diverse mechanisms that link early and later life course (for health, Avison, Hayward; for attainment,
Blossfeld, Kalil; reecting spatial differences, Browning)
5. Identify pathways of diversity in early life experiences (reecting gene-environment patterns and development of
soft skills, Diewald; health, Dornan; adversities, Ferraro; family, Hofferth)
B. Cohorts and life course studies
1. Empirical study of demonstration effectshow birth cohorts create contexts for subsequent cohorts,
facilitating social change (Elder)
2. Increased attention to both inter- and intra- cohort heterogeneity, and explanatory mechanisms (Elder,
Hermanowicz)
3. Increased awareness of likely cohort effects (in the study of disasters, DeWaard; health, Hayward; incarceration,
Wakeeld)
4. Joining of cohort and spatial data to study people in time and place (Elder)
5. Study of structural and cultural lags and ambiguities, difculties, and possibilities created by them (Dannefer,
Moen, Mortimer, ORand)
6. Use of age-cohort panel data to study age, period, cohort mechanisms (Elder, Hitlin, Johnson, Hayward)
C. Age phases of the life course
1. Study of different age phases experienced by groups (dened by sex, race, social status, etc.) in a society
(Mortimer)
2. Identication of social, economic, legal, and political forces that create, modify, and destroy age phases
(Hagestad, Moen, Mortimer, ORand)
3. Imagining different legal-regulatory structures that shape the life course and studying them with program eld
experiments, quasi-experiments (Dannefer, Herd, Moen, Mortimer)
4. Identication of experiences that fundamentally alter pathways through life (mental health, Avison; aging and
health, Ferraro; disasters, DeWaard; military, London; incarceration, Wakeeld)
D. Intergenerational studies
1. Consequences of changing demographic and institutional contexts for intergenerational ties (using cross-national
designs, Hagestad; with respect to family diversity, Hofferth)
2. Studying intergenerational transmission of behaviors, attitudes, etc. as a developmental problem (Thornberry)
3. Studying parental involvement in institutions on life courses of children (education, Johnson; the military,
London; incarceration, Wakeeld)
4. Connections between biological ties, cultural understandings, legal-policy settings of families, and
intergenerational supports (Diewald, Hagestad, Herd, Hofferth)
E. Intersection of human development and the life course
1. Turning points and consequences for trajectories (generally, Alwin, Macmillan; cognitive functioning, Alwin;
mental health, Avison; body mass index, Macmillan)
2. Connections between short- and long-term consequences of major life course experiences (education, Johnson)
3. Subjective experience of the life course (and agency, Hitlin; with respect to professional careers, research
methods to study, Hermanowicz)
(continued)
6 M.J. Shanahan et al.

Table 1 (continued)
F. The importance of decision-making and soft-skill processes in the life course
1. Knowledge of educational opportunities and constraints such as curricular structures, time demands, etc.
(Blossfeld, Crosnoe, Zapata-Gietl)
2. Financial literacy (ORand)
3. Cognitive competencies (Alwin, Blossfeld)
4. Soft skills in attainment and overcoming adversities (Diewald, Kalil)
5. Decision-making about the life course as culturally-conditioned deliberation, reecting societal values, norms
(Hitlin)
III. Major methodological challenges
A. Leveraging survey research
1. Increased understanding of respondent burden, retrospectively-recalled data (Bynner)
2. Availability of data, transparency in analyses, replication (Editors)
3. Merging of administrative records to participants in panel studies (Bynner)
4. Increasingly sophisticated use of biological data, including biomarkers, diverse forms of genetic information,
microbiome, imaging, and animal models (Avison, Bynner, Dannefer, Diewald, Ferraro, Hayward)
5. Increasing use of spatial data that describes social and physical contexts over time (Browning, Elder, Ferraro)
B. Research methodologies
1. Strengthen causal inference by experiments, naturally occurring experiments, quasi-experiments, and statistical
methods (Avison, Hayward, Herd, Johnson, Kalil, Moore, Moen)
2. Increased use of qualitative strategies to identify life course patterns and mechanisms (Hermanowicz)
3. Models that capture temporal complexities of both context and outcome (Alwin, Browning, Macmillan)
C. Measurement
1. Cascades of health; health portfolios; symptoms versus diseases and disorders (Avison, Hayward)
2. Culturally sensitive measures of agency (Hitlin)
3. Political economies for cross-national comparative research (Diewald)
4. Techniques for dening and delimiting age phases and generations (Elder, Mortimer)
IV. Life course and public policy
A. Increased collaboration between policy-makers and life course researchers through all stages of the research
process (Bynner, Laub)
B. Increased study of specic laws and regulations from a life course perspective (all chapters; see especially, Herd)
and cross-nationally (Bynner, Diewald, Dornan, Hagestad, Wadsworth)
C. Development and test of theory of life course risk management and mitigation cross-nationally (Diewald)
D. Study of policy and the life course in broader framework of rapid global changes with respect to, for example,
women and the economy (Moen, ORand)
a
For the sake of concision, listed names refer to rst authors of chapters. Names listed alphabetically

Bynner suggests that the future of life course advances with respect to causal analysis, attri-
studies will continue to depend on advances in tion, respondent burden, and theory. To these
data collection efforts, and this insight leads to many excellent points, we would add that the
several recommendations for future work: availability of data, transparency in sharing anal-
increasing the availability of sophisticated forms yses, and issues of replication are now concerns
of diverse data (biological, spatial, administra- cutting across the social and behavioral sciences;
tive, psychological and behavioral); the develop- increased attention to each of these issues will
ment of international standards with respect to strengthen the knowledge base.
the ethics of data collection, storage, and the pre- One of the foundational concepts of the life
vention of deductive disclosure; strengthening course is the birth cohort because, as Elder and
connections between basic research and policy; George note, cohorts locate the aging person in
the extension of data collection efforts beyond historical context, and they also can provide
the United States and Europe; and scientic insight into the genesis of social change. The
Introduction: Life Course Studies Trends, Challenges, and Future Directions 7

early statistical work of Schaie and the concep- to individual-level phenomena such as health.
tual work of Ryder laid the foundations for sub- Yet the authors argue that life course scholars
sequent research. Elder and George provide an have too readily (and perhaps unwittingly)
incisive review of these and related foundations, adopted a functional-developmental paradigm
noting that although cohort research was origi- that compromises the social richness of their
nally formulated with reference to between- explanations. Individual behaviors are socially
cohort comparisons, the eld quickly realized constituted and, as such, need to be studied in
that within-cohort heterogeneity was equally terms beyond the free-standing individual and
important in locating people in history. They also with a skeptical stance toward existing social
argue that although cohort research has produc- arrangementsranging from institutions to the
tively focused on a vast array of substantive top- life course itselfas natural. Their thesis is
ics, two processescohort replacement and illustrated with three probative examples: agency,
cohort diffusionhave emerged as fundamental gene-environment interactions, and early inu-
mechanisms that cut across many of these areas ences on outcomes later in life. In all of these
of research. They illustrate these pointsthe instances, Dannefer and his colleagues argue that
importance of juxtaposing inter- and intra-cohort sociology has adopted explanations based on
comparisons and of replacement and diffusion individual attributes (e.g., biological embedding)
by considering research on secularization, the and lost perspective on the fundamental impor-
Easterlin hypothesis, family structure, and gen- tance of peoples position in society and the
der role orientations. Alwin and McCammons malleability of both social institutions and the
essay on Generations and Glenns essay on age/ person throughout life. The implications of their
period/cohort models in Handbook I are excel- chapter for future research are far-reaching: the
lent companion essays; these three essays jointly considerable efforts devoted to the study of
provide superb foundations and future directions individual outcomes need to re-focus, shifting
for thinking about aging in historical context. attention to the social forces that create patterns
Elder and George identify several areas for in the life course, social forces that are often
future research. First, they discuss a neglected dysfunctional.
aspect of cohort research: how cohorts create Mortimer and Moen address the fundamental
social change. They elaborate Modells point that concept of age. The German word bauplan means
each succeeding cohort creates a new context for building plan and is used broadly to refer to the
subsequent cohorts, a demonstration effect fundamental structural features of something
according to which social change inuences the (load-bearing structures of a building, morpho-
life course of individual members of the cohort, logical features of animals, etc.). Age phases are
and these individual experiences aggregate to indeed the bauplan of the life coursethe basic
constitute a new context for others. Second, age-graded segments (e.g., adolescence, old-old
future research can protably explore connec- age) that are characterized by unique sets of
tions between inter- and intra-cohort heterogene- opportunities and limitations, that are widely rec-
ity. As longitudinal designs spanning many ognized by society, and that are regulated by
cohorts become increasingly available (e.g., the norms and societal and economic structures.
Health and Retirement Study), such studies will Mortimer and Moen note that the age phases of
be possible. And third, they recommend the addi- the life course have become increasingly differ-
tion of an ecological (spatial) focus to cohort entiated. Indeed, the pace of social change (prin-
research. Many life course processes (e.g., migra- cipally, demographic and economic) and the rise
tion, health, poverty) are known to reect cohort of individualism may be creating new age phases
membership and spatial processes, but studies much more quickly than social institutions
have not bridged these two perspectives. (especially governmental policies) and expecta-
Dannefer, Kelley-Moore, and Huang note that tions adapt, vivid examples of both cultural and
life course studies devote considerable attention structural lag.
8 M.J. Shanahan et al.

The result is increasing ambiguity about icy for the gendered life course. Key to this bridg-
appropriate age-graded behaviors, an ambigu- ing effort is the mediating role of the family,
ity further compounded by the very different life especially intergenerational family networks.
course experiences of women and men (the gen- Hagestad and her colleagues explore the many
dered life course), of different social classes (the complexities that arise from this basic insight: the
stratied life course) and of different races (the consequences of age segregation for intergenera-
racialized life course). The concept of intersec- tional relationships, shifting roles and responsi-
tionalitythat inequalities are best understood in bilities within intergenerational networks, and
terms of complex interactions involving gender, the increasing absence of such networks among
socioeconomic status, and race through time childless men. Traditionally, roles, opportunities,
raises the possibilities that different subgroups in and limitations in intergenerational networks
society have unique age phases (e.g., poor black have varied by gender. Yet, as the authors note,
males often experience an imprisonment phase) public policy is just beginning to acknowledge
or, in any event, quite unique experiences of age this reality by, for example, encouraging involved
phases that are common to the broader society dads in Scandinavian countries, Portugal, and
(e.g., adolescence). These themes are discussed Germany. These themes are explored with refer-
in terms of two specic examples, early adult- ence to the considerable diversity among nations
hood and encore adulthood, though they discuss in their policies governing rights and duties
several others as well, including midcourse (see within intergenerational networks.
Moen, Handbook I) and old age.
Mortimer and Moens analysis raises a series
of questions for future research that aregiven 4 Section II: Changing Social
how much progress has been made in life course Contexts and Life Course
studiessurprisingly foundational. How many Patterns
phases of the life course are there, and how do we
know? Or are age phases heuristic devices or Hofferth and Goldscheider review social changes
ideal types that should not be reied? What social since the mid-twentieth century that have affected
and economic forces maintain, destroy, and cre- family structures, family processes, and the
ate new phases? What are the implications of new socialization of children. The two stages of the
age phases for policy-makers? And, nally, what gender revolution (the increasing employment of
consequences do new age phases have for the women followed by increase in male participa-
individual, who, as John Meyer (1987) has tion in domestic work), lengthening education,
observed, negotiates the tension between having and shifts in societal norms have dramatically
a legitimate biography in the eyes of other peo- altered the family life course and produced
ple, while at the same time creating an individu- increasing diversity in family forms and child
alized biography as well? socialization. These trends are evident in aggre-
The nal chapter in Part I is an excellent com- gate statistics that document increases in single
panion piece to Mortimer and Moens essay. parent families, non-marital child-bearing,
Hagestad and Dykstra (with van Baarle) begin divorce, and the ages of marriage and childbirth,
with the observation that North American life as well as in time use studies that describe con-
course research has tended to focus on micro- siderable change in childrens lives (e.g., gender
level experiences of men and women, while convergence in childrens time spent in house-
European efforts have focused on macro-level work and sports).
phenomena, especially the implications of policy As the traditional family life course recedes
settings and demographic changes for life course in prevalence, children are increasingly reared in
patterns. The next generation research should dual earner families, stepfamilies, single parent
bridge this continental divide by examining the families, and with same sex parents. Moreover,
implications of changing demographics and pol- they are increasingly exposed to multiple family
Introduction: Life Course Studies Trends, Challenges, and Future Directions 9

structures before they themselves reach maturity process, as initial differences in cognitive skills
and form independent households. The authors and knowledge become increasingly pronounced
review substantial evidence that the character of through time. Crosnoe and Benner make an
experience in the highly formative years of child- important distinction between primary effects
hood and adolescence has major consequences that occur by virtue of differences in cognitive
for patterns of family formation and parent-child skill and preparation, and secondary effects that
relationships in the next generation. Their empha- arise as a consequence of differences in know-
sis on family change nicely complements chap- how and decision-making. Thus, interventions to
ters in Handbook I by Uhlenberg and Mueller, erase educational differentials through early skill
and by Putney and Bengtson. development (e.g., preschool programs) will only
Hofferth and Goldscheider emphasize that be partially successful if differences in informa-
these changes are manifest quite differently by tion (the importance of taking math and science
social class, racial/ethnic and immigrant status, coursework for college acceptance, how to navi-
and gender. Increasing diversity in family struc- gate nancial aid) are not also addressed.
tures and the timing of family formation and The consequences of education are pervasive
change (e.g. partnership dissolution and re- throughout the life course, promoting success
partnering) raise the potential for greater indi- across numerous adult domains (e.g., occupational
vidual choice. At the same time, low income and status, earnings, health and mortality, marital sta-
minority families face considerable constraints in bility and parenting quality). Increasingly, the cen-
constructing stable family lives. Disruption in tral distinguishing feature of haves and have
family life is more likely in less educated, lower nots in contemporary society is the acquisition of
income families that are the least able to cope a 4-year college degree. The authors identify
with such change. phases of heightened vulnerability and responsive-
Hofferth and Goldscheiders essay shows that ness that suggest points of effective intervention to
in the face of great heterogeneity and social reduce gross educational inequalities.
change, understanding family life will become In the next chapter, Zpata-Gietl, Rosenbaum,
increasingly challenging to life course scholars. and their colleagues draw attention to a much-
They call for future research on social class dif- neglected sector of higher education, the commu-
ferences in the consequences of cumulative expo- nity college, which enrolls vast numbers of
sure to distinct family structures and experiences; minority and rst-generation college students.
processes of partnering and marital/relationship Despite its rapid growth (now constituting nearly
sorting in youth, young adulthood, and adult- half of all U.S. college enrollment) and increas-
hood; immigrant families; same sex couples; and ing role in the transition from school to work and
gender roles in retirement. to adulthood in general, community colleges
The following three chapters on changing remain understudied. The authors document the
educational contexts indicate the promise of extent of community college expansion and iden-
greater interaction between scholars of education tify key problems for students and their families.
and those who study the life course. These chap- Like 4-year college students, they typically take
ters extend earlier considerations of education longer than expected to complete their degrees
found in Handbook I by Kariya and Rosenbaum, and many never do so; over 40 % of community
Pallas, and Entwisle, et al. Crosnoe and Benner college entrants do not have an Associates
describe the many ways that disparities in educa- degree or certicate 8 years later.
tional achievement and attainment (e.g., by race/ Students in community colleges face distinct
ethnicity and social class) develop and increase problems, largely due to their status as rst gen-
as the individual moves through childhood, ado- eration college students, lacking parents and
lescence, and adulthood. As illustrated by other relatives who can guide them through the
Entwisle, Alexander and Olsens chapter in college entrance and completion process. The
Handbook I, this is a quintessentially cumulative disadvantages of community college students
10 M.J. Shanahan et al.

well illustrate the secondary effects noted in substantially in modern societies, in accord with
Crosnoe and Benners chapter. That is, commu- global shifts in labor force composition and
nity college aspirants often lack basic informa- needs. Given universal gains in education, they
tion about what they need to do in high school to ask: Which groups (dened by educational origin
prepare for community college entrance, how to and gender) have beneted most strongly from
acquire nancial aid, how to select programs that educational expansion? How might educational
mesh well with their career goals, etc. Delays in advantages be transmitted intergenerationally?
degree achievement lead them to postpone nor- Are the answers to these questions the same
mative transitions to adulthood, such as marriage, across national contexts? Can prominent theories
parenthood, and entry to careers, while taking on of educational upgrading and expansion, posed
considerable debt. Those who begin family for- as universal trends in contemporary societies,
mation face role conicts among their student, explain the patterns of change across countries?
marital and parental roles. With delays in degree PIAAC (Program for the International
completion, their advancing age and assumption Assessment of Adult Competencies) data from
of adult statuses (e.g., as marital partner and 22 countries collected in 2011 and 2012 enable
parent) reduce parents tendencies to support cross-cohort comparisons of intergenerational
them, nancially and otherwise. inequality in educational attainment in succes-
The authors insights have strong policy sive birth cohorts from 19471952 to 1978
implications. For example, failure to pass the 1982. Gender is a strong theme in their chapter,
remedial placement test prior to community col- as rising educational attainments of women have
lege entry leads to remedial classes for the major- reduced male advantage in some contexts and
ity of students, which take time and money but do led women to surpass men in others. While girls
not earn credits towards a degree. The authors have advantages over boys in some cognitive
recommend that high schools administer such skills, in self-regulation, and teachers evalua-
tests in the junior year so that students may be tions, these advantages cannot account for wom-
alerted to deciencies and have their senior year ens steep rise in educational attainment in recent
to address them. Timely advising and clearer cur- decades. The authors instead link girls rising
ricular structures are also advised. The lack of educational attainments to concomitant shifts in
institutional bridges from school to work makes gender roles and time usethe increasing preva-
for severe challenges for students (see also lence of dual earner families, reduced gender
Kerckhoff, Handbook I), decreasing their moti- discrimination, and expanded opportunities for
vation and efforts while pursuing community col- women in occupations requiring higher educa-
lege programs and reducing their ability, after tional degrees. In a previous era, educating girls
graduation, to nd suitable work. The authors was considered less important given their lesser
offer several recommendations to increase the contribution to intergenerational social class
integration of community colleges with local standing. Increasingly, however, family socio-
employers and call for research on change in economic status depends on two earners. Mens
expectations and actual decisions regarding fam- roles, in contrast, have changed much less over
ily formation and other transitions in the face of time; many men still seek employment in blue
extended educational timetables. We need to collar and technical occupations for which voca-
know much more about what enables community tional training is sufcient. Complicating this
college students to succeed despite the odds. explanation, more highly educated families, with
In a remarkable empirical study highlighting their egalitarian norms about education, have
the life course principles of time and place, historically promoted girls and boys attain-
Blossfeld, Blossfeld, and Blossfeld examine ments more equally.
changes in educational inequality in cross- The investigators nd distinct patterns of
national perspective. They start with the observa- change in inequality across both secondary and
tion that educational attainment has increased tertiary levels. With respect to tertiary educa-
Introduction: Life Course Studies Trends, Challenges, and Future Directions 11

tional expansion (i.e., college graduation), in nine mystique for women, was founded on the
some countries inequality of educational attain- assumption that mens attention should be pri-
ment has risen, supporting cultural reproduction marily, if not exclusively, focused on achieving
theory. In the U.S., women in the most highly success in the paid work role, while womens
educated families are clear winners, and chil- focus should be conned to the home and rearing
dren of the least educated parents have not bene- children.
ted at all. A similar pattern is found in Poland. Though not all families could successfully
In contrast, South Korea belongs to a group of enact this structure, the single provider model of
countries in which growth in higher educational the family constituted the blueprint for life rou-
attainments has drawn in children from all educa- tines and objectives, the way things should be.
tional origins, reducing inequality, consistent For many male workers, structural features of
with modernization theory. The authors relate employment promoted this model of life and pro-
these differences to the rapidity of educational vided safety net provisions when it was jeopar-
expansion, ideological shifts, and the individual- dized. A living wage, more or less stable
ization of risk. employment and career ladders, unionization,
While this brief synopsis cannot convey the pensions, unemployment compensation and dis-
richness and scope of this cross-national longitu- ability insurance provided structural scaffolding
dinal comparison, sufce it to say that the to this regime, promoting family economic secu-
Blossfelds analysis illuminates the trajectories rity and well-being. In the present era of global-
of educational careers across nations and over ization, increasing competition, non-standard
time, the intergenerational transmission of edu- employment contracts, rapid technological and
cation, and the multifaceted interrelations of occupational change, wage stagnation and grow-
cross-national cultures, educational structures, ing inequality, the security of employment, and
social class mobility, and gender. The authors call lockstep movements from education to work to
for future research to enable a fuller understand- retirement have vanished.
ing of the manner in which parental education At the same time, echoing Hofferth and
conveys advantages to children (e.g., through Goldscheiders discussion, Moen notes that fami-
cognitive competencies, non-cognitive traits, lies have become much more heterogeneous. The
parental information and decision-making, and need for two earners to secure a satisfactory stan-
teachers evaluations). They underscore the need dard of living, along with womens widening
for detailed prospective and cross-national longi- aspirations, have propelled women into the work-
tudinal data on the long-term relationships force. Many employed women seek to enact the
between parents and children and how they inu- traditionally male career mystique, seeking
ence educational careers through the life course. parity with men in occupational and income
Moens essay, Work over the Gendered Life attainments. Increasing longevity and extension
Course, draws our attention to the increasingly of the active life span, coupled with the grow-
precarious and unpredictable nature of work and ing cost of living, have led many workers, both
its implications for the life patterns of men and male and female, to postpone retirement, further
women (see also Heinz, Handbook I). Although increasing labor force heterogeneity. But despite
the character of employment, the societal context the many societal and workforce changes render-
of work, and the attributes of workers have ing the traditional life-long full-time homemaker
changed drastically since the mid-twentieth cen- role nearly extinct, women are still expected,
tury, the normative framework and expectations even when working full time, to be primarily
surrounding work and careers have remained sur- responsible for child and elder care, as well as the
prisingly constant, constituting, according to care of their male partners. As a result, womens
Moen, a relic of times past. The 1950s labor force participation is intermittent, with
breadwinner-homemaker family, with its accom- interruptions dictated by family needs. This
panying career mystique for men and femi- pattern produces long-term disadvantage in
12 M.J. Shanahan et al.

career attainment and resource accumulation wealth of information about the character of mili-
both during and after their working lives. tary service and the attributes of those who enter
Moen argues for a re-thinking of basic the military. They make a convincing case for the
assumptions on the part of scholars as well as the importance of this experience, based on the num-
populace more broadly, e.g., that work always is bers of people who are serving at any given time,
the primary concern, that the male experience is who have served, or who are closely linked to
the template for the expected or ideal life course, persons with past or contemporaneous service,
that the lock-step three-pronged life course is and the potential for pervasive positive as well as
inevitable (preparation for work, work, and negative outcomes.
retirement), and that the individual is the focal London and Wilmoths essay illustrates the
unit. She advocates for replacement of these value of life course concepts (life long develop-
assumptions with recognition of the diverse goals ment, historical time and place, timing and
and circumstances of workers, who remain con- sequencing, human agency, and linked lives) in
nected to other family members whose changing studying selection to the military, contemporane-
needs may take precedence; the convergence of ous effects on health and well-being, and longer-
male and female employment and family care term consequences for educational attainment,
patterns; the need for greater exibility in rules labor force outcomes, family formation, marital
governing the acceptable work day, work week, stability, and health. They note studies of the
and work year; and greater exibility in the allo- military that have become classics in life course
cation of time to work, education, child rearing, studiese.g., Sampson and Laubs research on
and elder care to break up the traditional lock- service as a kning off experience, leading to
step life course (see also Moen, Handbook I). positive turning points away from trajectories of
Moen asks, how can we respond to the institu- deviance and crime, and Elders work demon-
tional inertia that is so much a part of the taken- strating more positive effects on later outcomes
for-granted character of modern existence, with when service occurs earlier rather than later in
women exerting monumental efforts to fulll their life, with the greater potential to disrupt ongoing
responsibilities in their rst (employment) and trajectories of family life and work. Drawing on
second (family) work shifts? To answer this ques- the extant empirical evidence, they illustrate the
tion, she advocates more experimental designs and impacts of military service on child and adoles-
eld experiments to demonstrate the benets (and cent offspring of active service members and vet-
costs) of organizational accommodation to work- erans; and on early and later adults, among those
ers changed lives. Then we will know which inno- who have served. Longitudinal studies suggest
vations promote life quality, organizational the need for long-term monitoring as detrimental
effectiveness and gender equity. Moens ideas are outcomes may dissipate over time, and positive
revolutionary in their scope and implications outcomes may take some time to emerge.
extending far beyond work to the ways we think Given recent wars in Iraq and Afghanistan,
about the gendered structure of time and the life military service in the context of a volunteer
course itself. Particularly intriguing is her empha- army has become a near-constant option for
sis on experimental life course innovation young people. London and Wilmoth argue that in
designing and implementing innovations that can the context of delayed transitions in emerging
test ideas about re-organizing the life course. adulthood military service might be aptly con-
While the ve initial chapters in Section II sidered a pathway to adulthood in itself rather
focus on changing contexts that have been much than (as framed in some research) an experience
studied by life course scholarsthe family, edu- that delays this transition. The authors call for the
cation, and workthe last three address contexts collection of nationally representative longitudi-
that have been given less attention: the military, nal data that enables study of selection into the
the criminal justice system, and disasters. military, the timing and diversity of military
In their assessment of the military as a life experiences, and both positive (educational
course context, London and Wilmoth provide a attainment, earnings) and negative (PTSD, injury,
Introduction: Life Course Studies Trends, Challenges, and Future Directions 13

depression, marital instability) outcomes Prison clearly has countervailing consequences


throughout life. Ideally, such study would include for crime. It removes offenders from society, pre-
others connected to military service members venting criminal activity while incarcerated. At the
and veterans, and address variation by gender, same time, it promotes recidivism by increasing
race/ethnicity, social class origin, and sexual criminal capital and by preventing the resump-
identity. The research they advocate, which, like tion, or initiation, of a conventional life course
the many studies they review, is likely to demon- upon release. As a result, the authors consider it
strate positive as well as negative impacts, has essential for future researchers to address the fea-
strong policy implications given recruitment dif- tures of the prison experience that may lead to bet-
culties and concerns that the all-volunteer mili- ter outcomes (job training and post-secondary
tary places undue burden on a very small portion education). Similarly, it is important to know for
of our population. whom rehabilitative efforts actually work so as to
With notable exceptions (see Handbook I be able to inform effective interventions and
chapters by Sampson and Laub and by Uggen reforms. To achieve this end, prospective longitu-
and Massoglia), the criminal justice system is a dinal studies that document experiences in prison
second institution that has been neglected by life and that follow ex-prisoners well after leaving
course researchers, despite the fact that crime and prison are necessary.
punishment are highly age graded and have Disaster, a third area neglected by life course
marked consequences for future trajectories. scholars, constitutes an extreme example of con-
Wakeeld and Apel highlight the life course prin- textual change. For DeWaard, disasters are a
ciples of context and timing: the development of most promising domain of life course research
mass incarceration since the 1970s, the dispro- given their potential to fundamentally alter
portionate concentration of criminal justice sanc- human lives as people suffer severe injury, dis-
tions in young, minority, and disadvantaged ability, homelessness, and other disruptions that
populations, and the consequences of removing threaten ongoing family and work trajectories.
individuals on the cusp of adulthood from pro- Since both natural and man-made calamities
gressing, along with their age peers, into adult fall under the rubric of disaster studies, a wide
roles. They document the pervasive impacts of range of phenomena are encompassed: wars, ter-
incarceration on the linked lives of family rorist attacks, earthquakes, hurricanes, oods,
members (e.g., physical and mental health, harsh droughts, and other events that wreak havoc on
parenting, marital dissolution, wages, homeless- people and places. DeWaard conceptualizes
ness, childrens educational achievement). disasters as processes, not discrete events, and
Finally, under the rubric of agency, they discuss points out that vulnerability to disaster, and the
the numerous draws of crime and devianceas a reasons for vulnerability, depend on the life
badge of honor in the criminal culture and a mode course stage in which it is experienced. For
of acquiring material resources. Nonetheless, example, for children and adolescents, disasters
felony convictions and incarceration are occu- threaten sensitive processes of emotional and
pational hazards posing numerous risks in terms cognitive development; for adults, disasters jeop-
of collateral consequences. Indeed, incarcera- ardize housing and other material resources; for
tion has the paradoxical effect of creating a large the elderly, disasters disrupt access to and conti-
population that is hidden, locked out, and living nuity of medical care. Life course consequences
on the margins of society, given formal exclu- depend on pre-existing vulnerabilities of persons
sion from employment, educational opportuni- and places and features of the disaster itself,
ties, and public housing, and informal penalties, especially the sudden or delayed character of
including stigmatization, surveillance following onset and desistance. To illustrate these ideas, he
prison, and secondary prisonization, the perva- estimates a dynamic population model showing
sive difculties inicted on partners and how disaster disrupts the age proles of migra-
children. tion and mobility.
14 M.J. Shanahan et al.

DeWaard calls for longitudinal research that with reference to the timing of poverty and its
inventories individuals pre-disaster characteris- consequences.
tics, measures the character of their exposure to Kalil and her colleagues emphasize study
disastrous events, and surveys long-term out- design and internal validity (see Moore and
comes. While extant research provides some Brand, this volume), and discuss results from
indication that persons are differentially vulnera- experimental and quasi-experimental studies.
ble to disasters as a function of their prior mental Their review points to several intriguing avenues
health and earlier adverse life events, we know for future research. First, research typically
little about what makes people especially at risk begins assessing children when they enter school
at distinct stages of the life course. His essay sug- systems and, consequently, very little is known
gests the potential for research focusing on expo- about the role of poverty (even less, its mecha-
sure to particular types of disaster at particular nisms) in the earliest stages of childhood. Second,
life stages. For example, might youth and young there appear to be long-term consequences of
adults be especially vulnerable to disasters linked childrens poverty for productivity in adulthood,
to climate change (e.g., droughts), since they will but these associations may not be mediated by
be the ones who are called upon to serve in wars schooling, wage rates, or behavioral indicators.
over scarce resources? They are also more apt to Thus, a major challenge is to identify the opera-
migrate across national borders as the acquisition tive mechanisms that link early poverty with life-
of resources fundamental to human life (food, long attainments. Finally, these two themes are
water) becomes problematic. Their attempts to complicated by both measurement strategies for
acquire adult role markers, with marked conse- assessing income and poverty, and by the wide
quences for future trajectories, will be beset with array of consequences that are studied.
all the difculties and disadvantages associated Hayward and Sheehan focus on adult health as
with immigrant status. a consequence of life-long experiences.
Particularly with the publication of Hayward and
Gormans (2004) classic essay on the long arm
5 Section III: Health of childhood, a considerable body of research
and Development Through suggests associations between childhood socio-
the Life Course economic circumstances and adult health. In the
present essay, Hayward and Sheehan identify
Section III applies life course perspectives to three avenues of research that will further eluci-
individual differences, including aspects of date these associations. First, studies will ideally
health, poverty, agency, and cognition. Kalil, shift from consideration of one indicator of health
Duncan, and Ziol-Guest begin this section with a to a holistic consideration of a portfolio of indi-
consideration of the extensive literature on the cators of health and well-being. This recommen-
effects of poverty on childrens development. dation follows from several ndings. Clearly,
They note that, already in the rst grade, teach- health is multi-faceted, but the authors discuss
ers ratings of students skills (e.g., paying atten- research showing that specic dimensions of
tion) are predicted by income, and that childhood health are differentially responsive to early expe-
poverty has life-long implications for status riences. Moreover, different groups in society
attainment. Thus, like educational achievement experience distinct patterns of health; they
(Crosnoe and Benner, this volume) the study of observe, for example, that older foreign-born
poverty and its effects on the life course raises Hispanics in the US have lower rates of some
issues about the initial mechanisms through chronic conditions and mortality (compared to
which children begin to differentiate and mecha- blacks and whites), but higher rates of disability.
nisms that maintain and magnify such differ- Finally, a multifaceted assessment of health is
ences. And these issues, in turn, are considered necessary to study the distinct possibility that
Introduction: Life Course Studies Trends, Challenges, and Future Directions 15

aging is associated with a cascade or career of provides resources that, by a wide variety of
health, possibly including pre-disease pathways, mechanisms, consistently promote health. They
symptoms, disease states and chronic conditions, review the principal mechanisms for this associa-
disability, frailty, and mortality. The development tion in recent decades, but note that returns on
and use of a health portfolio, particularly involv- education may depend on years of schooling and/
ing a cascade of eventualities, is thus a highly or degree attainment, and that increased attention
promising strategy to learn more about the long should also be devoted to the relationship between
arm of childhood. education and specic symptoms and disorders,
Second, Hayward and Sheehan urge the incor- both mental and physical. This recommendation
poration of biological models into life course is clearly consistent with Hayward and Sheehans
studies of health. Very early social experiences portfolio and cascade approaches.
may (or may not) embed in the biology of the Also consistent with Hayward and Sheehan,
person, but this possibility does not obviate the the authors urge a historical perspective that
need to study patterns of experiences beyond examines age and cohort patterns, although few
childhood. Trajectories of health must be datasets currently allow for such studies. They
described and explained in terms of origins, especially urge the study of the age-as-leveler
change, and the initiation and rate of change. At and cumulative advantage hypotheses: respec-
different points in the trajectory, different experi- tively, that health disparities decrease with age,
ences may become more or less salient, and the and that they increase with age. They suggest that
balance of childhood and adult inuences may both mechanisms can be operative, and they may
change. The authors explore these complexities also vary by cohort. A second key problematic in
with the example of atherosclerosis, which the study of social status and health remains
reects diverse factors at different points in life selection and causation mechanisms: respec-
and may (or may not) lead to increasingly serious tively, that health causes status attainment, and
compromises to health. that attained status causes health. Both possibili-
Finally, the authors note that the vast prepon- ties are likely true to some degree, at least for
derance of research on the social precursors to some aspects of health, although the precise dis-
health, even among life course scholars, has not entangling of these processes remains a formida-
attended to cohort differences (see Elder and ble challenge (Moore and Brand, this volume).
George, this volume). Lack of attention to cohort The study of both age/cohort and selection/cau-
differences is surprising given the many social sation may be further enriched by framing them
factors that are both related to health and that in terms of familial generations given that health
have changed markedly in recent decades; for and education tend to reproduce across the
example, the associations between dimensions of generations.
social status (e.g., occupation) and demographic Johnson and her colleagues urge greater atten-
composition and access to health resources and tion to short- and long-term mechanisms by
medical technology. The authors discuss major which education inuences health. Studies typi-
technological developmentsbeautifully cap- cally examine short-term or long-term implica-
tured in a tablethat surely have created disease- tions of educational experiences, but rarely
specic cohort patterns. Taken together, Hayward attempt to examine connections between short-
and Sheehans recommendations for future and long-term consequences. This opportunity
researchcohort-sensitive studies that use a seems especially promising with respect to col-
portfolio-approach to health and that consider the lege attendance, which likely increases some
full temporal complexity of both health and its unhealthy behaviors in the short-term, but has
antecedentspoint the way to highly nuanced long-term positive effects on health. More
research on health disparities. broadly, the authors explain the usefulness of
Johnson, Staff, Schulenberg, and Patrick begin thinking about educational careers extending
with the observation that education is a funda- from the rst years of schooling to educational
mental cause of health, meaning that education completion and, on the other hand, health
16 M.J. Shanahan et al.

behaviors and well-being as extending from birth of this research is that childhood adversity
to death. That is, they urge a re-focusing from ranging from low SES to maltreatmentis a
educational completion and later health, to one of potentially powerful antecedent to a surprisingly
life-long, co-occurring experiences comprising wide-range of disease states in later life. Ferraro
trajectories and transitions. Indeed, some schol- urges continued study of these associations but
ars have suggested that both health and education with an emphasis, once again, on resilience: what
are endogenous to traits such as social skills and are the truly salient features of adversity in child-
conscientiousness (e.g., Ross and Mirowsky hood? And what resources promote recovery
2011; Shanahan et al. 2014). from such experiences? These questions encour-
Ferraro adopts an aging framework that age the study of early adversity and later health not
considers health in old age as resulting from life- in terms of two points in time (childhood, adult-
long experiences beginning at conception (and, hood) but rather in terms of life-long experiences.
indeed, in prior familial generations). Although a Third, Ferraro discusses aging and health in
medical perspective on health in old age has terms of familial generations, the study of which
gained considerable traction among scholars, so, provides opportunities to study the transmission
too, has a life course lens. Ideal datadiverse (or reproduction) of three key sets of variables
birth cohorts of people studied from conception from grandparents to parents to children: genetic,
to 100 years of age and beyondwill likely never socioeconomic, and behavioral (see Thornberry,
exist, and so Ferraro considers the strengths, this volume; Hagestad, this volume). How these
weaknesses, and opportunities that attend to three different types of inuences coalesce, mediate,
practical strategies for studying health and aging. and moderate one another to promote healthy
First, several datasets provide insights into the aging has scarcely been studied and represents a
lives of centenarians, suggesting, for example, daunting level of complexity. Ferraro urges that,
three health career patterns among them: escap- across these three research problematics,
ers (people who reach 100 without any major dis- researchers draw on cumulative inequality theory
ease or illness); delayers (major disease onset at (with its emphasis on longitudinal models of sta-
age 80 or later); and survivors (major disease tus and health trajectories of different groups in
onset before age 80). Although the centenarian the population), biomarkers that may cast light
studies do not use prospective control groups on the black box common to research on early
of comparable people who die before the age of experience and later health, and spatial and phys-
100, this body of research has not identied any ical contexts of aging (see also, Browning et al.,
one factor (e.g., genetics) as decisive in reaching this volume). His recurring message, however, is
this milestone. Indeed, many centenarians report the value of increased attention to sources of
lives marked by hardships (including, most dra- resilience across the life course, both for the indi-
matically, periods of starvation). This surprising vidual and social groups.
conclusion leads to Ferraros rst recommenda- Signicant cross-fertilization has occurred
tion for future research: the study of how multi- between the sociology of mental health and life
ple, diverse factorsranging from the biological course studies in the past 15 years. Avison pro-
to the socialcombine over the life course to vides an overview of this synthesis, beginning
promote long lives, especially how such factors with the seminal papers of Pearlin and his col-
promote recovery in the face of extreme leagues, who argued that stress is a multifaceted,
challenges. dynamic process that encompasses life-long
Second, prospective panel studies, involving experiences of social stressors, coping mecha-
repeated assessments of the same people over nisms and social supports, and both improve-
time, are now common; many such studies ments and declines in well-being (eustress and
include retrospective data and, increasingly, distress, respectively). Avison identies several
administrative records describing experiences dividends that have resulted from this synthesis,
prior to the initiation of the panel. A major theme including studies of the long arm of childhood
Introduction: Life Course Studies Trends, Challenges, and Future Directions 17

(see Hayward and Sheehan, this volume; Ferraro, (or discontinuous) change in mental health, and
this volume), which involve an appreciation for the identication of social experiences that can
longitudinal patterns in stressors, often reecting have such remarkable implications. Avisons
age-graded roles; for trajectories of mental essay thus provides a superb overview of this sub-
health; and, increasingly, for complex connec- eld, identifying accomplishments, limitations,
tions over time between stressors and both physi- and future directions.
cal and mental health. Two essays in this section focus on individual
Trajectories of mental health have been the sub- differences that, while having clear connections to
ject of extensive research, with a focus on inter- health, are often studied in their own right. Hitlin
connections between changing social experiences and Kwon discuss the long-standing interest in
and changing symptoms. Avison notes several agency (see also Gecas, Handbook I). They focus
limitations to this research, including limited span attention on agency as a concept that has been tra-
of age coverage in most datasets, lack of substan- ditionally measured in ways that lack content
tively meaningful change in many symptoms, and validity. They note that most studies of agency
inattention to specic disorders. Conceptually, have used survey instruments such as self-efcacy,
issues of timing are especially challenging, includ- mastery, future orientations and, less commonly,
ing the relative importance of past and present cir- planful competence; psychological cognates such
cumstances, mental health at a point-in-time as conscientiousness are discussed in terms of
versus trajectories of symptoms, careers of men- agency as well. Hitlin and Kwon review research
tal health that include rst occurrence of a disor- on these topics, showing that a considerable body
der, recovery, and later episodes, and the of knowledge has emerged on how they are
long-standing challenge of selection and causa- socially distributed by social class, gender, and
tion. Like many authors in Handbook II, Avison race, and how they are associated with common
discusses several natural experiments and urges social experiences (e.g., school performance).
researchers to search for such opportunities. They identify several avenues for future
Despite these challenges, Avison urges the research. First, they explain the value of cross-
continued study of trajectories of mental health, national and cross-cultural studies of agency. A
and he suggests several themes for future research. limited body of research suggests that age-graded
First, research in psychobiology and developmen- trajectories of constructs like mastery differ by
tal psychology points to the importance of child- political economies, although results are thus far
hood experiences of stressors and their enduring only suggestive. The study of connections
social psychological consequences. Avison urges between political economies and agency through
that studies of trajectories of mental health life is thus one avenue for future research (see
which typically focus on adolescents and adults Diewald, this volume), which is becoming
extend to include childhood exposures and increasingly possible with the proliferation of
resulting behavioral tendencies. Second, he urges sufciently harmonized cross-national data. With
continued study of the complexities of the stress respect to culture, they note that research has
universe, especially the extension of research to been heavily inuenced by the long-standing dis-
multilevels of stress (e.g., neighborhoods and tinction between collectivistic and individualistic
other spatial distinctions; see Browning et al., this societies, but that this distinction has been criti-
volume; biomarkers, the Barker hypothesis, and cized as too simplistic. There is considerable
diverse genetic mechanisms; see Shanahan et al., variability within countries (often between, for
Handbook I), the meaning of stressors for the per- example, rural and urban dwellers), and younger
son, and the surprisingly unexplored issue of cohorts are being inuenced by a global culture
cumulation (i.e., the functional form of the effects that promotes individualism. Moreover, it is
of stressors). A third challenge for researchers is unclear whether common measures of agency
the possibility of turning points that re-direct developed in the Westadequately capture the
mental health trajectories, suggesting non-linear content validity of the concept in other societies.
18 M.J. Shanahan et al.

Finally, Hitlin and Kwon bridge these con- age-related stability to how social location and
cerns for cross-national and cultural differences events in the life course may alter age-graded
with issues of timing. They note that, while good trajectories.
progress has been made (e.g., the work of Ross Growth models of cognitive development
and Mirowsky), there is much about the age- almost always examine age-graded change. In
graded nature of agency that is not well under- contrast, they propose a latent difference score
stood. Evidence suggests that constructs such as model according to which the observations are
mastery follow an age-graded, inverted U-shape organized with respect to the experiencing of an
curve, increasing through earlier adulthood and event such as retirement. This event-centered
declining in later adulthood. Yet many segments growth model can thus reveal change patterns
of the curve remain unexplained, particularly in leading up to and subsequent to the event. The
adolescence. Moreover, some evidence suggests authors expertly review extant knowledge about
that mastery may often reect notable cohort cognitive functioning across the major phases of
effects, and they thus urge the study of agency as life, and then suggest extensions, for each phase,
an age-period-cohort problem. These future that would apply their event-centered approach.
directions suggest exciting new lines of research This strategy can also be used to study the same
on connections between macro and micro phe- event (e.g., transition to retirement) for different
nomena. First, basic issues of measurement need subgroups of the population who experience it at
to be addressed to capture the content validity of different ages, although selection into groups
agency in diverse countries and cultures. Second, may limit this design somewhat.
to the extent that such measures are comparable, As Alwin and his colleagues note, there are
we need research on how age-trajectories of many widely-experienced events across the life
agency vary by political economic systems, and course that can be studied in this framework,
the factors that explain these age-related patterns. including transitions involving school, work, and
And third, attention should be directed to cohort family. Thus, broadly viewed, their event-
and period effects that shape agency both within centered model provides insight into the develop-
and between societies. mental patterns of individual differences, but also
Alwin, Thomas, and Wray consider cognitive how these patterns can be altered by the many
development from a life course perspective, transitions and events of the life course. Growth
which is an intriguing focus given the strong models that are event-centered could be applied
developmental orientation of most research on to the study of diverse outcomes; indeed, their
this topic. We close this section with their essay approach meshes well with Avisons call for
because, in addition to expertly considering cog- research on turning points in trajectories of well-
nition, their work has much broader implications being and Dannefers interest in moving away
that extend to the study of any individual differ- from the functional-developmental paradigm.
ence (e.g., personality, health, agency). They note
that while life course scholars often refer to
human development in abstract terms, they 6 Section IV: Methods for Life
rarely focus on the ways in which life course Course Research
events, transitions, and trajectories actually
impinge on and are affected by developmental Section IV begins with an essay on qualitative
outcomes over the life span. The authors note longitudinal research in life course studies by
that stability in cognitive functioning is very high Hermanowicz, who draws on his extensive expe-
but that this may reect increasing stability in riences in the study of scientic careers. As he
social circumstances, especially in adulthood. notes, long-term interview-based projects are
Echoing Dannefer et al. (this volume), a more rare, but they offer enormous potential for the
probative consideration of the malleability of description of trajectories, transitions, and the
cognitive functioning would shift focus from interpretive stances of life-long experiences. This
Introduction: Life Course Studies Trends, Challenges, and Future Directions 19

potential is clearly illustrated in his award- sures of an outcome of interestthe authors use
winning book, The Stars Are Not Enough, and he the example of body mass indexhow does the
discusses this methodology with additional analyst model level and change (i.e., trajectories)
examples from family, health, education, and in terms of social structures, transitions, turning
crime. While showing the promise of this points, accumulation, accentuation, and leveling?
approach, he also points to several avenues for Macmillan and Furstenberg work through the
methodological renement, discussing issues of essential meaning of these concepts and then
design, execution, and analysis. Hermanowicz show, in intuitive terms, how they can be tested in
shows that this methodological strategy is well- a random-effects framework. Their contribution
suited to the study of between- and within-cohort is an excellent example of developing tight con-
processes, as well as capturing processes related nections among concepts, hypotheses, measures,
to age. Handbook I included an essay on life and models. Handbook I included an excellent
course narratives by Cohler and Hostetler that companion chapter by Macmillan and Eliason
largely focused on cohort-related mechanisms. that explored trajectories as scalar variables (i.e.,
A common criticism of life course research as pathways).
indeed, of the largely non-experimental subelds Thornberry discusses the increasingly com-
of the social and behavioral sciencesis that mon possibility of conducting research that
ndings to date are largely associational or, put examines trends across three familial genera-
differently, they are lacking the internal validity tions. He proposes a set of criteria that jointly
necessary to make solid causal inferences. For dene three-generation research, as well as its
example, Adler et al. (2012) notes that although promises and limitations. The three-generation
socioeconomic status and health is among the framework becomes especially powerful with
most studied topics in the social sciences (see prospective data for each generation and compa-
Frytak, Harley, and Finch, Handbook I), very few rable measures at the same age or developmental
causal conclusions are warranted in this subarea. stage for (at least) the parent and child genera-
Section IV includes an essay by Moore and Brand tions. As Thornberry explains, the resulting data
on causal inference in life course studies. allow for the study of the intergenerational repro-
They present a series of methods (multiple duction of behaviors at the same and differing
regression, matching models, instrumental vari- ages, their precursors and consequences. He
ables, and random and xed effect models) in especially focuses on the example of parenting,
largely intuitive terms and provide illustrations exploring the many unique research questions
from life course sociology. They emphasize these that can be addressed with three-generation
methods strengths, limitations, and assumptions designs, as well as challenges, particularly sam-
as they pertain to causal inference. They also pling and measurement issues. Many examples
briey consider experiments and quasi- are drawn from the Rochester Youth Development
experiments, which will likely become more com- Study, with which Thornberry has worked
mon in the years to come. Moore and Brand then extensively.
turn to the crucial issue of heterogeneity in treat- Section IV closes with an essay by Browning,
ment effects, meaning that the independent vari- Cagney, and Boettner on another type of data that
able has an effect, but that it varies by probability is becoming increasingly common: data that
of exposure to the independent variable. They then describe place, or sociospatial contexts, espe-
illustrate the methods with simulated data that cially neighborhoods. The authors present a sub-
allow for comparisons between, on the one hand, stantively rich overview of basic life course ideas
the true counterfactual effects, and estimates of and their application to neighborhood and place
causal effects from the methods that they present. effects. They distinguish between activity
Macmillan and Furstenberg present a unique spacesthe places of everyday, routine activi-
discussion of growth curve models that details tiesand ecological networks, which refer to the
how many core ideas of life course sociology can aggregation of activity spaces into larger mean-
be tested in this framework. Given repeated mea- ingful spaces such as cities.
20 M.J. Shanahan et al.

Browning and his colleagues discuss the larger conceptual framework within which to
promise of and current limitations to life course conduct research and to link ndings with policy.
studies of place: mechanisms (e.g., collective He then provides a concise overview of his work
efcacy, organizational resources, and social and with the Glueck data (with Robert Sampson),
physical disorder); differential exposure to set- which resulted in numerous important ndings
tings, including what the authors term eco- about desistance from criminal behaviors. Laub
networks that describe people and places and draws on his experiences at NIJ to reveal the
place trajectories; differential reception (or potential and challenges associated with applying
experiencing) of contexts by their inhabitants science to policy, including the inuence of non-
based on, for example, age, gender, and race/eth- scientic considerations such as values and poli-
nicity; concerns about dening (or bounding) tics. He advocates for research-practitioner
spaces such as neighborhoods; selection and cau- partnerships and illustrates this concept with his
sation; the use of new technologies to capture experiences with NIJs 4-year project with the
more detailed data; theoretical developments that Harvards Kennedy School of Government on the
help explain how multiple contexts simultane- Executive Session on Policing and Public Safety.
ously affect the person; and the extension of stud- Such a venture is rare, but the project was suc-
ies from their traditional purviews of children cessful and may serve as a model for future
and adolescents. By juxtaposing extant research efforts to bridge basic science and policy.
with these many challenges, the authors clearly Wadsworth and Kuh draw on a lifetime of
show that the study of place in the life course research in life course epidemiology to outline
holds exciting intellectual opportunities, both the basic approach and show its long-standing
substantively and methodologically. concerns with applied problems of public health,
ranging from infectious diseases to childhood
allergies. Indeed, their concise historical review
7 Section V: The Life Course shows how life course epidemiologyperhaps
and Policy, Building unique among the many subelds of life course
the Nexus researchhas always been driven by the
requirements of the medical sciences, policy
Handbook II concludes by returning to a point thinkers, and the social sciences. Paralleling
made in the opening chapter by John Bynner: Laubs argument, Wadsworth and Kuh argue that
given the very large investments that societies are life course epidemiology is an excellent concep-
making in the collection of longitudinal data, life tual and methodological framework that provides
course studies must convey their relevance to a context to both research and policy.
policy and interventions. The contributions in The life course epidemiology framework
Section V address many vexing social problems encompasses several themes that are especially
that are both the subject of life course research relevant to the application of basic science to the
and public policy: crime, health, poverty, cross- improvement of population health: a life-long
national differences in work and family, youth in view of the persons health extending from prior
the developing world, and implications of large- generations to old age, changing attributes of pop-
scale demographic and economic forces on the ulations, stress processes, and the use of big data
life course. that encompasses diverse forms of sophisticated
Laub begins Section V from a highly unique data to describe populations. Wadsworth and Kuh
vantage point: as a leading scholar in criminol- note that the bridging of science and policy is
ogy and as a former Director of the National becoming increasingly promising and challenging
Institute of Justice (NIJ) in the Ofce of Justice with advancements in technology, particularly
Programs in the Department of Justice during the involving the use of social and biological data.
Obama administration. Laub explains how a life Echoing Bynner, they urge increased international
course perspective is ideally suited to providing a cooperation in study design and measurement.
Introduction: Life Course Studies Trends, Challenges, and Future Directions 21

Herds starting point is that research on the differences and experiences in the family, result-
effects of policies and life course processes ing in embodied behaviors and decision-
often, unfortunately, do not intersect. She illus- making biases, which then alter the likelihood of
trates this point by considering research on the risk and adversity.
effects of social welfare policies in the United These developmental processes are compli-
States on health, and the largely distinct litera- cated by the fact that this basic templatebio-
ture on socioeconomic status and health across logical differences, experiences in the
the life course. She argues that the two litera- family-of-origin, and behavioral and
tures can inform one another, thus creating an decision-making tendenciesdiffers greatly
incentive for closer collaboration. On the one among groups in a society (men and women,
hand, life course research has paid little attention migrants and native-born, low and high status,
to social welfare policies, despite the fact that etc.). Diewald proposes that these developmental
social welfare spending consumes 60 % of both processes should be the true focus of policy-mak-
federal and state budgets. On the other hand, ers. And scholars who study the intersection
policy research has tended to focus on the short- between policy and the development of risk and
term implications of policies for health and also disadvantage should maintain three distinctions:
on health during the early life course, with rela- the extent to which the state prevents risk and/or
tive inattention to later life. Herd expertly weaves disadvantage; the types of risk and disadvantages
these two bodies of research to consider a wide that are especially worthy of public investments;
range of specic social welfare spending pro- and how differing groups in society benet, more
grams, including childcare, Medicare and and less, from policies. He illustrates his model
Medicaid, and income support programs such as and these distinctions with many examples from
Social Security. Highlighting the possibilities for Europe. A major challenge for the study of policy
cross-fertilization between life course and policy and the life course, however, derives from the
research, she identies several areas for future relative strengths and limitations of focusing on
research, including the long-term implications of one policy between countries (or states) and pol-
welfare programs directed at children, the impor- icy regimes.
tance of linked lives in the study of welfare pro- Dornan provides an overview of the highly
grams, the creation of quasi-experiments with unique Young Lives study, which brings a life
the introduction of new policies and differences course perspective to the study of youth in mid-
among states, and the role of policies in re- dle- and low-income countries, Ethiopia, India
directing transitions, trajectories, and pathways (Andhra Pradesh), Peru and Vietnam. He shows
in peoples lives. that, across these countries, rurality, low socio-
Diewald draws on a central concept in con- economic status, and ethnicity (minority status)
temporary life course studies, risk, to develop a reduce access to basic services, which explains
conceptual model of how welfare policies differ much of the early emergence of inequalities.
cross-nationally (see also, three chapters by Further, access to and the structure of schools can
Leisering, Weymann, and ORand, Handbook I). mitigate or reinforce these emerging differences.
He notes that societies differ greatly in how they Dornan highlights trajectories of stunting, a key
seek to manage risk (the probability of a hazard- indicator of physical development that is, in turn,
ous event occurring) and adversity (disadvan- associated with cognitive development. As he
tages resulting from risk). This distinction notes, the life course framework used by Young
underscores the need to understand precursors Lives allowed for the discovery of patterns that
and consequences of risk, as well as the circum- ran contrary to common wisdom: the earliest
stances in which risk actually results in adversity. period of life is certainly important to stunting
Diewald proposes a life course risk analysis trajectories but later childhood can be a turn-
according to which risk reects multilevel devel- around phase and, in any event, experiences in
opmental processes that include early biological later childhood are needed to maintain normative
22 M.J. Shanahan et al.

growth. Research also reveals the high degree of tions that are interdependent and cumulative in
associations among domains of development, their impact with age. The key denitional shift
physical, cognitive, and psychosocial. is from the traditional emphasis on sequences of
In contrast to US policy-making (as observed roles to continuous processes involving, princi-
by Laub and Herd), Dornan notes that many pally, selection and causation, cumulation,
international organizations draw on a life course agency in context, and inequality. This shift, they
framework, including the World Health argue, is necessary because the phase-based
Organization, the United Nations Development conception has become too restrictive in the wake
Program, and UNICEF. Consistent with Laubs of rapid change and myriad forms of inequality
and Herds chapters, he notes that policy needs to both between and within societies (see Mortimer
have a rmer basis in evidence. He further adds and Moen, this volume).
that this is especially true in middle- and low- Second, they urge the study of new forms of
income countries, the policies of which are often literacy that guide decision-making in the face of
inuenced by tradition, ideologies, and politics. new forms of risk, especially nancial and health
Finally, consistent with Laubs observations, literacy. This theme complements nicely the con-
Dornan urges that life course researchers involve ceptual model proposed by Diewald, which
policy-makers at all stages of their research, emphasized the central-but-neglected role of
extending from problem formulation to interpre- decision-making in the life course. And third,
tation of results and their implications, and com- reecting the mass movement of peoples across
munication of science to all of societys borders, the authors urge the study of migration
stakeholders. from a life course perspective (See DeWaard, this
Finally, ORand and Bostic argue that life volume; Jasso, Handbook I). Such research
course studies should be placed explicitly in the would attend to origin and host countries, cohorts
broad context of global change. They begin with and aging, leavers, stayers, and returners, and
Rileys classic formulation of structural lag, that relationships among groups in the host society.
social institutions, organizational arrangements Despite our best efforts, we are aware that our
and cultural stereotypes [lag] behind demo- summaries and Table 1 fall short of conveying the
graphic change. ORand and Bostic argue that richness of insight offered by the many chapters
the original structural lags that were identied by in Handbook II. We thus urge readers to discover
Riley (e.g., the inadequacies of policies to reect this richness, in the context of their own research
gender differences in the life course) remain and interests, through careful study of these essays.
that new lags have been introduced with the rapid In closing, we express our gratitude to the
pace of social change, including demographic many people who have contributed to this project
changes (especially involving aging) and eco- since its inception. First and foremost, we express
nomic globalization. As they observe, these our enormous appreciation to the authors of the
large-scale trends are having on-the-ground chapters. The focus on future directions is an
impact on institutional arrangements that affect especially challenging intellectual endeavor, but
educational opportunity, job security, health care one that pays off by inuencing the eld, espe-
eligibility and access, and individual and family cially young scholars who are formulating their
well-being across the life span. unique syntheses of the life course paradigm and
ORand and Bostic suggest three themes that identifying strategic areas of study. All of the
facilitate the study of the life course against the contributors expertly draw on their substantial
background of rapid, large-scale changes. First, research experiences and knowledge, and we
they propose a re-denition of the life course believe that their impressive efforts will indeed
itself, shifting from a sequence of age-graded be inuential. We also thank Esther Otten,
roles to a manifold cumulative phenomenon Publishing Editor at Springer, who has been most
consisting of intertwining processes associated encouraging, and Hendrikje Tuerlings, also at
with human development from birth to death, that Springer, who provided superb support through
include biological, cognitive, and social transi- the development of this project. Finally, our
Introduction: Life Course Studies Trends, Challenges, and Future Directions 23

heartfelt appreciation to our family members Meyer, J. (1987). Self and the life course. In G. M.
Thomas, J. W. Meyer, F. O. Ramirez, & J. Boli (Eds.),
Lilly, Marie, Liam (Shanahan), Kent, Diane,
Institutional structure: Constituting state, society,
Eileen (Mortimer), Jeff, Eben, Angelica, Liana and the individual (pp. 242260). Beverly Hills:
(Broadbent), Corey, Clay, and Quinn (Johnson) Sage.
our necessary and sufcient conditions. Mortimer, Jeylan T & Michael J. Shanahan (Eds). 2003.
Handbook of the Life Course. New York: Springer.
Ross, C. E., & Mirowsky, J. (2011). The interaction of
personal and parental education on health. Social
References Science and Medicine, 72, 591599.
Shanahan, M. J., & Freeman, J. (2012). Bibliographic
Adler, N., Bush, N. R., & Pantell, M. S. (2012). Rigor, trends in life course research. Paper presented at the
vigor, and the study of health disparities. Proceedings 2012 Annual Meeting of the American Sociological
of the National Academy of Sciences, 109(S2), Association, Denver, CO.
1715417159. Shanahan, M. J., Hill, P. L., Roberts, B. W., Eccles, J., &
Hayward, M. D., & Gorman, B. (2004). The long arm of Friedman, H. S. (2014). Conscientiousness, health,
childhood: The inuence of early life conditions on and aging: The life course of personality model.
mens mortality. Demography, 41, 87107. Developmental Psychology, 50(5), 14071425.
House, J. S. (1977). The three faces of social psychology.
Sociometry, 40, 161177.
Part I
Foundations of Life Course Studies
and Future Research
Institutionalization of Life Course
Studies

John Bynner

The institutionalisation of life course study interest. The consequence was the transformation
reects the culmination of a number of develop- of longitudinal study from a research method
ments that were central to its establishment as a in the social and behavioural scientists tool kit
eld of scientic inquiry in the form that it takes into a massive generator of research resources
today. The key component of this transformation demanding substantial scientic investment. Such
are the longitudinal data that supply the empirical investment resulted in the large and complex data
basis for life course research. Hence, the devel- sets on which the major life course research pro-
opment and institutionalisation of life course grams are based. The nal sections of the chapter
study reects the massive growth of longitudinal document the evolution of institutions and infra-
inquiry. The distinguishing feature of such work structure as a response to the need for data
is the need for continuity to sustain the study quality, communication and collaboration. In
through the good times of high investment and return, there has been a growing obligation on
high scientic returns and the more difcult researchers to ensure that data are accessible and
times when funding runs out between surveys. It ndings are communicated in comprehensible
is during the latter periods, in the absence of any form and as widely as possible to those in the
protective infrastructure, when the survival of scientic and policy communities who need
the study requires the strong leadership and dedi- them. A nal section addresses future prospects
cated commitment of staff, often at personal costs and challenges ending with some thoughts about
to themselves, for the study to survive (Bynner next steps.
and Goldstein 1998).
In this chapter I chart such progress in the
form of a brief historical overview of the devel- 1 Origins and Development
opment of longitudinal survey research, then
move to the interrelated technological, method- Life course study is typically described as a
ological, and theoretical changes that led to the perspective or approach rather than a scien-
growth of scientic and especially, government tic theory in the conventional sense of linked
hypotheses deduced from postulates tested by
empirical evidence (Mortimer and Shanahan
J. Bynner (*) 2003; Elder 1998; Elder and Shanahan 2006;
Centre for Longitudinal Studies, Heinz and Marshall 2003; Heinz et al. 2009).
University College London Institute of Education, Although originating in postwar sociology, the
Bedford Way, London WC1H0AL, UK full realisation of the perspective has come to be
e-mail: jbynner@slls.org.uk

Springer International Publishing Switzerland 2016 27


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_2
28 J. Bynner

seen as lying in the interdisciplinary understand- Accordingly, the early longitudinal studies tended
ing of human development gained from synthesis to be focused on specic topics in a particular
of ndings across a wide range of inquiry. The scientic discipline and restricted data collection
theoretical contributions -of particularly sociology, to a particular phase of the life course, especially
psychology, biology, criminology, demography, the early years. The maximum period to completion
epidemiology, geography and history, thus all was often no longer than 15 years. In contrast,
had a part to play. todays studies, in design and realisation, are typi-
Such holistic understanding, described by cally much broader in their attempt to bridge life
Lerner (2006) as developmental science, has in course domains, traverse all life course stages,
fact been recognised as the goal of the study of and to engage the concerns of many disciplines.
human development for some time. Paul Baltes In his historical overview of longitudinal
draws attention to the recognition of the value of research, Sontag (1971) reports the blossoming
the longitudinal orientation as early as the of US longitudinal studies through the early part
eighteenth century in the writings of the German of the twentieth century taking its stimulus from
philosopher Nikolaus Tettens. With an uncanny the series of meetings held in the mid-1920s
signalling of current life course preoccupations organised by the National Research Council and
Tettens also urged that such study should take involving such inuential social and behavioural
place from cradle to grave and be holistic in scientists as Margaret Mead, LL Thurston, Clark
conception, thus anticipating the need for multi- Wissler, Robert Yerkes, Richard Scammon, and
disciplinary approaches across the whole life Arnold Gesell. The outcome was the establishment
span (Baltes 1987; Baltes et al. 2006). Tettens of the Society for Research in Child Development
was followed in the nineteenth century by such and the journals Child Development and Child
writers as the Belgian polymath Adolphe Quetelet Development Monographs.
who, as Baltes (1995) reports, promoted the idea Leading institutes in the development of the
of social physics and similarly life span studies. eld included the Institute of Human Development
In A Treatise on Man and the Development of at Berkeley, the Brush Foundation at Cleveland,
his Faculties (1842), he also argued for cohort the Child Research Council of Denver, the Harvard
comparisons to separate age from historical Growth Study, the Fels Research Institute, the
(period) effects. Minnesota Child Development Institute and the
Iowa Child Welfare Station. The starting point
and mission was the need for multi-disciplinary,
1.1 Early Studies if not interdisciplinary, conceptualisation, but
this strategy was not maintained. Perhaps because
The feasibility of conducting such longitudinal of the early interest in child development, child
studies and realising their scientic potential was psychology soon took over as the dominant disci-
not a serious prospect until very much later. The pline to supply the main, if not only, theoretical
designer of a life course study that is to meet means through which human development was to
the ontogenetic requirement of explanations of be understood.
development based in human biology interacting The studies undertaken also suffered from all
with external circumstances across the whole the difculties of conducting such long-term
life span (Magnusson 1993) has a number of work at the time, and many such studies comprised
challenges to meet. Not least of these is that the no more than a leading Professor with support
mortality of those who start the study off means staff, most of whose time and energy was devoted
they will not be around to gain the scientic to data collection. Sontag (1971) sees the prob-
benet of its later stages, let alone its completion! lem as one of inadequately thought out scientic
The founder is therefore in effect bestowing a programs lacking hypotheses. There was a tendency
legacy on those that follow him or her pointing of staff to focus their efforts on data collection
to the need for infrastructure to ensure continuity. as a worthwhile end in itself while failing to
Institutionalization of Life Course Studies 29

see the need to exploit the data to justify their developmental systems approach to achieve the
collection in the rst place. integrative goals to which, in his terms, develop-
As Baltes et al. (2006) point out, the dening mental science should be directed.
feature of these longitudinal studies involved Of the few early studies that did survive long
separation between the separate stages of term, usually through the efforts of the dedicated
development childhood, adolescence, adulthood, individuals who took over responsibility for
old age rather than in formulating a holistic, them or through recognition from their archived
and in his terms, ontogenetic account embracing information of their scientic potential for life
all of them. The change came later with growing course studies of aging, some key studies became
interest in aging and old age to which psycholo- corner-stones of life course research (Elder 1974;
gist Bernice Neugartens Development and Aging Elder et al. 2003; Elder and Shanahan 2006).
program at the University of Chicago was a major Notable examples are the FELS longitudinal
contributor (Neugarten 1996). Matilda White study starting in the 1920s that followed all
Riley (1988) was another. From her position in children in Ohio and continues to this day; the
charge of Social Science Research in the US Oakland and Berkeley Growth and Guidance
National Institute of Aging, Riley promoted studies starting in the early and late 1920s respec-
longitudinal studies and the concept of aging as a tively; Termans 1921 study in Stanford California,
social construction, thereby opening the door to with birth years going back to 1900 that followed
interdisciplinary approaches. But In practice, as children in the top quartile range of scores on the
Baltes and colleagues point out, dialogue and Terman-Merrill IQ test (Sears 1977).
certainly integration between the two areas of The latter studys focus on the talented, as
child study and aging research was virtually opposed to low ability and disadvantaged poor
non-existent. performers, is of itself interesting in reecting the
There was therefore from the beginning, in preconceptions and interests in understanding tal-
effect, a closing off of interest in development ent and success, rather than failure that prevailed.
across the whole life span. Hence, for some time, This was at a time when genetic explanations of
there was neglect of the study of the long-term cognitive ability were dominant and eugenics in
consequences for psychological adjustment and social policy formation was a prominent concept.
physical health of early experience and circum- The focus on cognitive function and its possible
stances. Notable exceptions were: John Bowlbys decline continued to be reected in other long
(1952) World Health Organisation inquiry into term studies such as the Scottish Mental Surveys
the effects on adult mental health of war-time of children, starting in 1931 and 1947, that also
childrens separation from parents; Emmy Werners used the Terman-Merril test. The sample mem-
1950 Kauai birth cohort study (Werner 2004) bers were traced as recently as the 1990s for
focusing on long-term resilience in the face of follow-up studies of cognitive aging (Brett and
adverse child circumstances; and David Barkers Deary 2014). The priorities of the longstanding
(1992) studies that traced the origins of adult Malmo study in Sweden starting in 1937 (Husen
illness such as coronary heart disease back to et al. 1969; Furu 2000) were similarly directed
early fetal experience following conception. In towards the factors underlying the developing
fact as Elder and Shanahan (2006) note, the pos- intelligence of the followed children.
sibility of a life course perspective as signalled, The common feature of all these pre-war studies
for example, in Thomas and Znanieckis (1918 was their longevity and the collection of an
1920) classic text, The Polish Peasant in Europe ever-widening range of age-graded measures as
and America, 19181920 was largely ignored the samples matured. The ndings mainly from
until 40 years later. Lerner (2006) makes a similar secondary analysis of archived data supplied,
point in referring to the glacial speed with which accordingly, the basis for the break-through by
theoretical innovation in human development which John Clausen (1972, 1986), Director of the
usually moves, arguing for his interdisciplinary Institute of Human Development at Berkeley, and
30 J. Bynner

sociologist Glen Elder (1974, 1985), followed NSHD was followed in 1958 and 1970 by the
by others such as social historian John Modell rst of the series of British birth cohort surveys,
and sociologist Frank Furstenberg (Modell et al. starting at 12-year intervals. In each case the
1976; Modell 1989), developed the life course sample comprised this time all 17,000 births in a
perspective. single week in England, Wales and Scotland.
Following the tracing of study members in the
initial perinatal mortality survey of mothers and
1.2 Expansion children, for the purposes of a Government com-
mission on primary schooling, and through the
Following the second world-war, the focus of efforts of Neville Butler and Mia Kellmer Pringle,
longitudinal inquiry still remained xed on rela- the National Child Development Study (NCDS)
tively small, usually area-based inquiries, with became a longitudinal survey in 1974. The Child
only the beginnings of the much later shift to Health and Education Study (CHES) later to
large scale national epidemiological surveys such become the 1970 British Cohort Study (BCS70)
as the Norwegian and Danish birth cohort studies followed in 1970 when it was launched by Neville
(e.g. Olsen et al. 2001; Martin et al. 2006). Hence Butler as a longitudinal survey from birth. To
a central focus was the conditions of birth and continue the 12 year interval series, a new survey
their consequences for physical growth and later was expected to begin in 1992, but because of
health outcomes within the relatively restricted opposition from clinical research interests, the
frame of exposures and survival. The famous proposal for one was rejected. The grounds for
Framingham study, devoted to the aetiology of rejection was that little was to be learned from
heart disease, was one of these starting in 1947, the new survey, not already known from the ear-
and it is still collecting data. The origins of life lier ones; it was therefore not worth the cost a
course epidemiology can be found in the ndings notable rejection of the life course perspec-
of these early studies (Kuh and Ben Schlomo tive! The closest to the missing cohort (Martin
1997; Pickles et al. 2007; Halfon et al. 2014). The et al. 2006) was the Bristol area-based Avon
1955 Kauai longitudinal survey, extending from Longitudinal Survey of Parents and Children
birth to the late 40s, as the rst to identify predictors (ALSPAC) led by Jean Golding, which started in
of escape from disadvantage and adult mental 19911992 based on pregnancies rather than
health (Werner 2004), was also seminal in the births (Golding et al. 2009).
development of life course thinking. The series was revived in 2000 with the
government-funded Millennium Cohort Study
1.2.1 British Studies led by Heather Joshi based on 19,000 births, this
This period also witnessed the beginning of the time spread over a whole year in 400 electoral
British population-wide series of birth cohort wards rather than concentrated in a single week.
studies that started with post-war concerns There were sample boosts in Scotland, Wales and
about perinatal mortality (Williams 1997) before Northern Ireland and in wards of high economic
broadening out to a wider life course conception deprivation. In consequence differential weight-
(Bynner and Joshi 2007). The National Survey of ing was needed to achieve population representa-
Health and Development (NSHD), began by tion in analysis based on the survey. Data collection
James Douglas in 1946 (Douglas 1964), was was also extended to fathers as well as mothers
based on a 5,000 strong national sample taken for the rst time. The survey following, led by
from the 15,000 babies born in a single week in Carol Dezateux, is the Life Study due to begin in
England and Wales. And in the tradition of pre- 2015 is based on 100,000 rst trimester pregnan-
war inquiry the survey was weighted towards the cies in a design combining a number of area-
more able children through over-sampling the based studies with a national probability sample of
more educated families (Wadsworth (1991). 20,000 (Bynner et al. 2008; Dezateux 2014).
Institutionalization of Life Course Studies 31

1.2.2 Swedish Studies rst that permission should be sought from every
Building on the precedent of the 1937 Malmo individual cohort member to use any piece of
study, a number of longitudinal studies of cognitive register data about them and nally in 1986 that
and behavioural development began in Sweden the data tape containing personal identiers
(reviewed in Janson 2000), including: Evaluation should be destroyed. Subsequently the study
Through Follow-up, based on birth cohorts born relied on secondary analysis of what had been
in 1948 and 1951 that were rst contacted in extracted in anonymized form earlier.
1961 and 1966; the Individual Development and
Adaptation programme, starting in 19641965 1.2.3 Australasian and Canadian
and based on age cohorts starting at ages 10, 13 Studies
and 15 respectively; and Project Metropolitan Other countries contributed to the general expan-
based on a cohort of boys born in 1953; girls sion of longitudinal studies through the 1970s,
were added to the sample later. including the highly productive New Zealand
These longitudinal studies are distinctive in birth cohort studies begun in Dunedin (Silva and
capitalising on the rich stocks of register data Stanton 1996), and in Christchurch in 1977
collected in Nordic countries from birth onwards (Fergusson et al. 1989). The Dunedin study
across all life course domains for every individual beneted from major inputs from leading US
citizen. Project Metropolitan took advantage of developmental scientists including Avshalom
such resources comparatively, by means of what Caspi and Terrie Moftt, attracted to fellowships
were intended to be three parallel longitudinal offered by the director, Phil Silva, to analyse
surveys carried out on cohorts of boys born in the data. Moftt subsequently became deputy
1953 and living in 1963 in the three Nordic capital director. A later child study was established in
cities, Stockholm, Oslo and Copenhagen. A pro- Canada, the 19941995 National Longitudinal
jected Finnish study never materialized. The main Survey of Children, based on following up a child
foci of the project were social mobility, member- panel with an age range from birth to age 11. Canada
ship of groups and associations and positive and was also the rst country to undertake a continuing
negative deviance the last capturing positive longitudinal survey of immigrants.
deviation in which the individual does more than
is expected, e.g. becomes a prize winning athlete, 1.2.4 US Studies
a creative artist, or a moral guide to set against US studies also blossomed during this period but
negative deviation reected in a physical and this time typically with large national probability
mental disease, crime and delinquency, alcohol or samples and a move away from single starting
drug addiction, social isolation etc. (Janson birth dates to age-graded panels starting at
2000, p. 141). First contact with cohort members different life course stages ranging from early
took place in 1966 when the sample had reached childhood through adolescence to old age. These
age 13. Prior to that all data utilised were from the were located in a number of major university
registers. research centers including, Michigan (Institute
However, the linkage of the register data, for Social Research), Berkeley (Institute for
especially criminal records, to the personal data Human Development), Minnesota (Institute for
collected from cohort members, is where Project Human Development and also the Life Course
Metropolitan came undone. Adverse press reports Center, which houses the Youth Development
led to its termination rst in Oslo then in Study Mortimer, 2012), Chapel Hill, North
Copenhagen, where the restrictive permissions Carolina (Carolina Population Center), Wisconsin
procedures imposed by the authorities made the (Wisconsin Center for Demography and Ecology,
study no longer scientically and operationally which houses the Wisconsin Longitudinal Study
viable. In Stockholm, the 1966 survey and one Sewell et al. 2001) and Chicago, National Opinion
follow-up in 1968 went ahead but only on the Research Center e.g. Department of Labors
proviso from the Swedish Data Inspection Board, National Longitudinal Studies.
32 J. Bynner

The focus of these studies tended to be on starting at ages 4, 10, 14 and 19 in between. With
transitions at different ages such as school entry a rich range of follow-up data on circumstances,
and the move from education to employment. educational progress, and aspirations collected
The basic design varied from the single age from sample members, their parents and their
cohort, as in the ISR Michigan-run Youth in teachers, at frequent intervals, the study is in an
Transition Study (Bachman 1972) through quasi exceptionally strong position to supply evidence
sequential cohorts (age-graded panels) as in the to policy makers on the operations of every element
National Longitudinal Surveys of Youth (NLSY79, of the education system. This is as well as under-
NLSY97) conducted by Ohio State and NORC taking more long term investigation of educa-
University of Chicago and the rotating panel high tional dynamics at very age and stage through
school entrants of ISRs Monitoring the Future with individuals are passing (Leuze et al. 2011).
(Bachman 1986). The North Carolina Population Another different but complementary line of
Centers National Longitudinal Study of life course study emerged in Germany. The
Adolescent Health (Add Health) begun in 1994 Status Passages and the Life Course centre in
and also an age cohort design is a model exempli- Bremen led by Walter Heinz eschewed reliance
cation of the life course perspective embracing exclusively on quantitative longitudinal data in
biomarkers, including DNA, as well as a wide favour of a combination of survey and social
range of social, economic, psychological, and bio- biographical methods to elucidate the processes
medical measurement (Harris et al. 2009). A later of transition from education to the labour market
variant starting in 2001, the 5,000 strong Fragile (Heinz 1991, 1996). The ten books the centre
Families study, is based on samples of marital produced based on annual workshops that
and non-marital families, living in the hospital the centre ran include some of the classics in
catchment areas of 75 hospitals in 20 US cities this eld, e.g. Theoretical Advances in Life
(Reichman et al. 2001). Course Research Heinz (1991) and Society and
A major advance was made in the longitudinal Biography (Weyman and Heinz 1996).
study of aging with the US Health and Retirement
Study (HRS) begun in 1988 and again based at 1.2.6 Household Panel Surveys
the ISR (Michigan) (Juster and Suzman 1995). A long standing development paralleling the birth
The study provided the template, rst for the and age cohort studies was the multi-age-group
English Longitudinal Study of Aging (ELSA) and household panel study with annual or biennial
subsequently, comparable studies in a number of data collection initiated by the ISR (Michigan)
European countries and across the world e.g. Panel Study of Income Dynamics (PSID)
India, and Japan. All of these studies follow (McGonagle et al. 2012). Launched in 1968,
broadly the same strategy of inquiry but with PSID was followed by comparable panel surveys
some national variation reecting the different in Europe and internationally ranging from
policy contexts. Sample members are recruited in Russia to Mexico and Taiwan. The German
this case from cross-sectional time series surveys Socioeconomic Panel (SOEP) was the first
of health in the different countries. The strategy replication in 1984 (Wagner et al. 2006 ),
involves identifying sample members over the followed in 1991 by the British Household Panel
age of 50 from which the aging sample for Survey (BHPS) (Now named Understanding
follow-up is constructed. Society) (Buck and McFall 2012). The Canadian
Survey of Labour and Income Dynamics (SLID)
1.2.5 German Studies and the Australian Household Income and Labour
The National Educational Panel Study (NEPS) Dynamics (HILDA) survey are further notable
in Germany combines elements of all types of examples.
design in following since 2010 a series of six age Although economic questions drive the studies,
cohorts totalling over 16,000 individuals, the rst over time, as the multiple cohorts the panels
starting with newborns and the last, a pre-existing contain get older, they have increasingly become
panel of 2364 year-olds, and other cohorts multi-purpose and multi-module life course
Institutionalization of Life Course Studies 33

studies. Modules covering different areas of this one reveals the variety of sample designs while
content are rotated at time intervals through the also identifying the shared principles that have
survey of which biomarkers are among the most shaped the modern form of longitudinal research
recent examples. The main attraction of the studies methodology and commanding large scale research
is their coverage of the whole 16+ population resources. As we shall see, there has also been
over all ages and repeated measurement over some convergence within the eld of life course
1 year, or at the most 2 year, intervals. Their main study on the theoretical direction the effective use
limitation is the relatively small size of any given of these resources should take. Table 1 lists the
birth (or later age) cohort, even when expanded main surveys mentioned in this section organised
from the initial 5,000 to 40,000 households, as is by world regions. These are the prime examples
now the UK case. As a result of emigration and of a method developed often for small scale sci-
immigration, there is also loss of population entic inquiry cottage industry to a major
representativeness and sample size over time, tool that, in the longer term, serves both policy
necessitating periodic sample boosts. The precise and scientic need on an industrial scale. The
specication of what constitutes a household table paves the way for the next section of the
longitudinally also remains a problematic feature. chapter, which considers the innovations in tech-
Hence most long-term longitudinal analysis is nology, methodology, and theory that supplied
directed at individual household members as the the foundations of the changes through which the
unit of analysis rather than households. modern form of life course study came about.

1.2.7 Census- Based Studies


Finally another form of longitudinal survey does 2 Foundations of Growth
not involve the collection of new data at all, relying
instead on administrative record linkage to supply So what supplied the opportunity for the massive
the dataset. As noted earlier, Scandinavian studies lift-off from cottage industry to industrial scale
have always relied on the extensive Scandinavian production of longitudinal data? To understand
registers to supply the data in birth cohort studies the origins and institutionalisation of contem-
up to the fourth grade before direct survey mea- porary life course study, it is important to appre-
surement of the children takes over. A variant ciate the developments that lay behind it on a
conducted in a number of countries is to base the number of fronts: technological, methodological
whole study on a representative proportion of and theoretical.
participants in the national census. Vital statistics
registration data for such events as births and
deaths, and increasingly other administrative 2.1 Technological
data, can then be linked between censuses as the
relevant events occur. The major benet of such In the absence of the modern statistical software
studies is their huge potential size at relatively and massive storage now available on any research-
low cost with continuous augmenting and renewal ers desktop, constructing, maintaining, and
from data routinely collected for administrative updating a longitudinal database is very labour-
purposes. Their weakness, from the life course intensive and liable to error, as is analysing the
perspective, is the severe limitations on the range data: daunting as one informant told me in a
of content covered. study of the potential for secondary use of the
British National Child Development Study data
(Bynner 1984). In the early British cohort studies,
1.3 Review extending to the end of the 1970s, for example,
Holerith cards were the only means of extracting
This brief review of longitudinal studies develop- survey data for frequency and crosstab analysis
ment across the last century and the beginning of to be replaced by punched cards for supplying
34 J. Bynner

data rst to card sorters and then to main frame research literature replacing hard copy library
computers, where every minute of computer time holdings. A full listing can be found of the hundreds
used was charged for. Remote terminal access to of such archives that now exist across the world on
main frames was the rst step towards the digital http://www.sociosite.net/databases.phptop.
world, and desk top computers supported by But here another part of the story needs to be
off-the-shelf software. acknowledged. Traditionally the kinds of studies
Such enhancement of IT capability also under- launched in the early stages of birth cohort study
pins another development that was critical to development were concerned with newborns
later research investment: electronic archiving and were based on the clinical model of a scientist
and the access of the data for the burgeoning collecting data for a personal program of research.
secondary analysis that then became feasible The study would be set up on that basis with an
(Brooks-Gunn et al. 1991). Prior to electronic explicit agreement with the parents of the babies
archiving, such facilities as the Murray Archive or children involved that the data collected were
in the Radcliffe Institute of Advanced Studies at condential to the study personnel and only they
Harvard led the way in supplying a repository for had access to the data for research use.
the raw data of longitudinal studies. The archives As the interest shifted more towards multidis-
subsequent absorption into the Institute of ciplinary inquiry, and expansion to incorporate
Quantitative Social Science signalled completion direct government policy funding as well as sci-
of the electronic journey on which all data ence (public or research foundation) funding, the
archiving was going to go. alternative model developed as manifested in the
Other ventures that followed located longitu- term investment. That is to say, the deployment of
dinal date in archives to meet research resource large scale funding was used to support the pro-
needs across the whole spectrum of social and duction of a research resource accessible to all
behavioural science interests. Thus in the UK, for scientists with a proposed program of research
example, the Economic and Social Research based on longitudinal data. In such a scenario,
Councils Data Archive houses the Economic and little was demanded of the researcher accessing
Data service in the Universities of Essex and the anonymised archive data other than a formal
Manchester. The GESIS Leibnitz Institute for the request and, once released, the data set was in
Social Sciences based in Mannheim, Germany, practice usable for any research purpose. The only
performs the similar task for German data. Most proviso was that acknowledgment of the source
of, if not all, the US centers for research in human would be given in any publications based on
development where the longitudinal research research using the data.
enterprise was founded developed archiving However, the price to be paid for such a
facilities for housing their own and sometimes resource model lies more now in the demands
others data (e.g., Michigan, North Carolina placed on the researcher in formulating the
(Chapel Hill), Chicago, Ohio State, and proposal for funding and the expected output
Minnesota). Minnesotas Population Center, for from it. There has been growing pressure for
example, has a particularly long record in this accountability by demonstrating direct impact
area, housing the Integrated Public Use Micro of the ndings produced. In the case of the UK
Data Series (IPUMS-USA) that has distributed Economic and Social Research Council (ESRC),
harmonized micro data for the US Census and this takes the form of a section of the funding
American Community surveys from 1850. proposal being devoted to pathways to impact
Another development spreading even more and the requirement for a separate, post-end of
widely was that of the repository for the outputs of award, impact report. The other major change is
inquiry, research papers and monographs, in digital in the much more rigorous requirements placed
form. This facility is coupled with the massive on users for access to the data even in anonymised
development, via libraries, of online access to form.
Institutionalization of Life Course Studies 35

Although health science is not a major focus especially when heralded by adverse press
of this chapter, the origins of so many of the reports. And despite widening recognition of the
major studies in epidemiological inquiries, value, not to mention efciency, of using such
placing them at the intersection of medical, social data rather than collecting the data afresh from
and behaviour science, cannot sensibly be ignored. study members, the terms under which the dataset
For a period, there was a division between the is released thus become crucially important for
different policies of the medical as opposed to the research to succeed.
social science research funding bodies. The former The separation of a secure le that contains
endorsed restricted access limited to the research the name and address database from the (numeri-
team, except through collaborative arrangements, cal) data collected is thus now standard practice
and the latter endorsed availability of any anony- to ensure that protection from disclosure of study
mized data to the researcher. Such access was members identities can be maintained. But,
constrained only by policies concerned with data despite this safeguard, as geographical indicators
protection and prevention of disclosure of study are introduced at lower and lower aggregation
members identities. levels, the easier it becomes to identify study
The growth of national Biobanks for build- members. Moreover the longer any study contin-
ing repositories for DNA and other biomarkers, ues, the stronger the possibility that the patterns
as accessible resources for medical scientic use, in the longitudinal data across time can reveal
was another turning point that pushed policy the identity of a respondent. Such problems can
governing the use of longitudinal data in the be exaggerated as most study members trust the
social science direction. From then on, access researchers guarantees that the data will not be
to most publicly funded anonymised data became misused, but when that trust is broken, the result
increasingly unrestricted. The conditions of access can be devastating for the study.
could, however, as we shall see later, vary consid-
erably depending on the sensitivity of the data,
ranging from having to undertake the analysis 2.2 Methodological
on secure premises controlled by a government
agency (e.g. Stats Canada) or the researchers The transformation of individual IT capability
own university; having the analysis done for you through the high speed and vast storage capacity
(e.g. UK Ofce of National Statistics); having an of modern computers opened the door to the
individual license to use the data classied and second driver of development: making feasible
restricted for access according to sensitivity, with the use of high powered statistical methodology
very tough sanctions on anyone breaking its terms directed at modelling life course processes. Much
(Martin et al. 2006). of the statistical toolkit used today was developed
Such restrictions apply in particularly strong theoretically long before there was any adequate
form for linkage of government data with variation means of implementation. Technological advance
from one department and one region to the next, opened the door to actually implementing sophis-
presenting often layers of approval procedures ticated statistical models (Goldstein 1979).
that the research has to go through. For example, An important initiative that stimulated the
use of census data under most jurisdictions is application and development of new methods for
ruled out altogether and other data such as analysing longitudinal data, especially in Europe,
medical and criminal records released only under was the European Science Foundation-sponsored
strictly controlled conditions. The experience of European network of longitudinal studies on
Project Metropolitan, considered earlier, and Individual development and an accompanying
more recent examples such as the abandonment seminar series. Over the 5 years ending in 1991
of the Swiss SESAM study, shows how publicly during which the network operated, several
sensitive the research use of such data can be, hundred researchers from across the social
36 J. Bynner

and behavioural science disciplines attended (Muthn and Muthn 2014) and covering every
the seminars. The communication output was extension from the basic SEM model to latent
twofold. The rst was a directory of all longitudinal growth curve and latent trajectory modelling.
studies meeting certain criteria of sample size Some omnibus packages emerging through
and duration across the world (Schneider and the 1970s also played a crucial part in opening up
Edelstein 1990; Zentralstelle fuer Psychologische whole new vistas of basic and advanced statistics.
Information and Documentation (ZIPD 1994)- SPSS, rst available in 1970, led the way in
hundreds were listed. The second was a series of supplying the basic applications, including for a
8 seminal books including such inuential titles while, LISREL in its menu. More recently, and
as Problems and Methods in Longitudinal more comprehensively STATA created in 1985,
Research (Magnusson et al. 1991); Data Quality with provision for developing personalised soft-
in Longitudinal Research (Magnusson and ware options within the program, opened the
Bergman 1990); Analysis methods in Longitudinal door to the widest range of methods for modelling
Research on Individual Development (Magnusson social and econometric processes. Multiple
and Casaer 1993). imputation methods became available for dealing
Thus hazard modelling using the tools of with missing data and attrition in longitudinal
Event History Analysis came into its own and surveys (Abowd and Woodcock 2004; Brick and
accessible to every researcher as popularised by Kalton 1996; Rubin 2004; Goldstein 2009) thus
such books as Event History Analysis (Blossfeld complementing reliance on the weighting solutions
et al. 1989), Event history analysis in life course for missing cases that had been used previously.
studies (Mayer and Tuma 1990) and updated with The multilevel extensions of modelling meth-
a subtle shift in emphasis, Techniques of Event odology added the ecological dimension to the
History Modelling (Blossfeld and Rohwer 2002). analysis of variables by multivariate methods (e.g.
Multidimensional contingency table analysis Goldstein 2011). And integrative data analysis
using log linear and logistic modelling methods (Hofer and Picinin 2009) offers methods for analy-
in which time was one of the dimensions was also sis across independent cohorts within and between
for a period taken up widely. countries including those with missing data.
Structural equation modelling with latent In the rather different area of classicatory
variables and maximum likelihood estimation, and person-based methodology the more induc-
taking measurement error into account, is ideally tive techniques of Andrew Abotts sequence anal-
suited to the large scale timed sequential life ysis (Abbot 1995) using optimal matching (OM)
course data that longitudinal surveys produce methods also became available. The revival of
and, coupled with its theory testing through Lazarfelds 1968 latent class analysis, extending
model tting, became a staple of longitudinal to the panoply of different forms of multidimen-
researchers. The methods availability through sional scaling and cluster analysis, offered fur-
the LISREL programme (Jreskog and Srbom ther possibilities. Staunch critiques of the OM
1979), now in version 9.1, was the product of a approach e.g. Wu (2000), were addressed in a
collaboration through the 1980s between Karl second wave of methodological development
Jreskog (mathematical statistics) and Dag (Aisenbrey and Fasang 2010; Levy and Widmer
Srbom (mathematical programming) and 2013). There was now a means of grouping the
spawned a number of further less challenging myriad individual developmental patterns within
packages such as Peter Bentlers EQS (Bentler and across the major life course domains of edu-
1995) and AMOS (Arbuckle and Wothke 1999). cation, employment, family and health into an
The development included perhaps the denitive optimum number of major trajectory types.
SEM program, now in its seventh issue, MPlus, Associated typologies could then be constructed
written by Bengt Muthn, Joreskogs one-time of individuals falling within different trajectory
PhD student and his partner, Linda Muthn combinations.
Institutionalization of Life Course Studies 37

The standard packages like SPSS, with which different data sources and methods and reality
longitudinal researchers analysed longitudinal testing described by Donald Campbell (1991) as
data, have increasingly given way to new, and qualitative knowing (i.e. participant perceptions)
many would say, more liberating conceptions. to help in deciding one causal narrative as
The development of such software packages as opposed to another. The RCT supplies the gold
R using the S programming language, and standard for causal inference (Cook and
available free of charge to any researcher, offers Campbell 1979) and where feasible always adds
a programming environment in which the individ- value in addressing directly the causality question.
ual researcher can develop, alongside the stan- But as Lee Cronbach argued in Beyond the Two
dard techniques, his or her own statistical Disciplines of Scientic Psychology (Cronbach
methodology and applications. 1975) that criterion may not be sufcient on its
Finally, another development in line with the own to formulate a model of a life course process
creation of such an analysis environment is the in any given context that in a practical sense
growth of graphical methods such as Directed works. In challenging the idea of generality of
Acyclical Graphs (DAGs) which, through the causal conclusions across space and across time
visualisation of research problems, map more he makes the telling point: Once we attend to
effectively onto methods for modelling the interactions, we enter a hall of mirrors that
processes involved (Friedman 2008). Pioneers of extends to innity.
such work in developing standard packages like
LISREL and AMOS for structural equation mod-
elling linked a whole new world of graphical 2.3 Theoretical
realisations of statistical solutions to research
problems. The rationale for such developments The third driver of the new opportunities was
parallels that of life course study itself that is to theoretical: traditionally, as we have seen (chap-
say the desire to get away from the somewhat ter Introduction: Life Course Studies: Trends,
narrow connes of conventional multivariate Challenges, and Future Directions), sociologists
analysis and presentational techniques. and psychologists kept on separate sides of a the-
No development of methodology will of oretical divide in human development research.
course in itself meet the main epistemological The former focused on societal structures and his-
challenge for longitudinal research that of torical change as inuences on the shaping of
research design. The absence of randomised development and the latter more on individual
controls in longitudinal surveys categorises attributes and their biological basis, almost devoid
them in Cook and Campbells terms (1979) as of any historical and social context in which
quasi-experiments. That is to say, despite the development took place. Thus life-span develop-
wide range of data collected and repeated mea- mental psychology, while recognising the impor-
surement and multiple statistical controls applied tance of context to which the individual organism
in testing research hypotheses, the confounding had to respond, saw such contextualisation in
effects of unmeasured variables and selection ontological terms rather than as fundamentally
effects cannot ever be entirely ruled out. However, shaping and changing, through the exercise of
while the true experiment embodied in the ran- agency, the direction the life course takes.
domised control trial supplies the most robust Shanahan and Porfelli (2002) usefully illuminate
evidence of a causal effect, the price to be paid is the distinction between the two positions through
the loss of external validity; the generalizability identication of the different entry points.
of the ndings to all other non-experimental and The life course perspective thus takes synthesis
potentially interacting contexts remains uncer- further by bringing into the picture the historical
tain. Accordingly, validation has to be achieved component, which embraces the shaping inu-
by other means such as triangulation across ences, at different levels of societal changes and
38 J. Bynner

policy shifts, on the trajectories, transitions with the turning point not much later than the
and turning points through which peoples lives beginning of the 1990s. The different strands of
progress (Alwin 1995). The complementary developmental thinking, life span (ontogenetic),
contextualising life course features are the extrin- generational and temporal, began to come
sic factors identied with cohort, period, and together in a comprehensive theoretical frame-
age effects, control of which is central to longitu- work (Elder 1998; Elder et al. 2003; Elder and
dinal, and consequently life course, research Shanahan 2006).
design (Bynner 2005a). By the end of the 1990s a synthesis of
As we have seen, these ideas came to fruition approaches to conceptualising and conducting
initially through the work in Berkeley of John life course study in research areas as diverse as
Clausen (1972) and especially Glen Elder, whose returns to education, family formation and break-
Children of the Great Depression (Elder 1974), down, health inequalities and obesity, well-being
relying on the secondary analysis of long-term and cognitive aging and retirement, escape from
archived cohort data from the Oakland, Berkley disadvantage and social exclusion (Bynner 1996)
and Terman studies, effectively crystallised the was beginning to emerge. Methods of Life Course
life course concept. The ve dening principles - Research (Giele and Elder 1998) not only set
development in context, human agency, linked out the now classic diagram presenting the ve
lives, timing and location in time and space principles on which the life course is constructed
have been central to the life course approach ever but supplied a toolkit of accessible research
since (Elder 1985; Giele and Elder 1998). methods for applying them. The range extends
Further extensions followed from recognition from surveys through social biography, to the
of: (1) the need for analysis at different social practicalities of longitudinal follow-up and
ecology levels, where Bronfenbrenners (1979) integrating qualitative, quantitative and historical
ve tier bio-ecological conceptualisation of methods. The updated version 11 years later
macro, exo, meso, micro, individual levels as (Elder and Giele 2009) notably reorganises the
the basis of social interaction pointed the way; ve principles in hierarchical form. The book
(2) the value of extending data collection to also expands the scope of inquiry putting into
embrace intergenerational transfer of resources the frame not only more on archived and admin-
or capital within families from one generation to istrative data but biomedical measurement,
the next; (3) benets of international collabora- including DNA and intergenerational compara-
tion in cross-national inquiries to study the interac- tive multi-level longitudinal inquiry.
tion between social institutions and developmental Complementary texts include Looking at
processes in different societal contexts. Lives (Phelps et al. 2002), which turns attention
European contributions came from such work to the autobiographies of those who conducted
as that from Walter Heinz (1991, 1996; Heinz the major US studies. The book is unique in
and Marshall 2003) on the agency component showing how researcher biography itself may
and self-socialization; from Karl Ulrich Mayer become effectively a component of the study
(Mayer and Schoepin 1989) on the role of itself, not only in the shape it took, and whether it
Institutions and the state; from Ingrid Schoon continued, but in the ways it was communicated.
(2006) on accumulated risk across the life course Seven Swedish Longitudinal Studies (Janson
and resilience to its effects; and from Hans-Peter 2000) focuses more on the biographies of the
Blossfeld (2006) on the life course and globali- studies themselves, their histories and ndings,
sation processes. with much technical and practical detail sup-
With the advent of desk-based modern ICT plying helpful insights for researchers coming
there was an expansion of life course study new to the eld.
through the exploitation of longitudinal data,
Institutionalization of Life Course Studies 39

An organisational counterpart to these books 2006) describes it. The further development of
with a psychological orientation was the Life the system is considered ext.
History Society a loosely connected network with
foundations in psychopathology and behavioural
science that meets every 2 years in different 3 Consolidation and Growth
cities. Its objectives are: (1) to share information
that will advance the study of human development Examining the foundations of contemporary life
and inuence the next generation of longitudinal course study gives way in this section to the next
studies; (2) promote multidisciplinary research stage in its growth and institutionalisation. The
and interdisciplinary dialogue/collaboration section considers the mainly post-millennium
around a common theme (longitudinal research); further enhancements to longitudinal research
and (3) foster communication between senior and strategy and protocols. These contribute to the
junior researchers, and researchers and students. ever expanding eld of longitudinal survey
Straight and devious pathways from childhood to research now extending globally. The move
adulthood, based on the 1987 meeting (Robins towards synthesis and the increasing interest
and Rutter 1990), and involving contributions shown by governments in longitudinal research
across the biological, social, and behavioural ndings is reected in the latest generation of new
sciences, is another classic in the eld. The studies. I briey consider these and then move
International Journal of Behavioral Development, to the widening scope of longitudinal data
established in 1976, though having no direct con- collection bringing challenges as well as benets
nections with the LHS, is the major platform for to the life course researcher.
the kinds of output that its contributors produce.
The focus also changed in life course study,
with emphasis moving increasingly away from 3.1 New Studies
normative patterns towards multiple developmen-
tal pathways and outcomes, including the transi- Although the millennium year was no more than
tions, turning points and interactions between a date in a calendar, its symbolic signicance
life course variables and socio-political contexts provided a stimulus to investment in longitudinal
(Schulenberg and Schoon 2012). Analysis of studies across the world, prompting consolida-
development was not restricted to any one domain tion of what had been until then largely national
but extended to all of them. That is to say, family or local enterprises. The Norwegian1 and Danish
background, partnership and parenthood, educa- birth cohort studies, with an epidemiological
tion, employment, health and citizenship were orientation, got underway, together with a host of
increasingly pursued conjointly. more explicitly holistic life course studies around
This is not to deny the signicance of any single the theme of Growing up in (most prominently),
strand of the whole account; nor the centrality Ireland, Scotland, New Zealand, Australia, the
of the five principles of life course analysis; Netherlands (Generation R),2 France, (ELFE)3
nor interactions between them. Epigenetics the last, a birth cohort study modelled on the US
postulates environmental shaping of the way
genes are expressed and consequently how the 1
Olsen et al. (2001). The Danish national birth cohort its
social becomes the biological (Blane et al. 2013). background, structure and aim; Norwegian Institute of
All components of the life course are seen as Public Health (2005). Mother and Child Cohort study.
potentially in constant dynamic interaction with www.fhi.no/default0.asp
2
each other and the ever-changing social and Hofman et al. (2004). Growth, development and health
from early fetal life until young adulthood: the Generation
physical environment as the life course unfolds. R study. Paediatric Epidemiology, 18, 6172.
We can identify here an evolving developmental 3
The ELFE study: how do our children grow? www.
system as Lerner (Lerner et al. 2000; Lerner elfe-france.fr/index.php/en/
40 J. Bynner

National Childrens Study. There were also calls NIH cancelled the study, however, after an evalu-
for widening collaboration across countries to ation concluded that the study would be too
harmonise data and establish common standards costly and not exible enough to capitalize on
for data collection and use (e.g. Kogevinas et al. emerging scientic discoveries (e.g., gut micro-
2004). biome). NCS may well be an example of how
A series of strategic reviews in different very big science has disadvantages. Indeed NIH
countries both identied the importance of the Director Francis Collins noted that the main point
life course data the longitudinal studies produced was not that the study was cancelled, but rather
to the policy process and the need for the secure that the same research aims should be encouraged
funding, increasingly called investment, to insure and studied via other means.4
that the long-term returns sought would be forth- The NCS was also unusual in certain respects
coming. Typical funding, if not coming directly in attempting to begin study of the lives of the
from government, was awarded through a science cohort members as closely as possible from the
funding council, often combining support across moment of conception, so that a full pre-term
the medical and social science disciplines to record would be included in the data for analysis.
ensure multi-disciplinary approaches. The mod- Its UK counterpart, the 2015 Life Study, has
els developed were increasingly shared between followed suit, targeting 80,000 births for follow-
countries, re-enforcing the prospects for research up from the rst clinical verication of pregnancy
collaboration and comparative study. and data collection. The study was rst scheduled
Many countries initiated their own life course to begin in 2013 but will now be registering the
oriented millennium cohort studies in some rst pregnancies for the study in 2015. The study
cases, as for the French ELFE study, taking years combines a national probability sample with a
to negotiate with government the terms of the number of area case studies (Dezateux 2014).
funding that was needed. Others, such as a In parallel with new studies that had the ben-
proposed Swedish birth cohort study failed to get et of the developing infrastructure expertise and
government backing, because it was thought data collection protocols inherited from previous
the wealth of Swedish register data rendered it and established studies, there was development
unnecessary. Yet others began but did not get across the whole life course study enterprise. The
very far, such as the Swiss SESAM study focused scope of data collection increased together with
on the aetiology of adjustment and mental health. the development of collaborative arrangements to
Resonating with the Nordic countries Project take best advantage of it.
Metropolitan experience discussed earlier, this
well- developed birth cohort study foundered at a
late stage on ethical grounds concerned with the 3.2 Intergenerational Studies
inability of study members (i.e. new born babies)
to give direct consent for the collection of their As data collection progressed to the adult stage
DNA. It is not surprising that the study failed to of the life course there had been increasing
recruit mothers to the sample (Kummer 2011). recognition in many studies of the need to extend
In the case of the massive US 100,000 the scope of enquiry to the offspring of study mem-
pregnancies National Childrens Study (NCS), bers as a basis for understanding intergenerational
with a major focus on the health effects of continuities and transfers (see Thornberry, this
environmental exposures (Michael and volume). For the longest continuing longitudinal
OMuircheartaigh 2006), the problems were study, that of the Fels Institute, data extended to
operational and nancial. The study was sub- embrace not only children but grandchildren and
jected to a series of re-appraisals, with new
design and implementation strategies brought in. 4
http://www.nih.gov/about/director/12122014_state-
ment_ACD.htm
Institutionalization of Life Course Studies 41

great-grandchildren.5 The four FELS generations children were born at the ages data were collected
all within a single family thus offered extraordi- and their generalizability is limited to that extent.
nary possibilities for intergenerational study
across societal changes spanning over 80 years).
The US Add Health study is one of the most 3.3 Cross-National Studies
recent examples where funding has been raised to and Collaborations
enable such investigations.
However there is much variation in the way The intergenerational aspect of longitudinal
such extensions are carried out. Longstanding, studies aligns with another development already
household panel surveys (such as PSID, SOEP, signalled cross-national investigations and
BHPS) have routinely recruited panel members the networking arrangements to facilitate them,
children to the study either from birth or from a including sharing experience with new studies.
set age before the adult entry point of the main The latter were often put in place largely for prac-
study, usually age 16, is reached. The 1958 and tical reasons, especially for new studies, directed
1970 British Birth cohort studies recruited to the at learning from experience and not reinventing
study the children of one third of the male and the wheel. But they also served the wider scientic
female cohort members who had reached aged 3. goals of extending the scope of coverage, this
But this was on a one-off basis without subsequent time to cross-national differences and similari-
follow-up. In contrast, since 1986, the US NLSY79 ties, capturing particularly the institutional and
has undertaken assessments every 2 years of all policy variations with cultural roots, prevailing
children of female study members. US examples across the countries involved.
show the kinds of topics addressed: Thornberry Such joint studies were pioneered by the
(2009), anti-social behaviour; Martin et al. household panel surveys as a basis for under-
(2010), problem behaviour; and Mortimer (2012) standing income and associated family dynamics
youth development. in different national settings. Although these
Intergenerational studies based on birth surveys traditionally varied in some features of
cohorts are often wrongly interpreted in terms of coverage and frequency of follow-up, reecting
creating a population sample representative of the particular research foci of inquiries based on
the next generation and analysing the data for it the data and funding constraints, they all adhere
accordingly. However there are challenges in to much the same kind of template with respect to
making correct inferences from the data such the socio-economic conditions of living and
studies yield, not least because the age range of economic activity of panel members. They also
the child sample depends on the mothers age collect comparable event histories comprising
when she gave birth to the sample children. Until jobs, house moves, domestic and socio-economic
child bearing is completed in the cohort, estimates statuses and in some cases, health, on an annual
based on the children will thus be biased towards or two yearly basis. Their national representative-
early maternity, which will tend to weight the ness, albeit of the population when the panel
sample towards the more disadvantaged families. survey began in the different countries, is another
Cohort members children are therefore more attraction.
appropriately viewed as an attribute of the cohort Increasingly the mode of operation is thus
members rather than as an independent sample in towards data harmonisation and consolidation
their own right. That is to say, the inferences drawn through the formation of networks to support
about intergenerational continuity and discontinu- comparative analysis by exploiting the survey
ity are restricted to the cohort to which the studied data jointly. Collaboration can even extend to the
inclusion of new supplementary measures in all
5
http://medicine.wright.edu/lifespan-health-research-cen- studies. A good example is CHER (Consortium of
ter/fels-longitudinal-study/history Household Panels for European Socio-economic
42 J. Bynner

Research) embracing 17 European countries next 15 years) by evaluating data from existing
(plus USA and Canada). Another that brings cohorts, registers and relevant European data-
together 20 European longitudinal surveys bases, identifying gaps in knowledge, and devel-
(plus Israel) modelled on the US Health and oping recommendations for targeted research
Retirement Survey is SHARE (Study of Health action at the European level.
and Retirement in Europe). More sociologically driven networks with
An important outcome of such collaborative an interdisciplinary orientation are to be found
work is the construction of cross-national les in Switzerland where the Centre for the
created through data harmonisation across Interdisciplinary Study of Life Courses (PAVIE)
different studies. The now expanded multi based in Geneva and Lausanne supports a network
country Cross National Equivalent File (CNEF) of longitudinal researchers from a dozen univer-
combined initially harmonized data from the US sities working on the theme of Overcoming
German, Canadian, UK, Canadian and Australian Vulnerability (to new risks): Life Course
household panel surveys and is the most well Perspectives. LIVEs counterpart is the produc-
established example (Burkhauser and Liddle tive TREE longitudinal survey based in Basel
2006, 2007). and devoted to transition from post compulsory
Other networks have developed this kind of education to the labour market.
international collaborative work more widely The Young Lives project beginning in 2002
seeking syntheses of again studies drawing on and based in Oxford University goes even further
diverse data sources. Apart from SHARE, the globally by linking longitudinal surveys of
UK Healthy Aging across the Life Course 12,000 children in 618 months and 78 years
(HALCYON) network and the US based old cohorts of child and adolescent development
Integrative Analysis of Longitudinal Studies of in Ethiopia, India, Peru and Vietnam. Anthropo-
Aging (IALSA) network are two that have now logical inquiry by means of Qualitative studies of
combined forces. selected children and their families of children in
In the case of new studies, the impetus for each country is another feature.
collaboration is mainly sharing experience and
learning from others. Hence EUCCONET,
funded originally by the European Science 3.4 New Data
Foundation and established in 2008, was
coordinated by INSERM, the French institute 3.4.1 Biomarker Data
responsible for the ELFE birth cohort study in An area of major advance aligned to the develop-
partnership with the London-based Centre for ment of cross-national studies is the collection
Longitudinal Studies and involved 12 countries. and storage of biomarker data, such as blood
The aim was to bring together leaders of national samples and genetic coding of the extracted DNA
or major regional child cohorts in order to compare (Pembrey 2004). In this instance, the developing
practices, exchange experience, share question- case for cross national study and standardization
naires and other tools, and encourage compara- of measurement is less a matter of replicating
tive analyses. and testing hypotheses under different cultural
CHICOS, established in 2010, was another and economic conditions than pooling the data across
such network, this time funded under the European studies to produce samples of the huge size that
Union 7th Research Framework program and gene-environment interaction and gene expres-
comprising 30 child health cohort studies in 19 sion studies require.
countries. The networks aims went wider than The development of networks of Bio Banks
the design and operational interests of in a number of countries is a parallel develop-
EUCCONET: to assess and prepare for the col- ment, in which the medical researchers involved
lection of robust health data for birth/mother- are only now beginning to recognise the value of
child cohorts over a substantial time period (the interaction, including data linkage, with the much
Institutionalization of Life Course Studies 43

more broadly based life course studies. As the longitudinal research requiring detailed attention
adaptability of instrumentation for use in different not only to what can legitimately be linked but
contexts at home as well as in a clinic continues, how the data should be organised in terms of levels
the inclusion of more demanding measurement of access. Canada permits such research access to
such as EEGs and brain scans becomes increasingly any personal data only on the Stats Canada secure
feasible. Used cross nationally, the large-scale premises in various locations that the government
pooling of such data will similarly often become agency provides. Other countries are more liberal
the goal. The challenges that arise here are likely in their approach, licensing universities to provide
to do with data protection and preventing disclo- secure environments in which such work can be
sure of the identities of the study members. done. In yet other cases, different levels of access
are codied depending on the sensitivity of the
3.4.2 Geocoded Data data: (1) release for desk top analysis; (2) release
Geo-coding of spatial and locational data is under special licence; (3) release only in a safe
another major area for augmenting cohort study setting; (4) never release but can be analysed by
content though raising further challenging issues Government employees according to specication
for data protection and disclosure. Decisions (e.g., Add Health).
have to be made about setting minimum levels of
aggregation to ensure that cohort members (ano- 3.4.4 Social and Psychological Data
nymised) identities remain protected (Elliot et al. Biomarker and associated physical data may be
1998; Elliot and Dale 1999; Boyle et al. 2004). sufciently objective and robust in measurement
terms to be collected in much the same way from
3.4.3 Administrative Data one country to the next another. But apart from
The data enhancement aims of all such collab- ethical considerations, the transferability of
orative projects connect with another major social, economic, behavioural and psychological
development in life course studies the increasing data can be more problematic for other reasons.
accessibility and use of the vast amounts of In cross cohort analysis within a single country
administrative data collected by governments the cohorts are contextualised by the historical
recorded on individuals in longitudinal form. periods in which the cohort members were born.
When linked to on-going scientically driven Cross-national analysis of cohort studies data
longitudinal surveys, such resources offer huge compares cohorts contextualised, in addition, by
research opportunities in two main areas one is the institutions, structures and cultural imperatives
to check and validate the data collected in the of each of the countries in which the studies were
study itself about the same life course experiences; carried out. Comparative study is thus bedevilled
the other is to expand the dataset in areas of by the problem of equivalence the extent to
routine data collection such as educational which a measurement taken in one country is
attainment, health, housing, taxes, and transfer strictly commensurate with the measurement of
payments. The latter saves space in the question- ostensibly the same variable in a different country
naire for more subjective and self-evaluative (Kohn 1987; Scheuch 1990). Educational stan-
kinds of information, including, attitudes, values, dards, for example, may mean different things to
personality measures, and so on. people in different countries. Even economics
The risks of disclosure from such data linkage variables like socioeconomic status are loaded
and the need for clearly specied data protection with culturally grounded meaning. Hence stan-
terms are obvious and are often judged as putting dardization, though a laudable aim, can only go
out of bounds some kinds of data enhancing so far. As Kohn argues, the point will often arise
exercise, such as linking criminal records to the in comparative study when a nding fails to repli-
survey, for example. The risks also raise again the cate across countries. In such a case, resort to his-
issue of ethical frameworks for conducting torical and cultural analysis is likely to be the only
44 J. Bynner

means of locating where the origin of the inconsis- 4 Institutionalisation


tency lies. Qualitative appraisal then becomes the
essential counterpart to the quantitative analysis in In the UK, where the whole concept of repeated
effective comparative life course study. birth cohort studies carried out at national level
had perhaps taken root earliest, Government
scientic funding is awarded direct in the form of
3.5 International Standards Government department contracts or channelled
through science funding agencies such as the
Although standardization can only go some of Economic and Social Research Council (ESRC)
the way towards a generalised consistency of and the Medical research Council (MRC), comple-
approach and meaning in cross national studies, mented by the charitable foundation, Wellcome
there is clearly a need to attempt international Trust. All began to see the merit of strategic plan-
agreement in all the measurement areas of the ning in committing funding to longitudinal sur-
study. The United Nations Educational, Scientic veys. The result was that the governments Large
and Cultural Organization (UNESCO) have taken Scale (Scientic) Facilities funds were made
the lead in the establishment of the International accessible to the ESRC for the first time but
Standard Classication of Education, ISCED, only if bid for in partnership with MRC. In
Achievement Standards Framework the latest effect, such data collection enterprises as house-
version of which was released in 2014. The World hold panel studies and new cohort studies were
Health Organisations (WHO), International coming to be seen as large scale observatories to
Classication of Diseases ISCED (10) performs address key questions and hypotheses concerning
a similar function for biomedical and epidemiologi- human development in changing times with anal-
cal measurement and standards. Extending the ogies to those driving astronomy and subatomic
idea of such schema more generally across all the physics (Bynner 1993, 1996; Mayer 1993;
life course domains including family, housing, Wolfson 2006).
employment income citizenship and socio eco- In parallel, following a collaboration among
nomic status research data becomes the obvious key institutions responsible for the national birth
next step. Standardization of ethical codes as cohort studies to support them by collaboration
applied to longitudinal data is another area where in the form of a Joint Centre for Longitudinal
such international agreement, perhaps coordinated Research,6,7 the idea began to take shape for a
by UN agencies, would seem essential. think tank to promote longitudinal study and
address issues concerning its coordination and
improvement. At a time when I had recently
3.6 Review retired from my fulltime post in the London
Institute of Education I was able to undertake
The chapter reveals both steady expansion of life the scoping study that set the parameters for the
course study capability coupled with growing organisation and operations of the think tank.
optimism about the prospects for its full realisa- The work involved mainly interviews and
tion though the expanded scientic investment in correspondence with longitudinal research
new longitudinal surveys that is now readily experts and policy people and identied a huge
available on an unparalleled scale. Such expansion
has many challenges which will be examined 6
Reported in Bynner, J. (2005) Challenges for longitudi-
in the nal section. An important aspect of the nal research: synopsis of ndings from a scoping study
picture is the development of the institutions that http://www.slls.org.uk/=!longview-reports/c8a5
7
Comprising Centre for Longitudinal Studies (Institute of
are needed to give the work a solid professional
Education), Department of Community Epidemiology
base and from which to support it. We turn to and Institute of Child Health, University College London
these in the next section. and National Centre for Social research).
Institutionalization of Life Course Studies 45

range of interest, experience, and enthusiasm the proposed new birth cohort study that would
for longitudinal and life course research across re-establish the 12 year interval, following the
the science/policy divide (Bynner 2005b). All 30 year gap between the 1970 and 2000 birth
respondents felt that key issues of co-ordination, cohort studies. Inspired in part by the ambition of
rationalisation, and cost effectiveness, in a eld the US NCS, the sample for the new study would
that had largely grown ad hoc, needed to be be ve times the size of its predecessors and
resolved. Gaps were identied in the provision begin in pregnancy. Longview was commis-
and exploitation of longitudinal research resources, sioned to prepare the scientic case for the proj-
in the communication and interpretation of longi- ect and set out options for its design including the
tudinal research ndings and in the capacity to case for biomarkers. The 80,000 strong LIFE
take full advantage of methodological advances Study, as it is known, began piloting the new
by creating effective mechanisms for building survey in 2014.11
competence, fostering commitment, and generat- The nal recommendation was a coordinating
ing intellectual rewards. The blue print for the mechanism for bringing all the British cohort
think tank Longview was drawn up with the aim studies together in a framework for collaboration
of enhancing the quality, productivity, and returns on improving the quality and scope of longitudinal
to be gained from the whole longitudinal research data to ensure best value from the investment
enterprise. involved. In 2014 the CLOSER (Cohort and
Longview was established in 2005 with Longitudinal Studies Enhancement Resource)
support from the National Centre for Social facility and network was established to harmon-
Research8 and the International Centre for Child ise, coordinate, and provide support for nine
studies.9 Its arrival coincided with the Economic British longitudinal studies based in different
and Social Research Councils newly formulated universities including the six major birth
strategic goals for longitudinal studies and cohort studies i.e. 1946, 1958 1970, 19911992
Longview was commissioned immediately to (ALSPAC area study), 2000 and 2015.
undertake a series of scoping studies for achieving In the latter part of Longviews programme,
them. The most signicant of these, the 2006 Longview trustees, comprising myself as Director
Strategic Review of Panel and Cohort Studies and the partner bodies, with all the British cohort
(Martin et al. 2006), carried out by a specially studies and the British Household Panel Survey
assembled expert group established the case.10 represented, decided to pursue the communica-
The outcome was a proposal subsequently made tion aims of the think tank more directly, with the
by the ESRC to government for substantial aim of lling what was perceived to be a gap in
investment in the household panel survey. The the scientic literature. The solution was to develop
new funding would support an eightfold increase an international journal devoted exclusively to
in the surveys size, i.e. from 5,000 households to Longitudinal and Life Course Studies (Bynner
40,000 the worlds largest household panel. The et al. 2009). Through the support of a scientic
second report, Scientic Case for a New Birth funder, the Nuffield Foundation, a 3-year
Cohort Study (Bynner et al. 2008) was devoted to development grant was awarded. Work on setting
up the journal began in 2007 and the rst issue
was published in April 2009.
8
The UKs leading Survey Research organisation.
9
Charitable foundation set up in to run the 1970 birth
cohort study age-16 follow-up survey and to support
research based on it. 11
John Bynner, Michael Wadsworth, Harvey Goldstein,
10
Jean Martin, John Bynner, Graham Kalton, Paul Boyle, Barbara Maughan, Susan Purdon, Robert Michael, Kathy
Vernon Gayle, Samantha Parsons, Andrea Piesse, 2006, Sylva, James Hall, 2007, http://www.slls.org.
http://www.slls.org.uk/=!longview-reports/c8a5 uk/#!longview-reports/c8a5
46 J. Bynner

Longitudinal and Life Course Studies, now in supporting, and supplying a platform for longitu-
its 16th Issue, is published online using as the dinal and life course research internationally. The
IT platform the Open Journal System (OJS). societys core function would be to hold a con-
An Editorial Board was constituted comprising ference annually in different countries at which
60 academic experts from across the world. With research ndings could be presented and the
myself as Executive Editor, four members of papers reporting them developed for publication
the Board took on the role of Section Editors12 in the journal.
with responsibility for handling anonymous A year was spent establishing the society
peer review for submissions in the four areas to through an interim committee that I chaired. The
which they generally related Health Sciences, rst SLLS conference was held in 2010, again in
Social and Economic Sciences, Development Cambridge, at which research papers ranging
and Behavioural Sciences, Statistical Sciences across the eld of life course study were pre-
and Methodology. Submissions spanning more sented. Karl Ulrich Mayer, Max Plank Institute,
than one section were encouraged and processed Berlin, and Carol Dezateux, Institute of Child
jointly. A journal manager was appointed to Health, London and Director of the new LIFE
manage the operational side of the peer review Study, were the Keynote speakers. Responsibility
process and the various production stages copy for the annual conferences and the publication of
editing, layout proof-reading and online publi- the journal was passed from Longview to the
cation. The readership soon began to grow, rising society. The second conference and the rst
to over 1,000 registered readers within 2 years with an elected president, Heather Joshi, was
and at the last count was approaching 2,000. held in Bielefeld (2011) followed by subsequent
Another feature of Longviews programme conferences in Paris (2012), Amsterdam (2013),
had been consultative 2 day conferences as an and Lausanne (2014) and Dublin (2015).
integral part of the strategic reviews. Five were Some indication of the societys success is the
held, with the rst three in Oxford and the last steadily increasing conference attendances now
two in Cambridge. Apart from setting the param- approaching 300 and membership of the society,
eters for the design of the expanded household 400 up to 600 including lapsed members.
panel study and the new 2012 birth cohort study, Representation spans 20 countries organised in
the aim was to highlight issues in the planning, 20 chapters across the world, each coordinated
development and communication of longitudinal by 1 of 20 Global Representatives.
research. These issues included longitudinal Throughout its short life the society has been
survey methodology and practice together with committed to continuing Longviews aims
key developmental science policy areas such as through such developments as establishing, at the
cycles of deprivation in which longitudinal Paris conference, a Policy Interest Group giving
research had a vital part to play. a continuing home to Longview in the society,
In the fth Longview conference in 2009, drawing up a directory of members policy research
devoted largely to the design and implementation expertise and organising collaborative events,
of the new birth cohort study, the last item on the including for the Lausanne conference, a series
agenda was to discuss the establishment of a new of symposia devoted to the policy/longitudinal
society both to meet the elds associational, research interface. A comparable Biomedical
communication, and training needs and, through Interest Group was established in Lausanne.
the membership fees raised, ultimately to cover Postdoctoral training is also supported through
the costs of the journal. The outcome was a new an annual methods workshop held between con-
international Society for Longitudinal and Life ferences and on the day preceding the confer-
Course studies with the mission of promoting, ence. A newsletter is also circulated every
2 months to all members. EUCCONET, the
12
Michael Wadsworth, Robert Erikson, Barbara Maughan, European Science Council network set up for the
Harvey Goldstein. new cohort studies established at the time of the
Institutionalization of Life Course Studies 47

millennium, joined forces with the society in their role was little more than that of data col-
merging its nal conference with the societys lector. The consequence was a general weaken-
conference in Paris and has also been given a con- ing of commitment leading within a few years to
tinuing home in the society. the studys demise.
This experience underlines the point that
longitudinal surveys charting the development of
5 Challenges and Prospects individuals over time tend to engender a much
stronger sense of personal afnity between
The establishment of the Society for Longitudinal subjects and investigators as the life courses of
and Life Course Studies was an important step both unfold than is the case with, for example,
towards meeting the professional and associa- the repeated cross sectional survey. Without that
tional needs of a growing international commu- incentive the longitudinal study may lose momen-
nity of longitudinal and life course researchers. tum to the point where it nally dies. On the
Its development reects the widening interest in other hand, as the US National Childrens Study
life course analysis, within individual countries experience shows, too much independence for
to reveal how successive cohorts progress through what were area based research teams can obstruct
life from childhood to adulthood under different progress towards the collective aims that the
socio-economic and institutional inuences. whole study has been set up to achieve. Finding
The existence of the society also plays to the the right balance is thus critical to success.
wider issues of cross-national collaboration and
comparison where these effects can be replicated
or challenged through, for example, identifying 5.2 Historical Context
differences in effects under different policy
regimes. This nal section draws together in Another challenging aspect, common to national
more general terms the prospects for life course and international studies, relying on secondary
study and the challenges to be met in realising analysis of archived longitudinal data, is the
them across all the fronts on which research changing socio political and broader historical
strategy operates starting with comparative study context (period effect) in which any longitudinal
and ending with conceptualisation. survey takes place. Changing Britain Changing
Lives (Ferri et al. 2003) reports comparisons of
pathways and outcomes for cohort members up
to the early 30s in the 1946, 1958 and 1970
5.1 Collaboration British birth cohort studies. Full interpretation of
the differences, however, demands comparative
International work brings many benets while contextual information about the historical
placing certain demands on those responsible landscape through which the cohorts lives, up to
for the individual studies involved to ensure that age, had passed. Yet such information is
success. Clearly data harmonisation enabling rarely if ever available as an accompaniment to
results of analyses to be compared directly the archived quantitative data.
across countries, is sought while maintaining Anticipating a development in data archiving
each of the participating studies unique iden- in which such information is routinely attached
tity. Another requirement is close collaboration to the deposited data, A Companion to Life
across different national research teams to Course Studies (Wadsworth and Bynner 2011)
ensure. But especially for longitudinal surveys, attempts to do just that in narrative form for the
such goals can be difcult to achieve. Somewhat British birth cohort studies. The text charts the
ironically, when EUROSTAT in the 1990s set in socio political, education, economic and health
place a xed design for a European Panel Study, policy histories across the period from the end
feelings grew among those on the ground that of the second-world war to the present day.
48 J. Bynner

The longer term goal is to develop the charting of next generation of researchers that the changing
such histories not only at the societal level but at situation demands.
the community (meso) level as well.

5.5 Respondent Burden


5.3 Research Ethics
Another less obvious problem is the growing
As the inclusion of ever-more demanding burden on survey respondents as the life course
measurement, such as EEGs and brain scans, data demands increase. As longitudinal studies
become increasingly feasible in home settings increasingly move towards the role of multi-
and access to ever wider banks of administrative purpose research vehicles, there is a widening of
data becomes possible, opposition to them is also core purpose continuously expanding the amount
likely to grow. Fears about the surveillance and scope of measurement done, especially on
state raise major challenges to all such scientic the newly demanded behavioural task front.
enhancements. Resistance to data linkage, for There is also pressure to supply the control group
example, may lead to prohibitively cumbersome for intervention studies in an experimental design
permissions procedures with respondents on each involving manipulating respondents experience
occasion when the linkage is needed. Again, the directly by using the study sample to supply the
international dimension comes into the picture. experimental as well as the control group in
As suggest earlier, United Nations institutions behavioural economics experiments. These are
like the World Health Organisation and UNESCO already in use in household panel studies such as
could have a major role to play in drawing up the German SOEP (Wagner et al. 2006) and are
ethical standards for longitudinal surveys that all on the agenda for many more.
member countries could be obliged to implement. Among behavioural scientists there is a com-
This development also makes the case that mon preference for using standardized clinically
public education about the value of life course developed instruments for the measurement of
studies is likely to become a growing obligation psychological attributes rather than short forms.
of those who undertake them. In fact such studies Restricted funding and tight timetables also push
need to be seen as shared enterprises in which research teams to use what is available rather
every citizen has as much stake in the outcome as than what is ideally needed on theoretical and
the scientist and policy maker. practical grounds. Judging the trade-off between
the length of an instrument and its reliability
needs to be done empirically. This makes the case
5.4 Attrition for much more expenditure of time and effort
than is customary on developmental work to
The perennial problem of sample attrition has optimise the properties of new measures before
prospects of worsening as the general public they are implemented in a new round of data
become increasingly disenchanted with participa- collection and especially in new studies (Bynner
tion in surveys of any kind. Financial incentives et al. 2008).
on a large scale may ultimately become inevita-
ble, which could add substantially to the costs of
life course studies. Recourse to statistical 5.6 Investment and Impact
solutions for missing data and attrition, such as
multiple imputation based on data already in the In one sense there was never a time in which the
longitudinal record (Rubin 2004), raises further monitoring of societies and individuals progress
issues of capability to implement them. A parallel through them was more necessary. However life
pressure then becomes substantial investment course studies rising importance occurs at a time
in methods training to build the capacity in the when economic crises, place increasing strains on
Institutionalization of Life Course Studies 49

public nances, leading to scarcity of scientic that need them most in the south. The Society
resources. The almost limitless expansion of gov- for Longitudinal and Life Course Studies (SLLS)
ernment investment in longitudinal studies is is well placed to facilitate the massive boost in
unlikely to continue indenitely. Not only do the capacity building that is needed.
studies become more expensive with time in terms
of database maintenance and tracing increasingly
dispersed populations, but each new study added 5.8 Conceptualisation
to the programme in a series of repeated cohorts,
for example, escalates the costs of the whole Analogies for the development of life course
programme. study moving from cottage industry to major
But the cost burden is not restricted to the industries make the point that the contemporary
collection of the longitudinal data that each situation of life course studies reects the coming
new follow-up brings. There is also an ever- together of societal and scientic challenges
strengthening accountability demand placed on with the technological advances needed to meet
those who use the data to demonstrate impact. To them. This convergence is manifested through the
maintain momentum, the pressure building up will collection, storage, propagation, enhancement,
have to be met in some way, possibly through much and analysis of longitudinal data that longitudinal
wider use of electronic means routinely for data studies produce. What of life course study itself?
collection rather than the standard survey inter- Its institutionalisation through stakeholder
view. It also demands again the need for knowledge investment adds policy development to what was
transfer expertise and engagement in public initially purely scientic interest.
education to make the case continually that public The contemporary life course study, while
investment in life course studies is worthwhile. under pressure to serve multiple purposes, needs
the integrity of a scientic program of clearly
specied projects underpinning it to ensure that
5.7 Distribution the data generated will contribute to a coherent
body of knowledge. That is to say, the risk must
Such demands need to be set against perhaps be accepted of losing sight of unexpectedness
the most pressing problem in life course study and serendipity in favour of theory testing and
development and use the unequal distribution development. Without hypotheses informing the
of longitudinal research resources across the design of the study, longitudinal study becomes
world. Life course studies grew up and became an exercise in empiricism full of blind alleyways
institutionalised in western countries, but the that are mistaken for opportunities. The price
need for longitudinal data to monitor the effect of to be paid is the lack of any clear conception of
socioeconomic and technological change on the processes underlying potentially the develop-
individual development is just as pressing, if not mental continuities and discontinuities that the
more so, in developing countries across the study is attempting to model and understand.
southern hemisphere. Not only are the resources These considerations raise the issue of where,
but the expertise lacking to develop the longitudinal theoretically, the development of the life course
surveys that are needed let alone the infra- perspective needs to go next. The Life Course
structure to sustain them. UNICEF, among other Perspective was crafted in main part from the
international agencies, has a major initiative consequences of pre-second-world war economic
going to promote such a development and some catastrophe affecting people differently at different
small scale enterprises directed at child birth and ages and at different times (Elder 1974). The con-
infancy, such as the Jamaica Longitudinal study, tinuities with todays world are obvious, so also
have begun. The challenge is to nd effective means are the differences. Ever-accelerating technological
of spreading the resources and expertise concen- change and globalisation of production and
trated in the northern hemisphere to the countries markets and the rise of social media need to be
50 J. Bynner

set against the evidence of ever-widening through which the life course is constructed (Mayer
inequality, the phenomenon of an expanding and Schoepin 1989; Mayer 2001). The dynamic
precariat (Standing 2011), the decline in psycho- interactions between, for example, agency as oper-
logical well-being and emerging health problems ationalised by such psychological variables as self-
ranging from rising obesity to drug abuse and efcacy, motivation, aspiration, and the social
new pathogens. All make the case for continual institutions in which agency is exercised, need to
reappraisal. embrace the relational context of choice in life
The ve core principles of development in course decision-making. The approach also needs
context, shaped by individual (and collective) to recognise the role of inuential actors such as
agency, linked lives, timing of events (described parents, teachers, employers and peers, not to
by age, cohort and period effects) and history and mention the timing of events and changes, in the
culture (Elder and Giele 2009), and now cast in broader social, institutional and policy context in
line with a multi-level bio social ecology frame- which human development takes place.
work, may need further renement in the radi- Life course sociology, life span psychology and
cally transforming electronic world in which we human biology, therefore, while recognising their
now exist. The consequence will be differential differences also need to seek convergence in an
effects dened in new ways of such key classi- integrated conceptual framework comprising theo-
ers as gender, ethnicity, religion, employment retical insights gained from all of them. Perhaps it
status, education, health, income and social class. is time though to move away from the conceptual
The dynamics of societal change are mirrored in disjunctions between disciplines to the practical
the body of data that is needed to understand problems and challenges of everyday living the
them (Marshall and Mueller 2003). aging population, health and lifestyle, polarisation
and discrimination and focus attention on those
that can only be resolved holistically. In relation
6 Conclusion to aging, the policy need for a truly holistic life
course approach is particularly pressing.
Such considerations take us back to the theoretical In this respect Richard Lerners idea of applied
issues with which this chapter began. Baltes (2006) developmental science (Lerner 2006; Lerner et al.
makes the convincing point that for life span study 2000) has much to recommend it as an integrating
to have heuristic value it must not begin and end in frame. Taking the concept a step further: to recognise
childhood but extend across the whole of life fully the historical and developmental power of
guided by a single unied theory. Extending the life course study the idea of life course science is
same perspective to life course study while recog- perhaps even better placed as the nal piece of
nising discontinuities and turning points (Elder and the jigsaw puzzle. In life course science, the
Shanahan 2006) and even backwards movement longitudinal surveys on which the whole edice
identied with different forms of status passage stands, become the social observatories through
(Heinz 1991), the same general rule applies. That is which the scientic models needed for under-
to say, life course study must be continuous in standing human development are realised and
scope and interdisciplinary in form rather than tested. Big data requires effective structures to
being reinvented from different disciplinary realise their value; technological and method-
sources at the different stages with which human ological advance applied within a continually
development is usually identied. developing theoretically driven programme of
Baltes own prescription, while seemingly inter- life course study provide the tools to achieve it.
disciplinary, appears to downplay the more social
as opposed to the individualistic components of Acknowledgement My thanks to Michael Wadsworth,
development particularly the role of social rela- Barbara Maughan, Harvey Goldstein and Robert Erikson,
whose contribution was crucial to the establishment of the
tions, social structures, social institutions and his-
Longitudinal and Life Course Studies international journal
torical change in the explanation of the processes and the Society for longitudinal and Life Course Studies.
Institutionalization of Life Course Studies 51

Appendix

Table 1 Selected longitudinal life course surveys


Country, region of study Year initiated Sample Follow-up ages
United Kingdom
National Survey of Health 1946 Nationally representative sample comprising 0, 2, 4, 6, 7, 8, 9, 10,
and Development (NSHD)a 4,689 births in a single week in England 11, 13, 15, 19, 20,
stratied to include more advantaged and 22, 23, 24, 26, 31,
therefore better educated families 36, 43, 53, 6065
National Child Development 1958 Nationally representative sample 0, 7, 11, 16, 23, 33,
Study (NCDS)a comprising 17,634 births in a single week in 42, 46, 50
Great Britain (England, Scotland and Wales)
1970 British Cohort Study 1970 Nationally representative sample 0, 5, 10, 16, 26, 30,
(BCS70)a comprising 17,287 births in a single week in 34, 42
Great Britain (England, Scotland and Wales)
Avon Longitudinal Study of 19931994 Representative sample comprising 14,120 Continuous
Parents and Children pregnancies in County of Avon (Bristol
(ALSPAC)a region)
Millennium Cohort Study 2000/2001 Nationally representative sample 0, 2, 4, 7, 11, 14
(MCS)a comprising 18,553 births selected from
random sample of 400 electoral wards
stratied to have relatively high
disadvantaged and ethnic minority
populations in United Kingdom (England,
Scotland, Wales and Northern Ireland)
Life Studya 2015 Nationally representative sample Annual
comprising 20,000 pregnancies in United
Kingdom (England, Scotland, Wales and
Northern Ireland) plus four clinically
intensive area studies each comprising
1520,000 pregnancies
Longitudinal study of young 2004 Nationally representative sample Annual
people in Englanda comprising 21,000 1314 year-olds and
their families in a random sample of schools
in England stratied to include more
disadvantaged and ethnic minority families
Growing up in Scotlanda 2005 Representative samples of 5,000 birth and Annual
3,000 2 year-olds resident in Scotland
British Household Panel 1991 5,538 households; 10,264 individuals (age Annual of all current
Survey (BHPS)/ 16+); sample boosts 1999 in Scotland and original
Understanding Societya (1,500), Wales (1,500); in 2001 Northern household members
Ireland (2000); from 1994 1116 year-olds
supply data direct
English Longitudinal Study 2002 12,100 individuals aged 50+ identied Every 2 years
of Ageing (ELSA)a though the (cross-sectional) English
survey of ageing
Nordic
Sweden: evaluation through 1961 Nationally representative cohorts Continuous relying
follow-up studya comprising 12,000 13 year-olds starting in mainly on
1961,1966, 1980, 1982/1985, 1987/1990, government register
1992/1995 data plus additional
assessment
Danish National Birth Cohorta 19972002 100,000 pregnant women Annual
Norwegian Mother and 20002005 100,000 pregnant women and partners Annual
Child Cohort Studya
(continued)
52 J. Bynner

Table 1 (continued)
Country, region of study Year initiated Sample Follow-up ages
Other European
German Socio-Economic 1984 Nationally representative panel comprising Annual of all current
Panel (GSOEP)a initially 7,000 households in the Federal and original
Republic of Germany; 15,000 individuals household members
(age 16+) boost: 1990 boost to 20,000
through bringing in post reunication East
German Lande into the survey, 2000 boost
to 24,000
German: National Education 2010 Five nationally representative age cohorts Annual up to age
Panel Study (NEPS)a starting in 2010 and at ages, 0 (in 2012), 67 in all cohorts
4,10,14,19, and an adult panel aged 2150
(part of an earlier age cohort survey
followed annually from 2007)
Netherlands: generation R 2002 Birth cohort study in a multi-ethnic urban Five times through
studya population comprising 10,000 pregnancy and from
pregnancies. Subsample Focus Cohort then on annually
more intensive study based on 1,000 births until age 20
between April 2003 and April 2004
France: ELFE study Francea 2011 Nationally representative sample Every few months
comprising 18,000 births from birth and
annually from age 1
New Zealand and Australia
New Zealand: Christchurch 1977 All 1,265 individuals born in Christchurch At 4 months and
Health and Development in 1977 annually from ages 1
Study (CHDS)a to 16 then at less
frequent intervals
New Zealand: Dunedin 1972 All individual born in Dunedin in 1973 At ages 3, 5, 7, 11,
Multidisciplinary Health and 13, 15, 18, 21, 26 and
Development Study from then on at less
(DMHDS)a frequent intervals
Australia: growing up in 2004 Randomly selected families. Cohort 1 At 2-year intervals
Australia Longitudinal Study (01 year-old) and Cohort 2 from 2004 to 2018
of Australian Children (LSAC)a (45 years-old)
Canada
National Longitudinal 1994, 1998, Three nationally representative cohorts of Every 2 years, with
Survey of Children and 2000 individuals (ages 011, 01, and 01) each some sample
Youth (NLSCY)a comprising over 20,000 individuals reductions to meet
cost constraints
Survey of Labour and 19932010 Composed of two panels, each consisting A panel is surveyed
Income Dynamics (SLID)a of two labour force survey rotation groups for a period of six
and includes roughly 17,000 households consecutive years. A
new panel is
introduced every
3 years, so two panels
always overlap
Youth in Transition Surveys 19981999 National samples of 15 year olds and Follow ups every
(YITS) 1820 year olds 2 years through
20082009
Longitudinal Survey of 2001 National sample of immigrants 15 years Follow ups in 2003
Immigrants to Canada and older, arriving between October 1, and 2005
(LSIC)a 2000 and September 30, 2001
(continued)
Institutionalization of Life Course Studies 53

Table 1 (continued)
Country, region of study Year initiated Sample Follow-up ages
Workplace and Employment 1999 National sample of employers from the Follow ups annually
Survey (WES) Business Register, freshening with new through 2006;
companies annually; sample of linked employers followed
employees from sampled employers up throughout and
employees retained
for 2 years before
being replaced
National Graduates Survey 1982; 1986; National sample of graduates from Each class
(NGS)a 1990; 1995; Canadian postsecondary education followed-up 2 and
2000; 2005; institutions 5 years after
2009/2010 graduation
National Population Health 1994/1995 National household based sample of 17, Follow-ups every
Survey (NPHS) 276 respondents age 12 and over 2 years
US studies
Early Childhood 20012002 Nationally representative sample of 14,000 Birth through
Longitudinal Study, Birth children born in 2001 kindergarten entry, 4
Cohort (ECLS-B)a waves
Early Childhood 19981999 Diverse sample of 22,000 kindergartners Five waves spanning
Longitudinal Study with data from child, parent, teacher, K to 8th grade
Kindergarten Cohort administrator, trained evaluators
(ECLS-K)a
National Longitudinal 1979 Nationally representative sample of 12,686 Followed up
Survey of Youth 1979 young men and young women aged 1421 annually,
(NLSY79)a i.e. eight age cohorts. Children of female
members of the cohort added from 1986
and followed up every 2 years
National Longitudinal 1997 Nationally representative sample of young Followed up
Survey of Youth 1997 people aged 1216 years i.e. ve cohorts. annually
(NLSY97)a Sample boost to increase representation of
black and Hispanic youth
Health and Retirement 1992 Nationally representative samples of Followed up every
Survey (HRS)a individuals and their spouses aged 2 years. Several
5161 years Merger with the AHEAD companion studies
survey sample aged 70 years and over in in other countries
1998 (China, Japan,
England, etc.)
National Longitudinal 19941995 Nationally representative cohort Followed up at
Survey of Adolescent Health comprising 27,000 individuals aged 1 year then rising to
(Add Health)a 1318 in a stratied random sample of 5 then 6 year
schools including a core sample of intervals
12,000 and a number of boost samples
based on questionnaire returns
Panel Survey of Income 1966 Nationally representative sample of 22,000 Followed up
Dynamics (PSID)a households and a boost sample of high originally Annually
enumeration districts with large non-white and now at 2 year
populations intervals.
(Companion studies
in other countries)
Fragile familiesa 19982000 Representative sample 5,000 individuals Followed up at ages
and their parents and siblings born in 75 1, 3 5
hospitals in 20 cities to families in which
three quarters were non-marital at the
time of the rst survey
(continued)
54 J. Bynner

Table 1 (continued)
Country, region of study Year initiated Sample Follow-up ages
Youth development studya 1988 Representative sample of 1,139 students Follow-up annually
attending St. Paul, Minnesota Public or biannually, with
Schools. Includes surveys of their parents 19 waves from the
and 422 adolescent children ages of 1415 to
3738
Monitoring the Future 1976-current Representative sample of 2,400 12th Followed up every
(MTF)a graders followed each year, selected from 2 years until age 35
larger school based survey and then every
5 years thereafter
National study of families 19871988 Primary adult respondent selected from a Followed up
and households cross-section of 9,637 households plus an 19921994 and
oversampling from certain racial/ethnic again in 20012002
groups and family types (N = 13,007)
National Longitudinal Study 1972 School-based, nationally representative Followed-up in
of the High School Class of study of 12th graders 1973, 1974, 1976,
1972 (NLS-72) 1979, and 1986
High School and Beyond 1980 School-based, nationally representative Followed up every
(HS&B) sample of 10th and 12th graders 2 years until 1986;
10th grade cohort
also followed-up in
1992
National Education 1988 School-based, nationally representative Followed-up in
Longitudinal Study of 1988 sample of 8th graders 1990, 1992, and
(NELS:88) 2000
Educational Longitudinal 2002 School-based, nationally representative Followed up in
Study of 2002 (ELS:2002) sample of 10th graders 2004; 2006; 2012
Midlife Development in the 1995/1996 National sample, ages 2574 with Followed up in
United States (MIDUS) metropolitan oversamples, siblings, and 2004/2006 with
twins MIDUS III planned.
Companion study in
Japan, MIDJA
a
Referred to in the main text

Arbuckle, J. A., & Wothke, W. (19951999). Amos 4 users


References guide. Chicago: SmallWaters Corporation.
Bachman, J. G. (1972). Blueprint for a longitudinal study
Abbott, A. (1995). Sequence analysis: New methods for of adolescent boys (Youth In Transition, Vol. 1). Ann
old ideas. Annual Review of Sociology, 21, 93113. Arbor: Institute for Social Research, University of
Abowd, J. M., & Woodcock, S. D. (2004). Multiply- Michigan.
imputing condential characteristics and le links in Bachman, J. G. (1986). From Youth in Transition to
longitudinal linked data privacy in statistical data- Monitoring the Future: A tale of two longitudinal
bases. Proceedings Annals of the New York Academy studies of youth in the United States. Paper presented
of Sciences, 3050, 290297. at symposium on national longitudinal surveys of
Aisenbrey, S., & Fasang, A. E. (2010). New life for old youth in industrialised countries: Perceptions of fund-
idea: The second wave of sequence analysis: Bringing ing agencies, researchers and administrators, American
the Course back into the life course. Sociological Educational Research Association conference, San
Methods and Research, 38, 420462. Francisco.
Alwin, D. F. (1995). Taking time seriously: Studying Baltes, P. (1987). Theoretical propositions of life-span
social change, social structure and human lives. In developmental psychology: On the dynamics between
P. Moen, G. H. Elder Jr., & K. Lscher (Eds.), Linking growth and decline. Developmental Psychology, 23,
lives and contexts: Perspectives on the ecology of 611626.
human development. Washington, DC: American Baltes, P. (1995). Summary of introductory comments. In
Psychological Association. National Research Council (Ed.), Careers in science
Institutionalization of Life Course Studies 55

and technology (pp. 4142). Washington, DC: Bynner, J. (Ed.). (1993). Use of longitudinal cohort
National Academy Press. studies in the policy process. London: Anglo German
Baltes, P. B., Lindenberger, U., & Staudinger, U. M. Foundation.
(2006). Life span theory in developmental psychology. Bynner, J. (1996). Use of longitudinal data in the study of
In W. Damon & R. M. Lerner (Eds.), Handbook of social exclusion. Report for the Organisation of
child psychology: Vol. 1. Theoretical models of human Economic Cooperation and Development.
development (6th ed., pp. 569664). New York: Wiley. Bynner, J. (2005a). Longitudinal cohort designs. In
Barker, D. J. P. (1992). Fetal and infant origins of adult dis- K. KempfLeonard (Ed.), Encyclopedia of social
ease. London: British Medical Journal MJ Publishing. measurement (Vol. 2, pp. 591599). San Diego:
Bentler, P. M. (1995). EQS program Manual Multivariate Elsevier.
Software, Inc. Bynner, J. (2005b). Challenges for longitudinal research.
Blane, D., Kelley-Irvine, M., DErrico, A., Bartley, M., & Synopsis of ndings from a scoping study. Report to
Montgomery, S. (2013). Social-biological transitions: the UK Joint Centre for Longitudinal Research. http://
How does the social become biological? Longitudinal www.slls.org.uk/#!longview-reports/c8a5
and Life Course Studies, 4, 119135. Bynner, J., & Goldstein, H. (1998). Three generations of
Blossfeld, H.-P. (2006). Globalife: Life course in the glo- children An edited transcript of a video recording
balisation process. Final Report 19992005. University made in March 1982 of the then directors of the three
of Bamberg. major British cohort studies in conversation. Paediatric
Blossfeld, H.-P., & Rohwer, G. (2002). Techniques of and Perinatal Epidemiology: Special Issue in Honour of
event history modelling: New approaches to causal Professor Neville Butler, 12(supplement), 114.
analysis. London: Lawrence Erlbaum Associates. Bynner, J., & Joshi, H. (2007). Building the evidence
Blossfeld, H.-P., Hamerle, A., & Mayer, K. U. (1989). base from longitudinal data. The aims, content and
Event history analysis: Statistical theory and appli- achievements of the British Birth Cohort Studies.
cation in the social sciences. New York: Psychology Innovation the European Journal of Social Science
Press. Research, 20, 159179.
Bowlby, J. (1952). Maternal care and mental health. Bynner, J., Wadsworth, M., Goldstein, H., Maughan, B.,
Geneva: World Health Organization. Purdon, S., Michael, R. (2008). Scientic case for a
Boyle, P. J., Cullis, A., Feng, Z., Flowerdew, R., & Gayle, new birth cohort study. Report to the Research
V. (2004). Adding geographical variables and Resources Board of the Economic and Social Research
identiers to longitudinal Datasets. A Report to the Council and Appendices, http://www.longviewuk.
National Longitudinal Strategy Committee and the com/pages/reportsnew.shtml
UK Economic and Social Research Council. Bynner, J., Erikson, R., Goldstein, H., Maughan, B., &
Brett, C., & Deary, I. (2014). Realising health data linkage Wadsworth, M. (2009). Editorial: Longitudinal and
from a researchers perspective: Following up the Life course Studies, 1, 310.
6-Day sample of the Scottish Mental Survey 1947. Campbell, D. T. (1991). Methods for the experimenting
Longitudinal and Life Course Studies, 5, 283298. society. American Journal of Evaluation, 12, 223260.
Brick, J. M., & Kalton, G. (1996). Handling missing data Clausen, J. A. (1972). The life course of individuals. In
in survey research. Statistical Methods in Medical M. W. Riley, M. E. Johnston, & A. Foner (Eds.), Aging
Research, 5, 215238. and society: A sociology of age stratication (pp. 457
Bronfenbrenner, U. (1979). The ecology of human 574). New York: Russell- Sage Foundation.
development: Experiments by nature and by design. Clausen, J. A. (1986). The life course a sociological per-
Cambridge, MA: Harvard University Press. spective. Englewood Cliffs: Prentice Hall.
Brooks-Gunn, J., Phelps, E., & Elder, G. H. (1991). Cook, T. D., & Campbell, D. T. (1979). Quasi-
Studying lives through time: Secondary data analyses experimentation: Design and analysis issues for eld
in developmental psychology. Developmental settings. Chicago: Rand McNally.
Psychology, 27, 899910. Cronbach, L. J. (1975). Beyond the two disciplines of scien-
Buck, N., & McFall, S. (2012). Understanding society: tic psychology. American Psychologist, 20, 116127
Design overview. Longitudinal and Life Course Dezateux, C. (2014). Life study: Design and vision. Paper pre-
Studies, 3, 517. sented at the conference of the Society for Longitudinal
Burkhauser, R. V., & Lillard, D. R. (2006). The contribution and Life Course Studies, October, Lausanne.
and potential of data harmonization for cross-national Douglas, J. W. B. (1964). The home and the school.
comparative research. Journal of Comparative Policy London: MacGibbon Kee.
Analysis, 7, 313330. Elder, G. H., Jr. (1974). Children of the great depression:
Burkhauser, R. V., & Lillard, D. R. (2007). The expanded Social change in life experience. Chicago: University
cross-national equivalent file. Hilda joins its inter- of Chicago Press. Chicago.
national peers. Australian Economic Review, 40, Elder, G. H., Jr. (1985). Perspectives on the life course. In
208215. G. H. Elder Jr. (Ed.), Life course dynamics:
Bynner, J. (1984). Secondary analysis of National Child Trajectories and transitions, 19681980 (pp. 2349).
Development Study. Report for the UK Economic Ithaca: Cornell University Press.
and Social Research Council. www.cls.ioe.ac.uk/ Elder, G. H., Jr. (1998). The life course and human devel-
Bibliography.aspx? opment. In R. M. Lerner (Ed.), Handbook of child psy-
56 J. Bynner

chology, (Volume 1 665715: Theoretical models of design, UNC Carolina Population Centre. http://www.
human development). New York: Wiley. cpc.unc.edu/projects/addhealth/design
Elder, G. H., Jr., & Giele, J. Z. (Eds.). (2009). The craft of Heinz, W. R. (1991). Status pasages, social risks and
life course research. New York: Guilford Press. the life course. In W. R. Heinz (Ed.), Theoretical
Elder, G. H., Jr., & Shanahan, M. J. (2006). The life course advances in life course research (922). Weinheim:
in human development. In R. E. Lerner (Ed.), Deutscher Studien Verlag.
Theoretical models of human development (The Heinz, W. R. (1996). Status passages and Micro-macro
Handbook of child psychology 6th ed., Vol. 1, linkages in life course research. In A. A. Weymann &
pp. 665715). New York: Wiley. W. R. Heinz (Eds.), Society and biography (pp 5166).
Elder, G. H., Jr., Johnson, M. K., & Crosnoe, R. (2003). Weinhiem: Deutscher Studien Verlag.
The emergence, and development of life course the- Heinz, W. R., & Marshall, V. W. (Eds.). (2003). Social
ory. In J. T. Mortimer & M. J. Shanahan (Eds.), dynamics of the life course. New York: Aldine de Gruyter.
Handbook of the life course (pp. 319). New York: Heinz, W. R., Huinink, J., & Weymann, A. (Eds.). (2009).
Springer. The life course reader: Individuals and societies
Elliot, M. J., & Dale, A. (1999). Scenarios of attack: The across time. Frankfurt: Campus Verlag.
data intruders perspective on statistical disclosure risk. Hofer, S. M., & Piccinin, A. M. (2009). Toward an integra-
Netherlands Ofcial Statistics, Spring, pp, 610. tive science of lifespan development and aging. Journals
Elliot, M. J., Skinner, C. J., & Dale, A. (1998). Special of Gerontology: Psychological Sciences, 65B, 269278.
uniques, random uniques and sticky populations: Hofman, H., Jaddoe, V. W. V., Machenbach, J. P., et al.
some counterintuitive effects of geographical detail (2004). Growth development from early fetal life until
on disclosure risk. Research Ofcial Statistics, 1, young adulthood: The Generation R study. Paediatric
5368. Epidemiology, 18, 6172.
Fergusson, D. M., Horwood, L. J., Shannon, F. T., & Husn, T., Emanuelsson, I., Fgerlind, I., & Liljefors, R.
Lawton, J. M. (1989). The Christchurch developmental (1969). Talent, opportunities and career: A twenty six
atudy: A review of epidemiological ndings. Paediatric years follow-up of 1500 individuals. Stockholm:
and Perinatal Epidemiology, 3, 302325. Almqvist & Wisell.
Ferri, E., Bynner, J., & Wadsworth, M. E. (Eds.). (2003). Janson, C.-G. (Ed.). (2000). Seven Swedish longitudinal
Changing Britain, changing lives: Three generations studies in the behavioural sciences. Stockholm:
at the turn of the century. London: Institute of Swedish Council for Planning and Coordination of
Education Bedford Way Papers. Research.
Friedman, V. (2008). Data visualization and infographics. Jreskog, K. G., & Srbom, D. (1979). Advances in factor
Graphics, Monday Inspiration, 14th Jan 2008. analysis and structural equation models. Cambridge,
Furu, M. (2000). The Malmo Study. In C.-G. Janson (Ed.), MA: Abt Books.
Seven Swedish longitudinal studies (pp. 4475). Juster, F. T., & Suzman, R. (1995). An overview of the
Stockholm: Swedish Council for Planning and health and retirement study. Journal of Human
Coordination of Research. Resources, 30(Supplement 1995), S7S56.
Giele, J. Z., & Elder, G. H., Jr. (1998). Life course Kogevinas, M., Andersen, A.-M., & Olsen, J. (2004).
research: Development of a eld. In J. Z. Giele & Collaboration is needed to co-ordinate 33 European
G. H. Elder (Eds.), Methods of life course research: birth cohort studies. International Journal of
Qualitative and quantitative approaches (pp. 527). Epidemiology, 33, 11721173.
London: Sage. Kohn, M. (1987). Cross-national research as an analytic
Golding, J., Birmingham, K. E., & Jones, R. W. (2009). A strategy. American Sociological Review, 52, 713731.
guide to undertaking a birth cohort study: Purposes, Kuh, D., & Ben Shlomo, Y. (Eds.). (1997). A life course
pitfalls and practicalities. Paediatric and Perinatal approach to chronic disease epidemiology: Tracing
Epidemiology, 23(Issue Suppl 1), 123. the origins of ill-health from early to adult life. Oxford:
Goldstein, H. (1979). The design and analysis of longitu- Oxford University Press.
dinal studies. London: Academic. Kummer, H. (2011). The Swiss etiological study of adjust-
Goldstein, H. (2009). Handling attrition and non-response ment and mental health: Anatomy of a failure of
in longitudinal data. Longitudinal and Life Course research project. www.sulm.ch/pipette_magazin//
Studies, 1, 6372. pipette_1-2011_S9-13
Goldstein, H. (2011). Multilevel statistical models (4th Lerner, R. M. (Ed.). (2006). Developmental science,
ed.). Chichester: Wiley. developmental systems and contemporary theories of
Halfon, N., Larson, K., Lu, M., Tullis, E., & Russ, S. human development. In R. E. Lerner (Ed.) Handbook of
(2014). Life course health development: Past, present child psychology, volume 1: Theoretical models of
and future. Maternal and Child Health Journal, 18, human development. New York: Wiley.
344365. Lerner, R. M., Fisher, C. B., & Weinberg, R. A. (2000).
Harris, K. M., Halpern, C. T., Whitsel, E., Hussey, J., Applying developmental science in the 21st century:
Tabor, J., Entzel, P., & Udry, J. R. (2009). The national International scholarship for our times. International
longitudinal study of adolescent health: Research Journal of Behavioural Development, 24, 2429.
Institutionalization of Life Course Studies 57

Leuze, K., Ludwig-Mayerhofer, W., & Solga, H. (2011). Modell, J. (1989). From youth to adulthood in America
The German national education panel study: A wealth 19201975. Berkeley: University of Chicago Press.
of potential for research in school-to-work transitions. Modell, J., Furstenberg, F. F., & Hershberg. (1976). Social
Longitudinal and Life Course Studies, 2, 346355. change and transition to adulthood. Journal of Family
Levy, R. L., & Widmer, E. D. (2013). Gendered life courses: History, 1, 732.
Between standardization and individualization. Mortimer, J. T. (2012). The evolution, contributions, and
Zurich: Lit Verlag GmbH & co. KG Wien. prospects of the youth development study: An investi-
Magnusson, D. (1993). Human ontogeny a longitudinal gation in life course social psychology. Social
perspective. In D. Magnusson & P. Casaer (Eds.). Psychology Quarterly, 75, 527.
Longitudinal research on individual development: Mortimer, J. T., & Shanahan, M. J. (Eds.). (2003). Hand
Present status and future perspectives. Cambridge: book of the life course. New York: Springer.
Cambridge University Press. Mthen, B., & Mthen, L. (2014). MPlus users guide.
Magnusson, D., & Bergman, L. R. (1990). Data quality in www.statmodel.com/ugexcerpts.shtml
longitudinal research. Cambridge: Cambridge Neugarten, B. L. (1996). The meanings of age: Selected
University Press. papers of Bernice L. Neugarten. Chicago: The
Magnusson, D., & Casaer, P. (1993). Longitudinal University of Chicago Press.
research on individual development: Present status Norwegian Institute of Public Health. (2005). Mother and
and future perspectives. Cambridge: Cambridge child cohort study. www.fhi.no/default0.asp
University Press. Olsen, J., Melbye, M., Olsen, S. F., et al. (2001). The
Magnusson, D., Bergman, L. R., Rudinger, G., & Trestad, Danish National Birth Cohort its background
B. (Eds.). (1991). Problems and methods in longitudi- structure and aim. Scandinavian Journal of Public
nal research. Cambridge: Cambridge University Press. Health, 29, 300307.
Marshall, V. W., & Mueller, M. M. (2003). Theoretical Pembrey, M. (2004). Genetic epidemiology: Some special
roots of the life-course perspective. In W. R. Heinz & contributions of birth cohorts. Paediatric and Perinatal
V. W. Marshall (Eds.), Social dynamics of the life Epidemiology, 18, 37.
course. New York: Aldine de Gruyter. Phelps, E., Furstenburg, F. F., & Colby, A. (2002). Looking
Martin, J., Bynner, J., Kalton, G., Boyle, P., Goldstein, H., at lives: American longitudinal studies of the twentieth
Gayle, V. Parsons, S., & Piesse, A. (2006). Strategic century. New York: Sage Foundation.
review of panel and cohort studies: Report to the Pickles, A., Maughan, B., & Wadsworth, M. (2007).
research resources board of the economic and social Epidemiological methods in life course research.
research council and Appendices, www.longviewuk. Oxford: Oxford University Press.
com/pages/publications.shtml Reichman, N. E., Teitler, J. O., Garnkel, I., &
Martin, M. J., Conger, R. D., Schoeld, S. J., et al. (2010). McLanahan, S. (2001). Fragile families: Sample and
Evaluation of the interactionist mode of socioeco- design. Children and Youth Services Review, 23,
nomic status and problem behavior: Developmental 303326.
cascade across generations. Development and Riley, M. W. (1988). On the signicance of age in sociology.
Psychopathology, 22, 695713. In M. W. Riley (Ed.), Social change and the life
Mayer, K. U. (1993). The German experience in longitu- course, volume 1: Social structures and human lives
dinal population studies. In J. Bynner (Ed.), The use of (pp. 2446). New York: Sage.
longitudinal cohort studies in the policy process. Robins, L., & Rutter, M. (Eds.). (1990). Straight and
London: Anglo German Foundation. devious pathways from childhood to adulthood.
Mayer, K. U. (2001). The sociology of the life course and life Cambridge: Cambridge University Press.
span psychology Diverging or converging pathways. Rubin, D. B. (2004). Multiple imputation for non-response
Paper presented to the Society for the Study of Human in surveys. New York: Wiley.
Development conference. Ann Arbor, Michigan. Scheuch, E. K. (1990). The development of comparative
Mayer, K. U., & Schoepin, U. (1989). The state and the research. In E. Oyen (Ed.), Comparative methodology
life course. Annual Review of Sociology, 15, 187209. (pp. 1937). London: Sage.
Mayer, K. U., & Tuma, N. B. (Eds.). (1990). Event history Schneider, W., & Edelstein, W. (1990). ESF inventory of
analysis in life course studies. Madison: University of European longitudinal studies in the behavioural and
Wisconsin. medical sciences. Berlin: Max Planck Institute for
McGonagle, K., Schoeni, R. F., Sastry, N., & Freedman, Education and Human Development.
V. A. (2012). The panel study of income dynamics: Schoon, I. (2006). Risk and resilience. Adaptations in chang-
Overview, recent innovations and potential for life ing times. Cambridge: Cambridge University Press.
course research. Longitudinal and Life Course Studies, Schulenberg, J., & Schoon, I. (2012). The transition to
3, 268284. adulthood across time and space. Longitudinal and
Michael, R., & OMuircheartaigh, C. (2006). Design pri- Life Course Studies, 3, 164172.
orities and disciplinary perspectives: The case of the Sears, R. R. (1977). Sources of life satisfaction of the
National Childrens Study. Journal of the Royal Terman gifted men. American Psychologist, 32,
Statistical Society A, 171, 465480. 119128.
58 J. Bynner

Sewell, W. H., Hauser, R. M., Springer, K. W., & Hauser, historical context of the British birth cohort studies.
T. S. (2001). A review of the Wisconsin longitudinal London: Routledge.
study 19572001. Madison: Department of Sociology, Wagner, G. G., Frick, J. R., & Schpp, J. et al. (2006).
University of Wisconsin-Madison. Enhancing the power of panel studies- the case of
Shanahan, M. J., & Porfelli, E. (2002). Integrating the life the German socio-economic panel study (SOEP).
course and life span. Formulating research questions Paper presented to the StatsCanada-Conference,
with duel points of entry. Journal of Vocational Longitudinal Social Surveys in an International
Behaviour, 61, 398406. Perspective, Montreal.
Silva, P. A., & Stanton, W. R. (1996). Child to adult. Werner, E. E. (2004). What can we learn about resilience
Oxford: Oxford University Press. from large scale longitudinal studies? In Handbook of
Sontag, L. W. (1971). The history of longitudinal research: resilience in children. New York: Kluwer Academic/
Implications for the future. Child Development, 42, Plenum Publishers.
9871002. Williams, S. (1997). Women and childbirth in the twentieth
Standing, G. (2011). The precariat: The new dangerous century. Stroud: Sutton.
class. London: Bloomsbury Academic. Wolfson, M. C. (2006). Thoughts on the nature of big
Thomas, W. L., & Znaniecki, F. (19181920). The Polish science. Paper presented to the StatsCanada
peasant in Europe and America (Vols. 12). Urbana: Conference, Longitudinal Social Surveys in an
University of Illinois Press. International Perspective, Montreal.
Thornbery, T. P. (2009). The apple doesnt fall very far Wu, L. L. (2000). Some comments on sequence analysis and
from the tree (or does it?): Intergenerational patterns optimal matching methods in sociology: Review and
of antisocial behaviour. Criminology, 47, 297325. prospect. Sociological Methods and Research, 29, 4164.
Wadsworth, M. (1991). The imprint of time: Childhood Zentralstelle fur Psychologische Information and docu-
history and adult Life. Oxford: Clarendon. mentation (ZIPD). (1994). Inventory of European lon-
Wadsworth, M. E. J., & Bynner, J. (Eds.). (2011). A gitudinal studies in the behavioural and medical
companion to life course studies: The social and Sciences, Update 19901994. University of Trier.
Age, Cohorts, and the Life Course

Glen H. Elder, Jr. and Linda K. George

The bond between age and time is central to a from the United Kingdom and its extraordinary
theoretical understanding of the life course and array of national longitudinal cohort studies, each
its foundational traditions, sociocultural and dened by birth in a single year: 1946, 1958, 1970,
cohort-historical (Elder 1975). In the sociocul- 2000, and 2011. Systematic comparison of such
tural perspective, age distinctions are expressed ordered cohorts in the rapidly changing society
as social expectations regarding the timing of has yielded valuable information for policy mak-
events and social roles, whether early, on time, or ers. It is important to note that a particular social
late. As normative age grades from childhood to change may inuence different subsets of a cohort
old age, these age groups constitute a basis for differentially according to social class, gender, and
self-other denition and evaluation as exempli- race-ethnicity. Its impact also tends to vary across
ed by the process of leaving childhood for the successive cohorts according to their career stage
transition to young adult status. From a cohort- at the time of social change. Thus age-related
historical perspective, chronological age as birth changes both across and within cohorts inform our
year locates individuals in historical context and understanding of the intersection of history and
time through membership in a particular cohort, personal biography.
such as the Americans born during the rst The past half century has witnessed major
decade of the twentieth century. advances in theorizing about cohorts as a way of
Unlike normatively dened age strata, birth understanding the biographical impact of social
cohorts are not socially recognized or specied, change and in studying the change process itself.
although they may develop a shared mentality. Conceptual elaborations of the diverse meanings
Cohorts may be dened by historical markers of of age and the exponential growth of longitudinal
social change, for example, or simply by available studies have contributed to this advance and to
data, as in a longitudinal study of people born in a the emergence of the life course as theory and
particular year. An example of the latter comes eld of study. By observing peoples lives over
many years, we have gained a deeper appreciation
of their variation in relation to a changing world.
G.H. Elder, Jr. (*)
In Part I of this chapter, we highlight the major
University of North Carolina at Chapel Hill,
Chapel Hill, NC, USA contributions to such developments.
e-mail: glen.elder@unc.edu These contributions provide a framework for
L.K. George taking up the two complimentary approaches to
Duke University, Durham, NC, USA cohort studies, an inter-cohort perspective that

Springer International Publishing Switzerland 2016 59


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_3
60 G.H. Elder, Jr. and L.K. George

provides a window to historical and social 1.1 Early Contributions


change, and an intra-cohort approach. Part II
examines inter-cohort studies, using a systematic Although the denition of cohort is straightfor-
survey of the literatures contributions toward ward, the history of this concept in the behavioral
understanding social change and the dynamic and social sciences in general, and in life course
transformations of broad social contexts within theories in particular, is less so. It is fascinating
which lives unfold. The studies in this survey are that two watershed explications of the concept
designed to indicate whether and how society is were published in the same year. Norman Ryders
changing, and do not typically relate social now-classic paper, The Cohort as a Concept in
change to the life course. In Part III, a more the Study of Social Change, was published in
focused explanatory approach in cohort research the American Sociological Review in 1965. That
is used to reveal the distinctive contribution of same year, K. Warner Schaies pioneering article,
intra-cohort studies. It demonstrates the ways in A General Model for the Study of Developmental
which social change can have very different con- Problems, appeared in Psychological Bulletin.
sequences for specic subgroups within These two journals were and are the agship
cohorts and their life course. The integration of journals of sociology and psychology respec-
inter-cohort and intra-cohort studies, as well as tively. Because these two articles set the stage for
directions for future research, are addressed in theory and research on cohorts and remain foun-
Part IV. dational for contemporary studies, we review
them in some detail.
Ryders primary purpose was to identify the
1 Conceptual Issues processes by which social change occurs and he
and Perspectives proposed that cohorts are the vehicles that bring
about social change. In his words, Cohorts do
A cohort is a group of individuals who experi- not cause change; they permit it. If change does
enced an event of interest at the same time. occur, it differentiates cohorts of individuals
Cohorts can be dened on the basis of any kind of from one another, and comparison of their careers
event (e.g., entering college, becoming a parent). or life trajectories then becomes a way to study
In the social and behavioral sciences, unless oth- change (1965: 845). Ryder not only posits that
erwise specied, cohort generally refers to a birth cohorts can trigger social change, but that it is
cohort to individuals born at the same or approx- virtually axiomatic that they do. The cohort
imately the same time. In this chapter, cohorts record is not merely a summation of a set of indi-
will refer to birth cohorts. Birth cohorts share one vidual histories. Each cohort has a distinctive
characteristic that is not typical for cohorts based composition and character, reecting the circum-
on other events. Members of birth cohorts are the stances of its unique origination and history
same or approximately the same age. This is (1965: 844).
untrue for the vast majority of cohorts based on Ryder emphasized that the nature and amount
other events. In a cohort of individuals who of social change ushered in by cohorts differ
become rst-time parents in the same year, for widely. He identied ve cohort characteristics
example, cohort members will vary substantially that play a role in determining the content and
in age probably covering an age range of 20 or size of social change that occurs as a result of
more years. This inherent link between cohort and cohort differentiation. First, and as implied
age for birth cohorts is the rst in a number of above, cohorts that experience major social dis-
complicating factors that arise when attempting to ruptions during adolescence and young adult-
accurately attribute differences across cohorts or hood will generate more social change than
across age to specic processes. cohorts that come of age during times of stability.
Age, Cohorts, and the Life Course 61

Thus, events such as wars and major changes in of technological innovations is targeted dispro-
economic conditions, such as the Great portionately at younger members of society. Thus
Depression, tend to generate large amounts of age interacts with social conditions to make ado-
social change. Second, cohort size will affect the lescence and young adulthood the focal point for
nature and volume of social change, with larger the history of a cohort and the nature of its dif-
cohorts not only ushering in more change but ferentiation from earlier and later cohorts.
also having the numerical power to imprint that A few of Ryders other observations are perti-
change more rmly on the rest of society com- nent. First, although Ryder mentioned the con-
pared to smaller cohorts. Third, immigration has founding of cohort, age, and period, he devoted
the potential to increase cohort differentiation. very little attention to methods of statistically
The number of immigrants who enter a society estimating their effects. Second, Ryder never
matter, with large numbers of immigrants arriv- used the phrase cohort effect, although that has
ing within a relatively short time creating more become the label of choice for understanding
cohort differentiation than smaller numbers of how cohorts differ. Instead, he used the term
immigrants. Ryder also suggested that internal cohort differentiation, which seems to be a
migration from rural to urban areas is compara- more accurate descriptor. Third, Ryder was
ble to immigration from other countries. Fourth, keenly aware of the substantial variability within
major technological innovations can create sig- cohorts, stating that Every cohort is heteroge-
nicant differences across cohorts. Finally, social neous (1965: 847). He recognized the impor-
movements consisting mainly of young people tance of studying within-cohort variability for
more signicantly affect cohort differentiation understanding individual lives and that within-
because young adults are more open to changes cohort differences may exceed between-cohort
in the status quo than their older counterparts. differences. For the purposes of studying social
Ryder believed that adolescents and young change, however, Ryder rmly believed that
adults constitute major agents of social change in between-cohort differences were the critical
society. He reviewed a number of studies indicat- evidence.
ing that historical events experienced during At the same time that Ryder was exploring
young adulthood have lasting effects on cohort appropriate methods for studying social change,
members attitudes and behaviors and that these K. Warner Schaie was grappling with the identi-
effects are stronger than those exhibited by older cation problem in age-period-cohort designs
and younger cohorts who also experienced the and the implications of age effects for develop-
historical events. He also noted that the factors mental science. Schaie endorsed Kessens deni-
that create more social change tend to be experi- tion that a characteristic is said to be
enced by and have greater consequences for developmental if it can be related to age in an
young adults than for other age groups. orderly or lawful way (cited in Schaie 1965: 93).
Wars mobilize the armed forces in which the Schaies contribution to studying developmental
vast majority of soldiers are young men (and now change was the recognition that orderly and
young women). At the time he wrote this article, lawful age changes need to be isolated from
wars also affected young women by delaying cohort and period differences.
marriage and childbirth and, in some cases, To achieve this, he proposed the now-familiar
reducing the number of men available for family grid in which age, cohort, and time of measure-
formation. The majority of immigrants and inter- ment are arrayed in its rows, columns, and diago-
nal migrants from rural to urban areas are young nals. Most of the paper focuses on what Schaie
adults. With regard to cohort size, it is during calls three sequential forms of research design
adolescence and young adulthood that members and decision rules to use in identifying age
of very large cohorts become painfully aware of changes that have been disentangled from cohort
the mismatch between their numbers and open- and time of measurement or period differences.
ings in the labor market. And lastly the marketing The three sequential research designs are labeled
62 G.H. Elder, Jr. and L.K. George

cohort-sequential, time-sequential, and cross- Concepts of the life course and life span were
sequential. They differ only in the placement of undeveloped at the time Ryder and Schaie had
age, cohort and time of measurement in the col- their papers published, but they made substantial
umns, rows, and diagonals of the grid. The deci- contributions to the role of cohorts in life course
sion rules involve estimating the effects of age, and life span development. Ryders thinking
cohort, and time of measurement in two of the about cohorts inuenced Matilda White Rileys
three sequential designs and using the patterns of conceptualization of cohorts and their bio-social
results to estimate the extent to which the distri- processes in aging and the life course, as repre-
bution of the dependent variable of interest sented by Aging and Society (with Johnson and
reects those three competing forces. Foner 1972). Later on she developed greater
Today it is generally recognized that there is awareness of the structural lag between social
no way to totally disentangle age, cohort, and institutions and the life course of aging cohorts.
time of measurement effects (Yang and Land In the late 1970s, Riley organized and chaired a
2013). However a number of statistical strategies Social Science Research Council committee on
permit estimates of those three forces if the inves- the life course and human development.
tigator is willing to make an assumption that During the late 1960s and 1970s, Schaie
frees one degree of freedom for estimation, with played a central role in promoting an analytic
the nature of that assumption differing across framework for studying trajectories of aging
strategies. Schaies specic strategy is not popu- across birth cohorts. He did so in part as director
lar now, but his depiction of the confounding of of a training program in life-span development at
age, cohort, and period or time of measurement is West Virginia University that included postdoc-
accurate and compelling. Most importantly, toral students who became inuential gures in
Schaie demonstrated that longitudinal studies are the cohort study of human development and
not the panacea for solving this problem that aging. One of these students, Paul Baltes, became
many investigators had assumed in the 1960s. a leading gure in the emerging life-span study
Ryder and Schaie approached the issue of of cohort effects in developmental psychology
cohorts from very different disciplinary perspec- (Baltes et al. 1979).
tives. For Ryder, cohort differences are the key to
investigating social change. For Schaie, it was
important to examine the effects of age with 1.2 The Challenges of Heterogeneity
cohort differences taken into account i.e., to and Context
substantiate that relatively uniform age changes
are observed regardless of when individuals were Two issues of theoretical and analytic signi-
born or when they were measured. Ryders arti- cance became major themes in the social and
cle is primarily conceptual; Schaies method- behavioral sciences during the 1970s and 1980s
ological. But there are substantial similarities as and remain so: heterogeneity and context. Social
well: both authors wanted to isolate one of three and behavioral scientists were searching for ways
processes that are confounded in nature. They to generate more nuanced and ne-grained
shared the view that all three processes are pow- understanding of human attitudes and behaviors.
erful, thus raising the ante for disentangling their The call for increased attention to heterogeneity
effects. Both were interested in aggregate phe- came from many sources, from a new focus on
nomena as a population-specialist in sociology, gender and socioeconomic variability within
Ryder focused more on between-cohort differ- racial and ethnic groups to recognition of
ences and their insights regarding social change heterogeneity in the outcomes of stressful life
in society whereas Schaies career in studying events. Other investigators expressed dissatisfac-
how people age directed his efforts to isolating tion with survey data that provided little or no
age changes that are observed across time and information about the social structural and inter-
cohort succession in the lives of individuals. personal contexts in which individuals live as
Age, Cohorts, and the Life Course 63

well as on their life patterns. Studies began to tal research became focused on variability within
show that time, place, and macro-, mezzo-, and cohorts/age groups.
micro environments had strong effects on human
lives and that those effects had been understudied
and were not well understood. 1.3 The Emergence
Consider the convergence of historical and and Consequences of Life
ecological perspectives in longitudinal studies of Course Theory and Research
birth cohorts. While birth year denes the histori-
cal location of a cohort, its specic socio-cultural Although there were previous and notable
and historical context is derived from an under- attempts to study personal biography and life his-
standing of its social ecology. This interdepen- tories, life course theory and research began in
dence is most obvious in national birth cohort earnest with Elders now classic Children of the
studies which include adults from vastly different Great Depression: Social Change in Life
socioeconomic areas of a large country, such as Experience (1974). In the 40 years since its pub-
the United States. But the most compelling docu- lication, life course research has become a staple
mentation of this convergence comes from Robert of social and behavioral research. It laid the foun-
Sampsons cohort-historical and ecological stud- dation for literally hundreds of studies, spurred
ies (2012) that evolved from the Project on the development and adoption of analytic tech-
Human Development in Chicago Neighborhoods niques suited to testing its principles, and been
during the 1990s. Data were obtained from longi- usefully linked to a multitude of mainstream
tudinal cohorts of young people, community sur- sociological and psychological theories (e.g.,
veys, the assessment of social networks and stress process theory). There is now general con-
systematic social observations of neighborhoods sensus that the life course perspective includes
and local city governance. He found that the cul- ve principles: the principle of life-span develop-
ture and residents of a neighborhood were typi- ment, the principle of agency, the principle of
cally inuenced by the quality of surrounding time and place, the principle of timing, and the
neighborhoods. Nearly all neighborhoods in the principle of linked lives (Elder et al. 2003). Life
study were discovered to be linked across the course research typically does not address all ve
years through the intra-city migration of people. principles but all of them are central to the life
This ow of people, young and old, typically course paradigm.
connected neighborhoods of advantage and dis- Importantly for our purposes, Children of the
advantage, thereby perpetuating their inequality Great Depression (1974/1999) documented the
across the generations. importance of both cohort-historical and devel-
The increased attention to heterogeneity and opmental (or within-cohort) perspectives on
context had relevance to cohort-historical analy- human lives. An important theme of this seminal
ses that sought to identify social change through work was that the social and economic disrup-
cohort differentiation, but it was not a match for tions caused by the Great Depression had very
their powerful implications in developmental sci- different consequences for cohorts of boys and
ence. Once age changes were examined through girls born only 7 or 8 years apart in the 1920s.
the lenses of heterogeneity and context, develop- The Oakland study members were born in 1920
mental scientists quickly recognized that devel- 1921 and thus were old enough when hard times
opmental processes occur in the context of came to assist their parents. The Berkeley study
environments, resulting in substantial heteroge- members were born just before the Great
neity in the processes themselves and their out- Depression and consequently were dependent on
comes. In short, although developmental parents during the most stressful time of family
scientists did not abandon the search for age hardship.
changes that are universal or nearly universal The Berkeley boys were more adversely inu-
across time and place, a majority of developmen- enced by hard times than the Oakland boys, and
64 G.H. Elder, Jr. and L.K. George

their psychological disadvantage extended into understanding of the life course requires an
the adult years. However those who entered mili- emphasis on within-cohort variability.
tary service during or after World War II tended to A caveat about the concept of generation is
recover their self-condence and well-being. The needed at this point. The term generation has
Berkeley girls were protected in the Depression been used in two distinct ways in social and
years by the emotional support and example of behavioral research (Alwin and McCammon
their mothers. In the Oakland cohort, girls in 2003). First and as most frequently used, genera-
hard-pressed families were drawn into household tion refers to statuses in the family (i.e., grandpar-
chores during early adolescence, a time of physi- ents, parents, and children). Research dening
cal and social maturation, and they often lacked generations as positions within the family faces
attractive clothes for peer activities. As a result, some of the same issues that studies of cohorts do.
they were more adversely affected by family hard Observed differences between generations may
times than the boys. Taking gender roles into reect differences based on family roles per se or
account, these cohort differences are consistent they may reect age or cohort differences. Thus,
with Ryders observation about life-stage at time research on family generations confounds four
of social change and its consequences (1965; factors: age, time of measurement, cohort, and
846). As each cohort encounters historical events, family roles. Mannheim (1952) is generally cred-
it is distinctively marked by the career stage it ited with the second meaning of generation. He
occupies. Moreover, these cohort by gender dif- posited that persons born at approximately the
ferences persist well into middle age. same time (i.e., cohorts, although Mannheim did
In addition to cohort and gender differences, not use that term) have the potential to become
large amounts of within-cohort variability also generations in actuality, although most cohorts
were observed. Social class before the Great do not. A cohort becomes a generation in actual-
Depression and the amount of economic depriva- ity when it develops a distinctive conscious-
tion caused by the Depression, for example, gen- ness. Several conditions are required for the
erated differences in adult achievements, roles, development of a distinctive consciousness: (a)
and attitudes within cohorts. The observed a key historical event must occur, (b) the cohorts
within-cohort heterogeneity demonstrated that at-risk of becoming a generation in actuality will
although cohort differences, especially those be young, (c) and the historical event is suf-
generated by signicant social disruptions, have ciently disruptive as to shape the lives of adoles-
important consequences for biography, they are cents and young adults. According to Mannheim,
not equally important for all cohort members. As then, generation formation begins with cohorts,
noted by gender, cohort members are situated in but requires an additional intersection of age and
numerous contexts only one of which is social/historical change. Interesting research
cohort that combine to provide both opportuni- exists that is based on both denitions of genera-
ties and constraints for individual development. tion. This chapter, however, does not include
This chapter is organized around the two major attention to that research. Our focus is on the con-
ways in which cohort is important to life course cept of cohort per se and its links and contribu-
theory and research. The next section reviews the tions to life course theory and research.
contributions that cohort-historical studies have
contributed to our understanding of both social
change and, in broad brush, widespread changes 2 Cohort Differences:
in the life course. The following section reviews A Window to Historical
research that focuses on within-cohort differences and Social Change
and their consequences for personal biography
over the long-term. Cohort is an important con- The primary purpose of studying cohort differ-
textual factor in understanding within-cohort dif- ences is to monitor historical and social changes
ferences, but a more complete and nuanced over time. Some studies of cohort differences are
Age, Cohorts, and the Life Course 65

based on age-period-cohort (APC) analyses; oth- multiple times, few inter-cohort studies are longi-
ers are not. The advantage of APC analyses is tudinal. Cohort studies do not typically use panel
that, to the extent possible, cohort differences are data. Even large panel data sources rarely include
estimated with the potentially confounding inu- sufcient numbers of cohorts and times of mea-
ences of age and period (i.e., time of measure- surement to permit inter-cohort studies of the
ment) taken into account. APC studies are scope needed to observe historical trends.
sometimes criticized as being descriptive rather Consequently, most inter-cohort studies use
than theory-driven and, indeed, many of them repeated cross-sectional surveys, such as those in
focus on describing time trends in phenomena of the General Social Survey and National Health
interest. But this is not necessarily the case and Interview Survey. It is interesting that Ryder
other APC analyses are based on theories of (1965) noted that inter-cohort and intra-cohort
social change and/or test specic theory-based studies would typically rely on different kinds of
hypotheses. longitudinal data in his classic essay.
The links between studies of cohort differ- Another common difference between inter-
ences and the life course perspective are indirect. cohort and intra-cohort studies is the way that
That is, cohort studies are not life course research factors other than the independent variable of pri-
in the sense that investigators set out with the mary interest are conceptualized and analyzed.
objective of applying the ve principles of the By denition, variability in social status, life
life course perspective to the substantive topic of course transitions, and social resources and de-
interest. Inevitably, inter-cohort studies will cits is important in intra-cohort studies. These are
accomplish less in terms of advancing life course precisely the kinds of factors that generate differ-
research than intra-cohort studies, which are ences in life course trajectories. Thus, intra-
addressed in the next section. Nonetheless, inter- cohort studies focus on understanding variability
cohort studies (also termed here simply cohort across individuals and subgroups. The focus is
studies) are not irrelevant to life course research. typically much narrower in inter-cohort studies.
Cohort studies paint a broad picture of social and The goal is to identify differences across cohorts
historical changes that is important for under- and the focus is on differences in percentages or
standing individual lives. As an example, the life averages (i.e., point estimates) across cohorts.
course implications of cohabitation are far differ- APC studies usually control on basic demo-
ent when 3 % of couples cohabit at some point in graphic characteristics that may differ across
the life course than when 40 % of couples cohabit cohorts e.g., gender, race/ethnicity, education.
before their rst marriage and another 10 % Only if cohort studies address a research question
cohabit after divorce. Cohort studies provide that requires comparisons will the relationships
valuable information about the social and histori- between personal characteristics and the outcome
cal contexts in which lives are embedded. of interest be examined. Taking attitudes toward
Inter-cohort studies also differ from intra- traditional gender roles as an example, most APC
cohort studies in important methodological ways studies would examine differences in the percent-
(the terms inter-cohort and intra-cohort will be ages of cohort members endorsing traditional
used when they are being compared). Both inter- gender roles across cohorts, controlling on basic
cohort and intra-cohort studies rely on longitudi- demographic variables. Only if theory suggested
nal data, but the meaning of longitudinal that the level of endorsement or the rate at which
requires specication. If longitudinal data refer attitude change occurred differed for men and
to comparable data collected over multiple times women would gender contrasts be examined.
of measurement, both inter-cohort and intra- Thus, both inter-cohort studies and intra-cohort
cohort studies rely on longitudinal data. If, how- studies pay attention to basic demographic char-
ever, longitudinal data refers to panel data in acteristics, but they are typically of more substan-
which the same sample members are measured at tive interest in intra-cohort studies.
66 G.H. Elder, Jr. and L.K. George

2.1 A Note on Theory the changing social and historical contexts within
and Processes which individual lives unfold.
Another common criticism of APC studies is
A common criticism of APC studies is that they that even when age, period, and cohort are disen-
are not tightly linked to solid social science the- tangled, each remains a black box which offers
ory. A corollary of this critique is that cohort little, if any, explanation for the antecedents or
studies are descriptive rather than explanatory. meanings of those effects. We argue that this is a
There is some truth in these criticisms. The rea- false critique. In fact, the processes or mecha-
sons that cohort studies are less theory-driven nisms that explain cohort differences are strong
than other areas of research merit consideration. relative to other research traditions. Two pro-
First, the extent to which theory guides APC cesses explain virtually all cohort differences in
studies is in part a function of the volume and domains in which cohort differences have been
quality of available theory about social and his- examined: cohort replacement and diffusion.
torical change. With the exception of macro-level These two processes admittedly do not explain
theories focused on social and historical changes why two individuals from the same birth cohort
that are decades in the making such as industrial- respond differently to social change that is the
ization and urbanization, theories about more job of intra-cohort studies. These two processes
modest social and historical changes are quite do, however, explain the occurrence of cohort
rare. Thus, there are few theories on which cohort differences.
studies can rest. Second, at the same time that Cohort replacement, also referred to as cohort
theories focus attention on specic relationships succession, refers to the straightforward fact that
and processes, they deect attention from other cohorts are in a permanent state of ux as old
relationships and processes. In the social sci- cohorts die and new cohorts are born. The role of
ences, a large proportion of theories focus on cohort succession in social change was rst artic-
social reproduction rather than social change. ulated in classic work by Matilda White Riley
There is abundant evidence that social reproduc- (1987). White recognized that social change does
tion occurs in a number of life domains at both not result primarily via changes in attitudes and
the individual and population levels. Indeed, a behaviors in individuals, but rather by the replace-
common theme is this research is to demonstrate ment of cohorts whose members share specic
that despite what appear to be changes, deeper attitudes and behavioral tendencies by new
level structural forces manage to reproduce status cohorts with different attitudes and behavioral
hierarchies and the distribution of valuable preferences. As documented below, if the role of
resources. The commitment to identifying pat- cohort replacement in social change is not care-
terns of social reproduction deects attention fully examined and estimated, attributions about
from theories that focus on social change. the causes of social change are often simply
The extent to which cohort studies are theory- wrong. We will see this type of misattribution
driven, however, is highly variable. Cohort stud- with regard to the Womens Movement and its
ies addressing four domains of life experience are effect on gender egalitarian attitudes. Although
reviewed below. Well-developed theories form the Womens Movement played a role in chang-
the conceptual bedrock for two of these domains. ing attitudes about gender roles, the appearance
The two other domains are based on predictions of increasingly egalitarian attitudes over time
about the amounts and directions of hypothesized was largely a result of cohort replacement in
social and historical change, but are not linked to which cohorts with more traditional gender atti-
solid theoretical frameworks. This variability is tudes died.
representative of cohort studies more broadly. The other primary mechanism resulting in
Despite variability in the extent of theoretical jus- social change is diffusion. Cohort replacement
tication, we argue that cohort studies in all four focuses on the entrance of new and the exit of old
domains produce important information about cohorts, demonstrating that most social change
Age, Cohorts, and the Life Course 67

results from personnel changes rather than techniques for disentangling the effects of age,
within-individual change. In contrast, diffusion period, and cohort. In addition, because of their
focuses on differences within cohorts in the rate recency, ndings from these studies are more
or pace at which social change occurs. Diffusion likely to generate new research questions about
occurs in a sequence, with early adoption of new cohort differences and their meaning.
attitudes or behaviors by cohort members of high
socioeconomic status. Specically, in early years
of observation, SES is strongly and positively 2.2 Secularization: Cohort
correlated with increased support for new atti- Changes in Religious
tudes or behaviors; in subsequent years, however, Participation and Authority
correlations between SES and those attitudes or
behavior weaken or disappear. These relation- A dominant issue in the sociology of religion for
ships weaken because large proportions of the more than half a century has been secularization.
younger cohorts endorse the new attitude or Along with industrialization and urbanization,
implement the new behavior. Thus, persons of secularization is viewed as a key component of
high SES tend to be the avant garde in social the transition from the pre-modern to the modern
change. Over time the attitude changes are dif- world. These social and cultural processes liter-
fused to the broader population. ally transformed every aspect of human life.
To demonstrate what studies of cohort differ- What is secularization? In general terms secu-
ences offer to our understanding of historical and larization is the transformation of societies from
social change, this section provides exemplars of primary dependence on religious values, author-
research on four topics. We rst examine secular- ity, and institutions to non-religious or secular
ization. Because it rests on a well-articulated theo- values, authority, and institutions (Sommerville
retical framework and testable hypotheses, 1998; Taylor 2007). Scholars differ substantially
secularization is arguably the ideal topic for APC in conclusions about the extent to which secular-
analysis. We then turn to two types of demographic ization has penetrated modern and postmodern
change: cohort size, which represents a cohort societies. Three dimensions have received most
characteristic per se and family structure and attention. One dimension focuses on declining
dynamics, which have exhibited rapid changes rates of participation in religious institutions as a
over the past few decades. The nal domain exam- critical feature of secularization (Schwadel 2010;
ined is attitudes, for which we review cohort dif- Sommerville 1998). Rates of membership in reli-
ferences in gender egalitarian attitudes. These gious organizations and regular participation in
topical areas differ in several respects, including religious services declined gradually, but mark-
the consistency of ndings across studies, the edly over the twentieth century in industrialized
strength of cohort differences, the methods used to societies. A second dimension of secularization
detect cohort differences, the extent to which stud- is the decline in the authority that religious insti-
ies rest on solid theoretical foundations, and the tutions have over individual lives and societies
degree to which the investigators attempt to more broadly (Chaves 2011; Voye 1999).
explain observed cohort patterns. Religious leaders are no longer widely viewed as
The summaries of the topical areas are not the nal authorities on either private attitudes and
comprehensive, but illustrative of much larger behaviors or public policies and regulations. The
bodies of research. This is especially true where third dimension of secularization is the increas-
we use a single attitude as exemplar although we ing privatization of religion in the modern world
identied cohort studies of more than 40 specic (Lambert 1999). Scholars subscribing to this
attitudes. We briey trace the origins of research view argue that declines in participation in reli-
in the topical areas with emphasis on research gious institutions are only part of the story, and
published in the twenty-rst century. Recent that commitment to religious values and prac-
studies are typically based on more sophisticated tices remains strong in the non-public sphere.
68 G.H. Elder, Jr. and L.K. George

They also view the increasing proportions of None of the studies of cohort differences in reli-
people who describe themselves as spiritual, but gious service attendance reported signicant age
not religious as evidence of a more private, indi- effects. Thus, contrary to popular belief, there is
vidualized relationship with the sacred. Recent no evidence that individuals become more reli-
APC analyses have addressed these three dimen- gious at least in terms of service attendance as
sions of secularization. they age.
Some scholars have speculated that declining
2.2.1 Religious Service Attendance rates of religious service attendance have not
and Beliefs been accompanied by declines in religious
Without question, attendance at religious ser- beliefs a pattern termed believing without
vices declined precipitously since the rst half of belonging. Most of the studies that examined
the twentieth century in western countries. Two APC differences in attendance also examined
recent APC studies examined attendance data endorsement of traditional religious beliefs (e.g.,
from 1972 to 2006 in the U.S. (Schwadel 2010, belief in the Bible, belief in prayer). In all the rel-
2011) and reported continued decline across evant studies, religious belief declined to the
cohorts in the early years of the twenty-rst cen- same degree that attendance did (Crockett and
tury. Year-to-year cohort differences are quite Voas 2006; Kaufmann et al. 2012; Schwadel
small in absolute terms, but cumulatively they 2011; Voas and Crockett 2005).
add up to a substantial trajectory of declining
attendance. Literally dozens of other studies have 2.2.2 Religious Authority
documented declines in religious service atten- One dimension of religious authority is the faith
dance over much larger numbers of years than the of the public in religious leaders. One APC study
studies by Schwadel. But those studies are not examined condence in the leaders of religious
based on APC analyses and, therefore, do not organizations for the period of 19732010 in the
estimate the separate effects of age, period, and U.S. Condence in religious leaders was substan-
cohort. tially lower during the 1980s than in the years
Declining rates of religious service attendance before and since (Hoffman 2013). During the
are not restricted to the U.S. APC analyses of 1980s, several prominent religious leaders made
British data covering much of the twentieth cen- headlines as a result of adultery, child molesta-
tury show steady declines in religious service tion, and embezzlement of church funds. These
attendance (Crockett and Voas 2006; Voas and scandals were accompanied by fairly dramatic
Crockett 2005). Using data from ten European declines in condence in religious leaders. After
countries from 1981 to 2008, Kaufmann and col- the 1980s, faith in religious leaders leveled off,
leagues report cohort-based declines in atten- albeit at a considerably lower level than before
dance (Kaufmann et al. 2012). They note that the scandals. The analysis suggests that this pat-
declines have been steeper for cohorts born tern was initially a period effect that persisted
before 1945; since then, rates of attendance stabi- over time and came to persistently differentiate
lized in younger cohorts, suggesting that further cohorts exposed to that period effect.
declines (which would result in rates of atten-
dance below 10 %) are unlikely. It is often said 2.2.3 Non-religious Spirituality
that the U.S. is more religious than European Some scholars argue that in the transition from
countries. This is true in that larger percentages the modern to the postmodern world, highly-
of Americans than Europeans attend religious individualized forms of spirituality began to
services once a month or more a consistent pat- replace organized religion as a way for individu-
tern over at least 50 years. But the pattern of als to search for the sacred and for transcendence
declining attendance across cohorts is true of of the material world. A signicant proportion of
both the U.S. and Europe. Cohort replacement Americans, for example, describe themselves as
accounts for declining attendance over time. spiritual, but not religious. With data from 14
Age, Cohorts, and the Life Course 69

western countries, including the U.S. and Canada, Western societies. Indeed, the term baby boom
Houtman and Aupers (2007) examined the spread was coined to describe these very large birth
of non-religious spirituality from 1981 to 2000. cohorts. Long before the baby boom, demogra-
Although the time span was only 19 years, a phers worked with population pyramids which
strong and signicant pattern of increased spiri- provided graphic depictions of populations based
tuality was observed. The rate of increase, how- on the age and sex of population members. Thus,
ever, varied across countries. France, Britain, The social scientists had long histories of studying
Netherlands, and Norway experienced the high- age grades. This was a far cry, however, from
est increases; the U.S. and Canada were interme- acknowledging the conceptually distinct, but
diate; and southern Europe had the lowest rate of empirically tangled concepts of age, period, and
increase. Virtually all the differences were cohort. It was the arrival of the baby boom that
explained by cohort replacement. focused scientic and public attention on cohort
Taken together, APC studies tell a coherent size.
story of increasing secularization in developed The scholar most strongly associated with
countries over several decades and hint at the theory and research on cohort size is Richard
emergence of postmodern spirituality. Easterlin. Although he published several papers
Secularization theory is ideal for APC analyses on cohort size during the 1970s, Easterlins the-
because it focuses directly on a form of social ory and research on cohort size was best inte-
change best measured at the population level. grated in his book, Birth and Fortune: The Impact
Cohort differences are overwhelmingly the result of Numbers on Personal Welfare (1980, 1987).
of cohort replacement. There is no evidence that The central thesis of this book is that members of
individuals whose initial commitments were to large cohorts will experience lower well-being
organized religion turn to non-religious spiritual- than those in smaller cohorts, due primarily to
ity at older ages. Secularization is not the result of overcrowding in large cohorts. Cohort size is
aging. Rather it is the turnover in cohorts compris- expected to affect multiple dimensions of well-
ing the population that accounts for its expansion. being, including job stability, earnings, marital
Questions remain about the dynamics of secu- stability, fertility, crime, life satisfaction or hap-
larization. Although the effects of cohort succes- piness, political alienation, and suicide and other
sion on religious participation are strong, it also is self-damaging behaviors. A corollary of
possible that indicators of secularization exhibit Easterlins theory was his insistence that cohort
diffusion. We do not know whether reduced par- size should be measured as relative cohort size.
ticipation in religious services, decreased faith in Measuring relative cohort size is important
religious leaders, and increased participation in because it provides a continuous measure of
privatized religion occurred rst among cohort cohort size (as compared to categorizing cohorts
members of high socioeconomic status and then as large or small), which in turn permits investi-
diffused to the point that SES is no longer associ- gators to determine whether there is a
ated with secularized behaviors and attitudes. dose-response relationship between cohort size
This is a question worthy of investigation, both as and outcomes of interest.
a test of diffusion theory and as a means of gener- A substantial volume of research between the
ating hypotheses about the probabilities of late 1970s and early 1990s examined Easterlins
increased secularization as a result of diffusion. hypotheses about cohort size. Results were
mixed, although most studies supported expecta-
tions that large cohorts would fare worse than
2.3 Demographic Change small cohorts. These studies, however, were not
Associated with Cohort Size based on APC analyses. Thus, they compared
cohorts of different sizes, but did not model or
The baby boom lasted from 1946 to 1964 in the control on age and period effects. After the early
U.S. and at approximately the same time in other 1990s, research on cohort size was relatively
70 G.H. Elder, Jr. and L.K. George

dormant until recently. Earlier studies typically hypothesis. But there was an interesting differ-
examined outcomes with data that spanned the ence across countries as well. The relationship
1950s or early 1960s to the early or mid-1980s. between cohort size and earnings was higher in
The earliest years examined in recent studies are southern European countries than in northern
typically the 1960s through mid-1970s, but countries. Brunillo suggests that this may reect
extend the period of observation to the early size of the welfare state. That is, with their more
2000s. Most recent studies also are based on true generous welfare state benets, Northern coun-
APC analyses. tries may use public policy to reduce the size of
A novel outcome in the cohort size literature is cohort differences.
underemployment. Slack and Jensen (2008) APC studies of cohort size are relatively
examined the associations of cohort size and unique because they focus on a cohort character-
underemployment in the U.S. between 1974 and istic rather than the attitudes or behaviors of
2004. Being labeled as underemployed was based cohort members. The theoretical foundations for
on any of four conditions: being unemployed and studies of cohort size are relatively strong.
having given up after an extended job search; Although the Easterlin hypothesis is simple and
being unemployed and actively looking for work; straightforward it rests on a well-articulated the-
involuntarily working part-time, but desiring full- ory that explains why and how cohort size should
time work; and working, but earning less than affect the life chances of cohort members. An
125 % of the poverty threshold. This paper is advantage of Easterlins theory is that it predicts
based on a true APC analysis and all three factors that cohort size affects a wide range of outcomes.
are signicant. The age effect reveals that across Not all of the predicted outcomes have been
cohorts and times of measurement, young adults examined using APC analysis important topics
have signicantly higher rates of underemploy- for future research. In addition, some compo-
ment than their older counterparts. With regard to nents of Easterlins theory have not been ade-
period, not surprisingly underemployment rates quately studied. He hypothesizes that large and
are highest for all age groups during times of eco- small cohorts will cycle over time. This implies
nomic recession. The cohort effect is as expected: marked changes in fertility behavior across
rates of underemployment are higher for large cohorts. It is not clear whether there is a cycle of
than for small cohorts at all ages and times of cohort sizes over historical time or whether the
measurement. In addition, education interacted Baby Boom was historically-unique. Another
with relative cohort size such that underemploy- unaddressed question is whether the smaller
ment was especially common for members of cohorts born after an unusually large cohort fare
large cohorts with less than a high school educa- as well as small cohorts that were born just prior
tion. This paper illustrates two important meth- to the large cohort. It is possible that the pres-
odological points. First, the authors documented sures exerted by a large cohort result in more
that age, period, and cohort were all signicantly restricted life chances for cohorts following the
related to underemployment. This demonstrates large cohort than were available to equally small
the importance of disentangling APC to the cohorts born before the large cohort. Despite
extent possible. Second, testing for interactions these unanswered questions, studies of the asso-
between individual characteristics and cohort ciations between cohort size and various out-
size yielded signicant and meaningful ndings. comes provide useful information that is unlikely
Recent studies also include research based on to be revealed in any other way. Several scholars
non-U.S. samples. Brunillo (2010), for example, who examined cohort size state explicitly that
examined the relationship between cohort size within-cohort variability accounts for more of the
and earnings using data from 11 European coun- variance in the outcomes than cohort size does.
tries. A signicant negative relationship between But focusing on what variables explain the
cohort size and personal earnings was observed most variance misses the point. Examining the
in all 11 countries, as predicted by the Easterlin opportunities and constraints linked to cohort
Age, Cohorts, and the Life Course 71

size provides unique and useful information tively stable since 1990 (Raley and Bumpass
about, as Easterlin so aptly called it, birth and 2003). These trends are based on times of mea-
fortune. surement rather than birth cohorts and, indeed,
cohorts dened by year of divorce include indi-
viduals from multiple birth cohorts. Nonetheless,
2.4 Demographic Change the odds of divorce differ substantially across
in Family Structure cohorts and cohort differences map well with
and Dynamics period differences, as demonstrated in life table
analyses (e.g., Schoen and Canudas-Romo 2006).
Changes in family structure and dynamics are Note also that although divorce rates have pla-
arguably the most visible societal changes in teaued, marriage rates have declined, thus
recent decades. Much of the public discourse on decreasing the number of divorced individuals in
family changes focuses on cohabitation and non- the population.
marital births. Social scientists also are interested As rates of marriage decreased and rates of
in those issues, but survey the broader landscape divorce increased over time, rates of cohabitation
of family structure including marriage, divorce, increased dramatically, especially since 1980. By
and total fertility. Here we summarize cohort dif- 2010, almost half of American men and women
ferences in the U.S.; similar differences exist in cohabited before marriage; about half of these
all western countries. relationships culminated in marriage. According
Sizeable cohort differences in age of marriage to the National Center for Health Statistics, the
have occurred and changed direction since the percentages of U.S. women cohabiting at the
end of the nineteenth century in the U.S. Age at time of interview increased from 3.0 % in 1985 to
rst marriage between 1890 and 2010 exhibits a 11.2 % by 20062010 (Copen et al. 2012).
U-shaped curve for both men and women. In Similar percentages were reported by men. In
1890, the average age at rst marriage for men addition, the average duration of cohabitation
was 26.4 years. Age at rst marriage decreased to increased from less than 2 years in 1985 to nearly
its lowest level of 24.0 in 1950 and then increased 4 years in 20062010. In 1980 the vast majority
consistently to its highest level of 28.4 in 2010 of cohabitating couples had never been married.
(U.S. Census Bureau 2012). Corresponding ages In the 25 years since then, increasing proportions
for women were 23.5 in 1890, 20.4 in 1950, and of cohabiting couples include one or both part-
26.8 in 2010. The proportions of Americans who ners who were previously married and are from
never marry exhibit the same pattern. Thus recent earlier cohorts than couples for whom cohabita-
cohorts are signicantly more likely than older tion is their rst union. This pattern indicates
cohorts to marry at later ages or not marry at all. diffusion from younger to older cohorts over the
Substantial cohort differences in divorce past 25 years.
emerged over approximately the past 70 years. Easterlin (1980, 1987) hypothesized that fertil-
Historical events are often accompanied by ity occurs in cycles and that boom cohorts will
changes in family structure and dynamics. For be followed by bust cohorts. This hypothesis is
example, divorce rates (already low by todays supported by changes in fertility rates in the U.S.
standards) fell precipitously in the years of the since the baby boom ended. Beginning in the late
Great Depression, then increased substantially at 1960s and early 1970s, the U.S. fertility rate
the end of World War II (Schoen and Canudas- declined continuously until 1990. Between 1990
Romo 2006). Divorce rates fell somewhat in the and 2007, the fertility rate inched upward. Since
1950s and remained relatively stable until the 2007, the fertility rate declined to its historical
1970s. Divorce rates more than doubled during low, probably as a reaction to the Great Recession
the 1970s, increased more slowly in the 1980s, (Livingston and Cohn 2012). One way of describ-
and have continued to be high (about half of ing the size of the fertility rate uctuations is to
American marriages end in divorce), but rela- examine the total number of children born to
72 G.H. Elder, Jr. and L.K. George

women at different times. In 1910, average num- indicators of secularization and cohort size in two
ber of children/woman age 1544 was 3.4. important ways. First, very few studies of chang-
Corresponding numbers are as follows: 1930, 2.2 ing family structure are based on true APC analy-
children; 1955, 3.6 children; 1980, 1.8 children; ses. Thus, most estimates of cohort differences in
2000, 2.1 children; and 2011, 1.8 children (Haines family structure are not as precise or pure as
2008; Livingston and Cohn 2012). In addition to estimates in other domains. Second, theoretical
Easterlins hypothesis of fertility cycles, factors foundations for research on the dynamics of fam-
believed to contribute to the fertility declines ily structure are nearly absent. To the extent that
between 1970 and 2007 include increased avail- investigators speculate about the antecedents of
ability of contraception, legalized abortion, and trends in family structure, discussion is typically
public concerns about overpopulation. post-hoc. One common post hoc interpretation of
Although fertility rates declined considerably dramatic cohort and period differences in family
over the past 50 years, the proportion of non- structure is the tenant that the life course has
marital births to total births increased dramati- become increasingly individualized or deinstitu-
cally. Two studies examined cohort differences in tionalized (Dannefer 2011; Mayer 2009). This
rates of non-marital births (rather than reporting perspective is too ambiguous to be considered a
period changes). England et al. (2012) examined theory, but it implies that heterogeneity in life
rates of pre-marital births for women born course patterns has expanded to the extent that
between 1925 and 1959. The proportions of there is or soon will be no modal life course
women giving birth prior to marriage increased which is recognized by societal members as the
consistently across cohorts. Wu (2008) reported right way to sequence and/or time life
cohort differences in non-marital births by age 30 transitions.
for cohorts born between 1925 and 1969. The The combination of the absence of a theoreti-
proportion of women with non-marital births cal framework and the absence of true APC anal-
increased monotonically over time, increasing yses is especially problematic. For example, as
from 10 % of the 1929 cohort to 25 % of the 1969 noted above, several investigators attribute the
cohort. Note that pre-marital births represent a recent decline in fertility to the Great Recession.
subset of all non-marital births. In more recent There is precedent for this interpretation based
years, increases in non-marital births have been on the very low U.S. fertility rates during the
even larger. In 1990 28 % of all births were non- Great Depression. Nonetheless, this interpreta-
marital births; by 2010, non-marital births com- tion, while offered as an explanation for cohort
prised 41 % of all births. differences, is logically more likely to represent a
Without question, family structure changed period effect. True APC analyses are needed to
considerably since the early 1900s and this is disentangle the effects of age, period, and cohort
especially true in the more recent past. Although and, thus, permit more credible explanations.
cohort differences in family structure are dra- Some cohort differences in family structure
matic, they do not capture within-cohort differ- appear to represent diffusion. Recall that in diffu-
ences. Each source cited above makes the point sion, changes in attitudes or behavior occur ini-
that the overall cohort or period differences tially among persons of high socioeconomic
masked signicant within-cohort differences status. Then, as changes became widespread in
associated with education, race/ethnicity, and the population, the relationship of SES with atti-
nativity. With regard to the latter, for example, tudes weaken or disappear. With regard to family
women immigrants have the highest fertility rates structure, cohort differences in cohabitation fol-
in the U.S., although nativity differences have lowed a similar pattern (see Bhrolchain and
narrowed somewhat over the past decade Beaujouan 2013 for a careful analysis of diffu-
(Livingston and Cohn 2012). sion of cohabitation in Britain). Other changes in
Overall, cohort differences in family structure family structure, however, appear to involve dif-
and dynamics differ from those observed for fusion from younger to older cohorts. This is
Age, Cohorts, and the Life Course 73

especially true of cohort differences in non- 2.5.1 Egalitarian Gender Roles


marital pregnancies, which were initially nega- In response to the Womens Movement in the
tively related to SES. Recent increases in 1970s, social scientists consistently hypothesize
non-marital fertility, however, have occurred dis- increases over time in attitudes favoring egalitar-
proportionately among higher SES women ian gender roles. Related questions include how
(although non-marital fertility remains more rapidly and widespread changes were and
prevalent among low SES women). The condi- whether the move toward egalitarianism differed
tions under which diffusion is from higher status across demographic subgroups. Several studies
to lower status population groups vs. those under report increased egalitarianism; each also sheds
which diffusion is from lower status to higher sta- light on somewhat different issues and used dif-
tus groups merits both theoretical and empirical ferent periods of observation.
attention. There is substantial evidence that support for
egalitarian gender roles increased substantially in
the U.S. since the mid-1970s based on true APC
2.5 Cohort Differences in Gender analyses. Pampel (2011a) reported this based on
Role Attitudes repeated cross-sectional data obtained between
1977 and 2006 from 86 birth cohorts. Cotter et al.
The public has long-standing interest in the atti- (2011) reported the same general pattern using
tudes of societal members and the extent to which repeated cross-sectional data from 1974 through
attitudes differ for population subgroups. 2008, although there was no signicant change in
Politicians frequently quote opinion polls, attitudes toward gender roles between 1994 and
which measure attitudes, to show that the public 2008. They suggested that the Womens
favors their political platforms. Attitudes toward Movement generated signicant attitude changes
products and services are an essential part of across cohorts during its aftermath (Rossi 1985),
marketing research. And, of course, attitude for- but that more recent cohorts simply adopted the
mation and change are signicant issues for attitudes of their parents cohorts. Cotter and col-
social and behavioral scientists. In all of these leagues also document that half of the change in
venues, it is commonplace for attitudes to be attitudes observed over the 29 years of observa-
compared across age groups and in many tion resulted from cohort replacement. That is,
instances public attitudes and other phenomena half of the increased support for egalitarian gen-
are monitored over time to determine trends. der roles resulted from the deaths of older cohort
Age differences and time trends are often members who endorsed more traditional gender
explicit concerns in studies of attitudes, but the attitudes.
term cohort is almost never raised in public dis- Similar patterns of attitude change have been
course and is often not addressed in scientic observed in non-U.S. samples. Shu and Zhu
studies. Yet cohort differences are plausible alter- (2012) examined two specic gender role atti-
nate explanations to presumed age or period dif- tudes using repeated cross-sectional data from
ferences. All comparisons across age groups are China for the period of 19952007. For one atti-
also comparisons across cohorts, raising the pos- tude that it is appropriate for women to com-
sibility that these are cohort differences that have bine family and work roles there were no age,
nothing to do with age. Similarly, time trends period, or cohort effects because virtually all
may be period effects; but it also is possible that study participants strongly endorsed this attitude.
they reect the process of cohort replacement. Cohort differences were observed for a more
Thus, examining cohort differences in attitudes is general attitude endorsing gender equality in all
essential to permitting valid attributions about the aspects of life. More recent cohorts endorsed this
causes of historical change. attitude in greater proportions than older cohorts.
74 G.H. Elder, Jr. and L.K. George

In addition, within cohorts, education, income, consciousness-raising which the Womens


and occupational prestige were positively related Movement encouraged was important, but no
to more egalitarian attitudes. more important than cohort replacement.
Using 14 waves of data from the mid-1980s to
the early 2000s, Kraaykamp (2012) examined
attitudes toward gender egalitarianism in Norway. 2.6 Final Thoughts on Inter-
As in other countries, younger cohorts endorsed cohort Differences
egalitarian family roles more than older cohorts.
In addition, working women were signicantly Studies of cohort differences at the population
more likely to endorse egalitarianism than non- level are valuable because they reveal the overall
working women, although this difference nar- landscape of a society. APC analyses make it
rowed in recent cohorts. possible to estimate the inuence of aging, time
Finally, Pampel (2011b) examined support for of measurement, and cohort differences and each
egalitarian gender roles from 1988 to 2002 using is estimated with the other two effects taken into
data from 19 nations that included 84 cohorts. account. Cohort differences in particular permit
Again, younger cohorts favored egalitarian gen- observation of social and historical change.
der roles signicantly more than earlier cohorts. Given the many cohort differences observed in
The positive effect of education on gender role the topics reviewed here, social change is
equality rst strengthened over time, and then endemic. Some social changes are decades in the
weakened considerably, indicating diffusion. making; others arise quite rapidly. Given data
Regardless of cohort, women were more likely covering sufciently long periods of time, the
than men to endorse egalitarian gender roles. extent and pace of social changes can be observed.
Evidence of cohort differences in attitudes Perhaps the principle of the life course perspec-
toward gender egalitarianism is strong and consis- tive that is most ignored in inter-cohort studies is
tent, with younger cohorts endorsing signicantly place i.e., the ecological contexts within which
more egalitarian attitudes than older cohorts. In individual lives unfold. Virtually all APC studies
addition, research suggests that women and per- to date rely on national samples. This is true of
sons of high socioeconomic status are more likely both U.S. and non-U.S. studies. Yet age, period,
to endorse egalitarian attitudes than men and per- and cohort effects are possible for smaller geo-
sons of lower socioeconomic status, although the graphic units than the nation state. A case in point
SES difference appears to be narrowing as a result is the effects of Hurricane Katrina on life course
of diffusion. Two implications of this research trajectories. Research suggests that the effects of
merit note. First, these studies used APC analysis the disruptions and losses generated by Hurricane
and none of them observed signicant age or Katrina remain strong and pervasive for persons
period effects. Thus there is no evidence that indi- directly affected. There is no evidence, however,
viduals develop more gender egalitarian attitudes that Hurricane Katrina generated cohort and period
as they age. It appears that at least half of the effects for U.S. citizens living outside the direct
increase in gender egalitarian attitudes is the path of this natural disaster. Logically, there is no
result of the death of cohorts characterized by reason that APC studies could not be performed
more traditional attitudes toward gender roles. for samples from geographic units other than
Second, and related, much probably most nations (George 2014). Extending the geographi-
research on changes in gender roles and gender cal units to which APC analyses are applied is a
attitudes attribute changes to the Womens high priority for future research.
Movement. Other than APC studies, the literature Although valuable, cohort differences are
does not discuss the role of cohort succession in only one window to the life course and how it
changes in gender attitudes. Because of this, the changes over time. In many ways, cohort analysis
role of the Womens Movement is exaggerated in provides a view of the forest of life course pat-
most studies of gender roles and attitudes the terns; but it is intra-cohort variability that allows
Age, Cohorts, and the Life Course 75

us to see the trees. Intra-cohort variability is risk of being drafted in 1944 was experienced by
richer and more nely nuanced than can be males, single or married, between the ages of 18
revealed in cohort studies even those that pay and 35. Cross-cutting statuses among cohort
attention to within-cohort variation due to sex, members thus focus attention on high and low
race/ethnicity, nativity, and SES. In particular, risk subgroups of cohorts in terms of the change
intra-cohort studies are equipped to trace the at hand.
varying effects of broad social changes on indi- The temporal frame of a birth cohort extends
vidual lives. We turn now to intra-cohort variabil- in theory across the life span, and a similar per-
ity and how it advances our understanding of the spective is called for in considering the duration
life course. of the social change in question and its life course
effects, whatever the data limitations on available
measurements. For example, the rst half of the
3 Historical Variation Within 1930s decade include the peak years of the eco-
Cohorts: Life Course nomic decline in the Great Depression. However
and Developmental Effects in the United States only mobilization for World
War II brought an end to the prolonged economic
Social change tends to differentiate the life course hardship of the working class population. We
of adjacent cohorts. However change may also know more about families and children during
take the form of intra-individual trajectories the rst half of the 1930s when longitudinal data
within cohorts that are linked to the cross- were available on economic adaptations than dur-
cutting statuses of its members, such as those of ing the late 1930s. The long view of birth cohorts
social class and gender. Even age becomes a from the 1920s shows a historical path that
source of intra-cohort variation when the birth included the urban prosperity of the 1920s, the
cohort includes multiple years that indicate differ- hard times of the 1930s, and the challenges of the
ing life or career stages. This link to intra-cohort years marked by world war, postwar develop-
variation occurs through differential exposure to ment, and the end of an era of afuence, circa
social change. Ryder (1965, p. 847) refers to the 1974. The full impact of this life course journey
attractive simplicity of cohort membership and is beyond the scope of any study, but it is impor-
the potential attenuation of its inuence due to tant to be aware of the lifelong historical context,
such status congurations. In his view, every its continuities and disjunctures.
cohort is heterogeneous, and cohort compari- In what follows, we use empirical studies of
sons can be protably supplemented by relevant three historical changes in the United States
compositional variables. This intra-cohort per- Great Depression, World War II, and the decline
spective is designed for a developmental investi- of the rural farm population to illustrate an
gation of social change in life experience. intra-cohort perspective that investigates the
The cross-cutting statuses of a cohorts mem- inuence of social change on the gendered life
bers enable us to solve the puzzle of differential course trajectories of cohort members. We begin
exposure to change not all members of a cohort with birth cohorts from opposite ends of the
are exposed to change, such as a drastic decline 1920s (the Oakland and Berkeley studies, birth-
of the economy. And exposure is not uniform as dates of 19201921 and 19281929 Elder
to its impact. During severe economic declines in 1999). They encountered Depression hardships
the industrialized world, such as the Great at different times in their lives as well as World
Depression, the risk of unemployment was high- War II. The younger Berkeley men experienced
est among workers in the lower strata, and par- World War II as adolescents and were typically
ticularly among males. Employment was mobilized into military service after the war had
especially problematic among both young and ended, whereas nearly all of the Oakland men
older workers. In other national crises, such as served in the Second World War and returned
mobilization for WW II in the United States, the to civilian life with access to the educational
76 G.H. Elder, Jr. and L.K. George

benets of the GI Bill in the booming postwar and Schoon compared two cohorts with birth
era. Older recruits with marriages and jobs expe- dates of 1958 and 1970. Building on Children of
rienced the greatest degree of disruption in their the Great Depression and other research, she
lives. The later the entry, the greater the risk of a designed a comparative intra-cohort study of eco-
disrupted life course. nomic deprivation in life course development that
The third historical change involves the declin- spans 40 years and includes data on the major
ing rural farm population in the United States. economic recessions of the early 1980s and
The prosperity of the postwar era was followed by 1990s. The last adult comparison of the two
an expansion of agricultural production to the cohorts at the same age occurred when the study
point of excess in the declining market of the members of the cohorts were in their early 30s.
1980s recession. This economic crisis prompted Though socioeconomic disadvantage was more
outmigration from farm states, such as Iowa. In common among young people born in 1958, the
1989, a research team (Conger and Elder 1994) study found that growing up in a disadvantaged
launched a longitudinal study of families and their environment had more adverse life course effects
young adolescent members. The latter are mem- on members of the younger cohort. This may be
bers of a cohort born at the end of the 1970s who due to the more vulnerable developmental age of
grew up during the rural economic crisis of the this cohort in the severe recession of the 1980s.
1980s. Their families span a wide range of ties to Another possible factor is the increasing skill
the land, from full-time farming to families that requirements of middle class employment that
had recently lost their farm, to those that left the enhanced the penalty of disadvantaged origins.
farm some years ago and those with fathers who Ongoing longitudinal studies of this kind pro-
were born in an urban environment. vide the possibility of periodic studies that build
upon prior research on cohort trajectories. And in
fact, it was this option that enabled the identica-
3.1 Living Through Depression tion of cohort pathways through the Great
and War Depression and World War II for the Oakland and
Berkeley study members. But it is the dramatic
Children of the Great Depression (1974/1999) growth of national longitudinal studies that repre-
documents the importance of locating members sents the most exceptional opportunity for the
of longitudinal studies in their birth cohort and comparison of cohorts on social change effects
historical context, taking note of their develop- over the life course. Examples include the
mental trajectories and social histories. The National Longitudinal Surveys in the United States
Oakland and Berkeley study members were born and especially the pioneering national longitudinal
at opposite ends of the 1920s and consequently cohorts in the United Kingdom, from the 1946
experienced the same historical times whether cohort to 1958, 1970, 2000, and 2011. The eco-
economic depression, war, or postwar prosper- logical diversity of societal populations calls atten-
ity at different life stages in their gendered roles tion to the importance of taking this variation into
and class origin. Their particular sequence of dif- account in national longitudinal cohorts.
ferent historical times from childhood to mid-life When the 1920s came to an end, the Oakland
identies distinctive life histories. In what fol- children were completing grade school, while
lows we depict this sequence in life course devel- some of the last additions to the Berkeley cohort
opment through World War II and into the had just arrived in their families. The economic
postwar era. times seemed good overall and revealed no sign
One of the most impressive comparative intra- of the impending crisis, though it soon became a
cohort studies was carried out by Ingrid Schoon reality over the next 3 years for most of the
(2006) in Great Britain. The United Kingdom has Oakland and Berkeley families. Among the eco-
led the way in launching national longitudinal nomically deprived (a loss greater than a third of
studies that invite comparative cohort studies, 1929 income up to 1933) mounting pressures
Age, Cohorts, and the Life Course 77

ignited family tensions and conicts, especially assessments during adolescence, these boys were
in families with a history of troubles. The younger judged signicantly less ambitious, self-
boys in the Berkeley study were most often condent, and self-directed than the Oakland
exposed to the arbitrary and punitive behavior of boys at the same age. The Berkeley girls were
father, an abuse that undermined their self- spared this developmental outcome, owing to the
condence and sense of independence. The protective nurturance of their mothers. They were
younger girls were shielded from this abuse by notably more self-condent, ambitious and ener-
their mothers. getic than the Oakland girls.
The Oakland boys, by contrast, were old The historical paths of the Oakland and
enough to hold paid jobs in the community and to Berkeley men from the 1920s to the postwar era
help out at home. More of their time out of school were distinguished by experiencing hard times
was spent with friends away from parental super- and wars at different ages, a difference in time-
vision. They were also old enough to understand table that made a difference in their lives. The
the economic troubles their parents were having. Oakland men experienced adolescence across the
Most girls in this cohort assumed major responsi- economically depressed years of the 1930s and
bilities in deprived households and some also then were called to military service in the armed
held paid jobs. During their years of physical forces With few exceptions, they used their vet-
maturation, they tended to feel less attractive and erans benets to advantage by acquiring further
popular. It is noteworthy that other studies (Elder training and higher education. In contrast, the
and Caspi 1990; 226) document a life stage dif- early childhood years of the Berkeley cohort
ference in the vulnerability of boys and girls to were shaped by the uncertainties and stress of
environmental insults that corresponds with these Depression hardship, a disruptive time of height-
cohort ndings on developmental risk. Boys tend ened marital tensions, punitive discipline, and
to be more vulnerable during childhood, whereas resource scarcities. A number of these children
girls are more vulnerable in adolescence. grew up without self-assurance, goals or a sense
However neither the Oakland girls nor the boys of personal worth. Their wartime adolescence
show psychological effects of economic hardship may have reinforced this sense of incompetence
by graduation from high school. The boys were as well as their perception of military service as a
soon called to military duty and the girls entered possible bridge to opportunity.
the work force or college. World War II was winding down when the
The Berkeley study members were leaving Berkeley males were old enough to serve, but
elementary school for junior high in a commu- three out of four eventually joined the military and
nity that was soon to be mobilized for World War a substantial number served in the Korean War of
II. War mobilization became the ecology of ado- the early 1950s. As might be expected, the boys
lescence for the Berkeley cohort with block who chose a military occupation in 19431944
wardens, darkened windows at night, frequent were most likely to enter before the age of 21. This
collection drives involving young people, and the action had much to do with a background of dis-
constant ow of troop trains and ships to piers in advantage that motivated their joining at the earli-
San Francisco Bay. The work demands for est possible age. The military offered them a path
employed parents in defense-related jobs to greater opportunity through advanced education
increased signicantly across a 24 h day, mini- and a good job, and it was open to young men who
mizing the daily contact of parents and children. did not have an exemplary educational record or
In these ways, the war years further weakened the the economic resources for college. Two-thirds of
role of fathers in the lives of their children. But the Berkeley veterans with a deprived family ori-
the perceived manliness of military life and the gin and below average grades entered the service
stature of servicemen appealed to the Berkeley at the earliest possible age (Elder 1986); a percent-
boys who grew up in deprived families in which age that dropped to insignicance when neither
the mother had taken charge (Elder 1986). In factor was present.
78 G.H. Elder, Jr. and L.K. George

The Berkeley men who entered the military at This developmental path owes much to the
an early age were more likely to escape their life mens timely transition to military service. Most
history of disadvantage along this path when of the early entrants in the Berkeley cohort joined
compared to the later entrants and the non- the military before marriage, college, and a sub-
veterans. They were more likely to at least enter stantial career investment. They typically entered
college than the non-veterans, owing in part to the college after their service and thereby had access
educational benets of the GI Bill, and this educa- to educational funds provided by the GI Bill.
tional path enabled them to make the best use of Their post-military timing of marriage and
their personal resources in education and work careers avoided the disruptive effect of military
lives. This occurred in large part through the per- service, such as long separations and frequent
sonal assets they acquired in military service, residential change. For a picture of this disruptive
such as the self-discipline to cope with adversity effect, consider men who entered military service
and leadership skills. With these and other late in life during World War II (Elder et al.
resources, military service turned out to be more 1994). They are members of a gifted sample of
predictive of occupational attainment through California men who were born between 1903 and
higher education among men from deprived fami- 1920, known as the Stanford-Terman project. A
lies in the 1930s than among the non-deprived. At large number of the men were mobilized into
mid-life the men from deprived Berkeley families World War II at a late age, beyond the ages of
closely resembled the occupational status of the 3032. The men were rst surveyed in 1922 and
Oakland men from deprived families. data collections continued every 5 or so years up
But how did this achievement occur in the to the 1990s.
lives of men who were so unpromising in adoles- Mobilization after the age of 32 markedly
cence. According to assessments, they lacked increased the risk of personal and social disad-
ambition, goals and self-condence. Relevant vantages that persisted up to late life. Their edu-
insights come from evidence of personal change cation level exceeded that of other veterans, but
during military service. In adolescence, the they ended up with a signicantly higher divorce
Berkeley men who did not enter the military rate. The men also tended to experience a
scored notably higher on psychological compe- disappointing work life and income trajectory
tence than the early military entrants as mea- after the mid-40s to retirement, particularly in the
sured by indexes of self-inadequacy, goal professional class. Their physical health also
orientation, social competence, and submissive- declined after the age of 50, a misfortune not
ness. However, by the age of 40 this difference in associated with combat exposure. And not sur-
psychological well-being and competence faded prisingly, the late entrants were less inclined to
to insignicance. The men from deprived fami- report benets associated with their service
lies who entered the military as soon as they were experience.
old enough had made up their decit in psycho- Our focus on the developmental trajectories of
logical competence. The precise mechanisms of Americans born in cohorts at opposite ends of the
such personal change are unknown, except that 1920s has underscored the complexity and pay-off
the veterans talked about this change in postwar of linking their lives to the economic collapse of
interviews. They believed that they had learned to the 1930s and then to mobilization in the Second
manage wartime stresses and noted that this skill World War with the economic boom it generated.
had served them well in difcult situations. A The life course impact of Depression hardship for
medic in the Berkeley study recalled his trying a younger or older birth cohort could not be ade-
experience on the front line, caring for the quately assessed without taking into account its
severely wounded and dying. I managed to show path through war mobilization in the early 1940s,
a good deal of courage and good judgment. For as well as the cross-cutting statuses of age as life
the rst time in my life, I knew I could handle stage, class and gender. They all shaped the impact
extreme situations. of exposure to the change in question.
Age, Cohorts, and the Life Course 79

The Berkeley males illustrate this complexity. 19881989 to investigate this economic crisis in
As young children of preschool age in the early families with a young adolescent and a near sib.
years of the Depression, they tended to experi- The seventh graders (N = 451) represent members
ence the greatest disadvantage of Depression of a birth cohort dened by birth at the end of the
hardship through harsh punishment and the loss prosperous 1970s. They grew up during the farm
of a caring father. Though lacking ambition and crisis of the 1980s, completed high school in 1994
goals after high school, they were soon swept up and have been followed into their 1930s. All of the
in the military mobilization of the Second World families resided in eight agricultural counties of the
War. However the major battles had been fought north central region of the state.
before they could be recruited. Nevertheless The Iowa cohort is differentiated along a fam-
nearly three out of four entered the service by the ily gradient from farm to nonfarm of family types
early 1950s. in the economic crisis as of 1989. Approximately
The early military joiners were inuenced 30 % of the fathers were still engaged in farming
developmentally by the challenging training and represented survivors at the end of the 1980s,
demands and service roles they experienced. By although the psychological cost was often high.
the time they completed their military service and Some of the fathers worked off their land as did a
their education, they appeared to have changed in majority of their wives. A number of men had
many respects. They were more self-directed, farmed earlier in the 1980s but we see them in
condent, and agentic. For these young men, 1989 as a casualty of the economic crisis. During
their developmental life course had been turned the decade they lost their farms through bank-
around. A legacy of impairment from a hard ruptcy. These displaced farmers represent 13 %
times childhood had been replaced by a bene- of all fathers in the study and their average family
cial legacy of military service. income is a third lower than that of the farmers.
The remaining comparison group includes
families headed by men who grew up on a farm
3.2 Moving Off the Land but then followed a nonfarm career, and families
headed by men with a continuous history outside
The Great Depression and World War II sent a farming.
major shock through the life course of successive The men who retained their farms despite a
cohorts. Another major shock occurred 50 years very hard struggle were critical of their way of
later in the 1970s when a soaring world demand life that was so damaging to family life and
for agricultural commodities led to overproduc- health. Survival of the family farm often required
tion in the United States as market demand long days from both parents. The off-farm
declined during the economic recession of the employment of both partners became an impor-
early 1980s. Heavy indebtedness and declining tant factor in family survival in farming. But the
land values began to push families off their farms, psychological impact of losing the family farm
setting in motion the most severe rural economic stands out above all stressors among the Iowa
crisis in the Midwestern state of Iowa since the families, especially among the fathers. One of the
Great Depression, a crisis that accelerated the fathers who had to give up his family farm was
long-term historic movement of Americans out asked about recent life changes. He referred to
of farming. Twenty percent of Iowas farms did this loss even though he had lost his North Dakota
not survive this difcult time. farm over 7 years ago. His wife reminded him
The impact of this economic crisis on rural that the question asked for recent changes and
Midwestern families and the lives of their children that the Dakota farm was not a recent loss
provides an intergenerational example of a within (Conger and Elder 1994; 81). He replied that it
cohort perspective on rural socioeconomic change didnt seem that long ago.
in the life course. The Iowa Youth and Family Despite the emotional pain and hardship of the
Study (Conger and Elder 1994) was launched in crisis years, a follow-up of the study children in
80 G.H. Elder, Jr. and L.K. George

high school did not nd signicant evidence of an The work lives of Berkeley fathers were generally
adverse legacy from such experience (Elder and carried out some distance from their households,
Conger 2000). Even the most troubled youth in the and much of their childrens non-school time
past those in families that lost their farm occurred away from parental observation. Most
showed no unusual distress at this time. Farm fam- important is the stronger authority position of the
ilies had been on a downward income trend over Iowa farm fathers when compared to the Berkeley
this time, but their sons and daughters and even fathers. The hierarchy of social stratication was
those from displaced farm families were among less differentiated among farm families. They var-
the athletic, social, and academic leaders by the ied in wealth but not as much in life style.
time they entered high school. Family socioeco- Lastly, consider the issue of size of commu-
nomic status did not account for this outcome. nity. The Berkeley children came of age in a rela-
Access to a socially resourceful path through tively large community with corresponding social
high school made the difference in their lives. institutions, such as schools. By contrast, the
Parent investment in their future was especially Iowa cohort came of age in communities, schools,
relevant. Parents with ties to the land were active and churches that were small by comparison.
supporters of their childrens school-based activi- Small worlds tend to maximize participation
ties, from academics and dramatic arts to athlet- opportunities for the young and many of the Iowa
ics. Family connections also emerged as an youth from hard pressed farm families discov-
important factor. It refers to the shared family life ered condence-building opportunities in their
in farming working together and more generally schools. An Iowa boys family lost its land and
the routine of doing things together. The inter- livestock in the Great Farm Crisis, but he discov-
dependence of family life on a farm generates a ered a turning point in school theatre with the
sense of being counted on in daily chores and the recognition, support, and opportunities it pro-
harvests. Another distinctive feature of family vided. In his words, Drama changed my life.
resourcefulness involved the engagement and It gave me the condence to speak in front of
leadership of farm parents in community life in groups and then led to other opportunities,
church, civic organizations, schools and eco- including a university scholarship (Elder and
nomic organizations. Their children frequently Conger 2000; 182). A supportive family provided
followed the parental example through leader- an important foundation for this life change.
ship in student organizations. These social themes emerged from a study
What is it about socially engaged families that focused on young people in families with ties to
contributes to the educational success and leader- the land, but they are by means exclusive to them.
ship of their children? Typically these farm fami- This younger generation grew up in families that
lies are intact and have better educated parents, were faced with the constant uncertainty of a
but the enterprise of farming itself is itself family- livelihood in farming. Though some were still
based, a theme that carries over to community, attracted to the possibility of living on a farm,
church, and school involvement, regardless of they did not see a future in farming. In this way,
educational level. Socially-engaged parents know the Farm Crisis of the 1980s added fuel to the
more about issues that pertain to their children, historic decline of the farm population in the
and the latter realize that what they do matters a United States.
lot to parents and grandparents. On the farm and
even in the community, social life and work life
are shared experiences. Not all farm families 4 The Impact of Life Course
were characterized by this interdependence and Change
emotional connectedness, but most were.
This account of Iowa farm families stands in In Children of the Great Depression, the effects
sharp contrast to the urban Berkeley deprived fami- of economic hardship are traced through the lives
lies and the children who were born in 19281929. of adolescents to their middle years. In an early
Age, Cohorts, and the Life Course 81

review of this work (1975; 121), historian John CCC recruits convened at army recruiting sta-
Modell suggests that social change in the life tions; traveled to an army camp where they were
course establishes a new context for cohort tra- outtted in World War I clothing; were trans-
jectories and for members of the cohort itself. ported to the woods by troop-train; fell asleep in
The process follows a dialectical dynamic in army tents to the strain of Taps and woke to
which social change inuences the life course of Reveille. Despite concern over military control
individual members of the cohort, followed by and militarism in this mobilization, the program
the way that those individual experiences are was praised for its role of shepherding the
aggregated to constitute a new context for others nations youth as war shadows darkened in
living through these changes (Modell 1989; 22). Europe. By late 1940, the onset of a selective ser-
In this way, life course change represents the vice program across the country prompted a wave
emergence of a society that is new in some or of CCC enlistees.
many features. Accordingly, this perspective Most of the 16 million Americans who served
views the Oakland and Berkeley cohorts as col- in World War II came from the cohort of men
lective actors that shape their world for subse- born in the 1920s. Four out of ve men born dur-
quent cohorts. ing the decade became veterans of this war. In the
The following cases from American history Oakland cohort, birthdates of 19201921, virtu-
provide examples of contexts shaped by life ally all of the men served only 10 % did not. The
course change and the trajectories and transitions military induction of youth who grew up in the
that emerged from them. We begin with the lives Depression decade enabled a good many to
of young people patterned by the Civil escape impoverished backgrounds, dysfunctional
Conservation Corps of the New Deal. Through its families, and problematic academic records. As
training, the CCC established a path to military Brotz and Wilson (1946) described the process,
service in World War II. Next and more broadly military entry knifed-off the past through basic
considered, we view the cohort of men born in training that fostered equality and comradeship
the 1920s as Depression children and then mili- among the recruits. The enormous scope of such
tary recruits in World War II. The service time of training and the global diversity of military ser-
these recruits established a new life context that vice posed an unparalleled challenge at Wars
provided access to the generous educational ben- end, that of incorporating military personnel back
ets of the G.I. Bill with nancial support for in civilian society with resources and opportuni-
advanced education that in turn motivated the ties for a productive and fullling life. The G. I.
extraordinary civic involvement of non-black Bill of Rights emerged with the genius to a
veterans in postwar America. By contrast, black remarkably successful transition from soldier to
veterans focused most of their civic energies on citizen.
the achievement of their rights as American citi- No recruit in this Greatest Generation could
zens in the Civil Rights Movement. Most of these have imagined the extraordinary benets of this
linkages and life contexts varied by gender and G.I. Bill of Rights that had yet to be proposed in
race. the Congress and its transforming effect on their
Depression youth who were out of school postwar lives. Nearly half of all men from the
and work illustrate the agentic role of a social 1920s who served in the Armed Forces used the
aggregate when their increasingly dire circum- educational benets of the G.I. Bill, a gure that
stances prompted federal action through the Civil was more than twice the congressional estimate.
Conservation Corps of Roosevelts New Deal Other veterans chose instead to get on with a job
(Leuchtenburg 1964; especially p. 109). Perhaps instead of more training or education. In Soldiers
inspired by manpower development experiences to Citizens, Suzanne Mettler (2005) tells the story
in World War I, CCC camps were established and of how this legislation came to be and its power-
managed by the Army throughout the country to ful impact on the veterans who put it to use
instill martial virtues in the nations youth. in their training, higher education and life.
82 G.H. Elder, Jr. and L.K. George

Her pioneering study is based on surveys, both The rising level of education in postwar
retrospective and prospective, of veterans. Like America also contributed signicantly to its civic
the composition of the Armed Forces at the time, culture, and the ood of returning veterans was a
the respondents were mainly white men, along major component of this growth. Over half of the
with small samples of African-American men college students enrolled in 1947 were veterans,
and women. and a majority of those in Mettlers survey of the
Motivated by the all too meager provisions for Class of 1949 made use of the G.I. Bill in doing
veterans of the First World War, the American so. Its generous benets became a major turning
Legion became a vigorous advocate of generous point in their lives. Most claimed that they would
and inclusive benets during the Second World not have been able to afford college or more
War and mobilized to ensure congressional sup- training without such assistance. The most com-
port in 1944. The resulting legislation was shaped mon beneciaries of this benet were veterans
in part by a proposal from the Roosevelt adminis- who came from lower-income families who used
tration, along with Legion support. The legisla- it to pursue vocational/technical training that
tions content was inuenced by a soldiers enabled them to both work and continue their
request to his Legionnaire father regarding post- training. These men were two and a half times
war benets. When prodded on this issue by his more likely to use education benets than men
father, he and his buddies replied that they simply who entered college with such assistance. This
wanted the opportunity to get education or train- differential between college and training partly
ing, and to nd work. This request eventually reects the smaller number of veterans who were
became key components of the evolving G.I. Bill. qualied for higher education at this time.
Along with unemployment benets and access to The extraordinary civic involvement of the
low-interest housing loans, the G.I. Bill offered postwar generation also owes much to its years
educational benets up to 48 months for veterans of experience with a caring government (Mettler
with at least 3 months of service. Tuition and fees 2005). The 1920s birth cohort of veterans was
were covered up to a maximum of $500, along exposed to federal government in the 1930s and
with a monthly subsistence that varied according 1940s that was actively involved in responding to
to dependents. the needs of its citizens through social provi-
Veterans born in the 1920s became a large part sions. For returning veterans, G.I. Benets for
of what Robert Putnam (2000) called the most education and training represented by far the
civic generation in American history. Their civic most inuential aspect of such experience for
involvement may reect in part the impact of civic involvement. The generosity and inclusive-
basic training with its emphasis on team work ness of the benets fostered a desire to make their
and social responsibility an interdependence lives count in community service. These benets
expressed in the shared belief in military units markedly increased their civic involvement well
that you have my back and I have yours. The beyond the effects of education. It is noteworthy
linked lives in a perspective on the life course that the recipients of training from lower-income
thus represent a cornerstone of military training. families were more likely to make use of this
Ties to comrades are known to persist across the resource in civic activity when compared to the
years and are often expressed in reunions (Elder recipients of a college education. With these
and Clipp 1988). The power of this social bond in developments in mind, Mettler (p. 11) concludes
postwar America stems in part from the large that just as the G.I. Bill transformed the lives of
number of men and women who served in the veterans who used it, they in turn helped to
Second World War a total exceeding 16 million. change America.
When veterans returned to their home communi- The story of the G.I. Bill is largely centered on
ties after the war, they did so usually with other male veterans, and yet a full appreciation of its
veterans who often became members of their benets leads to the larger framework of the fam-
social networks. ily and generations. A substantial number of the
Age, Cohorts, and the Life Course 83

male veterans were married when they used the a second-class citizen. That was painful, very
G.I. Bill for access to support for the unem- painful. By the early 1960s, the cumulative
ployed, for assistance in paying college or train- experience of racial injustice had prompted
ing expenses, and for access to a loan to purchase greater involvement of black veterans in street
a home. The veterans interviewed by Suzanne demonstrations, protests, and marches in the civil
Mettler (2005) spoke about their gratitude for rights movement. Mettler found that over a third
what its benets meant for their families. Veterans of the African Americans in her sample who used
who entered college or technical training pro- the educational benets of the G. I. Bill were
grams as the rst family member to follow this involved in such mobilizations between 1950 and
path established a higher standard of educational 1964, with many in leadership roles, compared to
achievement for their children. 8 % of the black nonusers.
A relatively small number of women served in
the U.S. Armed Forces, and those who did were
less likely than male veterans to apply for bene- 5 Conclusion
ts of the G. I. Bill. As one of the women who did
serve noted, this benet was commonly thought This chapter brings to mind the progress achieved
to be for the men. A postwar cultural emphasis in recent decades toward a greater understanding of
on homemaking strengthened this theme. But age, cohorts and the life course, coupled with rec-
those who did take advantage of the educational ognition of the challenges that lie ahead. Over 50
benet were more likely than the non-users to years ago, Norman Ryder brought to publication a
obtain at least some college or training. And, eld-shaping essay that provided a conceptual
similar to the men, the women who took advan- framework for investigating social change in soci-
tage of the G.I. Bill were more likely to become ety by using an inter-cohort approach. This was
civic activists. soon followed by advances in an age/period/cohort
Gender clearly played a major role in structur- design that enabled research to identify the most
ing military service and access to the G. I. Bill in important sources of change, whether due to aging,
the 1920s birth cohort. But gender differences historical time, or cohort membership. More
were less pronounced than the striking contrast recently progress has been made in nding ways to
by race (Mettler 2005; Parker 2009). African use theory and observations to explain the observed
American men were barred from major roles in cohort or period effects.
World War II, though a large number served Ryders essay also noted that social change
approximately one million, typically in support may inuence cohort members differentially
for frontline troops. They reported being fairly by their gender, age, and social class an intra-
treated on benet access and in college or train- cohort dynamic. As pointed out in this chapter,
ing experience, but not at all in the labor market. this perspective can illuminate how a changing
Experience in the military empowered returning society inuences the individual over his or her
veterans to act on their own behalf and gave them life course. And the aggregation of such life
a sense of agency and equality in claiming the course effects provides insight on the changing
rights of an American citizen. This outlook trajectory and form of social change. Such work
clashed in the segregated South with mistreat- is underway in many regions of the globe, encour-
ment by employers and public ofcials. Such aged by the continued growth of cohort studies
actions and the denial of voting rights became that follow people over their lives through panel
intolerable, especially for those who had received research and retrospective life histories. These
a college education on the G.I Bill. studies have produced a more generalized case
One of the Tuskegee Airmen (Mettler 2005; for recognizing that all projects and their ndings
14041) noted that when they came home you are bounded by a specic historical time and
couldnt even be buried in the town cemetery, place. But we too seldom know whether the nd-
you had to be segregated. you were treated like ings are generalizable or indicate a trajectory of
84 G.H. Elder, Jr. and L.K. George

change. More comparative cohort studies are Elder, G. H., Jr. (1975). Age differentiation and the life
course. Annual Review of Sociology, 1, 165190.
needed to advance our understanding of social
Elder, G. H., Jr. (1986). Military times and turning points
change in the life course. in mens lives. Developmental Psychology, 22,
233245.
Elder, G. H., Jr. (1999). Children of the great depression:
Social change in life experience (25th anniversary edi-
References tion). Boulder: Westview Press.
Elder, G. H., Jr., & Caspi, A. (1990). Studying lives in a
Alwin, D., & McCammon, R. J. (2003). Generations, changing society: Sociological and personological
cohorts, and social change. In J. T. Mortimer & M. J. explorations. In A. I. Rabin, R. A. Zucker, R. A.
Shanahan (Eds.), Handbook of the life course Emmons, & S. Frank (Eds.), Studying persons and
(pp. 2349). New York: Kluwer Academic/Plenum lives (pp. 201247). New York: Springer.
Publishers. Elder, G. H., Jr., & Clipp, E. C. (1988). Wartime losses
Baltes, P. B., Cornelius, S. W., & Nesselroade, J. R. and social bonding: Inuences across 40 years in
(1979). Cohort effects in developmental psychology. mens lives. Psychiatry, 51, 177198.
In J. R. Nesselroade & P. B. Baltes (Eds.), Longitudinal Elder, G. H., Jr., & Conger, R. D. (2000). Children of the
research in the study of behavior and development land: Adversity and success in rural America.
(pp. 6187). New York: Academic. Chicago: University of Chicago Press.
Bhrolchain, M. N., & Beaujouan, E. (2013). Education Elder, G. H., Jr., Shanahan, M. J., & Clipp, E. C. (1994).
and cohabitation in Britain: A return to traditional pat- When war comes to mens lives: Life-course patterns in
terns? Population and Development Research, 39, family, work, and health. Psychology and Aging, 9, 516.
441458. Elder, G. H., Jr., Johnson, M. K., & Crosnoe, R. (2003).
Brotz, H., & Wilson, E. (1946). Characteristics of military The emergence and development of life course theory.
society. American Journal of Sociology, 51, 371375. In J. T. Mortimer & M. J. Shanahan (Eds.), Handbook
Brunillo, G. (2010). The effects of cohort size on European of the life course (pp. 322). New York: Kluwer
earnings. Journal of Population Economics, 23, Academic/Plenum Publishers.
273290. England, P., Shafer, E. F., & Wu, L. (2012). Premarital
Chaves, M. (2011). American religion: Contemporary conceptions, postconception (shotgun) marriages,
trends. Princeton: Princeton University Press. and premarital rst births: Education gradients in U.S.
Conger, R. D., & Elder, G. H., Jr., in collaboration with cohorts of white and black women born 19251959.
Lorenz, F. O., Simons, R. L., & Whitbeck, L. B. Demographic Research, 27, 153166.
(1994). Families in troubled times: Adapting to change George, L. K. (2014). Taking time seriously: A call to
in rural America. Social institutions and social action in mental health research. Journal of Health
change. Hawthorne: Aldine de Gruyter. and Social Behavior, 55, 114.
Copen, C. E., Daniels, K., Vespa, J., & Mosher, W. D. Haines, M. (2008). Fertility and mortality in the United
(2012). First marriages in the United States: Data from States. Economic history association. Retrieved at
the 20062010 National Survey of Family Growth. eh.net/encyclopedia/fertility-and-mortality-in-the-
National Health Statistics Reports. Retrieved at www. United-States.pdf on 18 Sept 2014.
cdc.gov/nchs/data/nhsr//nhsr049.pdf on 12 Sept 2014. Hoffman, J. P. (2013). Declining religious authority?
Cotter, D., Hermsen, J. M., & Vanneman, R. (2011). The Condence in the leaders of religious organizations,
end of the gender revolution? Gender role attitudes 19722010. Review of Religious Research, 55, 125.
from 1977 to 2008. American Journal of Sociology, Houtman, D., & Aupers, S. (2007). The spiritual turn and
117, 259289. the decline of tradition: The spread of post-Christian
Crockett, A., & Voas, D. (2006). Generations of decline: spirituality in 14 western countries, 19812000.
Religious change in 20th-century Britain. Journal for Journal for the Scientic Study of Religion, 46,
the Scientic Study of Religion, 45, 567584. 305320.
Dannefer, D. (2011). Age, the life course and the sociologi- Kaufmann, E., Goujon, A., & Skirbekk, V. (2012). The
cal imagination: Prospects for theory. In R. H. Binstock end of secularization in Europe? A socio-demographic
& L. K. George (Eds.), Handbook of aging and the perspective. Sociology of Religion, 73, 6991.
social sciences (7th ed., pp. 316). San Diego: Academic. Kraaykamp, G. (2012). Employment status and family
Easterlin, R. A. (1980). Birth and fortune: The impact of role attitudes: A trend analysis for the Netherlands.
numbers on personal welfare. Chicago: University of International Sociology, 27, 308329.
Chicago Press. Lambert, Y. (1999). Religion in modernity as a new axial
Easterlin, R. A. (1987). Birth and fortune: The impact of age: Secularization or new religious forms? Sociology
numbers on personal welfare (2nd ed.). Chicago: of Religion, 60, 303333.
University of Chicago Press. Leuchtenburg, W. E. (1964). The New Deal and the ana-
Elder, G. H., Jr. (1974). Children of the great depression: logue of war. In J. Braeman, R. H. Bremner, &
Social change in life experience. Chicago: University E. Walters (Eds.), Change and continuity in twentieth-
of Chicago Press. century America. New York: Harper and Row.
Age, Cohorts, and the Life Course 85

Livingston, G., & Cohn, D. (2012). U.S. birth rate falls to Sampson, J. R. (2012). Great American city: Chicago and
a record low: Decline is greatest among immigrants. the enduring neighborhood effect. Chicago: University
Pew Social Trends. Retrieved at www.pewsocial- of Chicago Press.
trends.org/2012/11/29/overall_trends.pdf on 18 Sept Schaie, K. W. (1965). A general model for the study of devel-
2014. opmental problems. Psychological Bulletin, 64, 92107.
Mannheim, K. (1952). The sociological problem of gen- Schoen, R., & Canudas-Romo, V. (2006). Timing effects
erations. In K. Mannheim (Ed.), Essays on the sociol- on divorce: 20th century experience in the United
ogy of knowledge (pp. 276322). London: Routledge States. Journal of Family and Marriage, 68, 749758.
and Kegan Paul. Schoon, I. (2006). Risk and resilience: Adaptations in chang-
Mayer, K. U. (2009). New directions in life course ing times. Cambridge: Cambridge University Press.
research. Annual Review of Sociology, 35, 413433. Schwadel, P. (2010). Age, period, and cohort effects on
Mettler, S. (2005). Soldiers to citizens: The G. I. Bill and U.S. religious service attendance: The declining
the making of the greatest generation. Oxford: Oxford impact of sex, southern residence, and Catholic afli-
University Press. ation. Sociology of Religion, 71, 224.
Modell, J. (1975). Levels of change over time. Periodicals Schwadel, P. (2011). Age, period, and cohort effects on
archive online, September, 116127. religious activities and beliefs. Social Science
Modell, J. (1989). Into ones own: From youth to adult- Research, 40, 181192.
hood in the United States, 19201975. Berkeley: Shu, X. L., & Zhu, Y. F. (2012). Uneven transitions:
University of California Press. Period- and cohort-related changes in gender attitudes
Pampel, F. (2011a). Cohort changes in the socio- in China, 19952007. Social Science Research, 41,
demographic determinants of gender egalitarianism. 11001115.
Social Forces, 89, 961982. Slack, T., & Jensen, L. (2008). Birth and fortune revis-
Pampel, F. (2011b). Cohort change, diffusion, and support ited: A cohort analysis of underemployment, 1974
for gender egalitarianism in cross-national perspec- 2004. Population Research and Policy Review, 27,
tive. Demographic Research, 25, 667693. 729749.
Parker, C. S. (2009). Fighting for democracy: Black veter- Sommerville, C. J. (1998). Secular society, religious popula-
ans and the struggle against white supremacy in the tion: Our tacit rules for using the term secularization.
postwar south. Princeton: Princeton University Press. Journal for the Scientic Study of Religion, 37, 249253.
Putnam, R. D. (2000). Bowling alone: The collapse and Taylor, C. (2007). A secular age. Cambridge, MA:
revival of American community. New York: Simon & Harvard University Press.
Schuster. U.S. Census Bureau. (2012). Median age at rst marriage
Raley, B. K., & Bumpass, L. (2003). The topography of by sex: 18902010. Washington, DC.: U.S. Census
the divorce plateau: Levels and trends in union stabil- Bureau. Retrieved at www.census.gov/hhes/socdemo/
ity in the United States after 1980. Demographic marriage/data/acs/ElliottetalPAA2012gs.pdf on 4
Research, 8, 245260. Aug 2013.
Riley, M. W. (1987). On the signicance of age in sociol- Voas, D., & Crockett, A. (2005). Religion in Britain: Neither
ogy. American Sociological Review, 52, 114. believing nor belonging. Sociology: The Journal of the
Riley, M. W., Johnson, M. E., & Foner, A. (1972). Aging British Sociological Association, 39, 1128.
and society (A sociology of age stratication, Vol. 3). Voye, L. (1999). Secularization in the context of advanced
New York: Russell Sage. modernity. Sociology of Religion, 60, 275288.
Rossi, A. S. (Ed.). (1985). Gender and the life course. Wu, L. I. (2008). Cohort estimates of nonmarital fertility
Washington, DC: American Sociological Association. for U.S. women. Demography, 45, 193207.
Ryder, N. B. (1965). The cohort as a concept in the study Yang, Y., & Land, K. (2013). Age, period, cohort analysis:
of social change. American Sociological Review, 30, New models, methods, and empirical applications.
843861. New York: Chapmen and Hall/CRC Press.
Opening the Social: Sociological
Imagination in Life Course Studies

Dale Dannefer, Jessica Kelley-Moore,


and Wenxuan Huang

1 Introduction ferment and vigorous debate, with a steady


growth in the number of published studies
It has now been a full half century since Leonard devoted to life-course questions in the sociology
Cains seminal essay Life Course and Social of age and in sociology more generally.
Structure appeared in 1964, and more than four In the decade since the rst edition of the
decades since the publication of Glen Elders Handbook of the Life Course was published, life
Children of the Great Depression (1974) and course scholarship has benefited from new
since John Clausen (1972) published The Life analytical tools and expanded data sources, both
Course as a chapter in Matilda Rileys initial national and international (see Bynner, this volume),
volume on age and social structure. In Germany, and has expanded its scope across multiple
Martin Kohli began developing the concept of the substantive domains, within and beyond the
life course as a social institution at about the discipline of sociology (Alexander et al. 2014;
same time (see e.g., Kohli and Meyer 1986). Dannefer 2013; Gluckman and Hanson 2008;
Since those beginnings, the concept of the life Smith 2005; Priestley 2001; Laub and Sampson
course has expanded its reach within and beyond 2003); And in this decade, a journal devoted
the sociology of age. In 1997, the American specically to the life course, Advances In Life
Sociological Associations Section on Aging was Course Research, was launched as was an inter-
renamed the Section on Aging and the Life national society dedicated solely to life course
Course. As the relevance of the life course scholarship, the Society for Longitudinal and
perspective has become recognized across a Life Course Studies (SLLS). Clearly, the life
broad range of other substantive areas of socio- course has become an established domain of
logical inquiry (e.g., family, criminology, work, study in contemporary sociology.
education, policy), the proliferation of research
on the life course has established the study of the
life course simultaneously as both a domain of 1.1 Common Phenomena,
normal science, and as an area of intellectual Divergent Explanations

Over recent decades, the energies of life-course


researchers have focused heavily on patterns of
D. Dannefer (*) J. Kelley-Moore W. Huang individual life course outcomes (e.g., health,
Department of Sociology, Case Western Reserve wealth, family structure) and their causal path-
University, Cleveland, OH, USA ways, fueled by the growth of quality longitudinal
e-mail: dale.dannefer@case.edu

Springer International Publishing Switzerland 2016 87


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_4
88 D. Dannefer et al.

data on nationally representative samples. While This is where a persistent paradox arises. For
there is broad agreement about the importance of what we repeatedly observe in life course
understanding such phenomena, the causal studies and associated literatures is that, despite
mechanisms and pathways remain much less broad new areas of potential explanatory terrain,
clear and an arena of considerable debate (see life course research often continues to avoid
Moore and Brand, this volume). It is our objec- such opportunities, in favor of more familiar
tive in this chapter to contribute to that debate by questions and problems that t within the
calling attention both to some undeveloped comfortable and established paradigmatic
implications of sociological theory as well as frameworks and empirical problem templates.
some important empirical discoveries and Even while sometimes expressing excitement
developments relevant to the life course. and intrigue at such discoveries, life course
The remarkable advance of empirical knowl- scholars and gerontologists have too often con-
edge and evidence in multiple elds has made tinued to use traditional assumptions that accept
clear that the power of social forces to shape and both social and individual-level arrangements
organize individual lives is in many cases greater and explanations as not only legitimate but as
than even social scientists anticipated. Such natural and inevitable. As has been noted ear-
advances include new levels of renement in sev- lier, this is precisely what happened in the after-
eral domains of biology, from brain sciences to math of the original discovery of the importance
the mapping of the genome. Despite an initial of cohort analysis (Riley et al. 1972; Schaie
aversion by many social scientists to embracing 1965). Cohort analysis brought the realization
such elds of study, it is becoming increasingly that social context matters for aging. However,
clear that the realm of G-E interactions is among instead of embracing the implications of this
those domains that are demonstrating the hereto- realization by looking more closely at how con-
fore unrecognized importance of social and envi- text effects aging, the dominant practice was
ronmental factors in explaining biochemical initially to equate cohort differences with con-
outcomes, e.g., in metabolic, hormonal and other textual effects, and treat intracohort variation as
physiological processes. error (Dannefer 2011). Synthetic cohort analy-
Such discoveries open new possibilities for sis, currently in vogue, reects yet a further
discerning the potential power of social forces retrenchment away from social causation, since
and social processes to account for individual- it implicitly assumes that there is a natural and
level phenomena, including some that have been true age trajectory that can be discerned by
traditionally assumed to be largely immune from amalgamating all cohorts together.
social inuence. For social scientists, this is a cir- The problem that we address here, then, is the
cumstance that brings both opportunities and continuing danger of containing the social
responsibilities, because it offers unusually fer- inhibiting key points of inquiry where possibilities
tile ground for sociological imagination that is, present themselves, to open the horizons of
for giving careful consideration to the possibility sociological explanation. We contend that this
that social forces and processes may be playing a tendency to limit the scope of the sociological
previously unnoticed or underestimated causal imagination emanates from the continued
role in shaping life course outcomes, even in the reliance of much life-course scholarship on the
realms of biological aging and genetics. These interlinked paradigms of functionalist sociology
empirically driven developments mean that the and developmental psychology for organizing the
social sciences are poised on the brink of fresh explanations for phenomena that occur over
and renewed horizons of inquiry, with new the life course (Dannefer 2011). In the following
opportunities to investigate the possibility of the sections, we seek to explicate this problem and
previously unnoticed power of social forces to then follow with examples from three relevant
shape life course outcomes. domains of life course inquiry.
Opening the Social: Sociological Imagination in Life Course Studies 89

1.2 Containing the Social: age-related and other features of existing institu-
The Functional- tions have been created as accommodations to
Developmental Nexus individual needs (e.g., 4-year-olds need kinder-
and Sociological Explanation garten, nonagenarians need nursing homes), an
assumption which implies that such institutions
Critiques of life course scholarship for its tendency warrant no critical analysis; (4) a paradigmatic
to rely on functionalist structural arrangements emphasis on social solidarity, consensus and nor-
and developmentally-based individual pathways mative order, and a concomitant disinclination to
have been presented elsewhere, including discus- integrate processes involving power and conict
sions of microcation (Hagestad and Dannefer at any level, including the micro-level, and (5) an
2001), individual-level reductionism (Dannefer assumption that human interests and institu-
1999b, 2011; Dannefer and Kelley-Moore 2008; tional interests are congruent. Although almost
Morss 1995) and ideologies of age (Baars 1991; always implicit and unstated, such assumptions
Katz 1996; Riley et al. 1994). While functional- underlie the paradigmatic logic and research
ism and developmentalism derive from work in questions that inform and guide most life course
two different disciplines, these two broad para- research.
digms share several key assumptions that inform To be sure, one can point to exceptions and to
the study of the life course, as well as related developments in the study of the life course and
elds such as human development and social ger- related areas that have challenged the dominance
ontology. At the outset, we acknowledge explic- of explanatory paradigms at the functional-
itly that many aspects of our accumulated developmental nexus. Examples include the
understanding of the life course have been Baltes-Dannefer debate over ontogenetic reduc-
discovered and achieved within the functional- tionism (e.g., Baltes et al. 1994; Baltes and
developmental framework. This includes discover- Nesselroade 1984; Dannefer 1984a, b, 1994;
ies that contained the seeds of its own critique, Dannefer and Perlmutter 1990; Featherman and
such as cohort analysis as we mention above. Lerner 1985); the Bengtson/Connidis debate con-
That is how science often proceeds. Yet despite trasting solidarity with conict and ambivalence
its contributions, it is equally clear that it has (Bengtson et al. 2002; Connidis and McMullin
important limitations, to which we now turn. 2002a, b; Fingerman and Hay 2004) and the
A full discussion of the assumptions and growing emphasis on processes generating pat-
operating principles central to the symbiotic terns of life-course inequality and stratication
relationship between functionalism and develop- such as weathering (Geronimus et al. 2006;
mental theory must lie beyond the scope of this Newman 2006) and cumulative dis/advantage
paper. For our purposes, it may be useful to map (e.g., Crystal and Shea 1990, 2002; Dannefer
key points of overlap that, for scientists working 1987, 2003a, b; Ferraro and Kelley-Moore 2003;
at their nexus, may serve to perpetuate operating ORand 2002) all of which represent challenges
assumptions that remain uncritically applied. to the dominant narrative underlying the
Some of the more important such shared notions functional-developmental nexus. Although such
are (1) an assumption of the legitimacy and challenges have enlivened and expanded the
functionality of the overall social order and of the range of questions and issues that are being fruit-
generally benign nature of social institutions fully pursued in the study of the life course, the
within which individuals develop and age; (2) an functional-developmental nexus has manifested a
emphasis on the explanatory integrity of the free- remarkable resilience. As we will show, its under-
standing individual, both through stable individual lying logic remains largely intact in key and
characteristics (e.g., temperament, personality, emerging areas of life course studies.
resilience) and accentuation processes, and through The key assumptions underlying the
self-efcacy as expressed in agentic action; (3) given functional-developmental nexus (such as those
#1 and #2, a further, implicit assumption that the just listed) have in common a tendency toward
90 D. Dannefer et al.

naturalization that is, a willingness to accept as interest. Rather, social science analysis must
rather permanent, inevitable and even natural, the begin, as Baars (1991), Berger and Luckmann
structural regimes that underlie both established (1967), Dannefer (1984a, b), Holstein and
social institutions and individual development, Gubrium (2000), and Montagu (1989) and oth-
and a concomitant assumption that existing social ers have made clear, with a recognition that the
structures and institutions operate as legitimate inuence of the social is constitutive in shap-
accommodations to human needs anchored in ing, regulating, and sustaining the individual,
developmental processes. This cross-disciplinary physically and psychically as well as socially.
symbiosis harmonizes our subject matter into a Although a review of the constitutive dimen-
comfort zone of relatively familiar and routine sions of social forces must remain beyond the
questions, without creating an opening for more scope of this chapter, such dimensions encom-
fundamental questions that are warranted by pass several key established domains of socio-
empirical evidence, concerning the forces that logical explanation and inquiry, including
underlie the structuring of the life course, bio- Meads analyses of the social genesis of the self
logically as well as socially. (1934), sociosomatic analyses of the role of
Thus, the functional-developmental nexus has culture in physical development, and recent dis-
had the effect of limiting, in subtle and often coveries concerning the extent of social and
unrecognized ways, attention to the active power environmental inuences on brain development
of social forces in shaping the lives of individuals. and on gene expression and perhaps even gene
Yet its underlying paradigmatic predispositions sequencing. For life course scholars, the expan-
comprise a heuristic logic to guide research. This sion of our recognition is especially daunting as
often-implicit logic restricts and contains the well as especially promising, because it adds to
effects of social forces a paradigmatic tendency these basic claims for social forces the inherent
that has elsewhere been termed a heuristic of dimension of temporality. Moreover, since
containment. We argue, instead, for a logic that many of the postulated effects are highly inter-
encourages a more expansive exercise of socio- active, tracking their interrelations through a
logical imagination in understanding the role of sustained period of time can be analytically and
the social, which has been called a heuristic of theoretically challenging.
openness (Dannefer 2011, 2012). In this chapter, we demonstrate the limits of
This is a situation that brings squarely into the functional-developmental nexus for advancing
focus the question of the actual signicance of inquiry in the current discourse in life course
the discipline of sociology and of the sociological studies by focusing on three areas of signicant
perspective. Too often, reading the life course developing research, each of which has encountered
literature, one might get the impression that the its own version of containment that is, of the
task of sociology is merely to add a few con- tendency to stick with the status-quo assumptions
textual variables that inuence or constrain the that tend to restrict inquiry. Instead, we propose
outworkings of causal forces that are located that what is needed is precisely the opposite: An
elsewhere e.g., in individual decision-making opening of the more expansive possibilities of
or developmental imperatives. sociological explanation that are implied by
Yet a key element of the meaning of sociol- recent discoveries or debates touching all three
ogy as a discipline is precisely in the capacity to of these areas of life course scholarship: (1) con-
investigate, discover, and scrutinize the role of cerns arising in formulating and applying the
the social in shaping such phenomena. That concept of agency in contemporary society, (2) early
cannot be done by an a priori surrender of life inuences on adult outcomes, and (3) life-course
explanatory terrain to putative individual char- implications of gene-environment interaction
acteristics such as temperament or volition studies. In each case, we will seek to demonstrate
nor by uncritically assuming established institu- the tension between sociological explanation and
tions can be counted on to operate in the human tendencies toward reductionism. We will suggest
Opening the Social: Sociological Imagination in Life Course Studies 91

that once again the heuristic of containment, inuential, as indicated by the frequent refer-
rather than being tempered by recent empirical ences and expressions of it in the life course lit-
discoveries of the powerful role of the social, erature. Relying on Elders approach, for
continues to operate to avoid or resist the full example, Gillespie and van der Lippe state that
implications of a confrontation with the explana- human agency refers to individuals ability to
tory potentials of social forces. navigate their lives within the constraints of their
social circumstances (2014: 2). Setterstens
(1999) reference to agency within structure
2 Containing the Social: Three articulates a similar view, as do numerous other
Examples formulations offered in life course research.
Despite its continued dominance in the accumu-
In the sections that follow, we review recent lated literature, this general approach suffers
literature in each of three lively domains of from at least two signicant problems, one meth-
discourse and inquiry in life-course studies. We odological and one theoretical, both of which are
demonstrate that the dominant approaches in grounded in the functional-developmental nexus.
each of these domains, while seeking to advance Both problems involve the assumption of an
sociological knowledge, risk constricting and unproblematic volitional agent navigating a
containing the scope of potential sociological largely xed and legitimate social context. The
imagination because of their grounding in the func- methodological problem is that while agency is
tional-developmental nexus. The three domains frequently invoked, it is seldom actually measured
are: (1) efforts to clarify the place of agency and or operationally dened; rather, it is simply
choice in the study of the life course (2) research assumed to be operative. In our review of studies
on the long-term consequences of early life expe- that use the term agency or synonymous words
rience; and (3) the growing interest in the impor- such as choice and decision-making in the
tance of gene-environment (G-E) interaction title or as keywords, in four major journals
studies. We then conclude by reecting on what (Journal of Aging Studies, Research on Aging,
the direction of research in each of these areas Journals of Gerontology Social Sciences, and
implies about life course studies. Advances in Life Course Research) from 2004 to
2014, only 9 of 49 articles actually attempt to
measure agency. To be sure, some of these studies
2.1 Agency: The Big Easy (mainly using survey questions about preferences
of the Life Course or qualitative interview discussions that probe
decision-making) treat agency in a thoughtful
In studies of the life course and in related elds way. Yet what concept can survive and thrive
such as life-span psychology, individual agency with such a track record of (lack of) empirical
and related concepts such as choice and scrutiny?
decision-making have regularly appeared as There is a reason for this. As Dannefer (1984a)
featured terms (see Hitlin, this volume). Agency and Marshall and Clarke (Marshall 2005;
is one of Glen Elders ve principles of life Marshall and Clarke 2010) have noted, agency
course analysis: Individuals construct their own is more typically used as the equivalent of the
life course through the choices and actions they error term. As Marshall puts it, agency
take within the opportunities and constraints of functions in this theoretical perspective in the
history and social circumstances (e.g., Elder same way that unexplained variance functions
1998a: 961962; Elder and Johnson 2003). Elder in statistical models (2005: 63). Thus, as it is
and associates have been quite consistent in artic- used in life course analysis, a remarkable thing
ulating this view as an enduring component of about agency is that it is somewhat atypical in
the American life-course perspective that involves having achieved its conceptual status without any
choice-making: This view continues to be very requirement that it be empirically measured or
92 D. Dannefer et al.

analyzed. From the perspective of modeling and agency. One way to see this is to ask the question:
analysis, it is frequently given a free pass, allowed Suppose there were, e.g., no constraints of social
into the discourse of life course studies, even circumstances would it then be plausible to
asserted as a cardinal principle, despite its lack of say that individuals formulate their plans of
a clear empirical referent. It may be thought of as action and act unscathed by social structure?
the big easy of life course research. For anyone who takes seriously sociological
Initially, this may appear paradoxical: how analysis, the answer is, of course not. It is a
can it make sense to relegate a key variable to the fundamental principle of basic social theory that
domain of unexplained variance? However, it consciousness is itself fundamentally organized
does make sense, has to make sense, if the by social structure. Individual consciousness
variable is located in the domain of volition. By where purposes and agentic intentions are formu-
denition, it cannot be predicted or analyzed. lated and then externalized in social action is
Thus, agency is rendered off limits to sociological rst of all irreducibly shaped by culture-specic
explanation. language (Berger and Luckmann 1967; Sewell
The second, theoretical problem concerns the 1992), by the multilayered complex of structures
reliance of the dominant approach on a social that pattern social interaction, expectations, taste,
control or moral integration framework (e.g., and aesthetics within every individuals con-
Wexler 1977) that regards agency and structure sciousness, beginning while she is still in the
as counterposed domains of freedom and con- womb. This is close to what Bourdieu intends
straint, with the limits of the former determined with the concept of habitus: the habitus
by the strength of the latter. While viewed as engenders all the thoughts, all the perceptions,
outmoded in social theory and in many substan- and all the actions consistent with those condi-
tive subelds of sociology, this framework tions and no others (Bourdieu 1977: 95). This
remains resilient in some areas, including in the recognition is consistent with Meads analysis of
study of aging and the life course. For example, the social genesis of the self, which has been
such an assumption is implicit in the assertion elaborated by numerous other scholars in the
that Within the constraints of their world, interactionist (e.g., Blumer 1969; Holstein and
competent people are planful and make choices Gubrium 1999, 2000; Schaie and Hendricks
among alternatives that form and can recast their 2000) and constructivist (e.g., Berger and
life course (Gillespie and van der Lippe Luckmann 1967) traditions.
2014: 2) and in statements like Familial or class- When the social structuration of intentionality
based cultural expectations can also constrain is recognized, it quickly becomes clear that it is
choice and action (Thoits 2006: 314), as though meaningless to treat agency and structure as
freedom would be greater for those who have no counterposed forces in a zero-sum game. The
cultural expectations. Making the general point agency of each individual actor is profoundly and
even more explicitly, Csizmadia et al. (2012: 1) unavoidably organized by the social structures
assert that weakening structural constraints within which that individual lives her life, and
permit overall higher levels of human agency most of her agentic action is inevitably directed
The implicit idea is that behavior can be adequately toward the reproduction of existing patterns of
understood as a kind of scrimmage or tug-of-war social life. This is not necessarily an unhealthy
between structure and agency, which are thus state of affairs, for without such social organiza-
assumed to be two independent and often opposed tion of agency, individuals would lack the degree
and competing forces. of social integration required for a foundation
In relation to a basic sociological understanding of mental health and a coherent sense of self.
of the genesis and nature of the self and self-society The feral human individual child who has no
relations, the social control approach is simplistic language is free of humanly produced social
and inadequate. Such formulations fail to appre- structure, and as a result lacks the basic enable-
hend the dialectical interrelation of structure and ment of symbolic engagement even to articulate
Opening the Social: Sociological Imagination in Life Course Studies 93

or perhaps even to formulate an agentic plan of that has been organized by their own linguistic
action (Lane 1976; Perry and Svalavitz 2006). and social experience, from the beginning of the
This is why Giddens emphasized that struc- life course onward.
tures must not be conceptualized as simply plac- Thus conceived, agency does not exist as the
ing constraints on human agency, but as error term, relegated to the caprice of free choice.
enabling (Giddens 1977: 161; see also Rather, it is recognized as it empirically exists
Dannefer 1999a; Marshall 2005; Marshall and as an expression of consciousness that is consti-
Clarke 2010). tuted by and typically integrated into the habitus
What this means is that the role of social struc- in which it operates (Baars 1991). As we have
ture is not merely to constrain agency, thereby noted elsewhere (e.g., Dannefer 1999a), agentic
dening and limiting the options among which expression also serves to create the social rela-
an otherwise free actor may choose. Rather, tionships that sustain the world. The task for life
what social structure does is to shape and dene course scholars as for other sociologists includes
the individuals consciousness, within which the need to understand how agency is shaped and
intentions and purposes are externalized into directed by the eld of interaction within which
agentic action. This is, of course what occurs con- the individual resides.
tinuously beginning in very early life and continuing These criticisms do not exist as mere arcane
on through the life course, through the learning theoretical abstractions; they have substantive
of language and culture (including, e.g., skills implications for designing research, interpreting
and aesthetic preferences) of ones habitus. ndings and extrapolating real-life implications.
It is worth noting that an emergent feature of They have implications for how realistically and
modern society is the rise of age consciousness responsibly we as life course scholars confront
(Chudacoff 1989) and a societal reliance on age the broader realities within which the individuals
as an organizing principle, which has been key to we study constitute their own lives.
the institutionalization of the life course (Kohli
2007; Dannefer 2012). Age consciousness 2.1.1 Agency and Life Transitions
refers to that complex of generally unreected As an example, consider the implications of these
but inuential assumptions about normal aging, principles for reframing the discussion of decisions
life stages and other social constructs that are or choices made during key life transitions. A
assumed to be part of human nature,, although prominent if not dominant theme in the literature
they are historically recent. For better or worse, on life transitions features choice making. Studies
such assumptions largely dene the terms within often present survey data suggesting that choices
which contemporary individuals envision the life (e.g., to retire or relocate) are largely volitional,
course. emphasizing what phenomenologically seem
Of course, the life-course expectations set by to be the agentic aspects of such choices, with
ones habitus are culturally specic and contest- phrasing like retirement decisions or relocation
able. As Francesco Duina (2014: 1819; Chapter options. An added noteworthy aspect of the
5) has recently documented, they take radically North American approach is an increased capacity
different forms in North America from the forms for self-blame for adverse life-course outcomes
they take in Japan and many European countries even if beyond the individuals control (Dannefer
in the former celebrating choice and opportunity, 2000a, b), which appears even more likely among
in the latter focusing more on institutionally those who encountered early adversity and disad-
supported continuities, connections and interde- vantage (Newman 2006).
pendencies, whether provided by extended fam- The same applies to other life transitions as
ily relations or by state policy. Does this mean well. Especially in North America, cultural nar-
that Americans are more agentic than Japanese ratives to provide a positive reframing of even
or Scandinavians? Clearly not, if one understands putatively adverse transitions such as losing ones
that citizens in each of these contexts are formulating job are available. Regarding job loss, Francesco
their intentions with an individual consciousness Duina writes:
94 D. Dannefer et al.

It would be reasonable to expect such a crisis to window the ambivalences in the reasoning
fuel a very negative interpretation of this transition.
processes that are part of the process through
In several respects, the evidence validates this
expectation. But in other regards we encounter a which individual actors formulate their inten-
fair amount of positive and hopeful language tions. Consider the case of Annie Kasinecz, with
what seems like an unwelcome punishment harbors whom Davidson begins his article:
exciting, even incredible, possibilities and
potential (2014: 87) Annie Kasinecz has two different ways of explaining
why, at age 27, she still lives with her mom. In the
The same optimism applies to the uncertainties rst version the optimistic one she says that she
and stresses of the transition to adulthood, where is doing the sensible thing by living rent-free as she
plans her next career move. After graduating from
the phenomena such as boomerang children Loyola University Chicago, Kasinecz struggled to
or the crowded nest (Shaputis 2011) are often support herself in the midst of the recession, work-
framed in ways that emphasize agency and ing a series of unsatisfying jobs selling ads at the
positive resolution, both in scientic and public soon-to-be bankrupt Sun-Times, bagging groceries
at Whole Foods, bartending in order to pay down
discourse. For example, Settersten and Rays her student loans. But she inevitably grew frus-
(2010) acclaimed study of the young adults and trated with each job and found herself stuck in one
the transition to adulthood, Not Quite Adults, nancial mess after another. Now that shes back in
effectively conveys both the diversity and the her high-school bedroom, perhaps she can nally
focus on her long-term goals.
common themes in the experienced dilemmas But in the second version the bleaker one
and stratied life chances confronting their young Kasinecz admits that she fears that her moms
respondents, demonstrating in many cases how house in Downers Grove, Ill., half an hour west of
decisions are economically or culturally con- the city, has become a crutch. She has been living
in that old bedroom for four years and is nowhere
strained. Settersten and Ray make clear that closer to guring out what shes going to do with
many if not most young adults encounter an edu- her career. Everyone tells me to just pick something,
cational system that lacks the resources in terms she says, but I dont know what to pick.
of both education and guidance to facilitate their
learning potentials. For many young adults, they In the experience of the many young people of
write because the system as set up simply has whom Ms. Kasinecz serves as an example, it
no room for them there were no other options, seems clear that such processes are characterized
at least not ones that could be identied (2010: less by an experience of volition than by ambiva-
17); others are excluded from access to educa- lence, uncertainty and temporal contingencies.
tional options because (t)he economy makes As Settersten and Ray note, it is important to
that choice for them. (2010: xiii). In the context locate the empirical analysis of such meaning-
of such sobering and informative assessments making processes in a consideration of the
that point to serious external challenges for many broader social context. For young people, the
teenagers and young adults, their parallel emphasis broader context is daunting. The unemployment
on choice, reected in the subtitle Why rate of workers under 25 in the USA is more than
20-Somethings Are Choosing a Slower Path to double the overall unemployment rate (14.5 % as
Adulthood, and Why Its Good For Everyone opposed to about 6 % overall) and more than triple
seems disconnected from the actual experiences (18.1 %) if one includes missing youth who are
of young adults, which form a central part of the not seeking work. New college graduates unem-
books message,. ployment rate is better (8.5 %), but still a third
Some indication of the salience of such higher than the overall average (Bureau of Labor
interpretive processes at the level of lived experi- Statistics 2014; Shierholz et al. 2014: 56).
ence may also be gauged by the reaction to Adam Moreover, overall wages of young workers have
Davidsons (2014) NYT article, Its Ofcial: been in a long-term decline (Shierholz et al.
Boomerang Kids Wont Leave. This article 2014: 15). Discussions of agency and the choices
quickly generated more than 1,600 comments young people make must acknowledge the
an inordinate reaction from readers. It offers a existential context imposed by such realities.
Opening the Social: Sociological Imagination in Life Course Studies 95

2.1.2 Agency and Precariousness their lives are generated. Such trends are often
The same realities apply to the factors inuencing portrayed as troubling and in conict with the
transitions in adulthood. The U.S. economy, like interests of most individuals, or at least most
those of all late modern societies, has been individuals in advanced industrial societies who
reshaped by globalization and by pressures to enjoyed high levels of security and a relatively
flexibilize labor and the accompanying high standard of living through the decades of the
eradication of the social contract between the late twentieth century.
owners of capital and corporate employees and The results of these changes are generally
home towns. Such shifts have been accompanied, argued to include the rapid growth of contingent
among other things by widespread warnings the labor, boomerang children, chronic and structural
idea of a stable one-career life is over. underemployment of middle aged workers, and
Precariousness is a word that has been used retirees who cannot afford retirement. At the
on both sides of the Atlantic to describe the same time, much is sometimes made of success
circumstances faced by individuals that derive stories of celebrated individuals who, in response
from these powerful trends. In the USA, Robert to a sudden termination, manage to start a success-
Reich (2000: 98) writes that the new precarious- ful business and become featured role models of
ness is reected in the eradication of the institu- the entrepreneurial spirit. Our point is not to
tionalized life course dened by the corporate offer an evaluation of such trends, but simply to
world through much of the Twentieth Century: indicate the situations confronted by contempo-
Steady work a predictable level of pay from year rary individuals and the individual and collective
to year has disappeared for all but a handful of life course patterns that result form their efforts
working people To stay competitive organi- cannot adequately be characterized by an assump-
zations have to turn all xed costs into variable tion that the domain of agency represents a
costs As a result, earnings have become less and
less predictable. (Even) a job that is formally straightforward volitional act that requires no
classied as full time but whose pay varies consider- interrogation. The existential position from
ably from month to month is not, as a practical within which the world is viewed and intentions
matter, a job one can rely on. are formulated is a key point of every individuals
life-course narrative.
Although as much as a third of the work force From our perspective, it is a disconcerting
were estimated to be temporary workers, part signal of the continued force of the heuristic of
timers, independent contractors, or freelancers in containment, that in the face of such massive
the year 2000, Reich (2000:98) argues that global trends that have diminished opportunities
the portion of employees uncertain about how for the young and that are reected the erosion of
much theyll earn is far bigger than even economic security, the discourse and narratives
the largest estimates due to the shift in modes of life course analysis continues to emphasize the
of compensation of corporate employees to com- importance of agency and choice with relatively
missions, time-limited contracts and so on. little interrogation either of the reasoning pro-
In Europe, the term precariat (Standing 2011) cesses through which decisions are made, or of
refers to a large and growing segment of the how to respond to such changes.
population which is excluded from the primary Interestingly, in public and sometimes scientic
economy and are identied by a collective lack discourse, volitional action is blamed for helping
of opportunity and economic uncertainty and to create the adverse social circumstances. For
vulnerability shared by such groups. These devel- example, Settersten and Ray (2010: 30) attribute
opments make clear the degree to which the indi- the economic downturn to the extravagant choices
vidual agency is shaped by economic, political of parents:
and technical developments at the societal or cor- As the 20082009 economic crisis exposed,
porate levels, where the opportunity structures Americans had been living beyond their means.
within which individuals must make meaning of While debt was once a four-letter word in America,
96 D. Dannefer et al.

adults- young and old alike had become too course, neither the goals nor results of collective
comfortable with red ink. Savings were at historic
agency are necessarily salutary; fascism, terror-
lows. Families had fed their spending habits by
renancing their mortgages ism and suicide cults are among the enterprises
to which individuals have joined together in
Thus, choice and volition are invoked not only collective agency. Nevertheless, to the extent that
to describe the agentic action of youth, but also of life course scholars are seriously interested in
parents whose choices are offered as an expla- agency as an explanatory concept (as opposed
nation of the problems to which youth must to a name for error variance, as Marshall
respond. Such claims, while not entirely inaccurate, observed), a typology of forms of agency may
omit acknowledgment of the broader political offer a useful way to move the understanding of
and economic forces that are widely recognized such issues forward. Clearly, such a development
as underlying the 2008 recession, as discussed by would resonate strongly with C. Wright Mills
Reich (above) and many others. call to study history and biography conjointly, as
In any case, the problems identied by Settersten is often and appropriately invoked as vision to
and Ray at the individual level and by Reich at which life-course scholars should respond
the structural level do indeed call forth agentic (Aronowitz 2003, 2012).
responses from all individuals, at every point in In sum, an empirical analysis of agency must, at
the life course and on a daily basis. minimum, recognize the social organization of the
A nal issue in a consideration of agency con- consciousness from within which intentions are
cerns the form of those agentic responses. As internalized into action. Such recognition is not
discussed within the life course framework possible from the assumptions of the functional-
(including in our own prior work), virtually all ref- developmental nexus, which centers on a choice-
erences to agency conceptualize agentic action as making individual, navigating a largely xed and
an individual matter, coping with the local micro- legitimate social context. And in such a context, the
world confronted on a daily basis. Yet especially very notion of collective action for social change
for dealing with crises that impact large segments has no place, since social change appears as neither
of a societys population, collective responses warranted (since the system is prejudged as by and
exemplied by large-scale social movements large legitimate) nor possible.
have often been decisive. In Europe, the concept of The implication for life-course researchers is
the precariat has derived its viability, at least in to incorporate the concept of agency carefully,
substantial part, from visible social movements conditional on two premises: (a) a recognition of
through which marginalized and excluded indi- the factors that organize agentic intentions in the
viduals have come together. The Occupy Wall actors consciousness, so that agency has the
Street movement obviously tapped into similar potential to carry explanatory power, thereby
concerns in the USA. While no such movements making it a meaningful independent variable or
on either of these continents have to date been ef- (b) if agency is positioned as a dependent
cacious, they represent a form of agentic response variable, so that the research objective is to show
expressed now by millions of individuals. how agency itself is organized by broader
For life course scholars and other social scien- forces (which may represent an interaction of
tists interested in the possibilities of agentic both physiological and social forces).
action, they represent an opportunity to consider
whether it may be useful to contrast different
types of agency (based on the extent to which it is 2.2 Early-Life Circumstances
individual vs. collective, spontaneous vs. formally in Life Course Context
organized, etc.) A recognition of the potentials of
collective agency in life course studies and Recently in life course research, scholars have
related elds is rare, although not altogether undertaken signicant new efforts to articulate
absent (see Elder 1974; Crockett 2002). Of the role of early-life circumstances for adult
Opening the Social: Sociological Imagination in Life Course Studies 97

outcomes. The early origins literature has The second family of explanatory mechansisms
exploded over a number of areas of social are labeled Path Dependency. Pathway or chains
science, ranging in both the childhood circum- of risk models frame social and health circum-
stances and in the adult outcomes under study. stances in very early life as the starting berth for
These include: the inuence of childhood family a lifetime of opportunities, or for stressors or
structure on adult wealth attainment (Kiester barriers (Dannefer 2003a; Kuh and Ben-Shlomo
2004); child abuse on adult religiosity (Bierman 2004; Ferraro and Shippee 2009). A unique fea-
2005); early-childhood education on adult utili- ture of the Path Dependent perspective is the
zation of social welfare (Cunha and Heckman operating assumption that these risk pathways or
2007); childhood adversity on adult life evalu- accumulative processes are generated and sus-
ations (Schaefer et al. 2011); and adolescent tained by a relatively xed opportunity structure
delinquency on long-term crime careers (Moftt through which an individual must navigate.
and Caspi 2001). [Think of tournament mobility (Rosenbaum
As scholars begin to sort through the bevy of 1978): there is a predictable order in the progres-
empirical ndings that link early-life circum- sion through brackets, and a set of probabilities
stances to adult outcomes, a number of theoreti- based on starting berth.] Following this logic, the
cal frameworks have emerged that differ in the predominant solution for escaping such a deter-
hypothesized magnitude of direct and indirect ministic structure lies in individual-level action
inuence of the early-life circumstances on later- and characteristics, such as positive coping or
life health, wealth, and well-being. These include, resilience (Landes et al. 2014; Zimmer-Gembeck
but are not limited to, fetal programming, critical and Skinner 2011), or attempts to improve ones
and sensitive periods, pathway or chains of risk, relative status (Elman and ORand 2004). Although
and cumulative dis/advantage. A more thorough not taken up explicitly here, the latent class
treatment of the family of these hypotheses appears perspective, which argues that ones early, mid-
in other places (Ferraro 2011; Shanahan 2013; dle, and later life socioeconomic circumstances
Kuh and Ben-Shlomo 2004) and we make no represent an underlying construct of placement in
attempt to summarize them here. For our pur- the social hierarchy, similarly relies on assump-
poses, most of the relevant explanatory frame- tions of a stable opportunity structure.
works can be sorted into two broad categories, These two frameworks have in common
each in its own way located at the nexus of func- the intent to locate the role of early-life social
tionalism-developmentalism. The rst we term circumstances in a broader life course framework.
the Latency framework, which is drawn from the Our attempt to distinguish between these two
critical/sensitive period model and focuses on categories is not intended to exacerbate a per-
acute exposures in very early life that have a last- ceived competition between them as explanatory
ing impact on growth, development, and func- frameworks. Indeed, conceptual and empirical
tioning during adulthood (Hertzman et al. 2001; work suggests strongly that both kinds of processes
Barker 1995; Kuh and Ben-Shlomo 2004). These are often operant simultaneously (Shanahan
crystalized individual-level characteristics (IQ, 2013; Montez and Hayward 2011; Ferraro and
organ function, stress sensitivity) get carried for- Shippee 2009). However, we will show that the
ward throughout life and their explanatory power research assumptions and analytical practices
is thus seen as residing within the individual, of each of these two approaches, in their own
with a presumption of relative permanency. distinct way, serve to contain the potential
Although exposures in other parts of the life course explanatory inuence of the social world. Below,
may have inuence, these are frequently framed as we take up each of these categories separately to
effect modiers, to use Kuh and Ben-Shlomo discuss their embeddedness in the functional-
(2004)s terminology, or as mediators. Thus the developmental paradigm, and then present some
presumed causal link of an early life exposure evidence pointing to the potential expansiveness
and an adult outcome remains intact. of the sociological imagination.
98 D. Dannefer et al.

2.2.1 Latency of Early Childhood the health-related development and life-course


Exposures outcomes. The problem is not that the inuence
Just as the life course framework was solidifying of the social world is not considered in the other
in the early- to mid-1990s, emphasizing not only parts of the life course, but that its causality is
the proximate inuences on health and well- often limited to early childhood when relatively
being, but also the more fundamental causes, new stable characteristics are assumed to be locked
epidemiologic work presented evidence that in to the individual. This is, of course, a clear
opened up fresh possibilities with regard to the example of Time 1 Encapsulation (Dannefer and
importance of very early life, even in utero, for Kelley-Moore 2009; Hagestad and Dannefer
understanding the causes of adult chronic dis- 2001), which implies that the impact of the social
ease. In 1995, Barker and colleagues published world is internalized early in life, and thus can be
an observational study showing that adults with thought of as a stable, obdurate characteristic for
cardiovascular disease had been disproportion- the individual that is carried forward through
ately low birth weight babies. time.
Now famously called the Barker hypothesis, When it is assumed that experience in adulthood
this study helped launch a rich line of inquiry in does not matter beyond mediating the effects of
epidemiology and biology referred to as Fetal very early life, or it serves solely as an effect
Origins of Adult Disease (FOAD). Drawing from modier, the potential dynamic, even transforma-
the critical or sensitive period model in epidemi- tive, inuence of social world is ignored. If, as we
ology whereby an early-life exposure may be argue at the beginning of this chapter, the social
more inuential on later-life health outcomes world is constitutive, continuously shaping (and
because it occurs during a period of vulnerability either stabilizing or changing) the individual at
or rapid development, early origins hypotheses all ages, then privileging the social worlds
started to appear in epidemiology, sociology, and impact at a particular age neglects the broader
public health (Gluckman and Hanson 2008; Kuh potentials of the social world to shape individuals
and Ben-Shlomo 2004; Doblhammer 2004). [see throughout the life course.
Skogen and Overland (2012) for a review of the This tension between the Latency perspective
development of FOAD research.] and one that purports a more dynamic role of the
As one marker of the explosion of interest in social world over the life course can be seen in the
the early-origins idea, Barkers (1995) original interpretation and policy application of the Nobel
study was cited more than 2,500 times between Economist James Heckmans research on the
1995 and 2014. The vast majority of the empirical impact of early-childhood education on long-term
papers citing Barkers work focused on physio- outcomes. Heckmans research has focused on the
logic or biologic mechanisms operant in neonates economic returns of intervention and investment
that may link fetal or early-life conditions to in disadvantaged children and adolescents, focus-
adult disease. Yet roughly one-quarter of those ing specically on the exact timing in young life
citing Barker have been social science studies when such intervention would yield the greatest
examining the potential link between childhood return in adulthood. For instance, he and his col-
circumstances and an observed health outcome in leagues (Cunha et al. 2010) undertook an evalua-
adulthood. tion of an early-childhood program, concluding
The Latency framework in life course studies that every dollar spent on a 4-year-old child from a
is an important rst step in specifying the impact disadvantaged family would range between a 7 %
of the early-life social environment for sociological and 10 % annual return to society in the form of
analysis because it has identied and documented less dependence on social welfare, less frequent
an array of exposures that may decisively inu- engagement with the criminal justice system, and
ence the individuals development or potential. increased income tax revenue. Compelling eco-
Thus, Latency-based research compels attention nomic data such this has fueled wide bi-partisan
to the importance of childhood social factors in support for early childhood education programs
Opening the Social: Sociological Imagination in Life Course Studies 99

such as Head Start. Interestingly, the recent eco- on continued and ongoing investment beyond the
nomic recession actually saw an increase in invest- program years that target ages 35. This is largely
ment in such programs, as part of a comprehensive attributed to the fact that the melt from the
strategy to combat the long-term effects of the eco- early-life intervention begins as soon as the chil-
nomic slide of American families (Business Week dren return to their disadvantaged environments
January 2014). and relatively lower levels of social investment
Heckmans interpretation of his work provides (Currie and Thomas 1995).
a strong endorsement for the critical/sensitive The greatest challenge to the argument that the
period model for developing cognitive and non- effectiveness of a critical/sensitive period invest-
cognitive (e.g., social interaction) skills. As he ment occurs solely in early childhood is data
wrote in a New York Times editorial, Quality from Cunha and Heckman (2007) showing that
early childhood programs for disadvantaged the returns in adulthood with regard high school
children foster human ourishing and they graduation, interactions with the criminal justice
improve our economic productivity in the pro- system, and receipt of public assistance are
cess. There is no trade-off between equity and the greatest, by more than three times for certain
efciency, as there is for other social programs. outcomes, when early intervention is partnered
Early investment in the lives of disadvantaged with continuous intervention through late
children will help reduce inequality, in both the adolescence. In Table 1, reproduced from their
short and the long run (Heckman 2013). report in the American Economic Review, we
From an economic point of view, which is observe indicators of several positive and adverse
Heckmans vantage point, his research certainly outcomes for a control group (no intervention)
supports early childhood investment as economi- and three experimental groups: early-childhood
cally efcient to the overall system. However, it intervention only; adolescent intervention only;
is a mistake to use his work as denitive evidence both early-childhood and adolescent interven-
of latency effects of early childhood on human tion. The rates of success, for every measure,
development. Indeed, a closer examination of were substantially higher for those who receive
Heckmans own evidence, as well as comple- the sustained intervention in both early childhood
mentary research, shows clearly that if early and adolescence.
investment in disadvantaged children (the target This means, of course, that the salutary effects
population for programs such as Head Start) is of Head Start or other effective early childhood
not followed up by later investment, its effect at programs are not internalized in ways that mean
later ages is lessened substantially (Cunha et al. they no longer require further support; rather it
2010). In other words, its success is contingent means that the social environment continues to

Table 1 Comparison of different investment strategiesa,b,c


Balanced early-
Early-childhood Adolescent only childhood and
Baseline only intervention intervention adolescent intervention
High school graduation 0.4109 0.6579 0.6391 0.9135
Enrollment in college 0.0448 0.1264 0.1165 0.3755
Conviction 0.2276 0.1710 0.1773 0.1083
Probation 0.2152 0.1487 0.1562 0.0815
Welfare 0.1767 0.0950 0.0968 0.0259
a
Reproduced from Cunha and Heckman (2007)
b
Intervention population: Disadvantaged Children: First Decile in the Distribution of Cognitive and Noncognitive Skills
at Age 6; Mothers are in the First Decile in the Distribution of Cognitive and Noncognitive Skills at Ages 1421
c
Proportion of sample
100 D. Dannefer et al.

be important after the pre-school period. Thus a circumstance to an adult outcome but are not
supportive social environment (or lack thereof) is explicitly explored in a given study. If, as
decisive well beyond early childhood. Shanahan (2013) states, such rigorous inquiry
To be clear, we are not arguing that the social requires more extensive multilevel and longitudi-
environment during the fetal period, infancy and nal data than are presently available to capture
early childhood should be viewed as unimportant. fully the dynamics between individuals and their
Clearly, they are extremely important. But the social worlds over the entire life course, then
true measure of the staying power of such early- even the most high quality empirical studies must
life effects independent of later environmental necessarily acknowledge their inability to engage
conditions with which they are likely correlated in full hypothesis testing. Thus, set in motion
cannot be gauged unless their relationship to/ may be used most often as a conceptual stand-in
interaction with later environmental conditions to convey the complexities of the dynamics over
is also measured. This is not just a technical the life course that are as-yet unobserved.
academic point. As we see with Heckman and There is no utility in complaining that suf-
colleagues work above, it is a matter with real- ficient data are not yet available to test fully the
life implications when we think about the chances life course models of path-dependency and that is
and opportunities for change of those who certainly not our intent here. However, phrases
endured disadvantage in childhood. such as set in motion, or chains of risk are
conceptually problematic, because they generally
2.2.2 Path Dependency do not include an attempt to specify the actual
The second framework, Path Dependency, relates structural forces in adulthood that sustain the
to a broader family of theories that attempt to path. Thus, social structure is both assumed to be
articulate how adult outcomes manifest from rigid and determinative of individual life chances,
experiences, opportunities, and reactions to these yet kept invisible, keeping social structure in a
that occur throughout the life course. Such an conceptual black box.
ambitious undertaking covers a relatively longer The net result is that one must accept the
period of time and inevitably yields numerous premises that opportunity structures are not only
hypotheses, ranging from mechanistic chains of stable and are part of the realities of the social
risk (Kuh and Ben-Shlomo 2004) to more abstract order, but also are homogeneous and undifferen-
concepts of cumulative dis/advantage processes tiated. Thus, that the exposure-outcome relation-
(Dannefer 2003a; Crystal and Shea 2002). Some ship will work relatively similarly for individuals
scholars have attempted to create comprehensive with the same early-life circumstances. In reality,
frameworks that blend these levels of analysis the opportunity structures of adulthood even
(Ferraro and Shippee 2009; Shanahan 2013; for those who begin with similar childhood
Slavich and Cole 2013). backgrounds, are not homogeneous and undif-
Most notably across this family of theories, ferentiated. Key sociological work already
and for our purpose here, these Path Dependence challenges the idea that adverse childhood cir-
frameworks are distinguished from the Latency cumstances set in motion pathways of disad-
framework based on their treatment, either vantage and has emphasized the important roles
explicitly or implicitly, of very early life (see Kuh of other social institutions such as family, educa-
and Ben-Shlomo 2004 for discussion). Indeed, tion, and the military. Alexander et al. (2014), in
some scholars who use these frameworks have their long-term study of disadvantaged youth in
referred to early-life conditions as setting in Baltimore, documented differences in adult
motion pathways or risks that play out over the education, employment, and family status and
life course (c.f., Haas 2008; Montez and Hayward emphasized important differentiators in these
2011; Ferraro 2011). Such phrasing is often outcomes. In sum, they found that interactions
invoked to refer to more complex accumulative with key social institutions at sensitive periods
processes that likely operate to link a childhood in the life course after childhood were the
Opening the Social: Sociological Imagination in Life Course Studies 101

predictors of long-term outcomes. Likewise, Laub large-scale studies such as The Health and Retirement
and Sampson (2003) found that early experiences Study (HRS) and National Longitudinal Study
with delinquency did not ubiquitously set in of Adolescent Health (Add Health) in the USA
motion careers in crime. Rather, some young and similar studies in other countries. Gene-
mens transformative encounters with social environment interaction can be broadly dened
institutions such as marriage or the military were as situations in which genetic effects connected
better predictors of their long-term trajectories. to a phenotype are dependent upon variability in
Relegating causality to opportunity structures the environment, or when genes modify an
in early life is, by default, to contain explanation organisms sensitivity to particular environmen-
for observed variation in outcomes to individual- tal features (Seabrook and Avison 2010). Over
level characteristics and action. Thus, individual the past decade, a number of intriguing sets of
characteristics, such as coping resources (Landes ndings have been published to indicate that
et al. 2014; Zimmer-Gembeck and Skinner 2011) there are few if any situations in which such G-E
or attempts to improve ones relative status dynamics are not operative in the life course.
(Elman and ORand 2004) are elevated in explan- We present the study of G-E interactions as a
atory power above adult social circumstances. third example of the pervasive assumptions of the
Some scholars have leveled this critique on functional-developmental nexus that are driving
the eld of life course sociology, particularly life course inquiry. Of particular relevance here,
regarding path-dependent formulations of the life we take up the concern that the individual genetic
course that include early-life structural causality characteristics are treated as xed, and as a result
followed by individual-level mediators in adult- the role of environment in G-E interactions is
hood (Hoffman 2008; Link and Phelan 1995). In limited to various modes social regulation or
short, while the path-dependent approach recog- control of genetic predispositions, for example,
nizes that the opportunity structures of adulthood through legal or normative constraints, through
frame risk and opportunity as a gauntlet through which environment either suppresses or allows
which the individual must navigate, it is unwar- for behavior that expresses specic genetic ten-
ranted to consider it an undifferentiated gauntlet dencies. This approach has yielded an array of
that affects all individuals with similar early interesting results (E.g., Boardman et al. 2010;
backgrounds in the same way. To do so is to miss Guo et al. 2009; Shanahan 2003; Shanahan and
the chance to observe the active social dynamics Hofer 2005) and has enjoyed considerable
that occur over the entire life course, relegating popularity. Yet it represents another example of
explanation back to the individual level. containing the social, because it restricts consid-
eration of the effect of social forces to behavioral
outcomes, and does not acknowledge how social
2.3 From G-E 1.0 to G-E 2.0: factors also affect gene expression at the molecu-
Gene-Environment lar level. As knowledge of epigenetic and related
Interactions, Social Control interactive processes through which environment
and the Social Regulation affects gene expression at the molecular level, it
of Genetic Expression is clear that the social environment could interact
with genes over the entire life course. We explicate
2.3.1 G-E Interaction 1.0: Social these points in more detail below.
Control of Gene-Based In the logic of social regulation, a key premise
Characteristics of G-E interaction is that the manifestation of
Interest in gene-environment interaction is expand- genetic predispositions, which are considered
ing rapidly among life-course scholars, driven xed and unchanging, will vary by the regulatory
in no small part by its potential relevance to a aspects of the environment. Specically, the focus
variety of life course processes and outcomes and is on whether an environment inhibits, encourages
also by the incorporation of genomic data into or is essentially indifferent to behavior that
102 D. Dannefer et al.

manifests or expresses a hypothesized or these and related approaches, individuals with


presumed genetic predisposition. The general different genotypes are hypothesized to react
argument is that gene-driven variability in activ- differently under different environmental condi-
ity patterns and behavior will be based on salient tions: an individual with sensitive genetic makeup
features of the environment in relation to xed is likely to have worse outcomes in an unfavorable
genetic predispositions. As Shanahan and Hofer environment, and better outcomes in a favorable
(2005) earlier put it: environment, while the environment will have no
[I]n settings marked by high levels of social effect on individuals with a less sensitive makeup.
control, [heritability] attenuates, whereas in contexts Such models may be thought of as an elaborated
marked by low levels of social control, [heritability] version of the trigger and/or compensation
increases. In other words, in circumstances marked models described by Shanahan (2005). The mod-
by high levels of social control, a large percentage
of the sample exhibits the same phenotype; in els continue to assume a tension between envi-
settings marked by low social control, peoples ronmental variation and a xed characteristics of
choices and behaviors are more apt to reect their the genome, in this case, genetic sensitivity. This
genotype. (p. 68) can be seen as an extension of Belsky, Steinberg
and Drapers (1991) earlier arguments postulat-
It should be noted that Shanahan and Hofer ing genetically xed differences in reproductive
offer a fourfold typology of G-E interactions strategies.
(trigger, control, compensation enhancement) in In such a case, gene-based characteristics
which the term control is used to describe only remain viewed as xed and unchanging, yet their
one of four alternative types. However, all four relevance for behavior is predicted to vary under
types share the key feature of a xed genetic differing environmental conditions. Therefore,
tendency whose manifestation in behavior is although one cannot understand or predict out-
regulated by environmental factors, and it is in comes without considering genes and environment
this broader sense that we use the term social conjointly, genes themselves are understood as
control. For example, Guo et al. (2009) tested independent of environmental inuence. Thus,
the hypothesis that the genetic contribution to despite their added renement, DST, GDSE and
adolescent drinking is subject to the inuence of related approaches, are, like the social control
friends drinking behaviors (2009: 221) and found approach, consistent with the Central Dogma of
that a highly controlled environment (i.e., low molecular biology, which assumes that genes are
level of friends drinking behavior) contributes to xed and hard-wired in James Watsons words,
the low level of genetic contribution to alcohol the central thing from which everything else
use. Presumably, a less controlled environment ows. (Rennie 2003: 69), and not subject to exter-
in which friends consume relatively high levels nal inuences upon their own internal processes of
of alcohol would enhance that genetic propensity, effects. While this familiar notion obviously
for better or worse. remains inuential, the assumptions on which it is
As research in this general area has progressed, based are facing a growing set of challenges.
numerous renements and novel forms of G-E
processes and effects have been proposed, For 2.3.2 From G-E 1.0 to G-E 2.0
example, differential susceptibility theory (DST) While social science research on G-E interactions
proposes that certain genetic variants may be has grown in the past decade, the same time
more sensitive that others to environmental period has also seen a concomitant and dramatic
conditions, whether positive or negative (Ellis expansion of discoveries in molecular genetics
et al. 2011). Such ideas are also being applied and related elds of demonstrating that genetic
and extended by social scientists. As one exam- effects are actually not so independent of external
ple Mitchell et al. (2013) describe a similar inuence as the Central Dogma assumed. Instead,
approach which they term genetic differential gene expression at the molecular level is recognized
sensitivity to social environment (GDSE). In as conditional on forces external to the cell,
Opening the Social: Sociological Imagination in Life Course Studies 103

including forces in the social environment. Thus, genome to shape complex behavioral phenotypes
it is increasingly recognized that social expe- and susceptibility to disease (2013: 2). Such
riences regulate genetic activity at the level of discoveries encompass the study of domains
intracellular chemistry (Shanahan 2013: 1; see including epigenetics (involving processes such
also Cole 2009; Dannefer 2011; Gluckman and as DNA methylation and histone modication),
Hanson 2008). The growing recognition of the retrotransposition, transcription, and other mech-
extent the chemistry of the cell itself is regulated anisms surrounding gene expression. Based on a
by the environment has led to a suggestion that recognition of the inadequacy of the assumptions
the study of G-E interactions has entered what of the Central Dogma, these approaches entail a
some scholars have called a postgenomic new order of magnitude for sociological expla-
(Plomin et al. 2002) or neogenomic (Charney nation in G-E interactions throughout the life
2012) era, and the emergence of several new sub- course.
elds of research, including environmental epi- Recognition of the importance of exogenous
genetics and social genomics. Environmental effects on gene expression has been growing
epigenetics has been dened as the study of fac- since the 1970s (Holliday 2006). The signicance
tors such as nutrition, pollution, and stress in of this broadly ecological approach is grounded
relation to gene regulation (Landecker and in the recognition that genetic effects can only be
Panofsky 2013: 334). In addition to the physical realized if genes are expressed, and gene expres-
environment that contains proximate determi- sion is conditioned by external inuences,
nants of gene regulation, more attention has been ranging from the cellular environment to cultural
attached to social environment which organizes symbols whose physical impact is mediated
individuals daily life and determining their through consciousness or, as Shanahan notes, by
exposure to gene-regulating conditions. complex mediating chains involving many lev-
To borrow Silicon Valley language, these els of analysis, possibly extending, for example,
developments can be thought of as heralding a from political economies to peoples reactions to
paradigm shift from G-E 1.0 to G-E 2.0 their immediate circumstance to intracellular
from research that assumes a xed and chemi- mechanisms (2013: 2).
cally impenetrable genome, to research that During cell differentiation and even day-to-day
recognizes the broad ranging environmental cell reproduction, genes can be turned on and off
impacts on gene expression. Thus, these new in response to the transcription factors which can
developments create a tension in social science be directly and indirectly inuenced by environ-
between whose who continue to treat the social mental signals (Holliday 2006; Slavich and Cole
environment as a functional, invisible press on 2013). With the advance in molecular biology and
individual genetic variation and those who seek the deepening understanding of the multiple levels
to expand the role of the social environment into of environment, it is recognized that genotypes
more dynamic, interactive and transformative and environment are mutually engaging in produc-
types of causality. In this way, we can view GE ing differentiated behaviors (Shanahan and
1.0 as is consistent with the functional- Boardman 2009; Slavich and Cole 2013). In real-
developmental nexus, while GE 2.0 refers to ity, the environment exerts more complicated
research and findings that invite an expand inuence on gene expression through an array of
the scope of sociological explanation by demon- mechanisms. For instance, Slavich and Cole
strating how social factors affect gene expression (2013) introduce social signal transduction as a
at the molecular level. mechanism that external social conditions get
Social genomics is described as the study of converted into genome-regulating biochemistry
processes by which the external social world (p. 6) characterized by the central role of central
gets not only under the skin but onto the neuron system. As Cole (2009) illustrates:
104 D. Dannefer et al.

Social-environmental processes regulate human biology. For instance, epidemiology has


gene expression by activating central nervous
embraced this perspective to study the developmen-
system processes that subsequently inuence
hormone and neurotransmitter activity in the tal process of health, and life course epidemiol-
periphery of body. Peripheral signaling molecules ogy has incorporated the dynamic social process
interact with cellular receptors to activate tran- that shape individuals health outcomes by study-
scription factors, which bind to characteristic DNA
ing both proximate and distal causes (Halfon and
motifs in gene promoters to initiate (or suppress)
gene expression. (Cole 2009: 133) Social signals Hochstein 2002; Kuh et al. 2003).
can be transduced through epigenetic processes Within sociology generally and specically
(e.g., DNA methylation and histone modication) within the study of the life course, a number of
which regulate gene expression.
scholars have called for more research attention
to the interaction of social processes and gene
Unlike the xed sequence of DNA which is expression (e.g., Dannefer 2011; Landecker and
determined in germ cells, epigenetic and other Panofsky 2013; Rose and Rose 2014; Shanahan
processes regulating gene expression are dynamic, 2013). Shanahan has traced out in detail the
partially regulated by social life, and occur potential payoffs of such possibilities. As social
throughout the life course. At the beginning of genomics entails the possibility of multilevel
the life course, such effects include fetal program- modeling of social context beyond the proximate
ming which occurs in response to a specic risk factors that may contribute to explanations
maternal environment as illustrated by the Barker based on critical and sensitive period effects, to
hypothesis (Barker 1995; see also Gluckman and consider also accumulation effects and more dis-
Hanson 2008) and then, possibly, by biological tal contextual factors.
embeddedness which is postulated to occur in Such developments notwithstanding, very
sensitive periods, mainly in childhood. Many little sociological research has yet confronted
such mechanisms also appear to operate through- this possibility. Apart from the conceptualization
out the adult life course, as individuals respond to and measurement difcult of genetic and envi-
environmental stimuli. Some such processes ronmental effects, researchers have raised cau-
are characterized by reversibility, which holds tions concerning unmeasured genetic variation.
the possibility for interventions to counteract the Gene-environment correlation (rGE) can inuence
adverse early experiences (Holliday 2006). the validity of causal inference, because mea-
While some have contended that epigenetics sured environments may be correlated with
is a process through which nurture shapes nature unmeasured genetic variation (if, for example,
(Powledge 2011), in many cases such effects may the perceived stress level which is measured as an
interact with specic features of the genome. environmental factor is actually a manifestation
Thus, it is more precise to recognize that such of certain genetic makeup) (Fletcher and Conley
outcomes cannot be possibly understood without 2013: e1). In other cases, gene-gene interaction
crossing or even dissolving such disciplinary has been proposed to account for a portion of
lines: The continuous interaction of the social the identied G-E interaction, in the form of
and the biochemical appears to be inherent in measured genotypes with unrecognized genetic
nature itself. In other words, the mechanisms variation in the environment moderator (Manuck
through which social experience of individuals and McCaffery 2014: 62). Other researchers
shape the performance of their organisms is an distance themselves from epigenetics and social
unavoidable feature of the phenomena being genomics for other reasons, for example, because
studied in the life sciences. it is seen as difcult, or of limited importance
As these developments have occurred, biolo- (see Bliss 2012). In a study of cohort differences
gists, geneticists and others working on such in smoking behavior, environmental effects have
problems from a biological perspective have been dismissed as noise that obstructs researchers
themselves articulated their own life course per- ability to see true genetic effects in their
spective (Adair 2007; Gluckman and Hanson presumed purity (see, e.g., Boardman et al.
2008) and for an ecological developmental 2010). However articulated, such responses have
Opening the Social: Sociological Imagination in Life Course Studies 105

in common a disinclination even to consider the studies (see also Shanahan 2013: 5), or perhaps
importance of social causation of gene expression. by denitional issues or by other factors in the
It is true that some researchers and others have relationship between social location and subjec-
made increasing reference to the need for further tive factors such as the perception of loneliness.
explorations of epigenetics and other processes If, indeed, no empirical evidence exists of a rela-
relevant to social genomics (e.g. Dannefer 2011; tionship between, e.g., social class and subjective
Guo et al. 2008; Mitchell et al. 2013). However, perceptions of loneliness or stress, while at the
among social scientists this possibility has yet to same time perceptions of stress are clearly related
gain much traction in the design of research. to gene expression in ways that are consequential
Instead, researchers continue to rely on the for health, it would raise intriguing questions of
assumption of a xed effect from a genetic vari- how to reconcile such a nding with the socio-
ant (e.g., Boardman et al. 2014; Guo et al. 2008). economic gradient.
Ironically, interest in the genetic consequences of Such a nding could pose an especially vex-
social forces seems, at least so far, to be of greater ing challenge for those who work within theoreti-
interest to biologists (Gilbert and Epel 2009; cal traditions that emphasize the role of ideology
Gluckman and Hanson 2008) and psychologists and false consciousness in accounting for health
(Slavich and Cole 2013) than it does to sociologi- differentials (through, e.g., unhealthy consumer
cal researchers. practices in nutrition, smoking, etc.) For if dont
More specic to sociological approaches to worry, be happy can be shown to have a salutary
theorizing the life course, research on social genom- health effect independent of ones objective social
ics is raising issues that prompt new problems location and circumstances, then false conscious-
and opportunities for life course theorizing. ness may be argued to be more adaptive, at least
Interestingly, social genomics has been launched in some circumstances than an objective
as a eld by psychologists, and from the vantage appraisal of ones circumstances. Such a nding
point of basic ideas in sociological theory and could return us to the questions of the conditions
the received wisdom of much social science, under which a long-term habituation to what
the implications of social genomics are anything objectively appear to be lower expectations may
but simple and straightforward. foster mental health better than do high aspirations
For example, in a pioneering line of research, (see., e.g., Campbell et al. 1976; Phillips 1978).
Cole and associates have demonstrated the impact of It must be acknowledged that the challenges
subjective awareness (e.g., of loneliness and stress) of studying the social regulation of gene expres-
on gene expression that impacts immune system sion at the molecular level is unquestionably a
functioning (e.g., Cole et al. 2011; Slavich and Cole daunting enterprise. Moreover, its complexity is
2013). Yet surprisingly at least from a social science amplied when considered in life course perspec-
perspective they nd that measurable social char- tive. Yet if a fundamental objective of science
acteristics such as socioeconomic status, network is, in Herbert Blumers terms, to be true to its
size or marital status (Slavich and Cole 2013: 67) subject matter, it is the complex reality that must
are not directly related to gene expression. be pursued. In this particular case, it has to be
Such a set of ndings raises a number of noted that these ndings point not only to a more
interesting sociological questions. First, how intricate set of G-E interactions, but also to the
should we understand this lack of association, potential of a considerably enhanced explanatory
given the centrality of the socioeconomic gra- role for sociology. Despite what some have seen
dient, marriage and other social circumstances in as a slow start (Bliss 2012), this appears to be an
accounting for variation in health, including area where an expanded interest in the explana-
mental health? What is the relationship between tory role of social forces is growing. It is perhaps
isolation and such social characteristics (socio- ironic that this has been driven most energetically
economic status, network size or marital status)? by discoveries in biology (e.g., Gluckman and
In part, this may be accounted for by the small Hanson 2008) and psychology (Cole 2009; Slavich
and non-representative samples used in these and Cole 2013).
106 D. Dannefer et al.

3 Conclusion Similarly, we are operating in a world in which


the overall political economy reinforces tenden-
The fresh opportunities for specifying sociological cies toward social exclusion that impacts life-
explanation in the study of life course phenom- course trajectories and the age-graded institutions
ena extend across wide domains of inquiry (e.g., schools, careers, nursing homes) within
from the biosocial (with the continued expanse of which they constituted. In this way, institutions
new discoveries of the importance of social cau- are operating in ways that are often not congruent
sation in G-E interactions) to political economy with the needs of the individuals who occupy or
(with its dramatic implications for the shaping of seek to occupy them. These are matters that are
agentic intentions and the possibilities of agentic not easily grasped within framework that assumes
action). the legitimacy of the social order and the congru-
The prospects for engaging the sociological ence of interests of individuals and of social insti-
imagination to pursue fully such opportunities tutions. It requires an intellectual approach that
will be enhanced by adopting a paradigmatic can apprehend the life course in the context of a
willingness to interrogate more fully the reach of more critical and thoroughgoing interrogation of
social forces into individual consciousness and how such institutions function operate. Thus,
individual biology, as well as the impact of forces both individual development and social systems
operating at the level of political economy. Such and institutions must be understood as constituted
an energetic interrogation will require an aware- through that human activity that daily reproduces
ness and willingness to reject long-standing the social world.
paradigmatic assumptions and theoretical com- It is hopeful to consider that the parallel but
mitments within the sociology of the life course seemingly unrelated events of (a) the contempo-
and related elds (including human development, rary global economic strains, reected in grow-
the sociology of age and social gerontology) ing inequality and social exclusion, and (b) new
which have systematically tended to inhibit or discoveries in the domain of G-E interations,
contain the possibilities of sociological expla- may jointly encourage, cajole and ultimately
nation. As we have demonstrated, assumptions compel life course theorizing to embrace an ori-
inherent to sociological functionalism and devel- entation that is truer to the goal of advancing a
opmental theory dene reality in ways that tend full understanding of our subject matter, and of
to restrict the depth with which either domain can simultaneously advancing the sociological
be explored. By examining recent developments imagination.
in three domains (1) agency and social action,
(2) early childhood effects on adult outcomes and Acknowledgment The authors wish to thank Elaine
(3) gene-environment interaction we have Dannefer and Neil Pendleton for comments on an earlier
version of this paper. We also wish to thank Luma Al
sought to demonstrate why a more paradigmati-
Masarweh and Sarah Shick for research assistance.
cally exible and open approach to the empirical
world of the life course is required.
The impact of social forces upon individuals
References
are now recognized as extending from the organi-
zation of perception and individual conscious- Adair, L. (2007). Developing a different perspective on
ness that shapes agentic intentions to the disease. Review of developmental origins of health
regulation of gene expression in early life and and disease. In P. Gluckman & M. Hanson (Eds.),
Developmental origins of health and disease (pp. 600
throughout the life course. This means that the
601). Cambridge: Cambridge University Press.
social gets not only into ones head and under Alexander, K., Entwisle, D., & Olson, L. (2014). The long
the skin but onto the genome, as Slavich and shadow: Family background, disadvantaged urban
Cole have put it. These are matters that are not youth, and the transition to adulthood. New York:
Russell Sage Foundation.
easily apprehended by traditional theories of
Aronowitz, S. (2003). A mills revival? Logos, 2(3),
individual development, change and aging. 6793.
Opening the Social: Sociological Imagination in Life Course Studies 107

Aronowitz, S. (2012). Taking it big: C. Wright Mills and the Clausen, J. A. (1972). The life course of individuals. In
making of political intellectuals. New York: Columbia. M. W. Riley, M. E. Johnson, & A. Foner (Eds.), Aging
Baars, J. (1991). The challenge of critical gerontology: The and society: Vol. III. A sociology of age stratication
problem of social constitution. Journal of Aging Studies, (pp. 454574). New York: Russell Sage Foundation.
5(3), 219243. doi:10.1016/0890-4065(91)90008-G. Cole, S. W. (2009). Social regulation of human gene
Baltes, P. B., & Nesselroade, J. R. (1984). Paradigm lost expression. Current Directions in Psychological
and paradigm regained: Critique of Dannefers por- Science, 18(3), 132137.
trayal of life-span developmental psychology. Cole, S. W., Arevalo, J. M., Takahashi, R., Sloan, E. K.,
American Sociological Review, 49(6), 841847. Lutgendorf, S. L., Sood, A. K., Sheridan, J. F., &
doi:10.2307/2095533. Seeman, T. E. (2010). Computational identication of
Baltes, P. B., Mayer, K. U., Helmchen, H., & Steinhagen- gene-social environment interaction at the human IL6
Thiessen, E. (1994). Reply to and reections on com- locus. Proceedings of the National Academy of
mentaries. Ageing and Society, 14, 589604. Sciences of the USA, 107, 56815686.
Barker, D. J. P. (1995). Fetal origins of coronary heart dis- Cole, S. W., Hawkley, L. C., Arevalo, J. M., & Cacioppo,
ease. BMJ, 311(6998), 171174. J. T. (2011). Transcript origin analysis identies anti-
Belsky, J., Steinberg, L., & Draper, P. (1991). Childhood gen presenting cells as primary targets of socially
experience, interpersonal development, and reproductive regulated gene expression in leukocytes. Proceedings
strategy: An evolutionary theory of socialization. Child of the National Academy of Sciences of the USA, 108,
Development, 62(4), 647670. doi:10.2307/1131166. 30803085, epub ahead of print.
Bengtson, V., Giarrusso, R., Mabry, J. B., & Silverstein, M. Connidis, I. A., & McMullin, J. (2002a). Ambivalence,
(2002). Solidarity, conict, and ambivalence: family ties, and doing sociology. Journal of Marriage
Complementary or competing perspectives on intergen- and Family, 64, 594601.
erational relationships? Journal of Marriage and Family, Connidis, I. A., & McMullin, J. (2002b). Sociological
64(3), 568576. doi:10.1111/j.1741-3737.2002.00568.x. ambivalence and family ties: A critical perspective.
Berger, P. L., & Luckmann, T. (1967). The social con- Journal of Marriage and Family, 64, 558567.
struction of reality: A treatise in the sociology of Crockett, L. J. (2002). Agency in the life course:
knowledge. Garden City: Anchor Books. Concepts and processes. http://digitalcommons.unl.
Bierman, A. (2005). The effects of childhood maltreat- edu/psychfacpub/361
ment on adult religiosity and spirituality: Rejecting Crystal, S., & Shea, D. (1990). Cumulative advantage,
God the father because of abusive fathers? Journal for cumulative disadvantage, and inequality among
the Scientic Study of Religion, 44, 349359. elderly people. The Gerontologist, 30(4), 437443.
Bliss, C. (2012). Race decoded: The genomic ght for Crystal, S., & Shea, D. G. (2002). Introduction:
social justice. Stanford: Stanford University Press. Cumulative advantage, public policy, and inequality in
Blumer, H. (1969). Symbolic interactionism: Perspective later life. Annual Review of Gerontology and
and method. Englewood Cliffs: Prentice-Hall. Geriatrics, 22, 113.
Boardman, J. D., Blalock, C. L., & Pampel, F. C. (2010). Csizmadia, A., Brunsma, D. L., & Cooney, T. M. (2012).
Trends in the genetic inuences on smoking. Journal Racial identication and developmental outcomes
of Health and Social Behavior, 51(1), 108123. among blackwhite multiracial youth: A review from
Boardman, J. D., Menard, S., Roettger, M. E., Knight, a life course perspective. Advances in Life Course
K. E., Boutwell, B. B., & Smolen, A. (2014). Genes in Research, 17(1), 3444.
the dopaminergic system and delinquent behaviors Cunha, F., & Heckman, J. J. (2007). The technology of skill
across the life course: The role of social controls and formation. American Economic Review, 97(2), 3147.
risks. Criminal Justice and Behavior. Cunha, F., Heckman, J., & Schennach, S. (2010).
doi:10.1177/0093854813514227. Estimating the technology of cognitive and noncogni-
Bourdieu, P. (1977). Outline of a theory of practice. tive skill formation. Econometrica, 78, 883931.
Melbourne: Cambridge University Press. Currie, J., & Thomas, D. (1995). Does head start make a dif-
Bureau of Labor Statistics. (2014). http://data.bls.gov/ ference? American Economic Review, 85(3), 341364.
timeseries/LNS14000000. Retrieved 22 Dec 2014. Dannefer, D. (1984a). Adult development and social the-
Cain, L. (1964). Life course and social structure. In R. E. ory: A paradigmatic reappraisal. American
L. Faris (Ed.), Handbook of modern sociology Sociological Review, 49, 100116.
(pp. 272309). Chicago: Rand McNally. Dannefer, D. (1984b). The role of the social in life-span
Campbell, A., Converse, P. E., & Rodgers, W. L. (1976). developmental psychology, past and future: Rejoinder
The quality of American life: Perceptions, evaluations, to Baltes and Nesselroade. American Sociological
and satisfactions. Russell Sage Foundation. Review, 49, 847850.
Charney, E. (2012). Behavior genetics and postgenomics. Dannefer, D. (1987). Aging as intracohort differentiation:
Behavioral and Brain Sciences, 35(5), 331358. Accentuation, the Matthew effect, and the life course.
Chudacoff, H. P. (1989). How old are you? Age con- Sociological Forum, 2(2), 211236.
sciousness in American culture (Vol. x). Princeton: Dannefer, D. (1994). Commentary (invited): Special issue
Princeton University Press. on the Berlin Aging Study (BASE). Aging & Society, 14.
108 D. Dannefer et al.

Dannefer, D. (1999a). Freedom isnt free: Power, alien- Handbook of child psychology (Vol. 1, pp. 939991).
ation, and the consequences of action. In New York: Wiley.
J. Brandtstadter & R. Lerner (Eds.), Action & develop- Elder, G. H., & Johnson, M. K. (2003). The life course
ment: Origins and functions of intentional self devel- and aging: Challenges, lessons, and new directions. In
opment. New York: Springer. Invitation to the life course: Toward new understand-
Dannefer, D. (1999b). Neoteny, naturalization and other ings of later life (pp. 4981). Amityville, N.T:
constituents of human development. In C. Ryff & Baywook Pub. Co.
B. Marshall (Eds.), Self and society of aging processes Ellis, B. J., Boyce, W. T., Belsky, J., Bakermans-Kranenburg,
(pp. 6793). New York: Springer. M. J., & van Ijzendoorn, M. H. (2011). Differential sus-
Dannefer, D. (2000a). Paradox of opportunity education, ceptibility to the environment: An evolutionary-neurode-
work, and age integration in the United States and velopmental theory. Development and Psychopathology,
Germany. The Gerontologist, 40(3), 282286. 23(1), 728. doi:10.1017/S0954579410000611.
doi:10.1093/geront/40.3.282. Elman, C., & ORand, A. M. (2004). The race is to the
Dannefer, D. (2000b). Bringing risk back in: The regula- swift: Socioeconomic origins, adult education, and
tion of the self in the postmodern state. In K. W. Schaie wage attainment. American Journal of Sociology,
& J. Hendricks (Eds.), The evolution of the aging self: 110(1), 123160.
The social impact on the aging process (pp. 269280). Featherman, D. L., & Lerner, R. M. (1985). Ontogenesis
New York: Springer. and sociogenesis: Problematics for theory and research
Dannefer, D. (2003a). Cumulative advantage/sisadvan- about development and socialization across the lifes-
tage and the life course: Cross-fertilizing age and pan. American Sociological Review, 10, 659676.
social science theory. The Journals of Gerontology Ferraro, K. F. (2011). Health and aging: Early origins,
Series B: Psychological Sciences and Social Sciences, persistent inequalities? In R. A. Settersten & J. L.
58(6), S327S337. doi:10.1093/geronb/58.6.S327. Angel (Eds.), Handbook of sociology of aging
Dannefer, D. (2003b). Toward a global geography of the (pp. 465475). New York: Springer.
life course. In J. T. Mortimer & M. J. Shanahan (Eds.), Ferraro, K. F., & Kelley-Moore, J. A. (2003). Cumulative
Handbook of the life course (pp. 647659). New York: disadvantage and health: Long-term consequences of
Springer. obesity? American Sociological Review, 68, 707729.
Dannefer, D. (2011). The life course, aging, and the socio- Ferraro, K. F., & Shippee, T. P. (2009). Aging and cumula-
logical imagination: New prospects for theory. In tive inequality: How does inequality get under the
R. Binstock & L. George (Eds.), Handbook of aging skin? The Gerontologist, 49, 333343.
and social sciences (pp. 316). New York: Academic. Fingerman, K. L., & Hay, E. (2004). Intergenerational
Dannefer, D. (2012). Enriching the tapestry: Expanding ambivalence in the context of the larger social net-
the scope of life course concepts. The Journals of work. In K. Pillemer & K. Luescher (Eds.),
Gerontology Series B: Psychological Sciences and Intergenerational ambivalences: New perspectives on
Social Sciences, 67, 221225, gbr148. parentchild relations in later life (pp. 133152).
Dannefer, D. (2013). Age and sociological explanation: Oxford: Elsevier Science.
Expanding horizons in the study of aging and the life Fletcher, J. M., & Conley, D. (2013). The challenge of
course. Contemporary Sociology: A Journal of causal inference in gene-environment interaction
Reviews, 42(6), 793800. research: Leveraging research designs from the social
Dannefer, D., & Kelley-Moore, J. (2008). New twists in sciences. American Journal of Public Health, 103(Suppl
the paths of the life course. In V. L. Bengtson (Ed.), 1), S42S45. doi:10.2105/AJPH.2013.301290.
Handbook of aging theory (2nd ed., pp. 389409). Geronimus, A. T., Hicken, M., Keene, D., & Bound, J.
New York: Springer. (2006). Weathering and age patterns of allostatic load
Dannefer, D., & Kelley-Moore, J. A. (2009). Theorizing scores among blacks and whites in the United States.
the life course: New twists in the paths. In Handbook American Journal of Public Health, 96(5), 826833.
of theories of aging (pp. 389412). New York: Springer Giddens, A. (1977). New rules of sociological method.
Dannefer, D., & Perlmutter, M. (1990). Development as a London: Constitution of Society.
multidimensional process: Individual and social con- Gilbert, S. F., & Epel, D. (2009). Ecological developmen-
stituents. Human Development, 33(23), 108137. tal biology: Integrating epigenetics, medicine and evo-
Davidson, A. (2014, June 20). Its ofcial: The Boomerang lution. Sunderland: Sinauer Associates.
kinds wont leave. The New York Times. http://www. Gillespie, B. J., & van der Lippe, T. (2014).
nytimes.com/2014/06/22/magazine/its-official-the- Intergenerational cohesiveness and later geographic
boomerang-kids-wont-leave.html?_r=0 distance to parents in the Netherlands. Advances in
Doblhammer, G. (2004). The late life legacy of very early Life Course Research. doi:10.1016/j.alcr.2014.07.001.
life. Berlin: Springer. Gluckman, P., & Hanson, M. (2008). Mismatch: The life-
Duina, F. (2014). Life transitions in America. Malden: Policy. style diseases timebomb. Oxford: New York.
Elder, G. H., Jr. (1974). Children of the great depression. Guo, G., Tong, Y., & Cai, T. (2008). Gene by social-context
Chicago: University of Chicago Press. interactions for number of sexual partners among white
Elder, G. H., Jr. (1998). The life course and human devel- male youths: Genetics-informed sociology. American
opment. In W. Damon & R. M. Lerner (Eds.), Journal of Sociology, 114(Sppl), S36S66.
Opening the Social: Sociological Imagination in Life Course Studies 109

Guo, G., Elder, G. H., Cai, T., & Hamilton, N. (2009). later life well-being. Journal of Gerontology:
Geneenvironment interactions: Peers alcohol use Psychological Sciences, 69, 942952.
moderates genetic contribution to adolescent drinking Lane, H. (1976). The wild boy of Aveyron. Cambridge,
behavior. Social Science Research, 38(1), 213224. MA: Harvard University Press.
doi:10.1016/j.ssresearch.2008.04.002. Laub, J. H., & Sampson, R. J. (2003). Shared beginnings,
Haas, S. (2008). Trajectories of functional health: The divergent lives: Delinquent boys to age 70. Cambridge,
long armof childhood health and socioeconomic fac- MA: Harvard University Press.
tors. Social Science and Medicine, 66, 849861. Link, B. G., & Phelan, J. (1995). Social conditions as fun-
Hagestad, G. O., & Dannefer, D. (2001). Concepts and damental causes of disease. Journal of Health and
theories of aging: Beyond microcation in social sci- Social Behavior, Spec No, 8094.
ence approaches. In R. H. Binstock & L. K. George Manuck, S. B., & McCaffery, J. M. (2014). Gene-
(Eds.), Handbook of aging and the social sciences (5th environment interaction. Annual Review of Psychology,
ed., pp. 321). New York: Academic. 65, 4170. doi:10.1146/annurev-psych-010213-115100.
Halfon, N., & Hochstein, M. (2002). Life course health Marshall, V. W. (2005). Agency, events, and structure at the
development: An integrated framework for developing end of the life course. Advances in Life Course Research,
health, policy, and research. The Milbank Quarterly, 10, 5791. doi:10.1016/S1040-2608(05)10002-1.
80(3), 433479. doi:10.1111/1468-0009.00019. Marshall, V. W., & Clarke, P. J. (2010). Agency and social
Heckman, J. J. (2013). Lifelines for poor children. structure in aging and life-course research. In
New York Times Opinionator. September 14. http:// D. Dannefer & C. Phillipson (Eds.), The SAGE hand-
opinionator.blogs.nytimes.com/2013/09/14/lifelines- book of social gerontology. London: SAGE.
for-poor-children/?_r=0 Mitchell, C., McLanahan, S., Brooks-Gunn, J., Garnkel, I.,
Hertzman, C., Power, C., Matthews, S., & Manor, O. Hobcraft, J., & Notterman, D. (2013). Genetic differen-
(2001). Using an interactive framework of society and tial sensitivity to social environments: Implications for
lifecourse to explain self-rated health in early adult- research. American Journal of Public Health, 103(Suppl
hood. Social Science and Medicine, 53, 15751585. 1), S102S110. doi:10.2105/AJPH.2013.301382.
Hoffmann, R. (2008). Socioeconomic differences in old Montagu, A. (1989). Growing young. Greenwood
age mortality. Dordrecht: Springer. Publishing Group.
Holliday, R. (2006). Epigenetics: A historical overview. Moftt, T. E., & Caspi, A. (2001). Childhood predictors
Epigenetics, 1(2), 7680. doi:10.4161/epi.1.2.2762. differentiate life-course persistent and adolescence-
Holstein, J. A., & Gubrium, J. F. (2000). Constructing the limited antisocial pathways among males and females.
life course. Dix Hills: General Hall. Development and Psychopathology, 13(2), 355375.
Holstein, J. A., & Gubrium, J. F. (2000). The self we live Montez, J. K., & Hayward, M. D. (2011). Early life condi-
by: Narrative identity in a postmodern world. tions and later life mortality. In R. G. Rogers & E. M.
New York: Oxford University Press. Crimmins (Eds.), International handbook of adult
Katz, S. (1996). Disciplining old age: The formation of mortality (International handbooks of population, Vol.
gerontological knowledge. Charlottesville: University 2, pp. 187206). New York: Springer.
of Virginia Press. Morss, J. (1995). Growing critical: Alternatives to devel-
Keister, L. A. (2004). Race, family structure, and wealth: opmental psychology. NY: Routledge.
The effect of childhood family on adult asset owner- Newman, K. (2006). A different shade of grey: Midlife
ship. Sociological Perspectives, 47, 161187. and beyond in the inner city. New York: New Press.
Kohli, M. (2007). The institutionalization of the life ORand, A. M. (2002). Cumulative advantage theory in
course: Looking back to look ahead. Research in life course research. Annual Review of Gerontology
Human Development, 4(34), 253271. and Geriatrics, 22, 1430.
doi:10.1080/15427600701663122. Perry, B., & Svalavitz, M. (2006). The boy who was raised
Kohli, M., & Meyer, J. W. (1986). Social structure and as a dog: And other stories from a child psychiatrists
social construction of life stages. Human Development, notebook: What traumatized children can teach us
29, 145161. about loss, love and healing. New York: Basic.
Kuh, D., & Ben-Shlomo, Y. (Eds.). (2004). A life course Phillips, D. L. (1978). Whats one to believe? Contemporary
approach to chronic disease epidemiology (Vol. 2). Sociology, 7, 392395. doi:10.2307/2064314.
Oxford: Oxford University Press. Plomin, R. E., DeFries, J. C., Craig, I. C., & McGufn,
Kuh, D., Ben-Shlomo, Y., Lynch, J., Hallqvist, J., & P. E. (2002). Behavioral genetics in the postgenomic
Power, C. (2003). Life course epidemiology. Journal era. Washington, DC: American Psychological
of Epidemiology and Community Health, 57(10), 778 Association.
783. doi:10.1136/jech.57.10.778. Powledge, T. M. (2011). Behavioral epigenetics: How
Landecker, H., & Panofsky, A. (2013). From social struc- nurture shapes nature. BioScience, 61(8), 588592.
ture to gene regulation, and back: A critical introduc- Priestley, M. (2001). Disability and the life course: Global
tion to environmental epigenetics for sociology. perspectives. Cambridge: Cambridge University
Annual Review of Sociology, 39, 333357. Press.
Landes, S. D., Ardelt, M., Vaillant, G. E., & Waldinger, R. J. Reich, R. B. (2000). The Future of Success. New York:
(2014). Childhood adversity, mid-life generativity, and Alred A. Knopf.
110 D. Dannefer et al.

Rennie, J. (2003). Celebrating the genetic jubilee: A con- Shanahan, M. J., & Boardman, J. D. (2009). Genetics and
versation with James D. Watson. Scientic American, behavior in the life course: A promising frontier. In
288(4), 6669. The craft of life course research (pp. 215235).
Riley, M. W., Johnson, M., & Foner, A. (Eds.). (1972). Shanahan, M. J., & Hofer, S. M. (2005). Social context in
Aging and society, volume 3: A sociology of age strati- geneenvironment interactions: Retrospect and pros-
cation. New York: Russell Sage Foundation. pect. The Journals of Gerontology Series B:
Riley, M. W., Kahn, R. L., & Foner, A. (1994). Age and struc- Psychological Sciences and Social Sciences,
tural lag: Societys failure to provide meaningful opportu- 60(Special Issue 1), 6576. Cambridge University
nities in work, family, and leisure. New York: Wiley. Press
Rose, H., & Rose, S. (2014). Genes, cells, and brains: The Shanahan, M. J., Hofer, S. M., & Shanahan, L. (2003).
promethean promises of the new biology (Reprint edi- Biological models of behavior and the life course. In
tion.). London/New York: Verso. J. T. Mortimer & M. J. Shanahan (Eds.), Handbook of
Rosenbaum, J. (1978). The structure of opportunity in the life course (pp. 597622). New York: Springer.
schools. Social Forces, 57, 236256. Shaputis, K. (2011). The crowded nest syndrome.
Schafer, M. H., Ferraro, K. F., & Mustillo, S. A. (2011). Olympia: Clutter Fairy Publishing.
Children of misfortune: Early adversity and cumula- Shierholz, H., Davis, A., & Kimball, W. (2014). The class
tive inequality in perceived life trajectories. American of 2014: The weak economy is idling too many young
Journal of Sociology, 116, 10531091. graduates. Washington, DC: Economic Policy Institute.
Schaie, K. W. (1965). A general model for the study of Skogen, J. C., & verland, S. (2012). The fetal origins of
developmental problems. Psychological Bulletin, 64, adult disease: A narrative review of the epidemiologi-
92107. cal literature. JRSM Short Reports, 3(8), 59.
Schaie, K. W., & Hendricks, J. (Eds.). (2000). Evolution Slavich, G. M., & Cole, S. W. (2013). The emerging eld
of the aging self: Societal impacts. New York: Springer. of human social genomics. Clinical Psychological
Seabrook, J. A., & Avison, W. R. (2010). Genotypeenvi- Science: A Journal of the Association for Psychological
ronment interaction and sociology: Contributions and Science, 1(3), 331348.
complexities. Social Science & Medicine, 70(9), Smith, R. C. (2005). Mexican New York: The transna-
12771284. tional lives of Mexican immigrants. Berkeley:
Settersten, R. A., Jr. (1999). Lives in time and place: The University of California Press.
problems and promises of developmental science. Standing, G. (2011). The precariat: The new dangerous
Amityville: Baywood Publishing Company. class. London: Bloomsbury Publishing.
Settersten, R., & Ray, B. E. (2010). Not quite adults: Why Thoits, P. A. (2006). Personal agency in the stress process.
20-somethings are choosing a slower path to adulthood, Journal of Health and Social Behavior, 47(4), 309323.
and why its good for everyone. New York: Bantam. Wexler, P. (1977). Comment on Ralph Turners The Real
Sewell, W. H., Jr. (1992). A theory of structure: Duality, Self: from Institution to Impulse. American Journal
agency, and transformation. American Journal of of Sociology, 83(1), 178186.
Sociology, 98, 129. Zimmer-Gembeck, M. J., & Skinner, E. A. (2011). The
Shanahan, M. (2013). Social genomics and the life course. development of coping across childhood and adoles-
In L. J. Waite & T. J. Plewes (Eds.), New directions in cence: An integrative review and critique of research.
the sociology of aging. Washington, DC: National International Journal of Behavioral Development, 35,
Academies Press. 117.
The Changing Social Construction
of Age and the Life Course:
Precarious Identity and Enactment
of Early and Encore Stages
of Adulthood

Jeylan T. Mortimer and Phyllis Moen

Though existing age periods and life stages seem the transition to adulthood; the Section on Aging
natural, they are socially constructed, that is, and the Life Course seeks unifying principles
shaped and dened by cultural beliefs, structural applicable to all stages of life while, historically,
arrangements, policies and practices that have many members of this section have focused on the
been institutionalized in particular societies at latter phases of adulthood. But there is recogni-
particular times and places (Buchmann 1989). tion that a wider purview can be useful. Both sec-
This social organization of the life course denes tions changed their names in the late 1990s in
the pool of possibilities and constraints that indi- attempts to signal a more expansive scope. The
viduals experience as they move through their Section on Aging became the Section on Aging
lives. Massive cultural and social changes (such and the Life Course in 1997. The Section on
as the industrial revolution) raise new challenges, Childhood became the Section on Children and
often resulting in a reworking of the character of Youth in 1998. Almost two decades later, these
the life course. In this chapter we argue that the two sections are again re-assessing boundaries
conuence of todays massive and rapid eco- and considering whether insights from studies of
nomic, technological, demographic and social children and youth might inform studies of mid-
transformations is reconstructing the twenty-rst life adults and the elderly, and vice versa. Scholars
century life course, including individuals (and afliated with each section are also considering
families) strategies and the quality of their expe- whether the recognition of new stages might be
riences as they move through time. useful, as education and economic realities delay
Nonetheless, scholars as well as social observ- the transition to adulthood (for some), and both
ers tend to take existing categories as given. For extended healthy life expectancy and the aging of
example, segmentation in the elds of sociology the large Boomer cohort (born from1946 to 1964)
and developmental psychology encourages them suggest a stage beyond conventional adulthood
to focus exclusively on widely recognized, long but before conventional old age. (Such reassess-
established age-graded life stages. Thus, in the ment inspired a session, co-sponsored by the two
American Sociological Association, the Section sections, at the Annual Meeting of the American
on Children and Youth examines childhood and Sociological Association in 2012.)
Our goal in this chapter is to underscore the
socially constructed character of age and life
J.T. Mortimer (*) P. Moen
stages. That is, the manner in which we concep-
Life Course Center and Department of Sociology,
University of Minnesota, Minneapolis, MN, USA tually organize and demarcate the life course is
e-mail: morti002@umn.edu not based on universal biological progressions.

Springer International Publishing Switzerland 2016 111


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_5
112 J.T. Mortimer and P. Moen

Rather, these age-graded stages emanate from omy, and declines in social protections changing
past historical conditions as modied by chang- expectations and options associated with age
ing institutional, demographic, and cultural reali- and the life course? And how do the changing
ties, altering the ways individuals perceive and features of age inuence inequalities across
experience their own life biographies. As a result, social groups?
socially recognized life stages differ across his-
torical periods, as well as across cohorts. It is These are big questions that implicate large
important to note, however, that in times of trans- bodies of literature, including investigations of
formation, those with material, social, and psy- social disparities across class, gender, race and
chological resources are usually the ones rst ethnicity, as well as across age divides.
exposed and best positioned to take advantage of The chapter is organized as follows. We rst
newly dened life course categories, accentuat- point to trends in the social construction of age,
ing within cohort inequalities based on the accu- noting increasing differentiation historically,
mulation of advantage and disadvantage in extending across the life span. Second, we sug-
distinct subgroups of the population. gest that such differentiation has been accentu-
We wish to raise several interrelated questions ated by the pace of social change, disruptive to
(not all of which we can address in this chapter) existing, seemingly natural norms and expecta-
in hopes of stimulating further discussion about tions about age and life stage. The unraveling of
the changing nature of the life course in the early age-graded expectations and protections invari-
twenty-rst century: ably generates ambiguities and uncertainties in
age-linked identities and behavior as old norms
To what extent do new age-related life course no longer apply and new ones have yet to evolve.
phases reect changes in statuses, positions, Third, we discuss the development of two new
and roles, as well as the quality of experi- life stages early adulthood and encore adult-
ences, in terms of commitments, identities, hood in part a consequence of the mismatch
and self-denitions, as traditional transition between outdated, institutionalized age-graded
markers (and their timing) become blurred? structures and key economic and demographic
How do we dene and study a life course in changes, as well as modications in the goals,
ux? Does the conceptualization and some- values and preferences of people in these two age
times reication of existing age categories groups. Shifting (often lagging) structural con-
contribute to research and theory development texts together with shifting individual orienta-
about the life course, or do they hamper the tions produce striking cross-cohort variation in
recognition of its uid and socially con- the experiences of individuals of the same ages in
structed nature? different time periods. Finally, we point to
What criteria can or should be used to delin- increasing diversity within age categories and
eate or assess the utility of new life stages in cohorts, accompanied by cumulative inequalities
the face of demographic, economic, and polit- across phases of life.
ical transformations?
What are the implications of new life stages
or the in-between-ness that they convey for 1 Increasing Differentiation
scientic study? For example, investigators in the Social Construction
might examine the extent to which periods of of Age
the life course are affected by the continuity or
discontinuity built into existing (and often The social construction of age (Buchmann 1989;
outdated) educational, organizational, and Elder 1975a; Mayer 2004, 2009; Settersten 2003;
governmental policies and practices. Settersten and Mayer 1997) is a dynamic,
How are large-scale social forces demogra- continuously evolving process in contemporary
phy, technology, a global information econ- societies, characterized by increasing age
The Changing Social Construction of Age and the Life Course: Precarious Identity 113

differentiation as new age categories come to be shops, and later in the emerging industrial
socially recognized. The social construction and economy. However, Osterman (1979) docu-
recognition of each newly identied age group mented how technological changes at the end of
parallel major macro-structural changes in social the nineteenth century lessened the demand for
institutions (e.g., in education, work, health care, unskilled youth in the textile, glass, and other
public policy, science/technology, etc.), which manufacturing industries, and in retail trade, even
prompt the development of a new social organi- as large numbers of immigrants increased the
zation of age-graded roles, expectations and unskilled labor supply. With the introduction of
activities encompassing newly dened segments mass production technology, both skilled crafts-
of the population. These changes prompt shifts in men and their unskilled child helpers were no
the culture, affecting myriad ideas surrounding longer needed. As a result of these trends,
what roles persons of different ages should take together with the contracting agricultural sector,
on, what attributes of character they should children were driven out of paid employment,
embody, and what activities they should engage and employers lessened their opposition to laws
in. Changing norms and values surrounding age that restricted child labor and extended compul-
mark the chronological ages that are considered sory schooling. The social construction of the
right and on-time for particular roles and childs role was thus fundamentally altered
activities (e.g., marriage, retirement); they spec- from productive worker to student. Instead of
ify age deadlines before which transitions assets to the family economy as farm hands and
should occur (e.g., parenthood); and they enable earners, children became priceless (Zelizer
identication of individuals who are clearly on 1985), valued in their own right. As education
or off time (Neugarten and Hagestad 1976; expanded historically, the central task of the
Hagestad and Neugarten 1985; Neugarten et al. childs life was increasingly seen as attending
1965; for a review, see Settersten 2003). When school (Mortimer 2003).
newly constructed age categories emerge and are In succeeding years, the age differentiation of
increasingly recognized, they become the tem- children has continued. Whereas traditionally
plate for and/or reect social policies that, in young children were not formally differentiated
turn, serve to institutionalize the revised age- until they entered school, advances in develop-
graded life course (Kohli 2007; Kohli et al. 1991; mental child psychology, as well as the prolifera-
Mayer 1986, 2004, 2009; Moen and Spencer tion of day care centers and other preschool
2006; Mortimer and Shanahan 2003; ORand and facilities, have fostered the recognition of multi-
Henretta 1999). ple early stages. Childhood is now separated into
In his classic albeit controversial study, early, middle, and late stages. Multiplication
Centuries of Childhood, Aries (1962) contended of age categories has affected even the very
that only two distinct age categories were socially youngest children. These are not unnoticed by
recognized through the middle ages: infants and children themselves. Jeylans granddaughter
adults. As soon as infants were able, they began Eileen, on turning 3, announced proudly that she
to assume adult-like functions, taking on various used to be a toddler (assigned to the toddler
work tasks in imitation of their elders. Though room in her day care center) but now she was a
differences in physiological capacity, skill, and preschooler. She went on to inform her grand-
maturity were surely noticed, according to Aries mother that babies come before toddlers (a
the age category of child only began to be iden- budding life course scholar!)
tied as a separate stage of life with the advent of By the turn of the twentieth century, when
schooling in the seventeenth century, affecting schools had become more age-graded and
upper-class groups initially, and boys earlier than increasing numbers of students were attending
girls. secondary schools, G. Stanley Hall (1904) her-
Historically in the United States, people of all alded the emergence of a new phase of life, in
ages were expected to work in agriculture, small between childhood and adulthood, that of
114 J.T. Mortimer and P. Moen

adolescence. One hundred years later, with the risk rather than exploration and opportunity
expansion of higher education and extension of (Cote 2014).
the time needed to acquire adult role markers At the other end of the age spectrum the cate-
(e.g., leaving home, becoming nancially inde- gory of old, up through much of the twentieth
pendent, completing education, obtaining full- century, was dened largely by retirement for
time work, marrying, and parenting), scholars men and employed women, and for homemaking
have introduced still another transitional phase wives, by the retirement of their husbands, and
that is increasingly recognized by the lay public. sometimes by grandparenthood (Kaufman and
Sociologists refer to this new phase as the tran- Elder 2003). As a result of enactment of public
sition to adulthood, early adulthood, or social security pensions (in response to the
youth, while some psychologists have adopted Great Depression in the U.S. and earlier in
the term emerging adulthood. Germany), by the mid-1960s age 65 had become
Like the age categories of childhood and the age around which retirement clustered in the
adolescence that were recognized long ago, U.S. and Europe. Retirement thus became an
many features of the lengthening transition to effective reference point for a variety of purposes,
adulthood are more characteristic of advantaged including the passage into old age. Retirement in
than working-class youth. Emerging adulthood the 1960s and 1970s became a scripted and dis-
(Arnett 2000), as a time of exploration, growth, tinct stage in the life course. However, paralleling
opportunity, freedom from rm commitments, the differentiation of younger age categories,
and extension of the valuable adolescent mora- increasing longevity and healthy life expectancy
torium, characterizes the experience of many have prompted categorical division of the previ-
middle and upper class young people. For them, ously old into ad hoc young-old, old-old
it is a time of exploration through higher educa- and oldest-old groupings (Youmans 1977;
tion, internships (often unpaid), tentative intimate Garfein and Herzog 1995). So too, came recogni-
relationships that prepare youth for more lasting tion of the frail elderly in accord with the
adult commitments, nancial dependence on par- reduction in their physiological capacities, health
ents, and extended co-residence with the family status, and social activities that often come with
of origin (Arnett 2000). advancing age.
However, working class and poor youths However, the retirement life stage has
extension of the post-adolescent period may become more problematic, as a result of recent
not be characterized by any of these experiences changes in the employer-employee contract, a
(Furstenberg 2008), though they may remain in global information economy, and economic
or return to the parental home for nancial rea- downturns, together with aging workforces and
sons. Contemporary young people at lower socio- populations (in Europe, Japan, and North
economic levels often experience a more rapid America). Traditionally dened-benet private
transition from school to work and acquire adult pension programs (offering income for life) are
responsibilities that come with early family for- disappearing. Older workers both advance and
mation. For the least educated, a dearth of work delay their retirement timing, leaving (or being
opportunities may drive them out of the labor laid off from) their career jobs at different ages,
force entirely. The most disadvantaged young and sometimes taking on new jobs (Han and
people are the most at risk; some are attracted to Moen 1999; Moen and Flood 2013; Rix 2011;
the risky fast life and criminal careers. Warner et al. 2010; Williamson 2011).
Incarceration has become a highly prevalent These demographic and labor market trends
experience in the transition to adulthood among have led some scholars to argue that a new phase
minority, especially African-American, male of the later life course has developed some-
youth (Wakeeld and Apel this volume). Thus, where between the family- and career-building
early adulthood has become, for many young years and the frailer years of late adulthood
people, a period of uncertainty, difculty, and (Flood and Moen 2015; James and Wink 2007;
The Changing Social Construction of Age and the Life Course: Precarious Identity 115

Laslett 1989; Moen 2003, 2011b; Moen and the Boomer population, producing some critiques
Flood 2013; Moen and Spencer 2006; Sadler of the focus on engagement and productivity as
2006; Silva 2008; Weiss and Bass 2002). This part of this life stage (Carr and Manning 2010;
encore adulthood, third age, or midcourse Estes 2004; Holstein and Minkler 2003; Liang
stage both emerges from and is fostering macro- and Luo 2012).
level transformations in society. This new stage The boundaries around early adulthood and
of life is reected in micro-level changes in the encore adulthood are blurred and somewhat
biographies of individuals and families (Moen contested. Millennials (coming to their 20s after
2013). Encore adulthood is being constructed in the turn of the twenty-rst century) may consider
part by scholars as a way of capturing the changes themselves adults far earlier than their parents (or
they are observing in the life course, but also by scholars) might acknowledge, despite their
individuals and families in the large cohort of delays, relative to the timetables of previous
Boomers now moving through their 50s, 60s, and cohorts, in acquiring the traditional markers of
their 70s. Older Boomers may have relinquished, adulthood buying homes, marrying, or having
for the most part, those roles that provided central children. Similarly, Boomers will be reluctant to
focus for their adult lives active parenting of acknowledge moving from encore adulthood to
young children and working full-time in career- old age until they join the ranks of the frail
related jobs but they do not consider themselves elderly, severely limited in their abilities and
any kind of old. The maintenance of relatively activities.
good health, with increasingly effective control There may be yet another emerging stage in
of chronic ailments, has enabled Boomers to pur- later adulthood. A nal increasingly recognized
sue a wide range of activities and remain socially age differentiation, thus far of greater interest to
connected well beyond the traditional ages of health professionals than sociologists, occurs at
retirement. the very end of life, as individuals move from the
Paralleling recognition of a new phase of early stage of frail elderly to hospice care.
adulthood, social observers describe men and
women in encore adulthood (roughly age
5575) as having freedom to adopt new roles, 2 Ambiguity and Uncertainty
taking up new jobs or starting their own busi- in Age Constructions,
nesses, volunteering or engaging in other service Interactions, and Identities
to their communities, developing new avoca-
tional pursuits like music, art, and writing (activi- In addition to the blurred boundaries of new life
ties that might have been hobbies earlier, when stages, with social change and new images and
there was scarce time to develop them), or realities surrounding children, youth, and older
becoming highly involved in their grandchil- adults come new cultural contradictions and
drens lives (Alboher 2013; Carstensen 2011; ambivalences (Mortimer and Moen 2012).
Freedman 2007). Growing numbers in this life
stage are choosing to prolong their work careers
or take on new jobs in light of new economic 2.1 Children, Adolescents,
uncertainties associated with retirement. Such and Early Adults
work provides a continuing income stream as
well as a sense of meaning or at least routine to Contemporary children are considered innocent
their lives. But, as with past life stage construc- and vulnerable, and at the same time precocious
tions like adolescence and contemporary and growing up too fast (Levin 2013). As Graff
emerging adulthood, the opportunities, health, (1995) points out, cultural images of young peo-
and other advantages associated with this new ple, though specic to their age, tend to stress
life phase are more likely to be experienced by problems and crises. Thus, youth are variously
relatively well-off, college-educated segments of described as rebellious, deant, delinquent,
116 J.T. Mortimer and P. Moen

drug-addicted, promiscuous, and a lost genera- and understood. For example, when the life stage
tion, while sample surveys show the vast major- adolescence was introduced, it was not clearly
ity exhibit conforming attitudes and behavior. In dened (Elder 1975b). But in time, adolescence
the 1960s, when the Boomers were children and became institutionalized in high school education
youth, and a consumer economy aimed at molding teen-
A romantic (and increasingly nostalgic) mystique agers tastes and spending habits. Still, it remains
that weaves together the Beatles, rebels with and as an in between stage, involving many uncer-
without a cause, cultures counter and alternative, tainties and unclear expectations, as young peo-
free love, political struggles for rights and for ple assume increasing autonomy from parents
peace only rarely culminated in what Kenneth
Keniston calls postmodern or postconventional and alternately take on behaviors characteristic of
youth. Most young people are normal,a cate- childhood and adulthood (Fine 2001). How
gory that embraces a diverse range of individuals much freedom should parents give their teenage
who are neither visionaries nor victims. (Graff children? If they ease up on their monitoring, will
1995: pp. 339340)
their children come to harm? How much, if at all,
Similarly, contemporary media portrayals and should teenagers be allowed to assume adult
commentaries depict the new stage of early or behaviors, like holding part-time jobs (Mortimer
emerging adulthood with negative terms such 2003)?
as adultolescents, boomerang children, and Youth in the early or transition to adulthood
failure to launch (Swartz 2009). The reality is phase are likewise caught between adolescence
far more complicated, with considerable diver- and adulthood. They are expected to engage in
gence in youth orientations, behaviors, and path- post-secondary education and often face a
ways to adulthood by social class, race/ethnicity, lengthy oundering period as they seek work
and gender. This diversity is sometimes reected that matches their abilities and skills. As a result,
in differential willingness on the part of adults to nancial and residential dependence on parents
support policies designed to help our own and often continues longer than either party would
other peoples children (Graff 1995). like (Swartz 2009; Furstenberg et al. 2004). The
Each time a new age group emerges, as fact that parents can now provide health insur-
scholars propose novel categorical designations ance until their children are age 26 (as a result
that reect new social circumstances, these new of the Affordable Care Act) reects recognition
labels lter outwards to inform lay understand- of this longer-term dependency. Youth in transi-
ings. Recognition that old labels do not t, and tion to adulthood only gradually relinquish
that new ideas about age are warranted, is often behaviors typically associated with adolescence
prompted by multiple ambiguities arising from risky and irresponsible behaviors, like party-
the mismatch between existing age-graded poli- ing, binge drinking, and driving too fast as they
cies, practices and expectations, on the one hand, acquire adult role markers (Massoglia and Uggen
and new experiences, problems, opportunities 2010).
and realities that have fostered the new life stages, Along with unclear expectations surrounding
on the other. appropriate behaviors for persons in newly
These ambiguities foster uncertainty and emergent age categories comes uncertainty
ambivalence on the part of those who nd them- regarding the subjective self-denition of age.
selves in the new life phases, as well as those in For example, given the problematic character of
other more established age groups who interact age identity for persons in transition to adult-
with them. For people in each new age cate- hood, researchers have examined the circum-
gory nd themselves between individuals in adja- stances that lead young people to take on the
cent stages that have been recognized for much identity of an adult (Johnson and Mollborn 2009;
longer periods of time, institutionalized in exist- Johnson et al. 2007; Benson and Furstenberg
ing cultural norms and organizational arrange- 2006; Shanahan et al. 2005). Whereas psycholo-
ments, and therefore more clearly demarcated gists emphasize youths own embrace of
The Changing Social Construction of Age and the Life Course: Precarious Identity 117

independence, responsibility, nurturance and many years of relatively good health (including
other indications of maturity as hallmarks and the successful management of chronic health
self-dening criteria of adulthood (Arnett 1997, conditions), in the absence of traditional social
2000), sociologists nd that demographic mark- obligations associated with full-time work and
ers, particularly those associated with family child-rearing. This hopeful prospect is paired,
roles, are predictive of actually considering one- however, with the risks of layoffs, economic
self an adult (Shanahan et al. 2005; Eliason et al. hardship, and the onset of new health problems in
2015). Clearly, both are intertwined, as the a tumultuous economic climate replete with
assumption of role markers of adulthood (espe- unraveling safety nets (Carr and Muschert 2009).
cially, becoming a parent) encourage, even But the ambiguities and uncertainties go
require, classic emblems of maturity. beyond economic security to the very lifestyles
of those in this new encore adult stage. As Moen
(2003) points out, people dont plan much nan-
2.2 Encore and Older Adults cially, but are much more likely to plan for their
nancial needs in retirement than to anticipate
Turning to later phases of life, encore adulthood how they will spend their time. The new bonus
is evolving as an in between stage between years are coming not at the end of life but between
conventional adulthood (replete with full-time career and family building on the one hand, and
jobs and raising children) and the frailties old age (dened as debilitating conditions
acquired as individuals approach old age. In the severely limiting social participation), on the
1980s, the British historian Peter Laslett (1989) other.
rst identied this as a third age (after the rst Marc Freedman (2007, 2011) and others (see
age of childhood and the second age of adult- Alboher 2013; Farrell 2014; Moen forthcoming)
hood, but before the fourth age of frailty). It was are writing about the opportunities for meaning-
marked by the transition to retirement and ended ful engagement in the form of encore careers in
with either frailty or death (see also Kertzer and which encore adults can work for the greater
Laslett 1995). good. To be sure, some Boomers follow the tradi-
But this formulation of a stage between fam- tional retirement transition to full-time leisure,
ily/career building and old age is only beginning but many say they want to work in retirement
to be recognized in the U.S., as the taken-for- (AARP 2014), typically in not so big jobs
granted naturalness attributed to the culture (Moen 2007) that provide them considerable
and structure of the lock-step transition to retire- exibility and less stress than conventional full-
ment has evaporated. Seniority is no longer time work.
accompanied by job security for increasing num- As is the case for early adulthood, language
bers of men and women (Moen and Peterson and ideas surrounding encore adulthood have yet
2009; Sweet and Meiksins 2013). The result is to be institutionalized, producing ambiguities in
risk and ambiguity in tandem with the unraveling the culture and in the policies, practices, identi-
of traditional career and retirement patterns. ties, and attributes associated with the latter part
Boomers now face great uncertainty regarding of adulthood. Despite similarities in the increas-
employment security, retirement timing, and eco- ing recognition of an in between status, orien-
nomic stability in the later adult years (Estes tations to early adulthood and encore
2004; Kotter 1995; Moen 2011b, forthcoming; adulthood differ. Young people approach their
ORand and Bostic this volume; Warren et al. new stage with ambivalence (Hartmann and
2012; Williamson 2011; Wong and Hardy 2009). Swartz 2007), as adult responsibilities are
Lacking cultural traditions and role models, indi- sometimes considered onerous and lessening of
viduals (and couples) may not know how to plan opportunities to engage in youthful and enjoy-
and prepare for the exigencies or the promise of able pursuits. However, youth generally welcome
this new life stage. They may look forward to the roles and identities linked to adulthood, given
118 J.T. Mortimer and P. Moen

the considerable resource and status advantages positive identities to counteract the negative
such age progression entails. In contrast, biologi- self-denitions and stereotypes widely associ-
cal aging and the structure of existing social insti- ated with this age group. People in this age range
tutions mean that age identities at the other end of do not consider themselves old, aging,
life often signify losses of career jobs and even senior citizens, elderly, or other labels, as
paid work, marriage partners, health, and inde- these are more tting to the frailties associated
pendence. As Rosow (1974) pointed out some with the next stage of inrmity (Moen forthcom-
time ago, retirement is a roleless role, even ing). The convention in data collection and analy-
moreso given the unraveling of conventional sis to categorize people as 65 and over or 55
gold watch exits and a golden years framing and over diminishes the possibility of capturing
of retirement security (Estes et al. 2007; Liang the realities of two life stages: encore adulthood
and Luo 2012). and a later old age.
Riley (1987, 1994) underscored that while
older people are healthier and living longer, soci-
ety has not yet caught up by delineating new 3 Linked Lives
social denitions and productive social roles for and the Gendered Life
recent cohorts of retirees. She called this disjunc- Course
tion structural lag. We are witnessing the dein-
stitutionalization of existing regimes that dened A focus on the social construction of particular
age boundaries in the past, such as the demise of age groups emphasizes individuals as the focal
private pensions, job security, and legislation that unit, obscuring the linked lives dimension of
contributed to a clearly dened age of retirement the life course (Elder et al. 2003). People in dif-
(Brckner and Mayer 2005; Moen 2013). ferent life stages constitute social convoys
As a result, age-linked identities are uncertain (Kahn and Antonnuci 1980; Moen and Hernandez
and often problematic among encore adults, 2009), leading lives that are intertwined with one
albeit for different reasons. We have already another. For example, their early adult childrens
commented on the multiple possibilities for new own stressful situations create distress for their
commitments and self-denitions for those enter- parents in the encore adult years as well (Milkie
ing what is emerging as an encore to adult- et al. 2008). The increasing numbers, longevity,
hood. Nonetheless, the period post 65 remains and visibility of the frail elderly have prompted
devalued, often associated with disability and a situation in which some of the same ambiva-
disease, and those who are no longer employed or lences (Willson et al. 2003) and fears that charac-
raising children are often stigmatized with labels terize parents vis--vis adolescent and early
and connotations many see as negative, such as adult children now typify the orientations of
old people, elderly, or aged (Levin 2013). adult and encore adult children toward their
This stereotyping is fostering efforts to substitute aging parents.
new images, identities, and denitions of what it Understandably, older adults are reluctant to
means to be beyond the peak responsibilities of move from identities and roles associated with
middle adulthood. Older adults tend to think of the second acts implied by encore adulthood to
themselves as less old than they are chrono- being elderly or senior citizens and on to the
logically, and, as they grow older, their subjective next oldest old or frail elderly status. Adult
ages (the ages they feel) become increasingly children (boomers) worry about their elderly par-
younger than their chronological ages (Kaufman ents increasing social isolation and risk, even as
and Elder 2002). As the large Boomer cohort their parents insist on living independently in
(born 19461964) moves through their 50s, 60s, their homes, while their mental faculties and
and 70s, new age categories, such as encore physical capacities erode. How much autonomy
adulthood, partially retired or third age are should their aging parents retain? At what point
likely to provide alternative and potentially quite should the adult children take over nancial
The Changing Social Construction of Age and the Life Course: Precarious Identity 119

responsibilities, becoming co-signers on their what lies ahead, to plan for upcoming role
parents bank accounts and trustees of their transitions and new circumstances and, when
estates? Will their adherence to parents wishes age-linked changes do occur, to satisfactorily
to remain independent increase the risks of falls, adjust to the new realities. Most important, rapid
res, or other calamities? Such problems arose social, technological, and economic changes
less often in previous eras when people typically have rendered existing age-graded blueprints
did not live long enough to join the ranks of the (such as marrying or retiring at a certain age)
frail elderly. A recent spate of magazine arti- obsolete, limiting the capacity of people at all
cles and self-help books have arisen in response, stages of the life course to set realistic goals, to
documenting the trials of such middle-aged and envision future challenges, and to actualize their
older children and offering advice (Alderman preferences. Whether in early or later adulthood,
2010; Considine 2012; Delehanty et al. 2008). In such preferences are often based on expectations
fact, a Google search on caring for elderly par- congruent with prior social conditions or needs.
ents yielded over one million entries. Issues of Outmoded age-related identities and templates,
quality vs. length of life arise as adult children, characteristic of earlier times, may preclude
along with medical personnel and other advisors, effective anticipatory socialization, so useful in
wrestle with difcult decisions regarding end-of- subsequent adjustment. At the same time, new,
life treatments and care. more realistic schema may not have yet taken
A linked lives framing also reminds us of hold. Individuals, as they move from one age sta-
the gendered nature of the life course (Drobni tus to another are on shifting sands what is age
and Blossfeld 2004; Moen 2011a; Moen and normative behavior and identity may be quite
Spencer 2006; Williams and Umberson 2004). unclear.
For example, caregiving for older parents and When new destinations, that is, new age-
spouses is gendered, with women disproportion- graded social roles, are unclear, it is difcult to
ately caring for their parents and ailing husbands know how to prepare for them or when one is
(Chesley and Moen 2006; Dentinger and ready for them. Instead of manifesting forward
Clarkberg 2002; Pavalko and Woodbury 2000). progress, trajectories may be experienced, and
Moreover, earlier ways of navigating work and objectively characterized, as oundering, turbu-
family obligations by women (still assumed to be lence, and failure. Individuals, especially those
the family care providers) have deleterious con- with less education and other resources, may feel
sequences for them. For example, taking time to a pervasive sense of being off-time, heading
raise their young children by scaling back work toward a precarious and uncertain future.
hours or moving out of the labor force for a time, While numerous examples could be offered,
or moving to follow husbands careers mean that let us take two cases in point: youth in transition
women in encore adulthood (and older) have less from school to work, a primary marker of adult-
savings, are less apt to have pensions, and have hood, and adults moving toward retirement. In
lower social security benets than men. both cases, there appears to be decreasing oppor-
tunity to take on normative and desirable age-
graded social roles. Expectations based on typical
4 Disjunctions Between conditions in the middle of the twentieth century
Structural Opportunity no longer hold, resulting in widespread anxiety,
and Individual disappointment, and uncertainty about what lies
Preference/ Need ahead.
Numerous and widely recognized economic
Superimposed on the dilemmas and uncertainties changes since the 1970s (accentuated since the
surrounding the forging of new age-linked identi- onset of the Great Recession by the nancial,
ties among younger and older adults are contem- housing and other crises) have heightened
porary social trends that lessen the capacity to see inequality in the distribution of income and
120 J.T. Mortimer and P. Moen

wealth, reduced real wages for most Americans, have diminished job opportunities for early and
squeezed families nancially, increased debt, encore adults alike and reduced the commitment
worsened employment opportunities across the of employers to employees. Career-like, full-time
board, and eroded publicly-provided safety nets. jobs and standard employment contracts have
These are the subjects of prodigious research and been overtaken by non-standard, short-term
continuous discussion in the media (for exam- arrangements.
ples, see Edsall 2012; Grusky et al. 2011; Reich The transition from school to work in the
2011). More to the point here, these macro- United States has historically taken longer and
economic trends exacerbate the uncertainties has involved more uncertainty than in other coun-
associated with movement into contemporary, tries with stronger institutionalized pathways
highly differentiated, age-linked social roles and from school to work, e.g., Germany, Austria, and
the acquisition of corresponding age-graded Denmark with their apprentice systems, or Japan,
identities. with linkages between schools and employers
Let us rst consider the plight of youth mov- (Mortimer and Krger 2000; Kerckhoff 2003).
ing through the transition from school to work. But under the labor market conditions of the
Traditionally, up through the mid-twentieth cen- new global economy, oundering has taken
tury, teenagers could anticipate a fairly rapid on new meaning as young people encounter
movement from high school or college, to full- exceedingly tight employment markets after
time work. As a result, a large part of vocational leaving school and long periods of job search.
development occurred during the teen years. While most attention has been focused on high
Adolescents held paid part-time jobs while they school dropouts, who have the poorest prospects,
attended high school, which enabled them to in recent cohorts even college graduates and
acquire familiarity with the workplace while recipients of Ph.D.s and professional degrees are
gaining a sense of responsibility, independence, experiencing more difculty making this key
and work readiness. They learned the impor- transition to adulthood the acquisition of full-
tance of timeliness, how to interact with supervi- time career-like work. Young people try to
sors and customers, and appropriate dress, styles obtain as much education as they can, but those
of speaking, and other behaviors on the job who obtain just some college (without any
(Mortimer 2003). This important venue of antici- degree receipt) may nd themselves in just as
patory socialization has now been lost due to the poor circumstances, with respect to their capacity
virtual collapse of the teenage labor market (Fogg to obtain adult-like career jobs, as those who
and Harrington 2011; Staff et al. 2014). Echoing only have high school diplomas (Vuolo et al.
some of the same social and economic trends 2009).
described by Osterman (1979) at the turn of the As a result, the transition from school to work,
last century, contemporary opportunities for teen a hallmark of transition to adulthood, is no longer
work have declined as a result of displacement by a discrete shift. Instead, youth move from school
newly-arrived immigrants and older workers, to work and back again, as they nd that their
changes in the distributions of occupations and educational credentials are insufcient for the
industries, e.g., disappearance of the paper boy kinds of jobs they want. They often pursue both
(or girl) and full-service gas stations, internet work and educational degrees in tandem, or
marketing, and increasing self-service in retail neither, as they suffer bouts of unemployment or
and other sectors. engage in other pursuits homemaking, or, for
At the same time, continuing occupational the most advantaged, travel abroad or voluntary
changes, rapid technological shifts, particularly service work. Many youth work without pay, vol-
related to ICT (information and communications unteering or in unpaid internships, in attempts to
technology), and a multiplex of trends associated build valuable employment networks or gain the
with globalization (increasing competition, cost- kinds of work-related experience that employers
cutting, out-sourcing, mergers, and downsizing) seek.
The Changing Social Construction of Age and the Life Course: Precarious Identity 121

In the new economy, early adults are told At the other end of the life span, deteriorating
that they must carve out their own careers rather economic conditions and retirement protections
than climb well-established career ladders (Heinz have exacerbated the uncertainties and difcul-
2003). They must become self-employed entre- ties that have traditionally characterized the
preneurs, free-lancers, and consultants, generat- movement into retirement. Adults in their 50s
ing their own start-ups and clienteles. A recent may nd it difcult to keep what were formerly
article in the New York Times (Rampell 2012) secure jobs, and may call themselves retired
declared that personal trainers had become one of when they are unable to nd employment.
the fastest growing occupations. Employment is Paralleling youths meandering movements
quite unstable as young people try their luck with between school and work, a smooth transition
various entrepreneurial pursuits. from work to retirement no longer characterizes
Lengthening education and difculties in many older workers expectations, experiences,
acquiring work that enables economic indepen- or preferences, even though existing time clocks
dence generally preclude independent residence and calendars constrain their options (Moen et al.
(resulting in boomerang children) and delay a 2005). Instead, they often move in and out of the
sense of readiness to marry and have children. labor force, from full- to part-time jobs, and back
Many young people settle on less stable cohabit- again, or from career-like to non-career employ-
ing and childless arrangements; others have chil- ment (Moen 2003, forthcoming). The outdated
dren out of wedlock. Along with these trends social organization of work (linking, for exam-
come further uncertainty and difculty in acquir- ple, health care and other benets to full-time
ing a rm identity as an adult (Eliason et al. 2015). jobs, or policies encouraging one-way, one-time
Difculties in the contemporary transition retirement exits) lessens encore adults capacity
from school to work, and to adulthood more gen- to enact their preferences for greater exibility
erally, could be offset by ameliorative social poli- and less time-consuming or less physically
cies. For example, public colleges have demanding work.
experienced major cutbacks in state education As work has become more precarious, adult
spending in recent decades, placing the economic work careers become more make-shift, more like
burden of college education more squarely on successions of jobs interspersed with periods of
families and increasing both student and parent unemployment, rather than progressive move-
debt. Restoration of such funding would lower ments, building on growing knowledge and skills
the high costs of post-secondary education, borne that heighten occupational allegiances and identi-
by students and their families, and lessen the ties (Moen this volume). Adult workers increas-
need for young adults to live with their parents, ingly view their jobs much like free-lance
with potential improvement in intergenerational musicians view their gigs, temporary respites
relationships. The disappearance of teenage work from unemployment. But as workers grow older,
in the free market economy could be offset, at a weak economy together with age discrimina-
least partially, by government sponsored work tion, age-based stereotypes, and an age-stratied
programs, more opportunities for volunteer work labor market (presuming entry level jobs go to
for teens, and closer connections between high young people) make it difcult for older workers
schools and workplaces (e.g., providing intern- who are laid off (or forced to take early retire-
ships, job shadowing, and more congruence ment packages) to become reemployed, or for
between knowledge and skills learned in school retirees to nd post-retirement employment.
and those needed in the workforce). Just as Contemporary older workers confront height-
investment in paid work during high school ened job insecurity, unemployment, and non-
reduced oundering during the school to work standard employment all of which have been
transition in an earlier era (Vuolo et al. 2014), shown to affect health (c.f. Price and Burgard
such experiences may likewise be benecial, 2008). Because unemployed older workers nd it
promoting anticipatory socialization to work more difcult to obtain new work, they suffer
among contemporary teenagers. from long-term unemployment more often than
122 J.T. Mortimer and P. Moen

younger workers who lose their jobs. Many are Decisions about marital and parental timing
forced into what becomes a too early retire- and divorce earlier in adulthood also have impli-
ment as repeated unsuccessful job searches give cations for those in later life course stages.
way to becoming discouraged workers, drop- Fundamental changes have occurred in the insti-
ping out of the labor force. Whereas unemployed tutions of marriage and the family (c.f. Cherlin
or discouraged workers in their mid-60s at least 2009) and in family demography, including fewer
qualify for Medicare and Social Security benets children, parenthood initiated at later ages,
(if they have sufcient work experience), unem- increasing divorce and remarriage rates, growing
ployed men and women in their 50s face wholly legitimacy of same-sex marriage, the high pro-
inadequate safety nets (e.g., food stamps and portion of unwed pregnancies, and marriage or
tight limits on welfare income). parenthood forgone. These changes are challeng-
ing established norms about lial responsibility
and reducing (or sometimes expanding) the net-
5 Increasing Diversity Within works of kin available for, and willing to care for,
Age Categories ailing older relatives. Extended durations of fam-
and Continuity Over Time ily relationships as a result of increasing longev-
ity also change the nature of kinship ties, meaning
Of course, the timing of earlier transitions has that the linked lives of kin are lasting longer than
long-lasting consequences, as earlier decisions ever (Saraceno 2008; Silverstein and Giarrusso
and circumstances play themselves out through 2011). Extended longevity also increases the
the individual and family life spans. As the pas- odds that older adults will become care-providers
sage to adulthood lengthens for those who extend for even older aging parents/inrm spouses, as
their educations to the highest levels, others drop well as sources of economic support for their
out of high school and begin child bearing in adult children. Marriage is an important form of
their teens and early 20s. Thus, late child-bearers social support that is linked to reduced illness and
may be sitting alongside parents half their age in increased longevity. But while many older adults
elementary school PTA meetings. The delay of continue to enjoy support provided by adult chil-
child-bearing until the early 40s and even beyond, dren and grandchildren, divorce and step-
coupled with a lengthening transition to adult- families, along with lower fertility rates, have
hood in the next generation, means that many increased the odds that older adults will move
contemporary parents may have to delay the free- through their later years alone, with little family
doms of the empty nest, the opportunities of a support. Women are more apt than men to be or
third age of reinvention, and the joys of grandpar- to become single (due to differential widowhood,
enthood until they are in their 70s. remarriage, and mortality rates), and to be more
The timing of occupational career acquisition at risk of social isolation.
has major consequences for subsequent attain- Despite such diversity in circumstances within
ments. A latent class analysis of data from the age categories and greater uncertainty and unpre-
longitudinal Youth Development Study (Vuolo dictability throughout life, many individuals still
et al. 2009) found that a bare majority (53 %) of experience continuity, over time, in attitudes,
youth who left high school in the early 1990s behaviors, and life circumstances. In his study of
were able to nd what they themselves consid- Americans born in the 1920s, Clausen (1993)
ered career jobs; these were likely to have found that adolescents who exhibited planful
obtained Associates or BA/BS degrees. Almost competence made better life choices and, as a
half the panel did not locate such work even by result, had smoother and more satisfying lives
the age of 30. What this means is that discrepancy (e.g., more stable careers and marriages).
in occupational circumstances and career prog- Similarly, in a recent cohort of youth born in the
ress early on continues through the work life, as mid-70s, those who manifested a sense of eco-
same-age individuals nd themselves in very dif- nomic efcacy during adolescence (Lee and
ferent phases of their careers. Mortimer 2009) were found to more successfully
The Changing Social Construction of Age and the Life Course: Precarious Identity 123

acquire the markers of adulthood. Agentic orien- scholars. In view of the continuing differentiation
tations and behaviors (Hitlin and Elder 2007; of age stages, historically and across cohorts, it
Marshall 2005) during the transition to adulthood cannot be assumed that the life course has the
were measured by the maintenance of high aspi- same meaning through time. Instead, as they
rations, career certainty, and active job search come of age successive cohorts develop dis-
(Mortimer et al. 2014; Vuolo et al. 2012). Youth tinct expectations about the course that their lives
who exhibited continuingly high agency had will follow, based largely on the experiences of
higher socioeconomic status, more stable work their own parents and grandparents and various
careers, and more intrinsically-satisfying work cultural cues.
experiences in their mid-30s, and were more suc- But in periods of rapid change, these expecta-
cessful in weathering the Great Recession. In tions may not be fullled, causing new opportu-
contrast, youth who exhibited low and declining nities for some and distress for others. Thus, the
agency during the transition were more likely to more advantaged emerging adults may relish
suffer unemployment and poor work conditions. the greater opportunities for exploration that their
These differences were not explained by the extended freedom from adult responsibilities
youths educational attainments. Since successes, allows. Less advantaged segments of the early
as well as adversities, are likely to cumulate adult population cannot take advantage of such
(Dannefer 1987; Schafer et al. 2011) over time, it opportunities (e.g., those attendant on extended
is likely that youth who exhibit these patterns of higher education, travel, unpaid internships, etc.).
agency will further diverge, as they move into They are greatly disadvantaged in competition
their middle and older years, in economic status for jobs that offer a livable wage, enabling for-
as well as other dimensions of well-being. ward movement with respect to the traditional
Furthermore, the advantages and disadvan- markers of adulthood (leaving home, marrying,
tages associated with long-term patterns of becoming a parent, etc.). Similarly, encore
agency, occupational attainment and career sta- adults may be able to nd new sources of fulll-
bility through the adult career affect economic ment in later-life entrepreneurial ventures or
standing in later adulthood. Pensions, social careers as volunteers when they are no longer
security payments, and personal savings and tied down by demanding work and family
investments depend on earlier career progress, roles. Others, subject to forced early retirement
stable employment, and economic well-being. as a result of layoffs and unemployment, may
The fact that individual retirement accounts are experience deprivation, lack of purpose, and ill
quickly replacing dened pension-like retirement health, without meaningful roles to structure their
benets (a trend that ORand 2003, calls the time and provide material sustenance. The most
individualization of risk) makes it incumbent disadvantaged elders may not even live long
upon workers to start their investment programs enough to experience either encore or deprived
early and to save consistently in order to accumu- adulthoods.
late sufcient resources to obtain a secure retire- At the same time, different cohorts, with their
ment. The increasing turbulence of individual distinct conceptions of what the life course is, or
work careers makes it less likely that workers should be like, come together in their continu-
will be able to do so (Arthur and Rousseau 1996). ingly linked lives. For example, Newman (2012)
documents the ambiguities and sometimes con-
icts that arise when young adults around the
6 Conclusion world continue to live at home well into their 20s
and even 30s, much to their parents dismay at
Full recognition of the socially constructed and their childrens failure to enact the parents
malleable character of the highly diversied expectations, based on their own experience, for
phases of life gives rise to several interrelated a more accelerated transition to adulthood.
questions with strong implications for life course Similarly, contemporary elderly people, who
124 J.T. Mortimer and P. Moen

have provided substantial care to their own, now ambiguities, ambivalences, and dilemmas that
departed parents, may have developed expecta- signal the obsolescence of prior age-graded con-
tions about lial caregiving based on their own cepts and norms may at the same time suggest
experience. In the present era of enhanced lon- that a new phase is emerging.
gevity, and with declining numbers of adult chil- We conclude that there are many reasons for
dren who can share this responsibility, these both scholars of children and youth and aging
elders may not comprehend why their own adult and the life course to come together to share their
or encore adult children cannot provide care insights. Through collaborative investigations,
for them in their own more lengthy period of they are likely to reveal parallels in the social
frailty. constructions of age, problematic self-denitions
What methodological implications follow of age identity, increasing difculties in assum-
from these changing and diversied concepts of ing and enacting age-linked social roles, diversity
age and life stage? This perspective argues for within age groups, and linkages between earlier
continued monitoring of age-graded shifts in and later events across all phases of the life
social roles, identities, and meanings in succes- course. The proliferation of socially-dened age
sive cohorts to reveal change and stability across categories affects both younger (most recently,
these dimensions over time. Macmillan and early adults) and older people (encore
Eliason (2003) make a strong case for multi-level adults). These new life stages provide strategic
latent class analyses to identify shifting pathways research sites for considering larger issues of
to adulthood; this method could also be applied social change, structural lag, identity, inequality,
to the latter phases of life. Congurations of roles diversity, socialization, linked lives, and transi-
at given ages and pathways across such congu- tions. Members of newly-emergent age catego-
rations through time can thus be monitored (see ries encounter difculties in nding their place
Eliason et al. 2015). Representative samples of and acquiring normative age-linked identities
the population, with oversampling of racial/eth- (e.g., as an adult, or one who is in encore
nic minorities, immigrants, and other groups, adulthood). While age differentiation continues,
which are likely to experience distinct life recent economic trends make it less likely that
courses, are of utmost importance. Systematic both younger and older individuals will be able to
assessment of the prevalence of pathways, by move into preferred age-graded social roles (e.g.,
socioeconomic status, race, and gender, will illu- as full-time adult worker with a career or as an
minate potentially changing bases of stratica- economically secure and engaged encore adult).
tion of the life course. To understand the changing Variation in the timing of key role transitions
structure of the subjective life course, this role- (e.g., parenthood, the acquisition of stable adult-
based approach requires supplementation by like work, retirement) results in a diversity of
studies, both quantitative and qualitative, of age- life circumstances for persons of the same chron-
based identities, meanings, and evaluations. ological age, with important implications for age
Not that it is easy, as social changes unfold, to group solidarity and political action. Continuing
identify precisely when a new phase has attention to agency and life conditions, structural
emerged, nor to understand the features that dif- constraints throughout the life course, and the
ferentiate it from contiguous earlier and later life potential outdated character of age-graded poli-
stages. New socially constructed age grades may cies is necessary to understand long-term pat-
come to light when old ways of thinking about terns. Advantages and disadvantages cumulate
the life course no longer ts, as new stages are over time, but in different ways for women and
recognized and named by those experiencing men (Bird and Rieker 2008; Hudson 2005, 2009;
them as well as by social commentators, the Marshall 2005; Moen forthcoming; Mortimer
media, and the organizations that alter their prac- 2003; Schafer et al. 2011; Willson et al. 2007).
tices in an attempt to accommodate new age- Long-term expressions of individual agency,
graded realities. The very contradictions, which are more or less effective given structural
The Changing Social Construction of Age and the Life Course: Precarious Identity 125

constraints and opportunities, produce stratica- more exible social policies, safety nets, and
tion across multiple dimensions of well-being interventions, less closely tied to chronological
(economic security, psychological and physical age and existing (increasingly obsolete) age cat-
health) in old age. egorizations, and better adapted to the new, more
These considerations make it necessary to exible and challenging life courses of contem-
rethink programs and policies geared to earlier porary cohorts. Both early and encore adulthood
times, in which socially constructed age catego- offer strategic opportunities to rethink and reset
ries were more clearly demarcated, more closely mindsets, research agendas, organizational
tied to chronological age, and less strongly dif- arrangements, and policies in ways that open up
ferentiated. For example, policies in the second rather than constrain the twenty-rst century life
half of the adult life course are designed to pro- course.
vide social protections around the transition to
retirement (e.g., Social Security and Medicare).
At the same time, rigid, age-determined benets, References
expectations, and practices may have adverse
consequences. Consider, for example, the effects AARP. (2014). Staying ahead of the curve 2013: AARP
multicultural work and career study: Older workers in
of taken-for-granted policies that limit access by
an uneasy job market. Washington, DC:
age (such as Medicare at 65), or age discrimina- AARP. Retrieved from www.aarp.org
tion that restricts the rehiring of older laid off Alboher, M. (2013). The encore career handbook: How to
workers, or the employment options of encore make a living and a difference in the second half of
life. New York: Workman Publishing Company.
adults seeking new ways of working or volun-
Alderman, L. (2010). Deciding on care for elderly parents
teering. Social policies need to be better cali- in declining health. New York Times. 12 March 2010.
brated to the exigencies and preferences of early Aries, P. (1962). Centuries of childhood. New York:
and encore adults as they navigate this uncertain Vintage.
Arnett, J. J. (1997). Young peoples conceptions of the
terrain.
transition to adulthood. Youth and Society, 29, 123.
Given the formative period of early childhood, Arnett, J. J. (2000). Emerging adulthood: A theory of
the growing importance of schooling in modern development from the late teens through the twenties.
economies, and the cumulative character of aca- American Psychologist, 55, 469480.
Arthur, M. B., & Rousseau, D. M. (1996). The boundary-
demic achievement, interventions for young chil-
less career: A new employment principle for a new
dren are largely focused on enhancing school organizational era. New York: Oxford University
readiness and the transition to rst grade (e.g., Press.
publicly funded universal preschool and the Head Benson, J., & Furstenberg, F. F., Jr. (2006). Entry into
adulthood: Are adult role transitions meaningful
Start program for children in low-income fami-
markers of adult identity? Advances in Life Course
lies). Kuh and Ben-Shlomo (2004, p. 458) argue Research, 11, 199214.
for the need for life course epidemiologists and Berkman, L. F., Ertel, K. A., & Glymour, M. M. (2011).
policy makers to move beyond childhood inter- Aging and social intervention: Life course perspec-
tives. In R. H. Binstock & L. K. George (Eds.),
ventions to identify opportunities to break
Handbook of aging and the social sciences (pp. 337
adverse chains of risk at other life stages (see 351). San Diego: Academic.
also Berkman et al. 2011; Dannefer et al. this vol- Bird, C. E., & Rieker, P. P. (2008). Gender and health:
ume). Their emphasis on the need for policies The effects of constrained choices and social policies.
New York: Cambridge University Press.
addressing adolescent and early adulthood transi-
Brckner, H., & Mayer, K. U. (2005). De-standardization
tions to provide not just safety nets but spring- of the life course: What it might mean? And if it means
boards to alter life course trajectories with anything, whether it actually took place? Advances in
benets for subsequent health could be applied Life Course Research, 9, 2753.
Buchmann, M. (1989). The Script of Life in Modern
to transitions throughout the life course.
Society. Chicago, IL: University of Chicago Press.
Increasing risk, turmoil, and uncertainty through- Carr, D. C., & Manning, L. K. (2010). A new paradigm for
out the life span, with frequent and often unpre- qualitative research in the United States: The era of the
dictable change in life circumstances, argue for third age. Qualitative Sociology Review, VI(1), 1632.
126 J.T. Mortimer and P. Moen

Carr, D. C., & Muschert, G. W. (2009). Risk and aging. In Farrell, C. (2014). Encore careers affect planning for
J. Powell & A. Wahidin (Eds.), Risk and social wel- retirement. StarTribune: Minneapolis. Retrieved,
fare. Hauppauge: Nova Science Publishers. August 11, 2015 at http://www.startribune.com/
Carstensen, L. L. (2011). A long bright future: Happiness, sept-21-farrell-encore-careers-affect-planning-
health and nancial security in an age of increased forretirement/275815821/
longevity. New York: Public Affairs. Fine, G. A. (2001). Gifted tongues: High school debate
Cherlin, A. (2009). The marriage-go-round: The state of and adolescent culture. Princeton: Princeton
marriage and the family in America today. New York: University Press.
Alfred A. Knopf. Flood, S., & Moen, P. (2015). Healthy time use in the
Chesley, N., & Moen, P. (2006). When workers care: encore years. Do work, resources, relations, and
Dual-earner couples caregiving strategies, benet gender matter? Journal of Health and Social Behavior,
use, and psychological well-being. American 56, 7497.
Behavioral Scientist, 49(9), 12481269. Fogg, N. P., & Harrington, P. E. (2011). The collapse of
Clausen, J. A. (1993). American lives: Looking back at the the labor market for 1624 year-olds. Cascade, 78
children of the great depression. Berkeley: University retrieved August 11, 2015 at https://www.philadelphi-
of California Press. afed.org/community-development/publications/
Considine, A. (2012). For millions, a poignant bond. New cascade/78/04_collapse-of-16-to-24-labor-market
York Times. 13 June 2012. Freedman, M. (2007). Encore: Finding work that matters
Cote, J. E. (2014). The dangerous myth of emerging adult- in the second half of life. New York: Public Affairs.
hood: An evidence-based critique of a awed develop- Freedman, M. (2011). The big shift: Navigating the new
mental theory. Applied Developmental Science, 18, stage beyond midlife. New York: Public Affairs.
177188. Furstenberg, F. F., Jr. (2008). The intersection of social
Dannefer, D. (1987). Aging as intracohort differentiation: class and the transition to adulthood. New Directions
Accentuation, the Matthew Effect, and the life course. for Child and Adolescent Development, 119, 110.
Sociological Forum, 2, 211237. Furstenberg, F. F., Jr., Kennedy, S., McLloyd, V., Rumbaut,
Delehanty, H., Ginzler, E., & Pipher, M. (2008). Caring R. G., & Settersten, R. A., Jr. (2004). Growing up is
for your parents: The complete family guide. harder to do. Contexts, 3, 3341.
New York: Sterling Publishing Company. Garfein, A. J., & Herzog, A. R. (1995). Robust aging
Dentinger, E., & Clarkberg, M. (2002). Informal caregiv- among the young-old, old-old, and oldest old. Journal
ing and retirement timing among men and women: of Gerontology: Social Sciences, 50B, S77S87.
Gender and caregiving relationships in late midlife. Graff, H. J. (1995). Conicting paths. Growing up in
Journal of Family Issues, 23(7), 857879. America. Cambridge, MA: Harvard University Press.
Drobni, S., & Blossfeld, P.-B. (2004). Career patterns Grusky, D. B., Western, B., & Wimer, C. (2011). The great
over the life course: Gender, class, and linked lives. In recession. New York: Russell Sage Foundation.
A. L. Kalleberg, S. L. Morgan, J. Myles, & R. A. Hagestad, G. O., & Neugarten, B. L. (1985). Age and the
Rosenfeld (Eds.), Inequality: Structures, dynamics life course. In E. Shanas & R. Binstock (Eds.),
and mechanisms. Essays in honor of Aage B. Srensen Handbook of aging and the social sciences (2nd ed.,
(Research in social stratication and mobility, Vol. 21, pp. 4661). New York: Van Nostrand Reinhold.
pp. 139164). Amsterdam: Elsevier. Hall, G. S. (1904). Adolescence: Its psychology and its
Edsall, T. (2012). The hollowing out. New York Times. 8 relations to physiology, anthropology, sociology, sex,
July 2012. crime, religion, and education (Vol. 1). New York:
Elder, G. H., Jr. (1975a). Age differentiation and the life D. Appleton & Company.
course. Annual Review of Sociology, 1, 165190. Han, S.-H., & Moen, P. (1999). Clocking out: Temporal
Elder, G. H., Jr. (1975b). Adolescence in the life cycle. In patterning of retirement. American Journal of
S. E. Dragastin & G. H. Elder Jr. (Eds.), Adolescence Sociology, 105, 191236.
in the life cycle (pp. 122). Washington, DC: Hartmann, D., & Swartz, T. T. (2007). The new adult-
Hemisphere/Halsted Press. hood? The transition to adulthood from the perspec-
Elder, G. H., Jr., Johnson, M. K., & Crosnoe, R. (2003). tive of transitioning young adults. Advances in Life
The emergence and development of life course theory. Course Research, 11, 255289.
In J. T. Mortimer & M. J. Shanahan (Eds.), Handbook Heinz, W. R. (2003). From work trajectories to negotiated
of the life course (pp. 319). New York: Springer. careers: The contingent work life course. In J. T.
Eliason, S. R., Mortimer, J.T., & Vuolo, M. (2015) The tran- Mortimer & M. J. Shanahan (Eds.), Handbook of the
sition to adulthood: life course structures and subjective life course (pp. 185204). New York: Springer.
perceptions. Social Psychology Quarterly, 78 (3). Hitlin, S., & Elder, G. H., Jr. (2007). Time, self, and the
Estes, C. L. (2004). Social security privatization and older curiously abstract concept of agency. Sociological
women: A feminist political economy perspective. Theory, 25, 170191.
Journal of Aging Studies, 18(1), 926. Holstein, M. B., & Minkler, M. (2003). Self, society, and
Estes, S. B., Noonan, M. C., & Maume, D. J. (2007). Is the New Gerontology.. The Gerontologist, 43(6),
work-family policy use related to the gendered divi- 787796.
sion of housework? Journal of Family and Economic Hudson, R. B. (2005). The new politics of old age policy.
Issues, 28, 527545. Baltimore: The Johns Hopkins University Press.
The Changing Social Construction of Age and the Life Course: Precarious Identity 127

Hudson, R. B. (2009). From industrialization to institu- Marshall, V. W. (2005). Agency, events, and structure at
tionalism: Theoretical accounts of aging policy devel- the end of the life course. In R. Levy, P. Ghisletta,
opment in the United States. In V. L. Bengtson, J.-M. LeGoff, D. Spini, & E. Widmer (Eds.), Advances
M. Silverstein, N. M. Putney, & D. Gans (Eds.), in life course research (Vol. 10, pp. 5791). Oxford:
Handbook of theories of aging (pp. 531554). Elsevier.
New York: Springer. Massoglia, M., & Uggen, C. (2010). Settling down and
James, J., & Wink, P. (2007). The crown of life: Dynamics aging out: Toward an interactionist theory of desis-
of the early postretirement period. New York: Springer. tance and the transition to adulthood. American
Johnson, M. K., & Mollborn, S. (2009). Growing up Journal of Sociology, 116, 543582.
faster, feeling older: Hardship in childhood and ado- Mayer, K. U. (2004). Whose lives? How history, societies
lescence. Social Psychology Quarterly, 72, 3960. and institutions dene and shape life courses. Research
Johnson, M. K., Berg, J. A., & Sirotzki, T. (2007). in Human Development, 1, 161187.
Differentiation in self-perceived adulthood: Extending Mayer, K. U. (2009). New trends in life course research.
the conuence model of subjective age identity. Social Annual Review of Sociology, 35, 493514.
Psychology Quarterly, 70, 243261. Meyer, J. W. (1986). The self and the life course:
Kahn, R. L., & Antonucci, T. C. (1980). Convoys over the Institutionalization and its effects. In A. B. Sorensen,
life course. Attachment, roles, and social support. In F. E. Weinert, & L. R. Sherrod (Eds.), Human develop-
P. B. Baltes & O. G. Brim (Eds.), Life-span develop- ment and the life course: Multidisciplinary perspec-
ment and behavior (pp. 254283). New York: tives (pp. 199216). Hillsdale: Erlbaum.
Academic. Milkie, M. A., Bierman, A., & Schieman, S. (2008). How
Kaufman, G., & Elder, G. H., Jr. (2002). Revisiting age adult children inuence older parents mental health:
identity: A research note. Journal of Aging Studies, 16, Integrating stress-process and life-course perspec-
169176. tives. Social Psychology Quarterly, 71(1), 86105.
Kaufman, G., & Elder, G. H., Jr. (2003). Grandparenting Moen, P. (2003). Midcourse: Navigating retirement and a
and age identity. Journal of Aging Studies, 17, new life stage. In J. T. Mortimer & M. J. Shanahan
269282. (Eds.), Handbook of the life course (pp. 269291).
Kerckhoff, A. C. (2003). From student to worker. In J. T. New York: Springer.
Mortimer & M. J. Shanahan (Eds.), Handbook of the Moen, P. (2007). Not so big jobs and retirements: What
life course (pp. 251267). New York: Springer. workers (and retirees) really want. Generations, 31(1),
Kertzer, D. I., & Laslett, P. (1995). Aging in the past: 3136.
Demography, society, and old age. Berkeley: Moen, P. (2011a). Living the gendered life course in time
University of California Press. and space. In R. Settersten & J. Angel (Eds.),
Kohli, M. (2007). The institutionalization of the life Handbook of the sociology of aging and the life course
course: Looking back to look ahead. Research in (pp. 655657). New York: Springer.
Human Development, 4(3), 253271. Moen, P. (2011b). A life course approach to the third age.
Kohli, M., Rein, M., Guillemard, A.-M., & van Gunsteren, In D. C. Carr & K. Komp (Eds.), Gerontology in the
H. (1991). Time for retirement: Comparative studies of era of the third age: Implications and next steps
early exit from the labour force. Cambridge: (pp. 1332). New York: Springer.
Cambridge University Press. Moen, P. (2013). Constrained choices: The shifting insti-
Kotter, J. P. (1995). The new rules: How to succeed in tutional contexts of aging and the life course. In
todays post-corporate world. New York: Free Press. L. Waite (Ed.), Perspectives on the future of the sociol-
Kuh, D., & Ben-Shlomo, Y. (2004). A life course approach ogy of aging (pp. 175216). Washington, DC: National
to chronic disease epidemiology. New York: Oxford Research Council.
University Press. Moen, P. (Forthcoming). Encore Adulthood: Boomers on
Laslett, P. (1989). A fresh map of life: The emergence of the Edge of Risk, Renewal, and Purpose. New York:
the third age. Cambridge: Harvard University Press. Oxford University Press.
Lee, J. C., & Mortimer, J. T. (2009). Family socialization, Moen, P., & Flood, S. (2013). Limited engagements?
economic self-efcacy, and the attainment of nancial Womens and mens work/volunteer time in the encore
independence in early adulthood. Longitudinal and life course stage. Social Problems, 60(2), 206233.
Life Course Studies, 1, 4562. Moen, P., & Hernandez, E. (2009). Social convoys:
Levin, J. (2013). The declining signicance of age. Studying linked lives in time, context, and motion. In
New York: Routledge G. Elder Jr. & J. Giele (Eds.), The craft of life course
Liang, J., & Luo, B. (2012). Toward a discourse shift in research (pp. 258279). New York: Guilford Press.
social gerontology: From successful aging to harmoni- Moen, P., & Peterson, J. (2009). A third path? Multiplex
ous aging. Journal of Aging Studies, 26, 327334. time, gender, and retirement encores in the United
Macmillan, R., & Eliason, S. R. (2003). Characterizing the States. Nova Acta Leopoldina, 106(370), 4158.
life course as role configurations and pathways: A Moen, P., & Spencer, D. (2006). Converging divergences
latent structure approach. In J. T. Mortimer & in age, gender, health, and well-being: Strategic selec-
M. J. Shanahan (Eds.), Handbook of the life course tion in the third age. In R. Binstock & L. George
(pp.529554). New York: Kluwer Academic/Plenum (Eds.), Handbook of aging and the social sciences
Publishers. (pp. 127144). Burlington: Elsevier Academic Press.
128 J.T. Mortimer and P. Moen

Moen, P., Sweet, S., & Swisher, R. (2005). Embedded Riley, M. W. (1994). Aging and society: Past, present, and
career clocks: The case of retirement planning. future. The Gerontologist, 34(4), 436446.
Advances in Life Course Research, 9, 237265. Rix, S. E. (2011). Employment and aging. In R. H.
Mortimer, J. T. (2003). Working and growing up in Binstock & L. K. George (Eds.), Handbook of aging
America. Cambridge, MA: Harvard University Press. and the social sciences (pp. 193206). San Diego:
Mortimer, J. T., & Krger, H. (2000). Pathways from Academic.
school to work in Germany and the United States. In Rosow, I. (1974). Socialization to Old Age. Berkeley:
M. Hallinan (Ed.), Handbook of the sociology of edu- University of California Press.
cation (pp. 475497). New York: Kluwer Academic/ Sadler, W. A. (2006). Changing life options: Uncovering
Plenum Publishers. the riches of the third age. LLI Review, 1, 1120.
Mortimer, J.T., & Moen P. (2012). The social construction Saraceno, C. (2008). Families, ageing and social policy.
of age through the life course: Precarious identity and Cheltenham: Edward Elger.
enactment of age-graded social roles. Paper presented Schafer, M. H., Ferraro, K. F., & Mustillo, S. A. (2011).
at the annual meeting of the American Sociological Children of misfortune: Early adversity and cumula-
Association. Denver, August 1720. tive inequality in perceived life trajectories. American
Mortimer, J. T., & Shanahan, M. J. (2003). Handbook of Journal of Sociology, 116, 10531091.
the life course. New York: Plenum Press. Settersten, R. A., Jr. (2003). Age structuring and the
Mortimer, J. T., Vuolo, M., & Staff, J. (2014). Agentic rhythm of the life course. In J. T. Mortimer & M. J.
pathways toward fulllment in work. In A. C. Keller, Shanahan (Eds.), Handbook of the life course
R. Samuel, M. M. Bergman, & N. K. Semmer (Eds.), (pp. 8198). New York: Springer.
Psychological, educational, and sociological perspec- Settersten, R. A., Jr., & Mayer, K. U. (1997). The mea-
tives on success and well-being in career development surement of age, age structuring and the life course.
(pp. 99126). Dordrecht: Springer. Annual Review of Sociology, 23, 233261.
Neugarten, B. L., & Hagestad, G. O. (1976). Age and the Shanahan, M., Porfeli, E. J., Mortimer, J. T., & Erickson,
life course. In R. Binstock & E. Shanas (Eds.), L. (2005). Subjective age identity and the transition to
Handbook of aging and the social sciences (pp. 3555). adulthood: When does one become an adult? In R. A.
New York: Van Nostrand Reinhold Company. Settersten Jr., F. F. Furstenberg Jr., & R. G. Rumbaut
Neugarten, B. L., Moore, J. W., & Lowe, J. C. (1965). Age (Eds.), On the frontier of adulthood: Theory, research,
norms, age constraints, and adult socialization. and public policy (pp. 225255). Chicago: University
American Journal of Sociology, 70, 710717. of Chicago Press.
Newman, K. (2012). The accordian family: Boomering Silva, L. R. F. (2008). From old age to third age: The his-
kids, anxious parents, and the private toll of global torical course of the identities linked to the process of
competition. Boston: Beacon Press. ageing. Histria, Cincias, Sade Manguinhos,
ORand, A. (2003). The future of the life course: Late 15(1), 155168.
modernity and life course risks. In J. T. Mortimer & Silverstein, M., & Giarrusso, R. (2011). Aging individu-
M. J. Shanahan (Eds.), Handbook of the life course als, families, and societies: Micro-meso-macro link-
(pp. 693701). New York: Springer. ages in the life course. In R. A. Settersten Jr. & J. L.
ORand, A., & Henretta, J. (1999). Age and inequality: Angel (Eds.), Handbook of the sociology of aging
Diverse pathways through later life. Boulder: (pp. 3549). New York: Springer.
Westview Press. Staff, J., Johnson, M. K., Patrick, M. E., & Schulenberg,
Osterman, P. (1979). Education and labor markets at the J. E. (2014). The great recession and recent teenage
turn of the century. Politics and Society, 9, 103122. employment trends in the United States. Longitudinal
Pavalko, E. K., & Woodbury, S. (2000). Social roles as and Life Course Studies, 5(2), http://dx.doi.
process: Caregiving careers and womens health. org/10.14301/llcs.v5i2.275.
Journal of Health and Social Behavior, 41(1), 91105. Swartz, T. T. (2009). Intergenerational family relations in
Price, R. H., & Burgard, S. A. (2008). The new employ- adulthood: Patterns, variations, and implications in the
ment contract and worker health in the United States. contemporary United States. Annual Review of
In R. F. Schoeni, J. S. House, G. A. Kaplan, & Sociology, 35, 191212.
H. Pollack (Eds.), Social and economic policy as Sweet, S., & Meiksins, P. (2013). Changing contours of
health policy: Rethinking Americas approach to work: Jobs and opportunities in the new economy (2nd
improving health (pp. 201228). New York: Russell ed.). Thousand Oaks: Pine Forge Press.
Sage. Vuolo, M., Mortimer, J. T., & Staff, J. (2009). Floundering
Rampell, C. (2012). A jobs boom built on sweat in an age during the school-to-work transition: A latent life path
of belt tightening. New York Times. 30 June 2012. approach. Detroit: Population Association of America.
Reich, R. (2011). Why inequality is the real cause of our April 30May 2, 2009.
ongoing terrible economy. New York Times. 4 Sept Vuolo, M. C., Staff, J., & Mortimer, J. T. (2012).
2011. Weathering the great recession: Psychological and
Riley, M. W. (1987). On the signicance of age in sociol- behavioral trajectories in the transition from school to
ogy. American Sociological Review, 52(1), 114. work. Developmental Psychology, 48, 17591773.
The Changing Social Construction of Age and the Life Course: Precarious Identity 129

Vuolo, M., Mortimer, J. T., & Staff, J. (2014). Adolescent Williamson, J. B. (2011). The future of retirement secu-
precursors of pathways from school to work. Journal rity. In R. H. Binstock & L. K. George (Eds.),
of Research on Adolescence, 24, 145162. Handbook of aging and the social sciences (pp. 281
Warner, D. F., Hayward, M. D., & Hardy, M. A. (2010). 294). San Diego: Academic.
The retirement life course in America at the dawn of Willson, A., Shuey, K. M., & Elder, G. H., Jr. (2003).
the twenty-rst century. Population Research and Ambivalence in the relationship of adult children to
Policy Review, 29(6), 893919. aging parents and in-laws. Journal of Marriage and
Warren, J. R., Knies, L., Haas, S., & Hernandez, E. M. Family, 65, 10551072.
(2012). The impact of childhood sickness on adult Willson, A., Shuey, K. M., & Elder, G. H., Jr. (2007).
socioeconomic outcomes: Evidence from later 19th Cumulative advantage processes as mechanisms of
century America. Social Science & Medicine, 75(8), inequality in life course health. American Journal of
15311538. Sociology, 112, 18861924.
Weiss, R. S., & Bass, S. A. (2002). Challenges of the third Wong, J. D., & Hardy, M. A. (2009). Womens retirement
age: Meaning and purpose in later life. Oxford: expectations: How stable are they? Journal of
Oxford University Press. Gerontology: Social Sciences, 64B(1), S77S86.
Williams, K., & Umberson, D. (2004). Marital status, Youmans, E. G. (1977). Attitudes: Young-old and old-old.
marital transitions, and health: A gendered life course The Gerontologist, 17, 175178.
perspective. Journal of Health and Social Behavior, Zelizer, V. (1985). Pricing the priceless child. Princeton:
45, 8198. Princeton University Press.
Structuration of the Life Course:
Some Neglected Aspects

Gunhild O. Hagestad and Pearl A. Dykstra

1 Our Point of Departure solving analytical puzzles if there are interna-


tional dialogues and collaboration across
Since both authors have studied and worked in research communities.
Europe as well as in North America, it is natural After a brief historical overview of approaches
for us to take a comparative view, exploring to the life course on two continents (Sect. 2), we
contrasts and similarities between European and highlight some classic work on mens and wom-
North American perspectives. We start from two ens lives and raise the issue of whether life
basic premises: First, life course studies are a course scholars on two continents have met the
perfect arena in which to raise and seek solu- empirical and analytical challenges of analyzing
tions to some fundamental analytical puzzles how gendered life courses are shaped by cultural
that have faced social scientists since the start of expectations as well as contrasting positions in
our disciplines. Central among them are rela- society (Sect. 3). We ask: to what extent do
tionships among levels and units of analysis. macro-level forces create different patterns of
Among major gures in classic social science, interdependence, divergent and contrasting life
Durkheim, through his emphasis on social facts, trajectories for men and women in the early
took the clearest stance with regard to levels: decades of a new century? As part of the focus on
collective life does not derive from individ- macro-level structures, we include a Sect. 4 out-
ual life, [and] the latter cannot explain the lining key demographic shifts. Subsequently, we
former (Durkheim 1895/1982, p. 134). explore how demographic context, laws and poli-
Second, there is a much better chance of build- cies create contrasting patterns of interdepen-
ing new insights, lling knowledge gaps, and dence (Sect. 5) and different transition patterns
for men and women (Sect. 6). In the nal Sect. 7
we attempt to restate and discuss key analytical
and empirical challenges facing life course
With invaluable assistance from Stephan van Baarle, who researchers who aim to understand the inuences
helped with literature and policy searches. of demographic contexts, welfare regimes and
G.O. Hagestad (*) methodological traditions.
NOVA Research Institute, Oslo and Akershus It is important to note that in attempting a
University College of Applied Sciences,
Oslo, Norway
rather panoramic view of complex issues, our
e-mail: gohagestad@gmail.com aim is not to provide a comprehensive overview
P.A. Dykstra
of facts and ndings, but rather to highlight
Erasmus University, Rotterdam, The Netherlands dimensions and perspectives that we feel have
e-mail: dykstra@fsw.eur.nl not received the attention they deserve.

Springer International Publishing Switzerland 2016 131


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_6
132 G.O. Hagestad and P.A. Dykstra

In several recent overviews of life course In the 1960s and 1970s, key publications
research, authors have argued that North marked a new epoch for the study of how social
American and European scholars have tended to structure and culture assign social meanings to
focus on different levels of analysis, emphasizing maturation and aging, as well as to metered
contrasting dimensions of social context. For biographical time, i.e., chronological age. In
example, Leisering (2004) suggests that US 1964, Leonard Cain published a pioneering paper
researchers have tended to take a social- on life course and social structure. He based his
psychological perspective, emphasizing the presentation on classic anthropological studies of
micro-and meso levels of individuals and their age differentiation, age grading and rites of pas-
families, contemplating culture and shared mean- sage. Focusing on age-status systems, Cain con-
ing as the main organizing force in shaping lives. centrated on a sequence of age-linked roles and
Many of them build on what is referred to as the emphasized the importance of historical genera-
Chicago school of sociology, with E. Hughes, tions and cohort ow. His discussion, however,
G.H. Mead, W.I. Thomas and F. Znaniecki among had little or no consideration of sex/gender.
the key gures. This tradition within life course Surprisingly, neither did Riley and collaborators,
research goes back to the mid-twentieth century. whose ground-breaking volumes on age strati-
In contrast, European scholars have focused on cation were published in the late 1960s and early
macro-level social structures and the institution- 1970s (e.g., Riley et al. 1972). These authors
alized life course. This tradition builds on classic elaborated how the social structuring of age is
sociological accounts of social change, with tied to the division of labour and a system of
M. Weber as a key gure. stratication. While both Cain and Riley empha-
sized the importance of cohort ow and historical
context, Elders (1974) landmark volume on
2 A Comparative Lens: Children of the Great Depression made this prin-
Perspectives on Lives ciple come to life. Furthermore, his book pre-
in Different Contexts sented evidence of how families are a critical
mediating force, between macro-level historical
2.1 Structuring the Life Course: events and changes, such as the Depression, and
Early North American the micro-level of individual lives. Inspired by
Contributions this view, researchers have later shown that the
family realm may not only intensify, but can also
In the 1950s, scholars of human development soften or diffuse the force of societal conditions.
conducted the pioneering Kansas City Study of For example, many African American parents,
Adult Life, collecting complex interview data. under conditions of racial segregation and dis-
Questions covered perceptions of the normal crimination, instilled condence and courage in
expectable life: culturally shared expectations their children, enabling them to reach occupa-
of age-appropriate behaviors and the right tional goals despite massive structural and cul-
timing of key transitions (e.g. Neugarten 1969; tural barriers in the wider societal context (e.g.,
Neugarten et al. 1965). Implicitly, the project Fields-Smith 2005; Loder-Jackson et al. 2007).
explored concepts that sociologist Robert Scholars focusing on the mediating role of fami-
Merton (1942/1979) called the three Ps: pre- lies clearly illustrate that knowledge of intra-
scription, proscription and permission, in this cohort variation is necessary for understanding
case linked to age status. Since then, similar inter-cohort contrasts.
studies have been carried out in the Chicago Work on family contexts also raises funda-
area (Settersten and Hagestad 1996a, b), in mental issues about gender. Elder (1974) sug-
Japan (Plath and Ikeda 1975) and in more than gested that boys were more negatively affected
20 European societies (Billari et al. 2011; by changes in the familys economic circum-
Spder et al. 2014). stances than were girls, because the latter were
Structuration of the Life Course: Some Neglected Aspects 133

better able to maintain a sense of continuity and meso-levels and structural constraints on the
stability as mothers helpers in the household. macro-level (Diewald 2000; Kohli 1986; Levy
Here, we touch on a question that has often been 2013a, b). Settersten (2003) speaks of agency in
raised in anthropological work: are womens structure, while Elder et al. (2015): argue that:
lives more structured on a micro-level of social Individuals construct their own life course
context, while mens life trajectories are more through the choices and actions they take within
heavily shaped by macro-level institutional the opportunities and constraints of history and
anchoring (Neugarten and Hagestad 1976; Young social circumstance (p. 29). Mayer (2003) takes
1965)? Building on Angrists (Angrist and a somewhat different view, stating that:
Almquist 1975) discussion of womens contin- Sociologists tend to believe more in selection than
gency orientation, Moen (2001) states that in choice..Within given institutional contexts,
womens lives are typically contingent lives, individuals are probably more frequently being
shaped around the experience of others: their selected than selecting themselves.If material
resources, power, authority, information and sym-
husbands, children, and parents (p. 189). In bolic goods are distributed very unequally within
other words, she argues that interdependence given societies, then it follows that more people
among lives is a more powerful force for women have to accommodate than have the opportunity to
than for men. We return to this challenging issue control (p. 466).
in a later section.
Following up his volume on the Depression, As he has done many times, Mayer points out
Elder (1979) worked to systematize perspectives that social groups, in which lives are interwoven,
and concepts for studying lives in changing his- are highly structured by the temporal dynamics
torical contexts. Central in his framework were of social institutions and organizations.
four organizing concepts: transitions, trajecto-
ries, interdependence among lives, and agency.
To describe transitions and trajectories, key con- 2.2 Shaping the Life Course
cepts are timing, sequencing and duration. The Through Laws and Policies:
strong emphasis on culturally shared expecta- European Perspectives
tions and proximal conditions in families, com-
munities and social networks highlights how time In 1976, Cain pointed to the role of law in shap-
and place shape lives. One might think that ing lives, but he later stated (Cain 2003): Most
Elders framework includes macro-level structur- scholarly literature on age-related phenomena
ing through social policy, but as Leisering (2004) has lacked curiosity about the legal basis for the
points out, the discussion of how lives were status of various age-categories (p. 310, empha-
marked by the Depression did not include the sis added). In the same year, anthropologist Fry
possible impact of the New Deal. Subsequent (2003) discussed the bafing variety of age/time
consideration of interdependent lives shows that concepts in modern western societies, suggesting
his perspective is mostly social psychological: that we use the term legislative time. Quite a few
Lives are lived interdependently and social- years earlier, Hernes (1987), in a collection of
historical inuences are expressed through this essays on Welfare states and woman power
network of shared relationships. Interdependent (within a Nordic context), used the term chronop-
lives highlight the role of signicant others in olitics and called for more scholarly attention to
regulating and shaping the timing of life trajecto- the politics of time, age and gender. By the late
ries through a network of informal control 1980s, European, mostly German, writing on the
(Elder et al. 2015, p. 31, emphasis added). life course examined how macro-level structural
Several authors have pointed to problematic conditions shape transitions and trajectories.
aspects of agency as an organizing concept, given Going back to Cains (2003) assessment of the
the realities of interdependent lives on micro-and US, one could argue that the central curiosity in
134 G.O. Hagestad and P.A. Dykstra

Europe was, indeed, legal and policy-bases for examined how welfare states create different
age markers and the denition of life phases. care regimes for young and old, which in turn
Much of this literature emphasized that the state, present starkly contrasting opportunities and con-
through laws and policies, deliberately creates straints in mens and womens life course and
opportunities and constraints, (in German structure parent-child ties in several generations.
Rahmenbedingungen) an overarching framework We return to the issue of care regimes in Sect.
for shaping lives. This point is emphasized by 5.4. In considering the organization of mens and
Leisering (2004), who explicitly states that Life womens lives, both European and US research-
course policies are intended to change the struc- ers could benet by returning to classic discus-
ture of the life course (p. 210). He goes on to say sions of how age and sex in combination present
that researchers must search for tacit objectives, the two sexes with different cultural expectations
not only those that are openly stated. Furthermore, as well as contrasting locations in societys divi-
he makes a distinction between positive and neg- sion of labor and social institutions.
ative life course policies. The latter occur when
policy-makers leave it to markets or charity to
form life course patterns. 3 Her and His Life: Revisiting
In the 1980s, German scholars, with Kohli Classics
(1985, 1986), Mayer and Mller (1986), and
Mayer and Schpin (1989) among the pioneers, In a presentation to the 1941 meetings of the
argued that the modern nation-state, through its American Sociological Society, Ralph Linton
laws, institutions and modern bureaucracies gave invited colleagues to enter new research terrain
rise to the institutionalized life course. States by studying what he called the age-sex system.
structure the life course by delineating phases He emphasized that the characteristics of age
through legally stipulated chronological markers and sex may be treated as a unit, since member-
and by dening rights and duties linked to chron- ship in a particular age-sex category......will be
ological age. States can also contribute to indi- found to be a prerequisite for the occupation of
vidual life course continuity through risk practically any status within a social system
management and safety nets. Some of these pol- (Linton 1942, pp. 589590, emphasis added).
icy efforts aim at supporting and minimizing The issue of the American Sociological
risks for families. A focus on state-provided risk Review in which Lintons presentation was pub-
reduction and support is illustrated in two more lished also has a paper by Talcott Parsons (1942).
recent comparative projects: Leisering and He highlighted age-sex roles in the nuclear fam-
Leibfrieds (1999) study of poverty in Germany ily, especially the asymmetrical relation of the
and the US and an examination of the extent to sexes to the occupational structure (p. 605) and
which nine contemporary European welfare its consequences for men and women in different
states provide support for mens and womens phases of adulthood. Later, Parsons worked with
key adult transitions (Anxo et al. 2010). social psychologist Robert Bales on the dichot-
Early work on welfare state reduction of life omy between emotional expressive and instru-
risks focused on wage labor (e.g., Esping- mental roles (Parsons and Bales 1955).
Andersen 1990), thereby neglecting the fact that Contemporary feminist critiques of work/family
participation in the labor market presupposes a arrangements still return to this distinction, espe-
support-system of unpaid family work. Feminist cially in discussions of parenthood as the big
scholars pointed out that programs developed to divide in differentiating her and his adulthood
protect workers from the vagaries of the market (Levy 2013a; OConnor 1996).
sometimes acted to reproduce and reinforce In the wake of the feminist movement in the
inequalities originating from the unequal distri- 1960s and 1970s came growing research recogni-
bution of care responsibilities at home (Lewis tion that men and women have different adult
2002; Orloff 1993). These scholars have lives. Based on their comprehensive studies of
Structuration of the Life Course: Some Neglected Aspects 135

American life, Campbell, Converse and Rogers than examining these social categories as produc-
(1976) concluded: To an important degree, men ing distinctive cleavages, the principle of inter-
and women grow up in different cultures, develop sectionality leads to scrutiny of how they
different expectations, learn different roles, and mutually interact with one another. Nevertheless,
live different lives (p. 395). When Cain (1976) we argue that even limiting attention to two sta-
argued for studying the role of law in age struc- tuses, little progress has been made in under-
turing, he reminded the reader that in classic legal standing age and gender as intertwined systems.
codes, such as the Code Napoleon, different age In a recent special issue of Signs on intersection-
limits were set for men and womens transitions. ality studies (Cho et al. 2013), age is not men-
For marrying, they specied age 15 for women, tioned at all.
18 for men. To a very limited extent have the calls Alice Rossi, as president of the American
for new knowledge, voiced in the 1970s, been Sociological Association three decades ago,
systematically followed up. McMullin (1995) selected age and gender as the theme of the
and Ginn and Arber (1995) expressed dismay annual meeting. The goal she expressed in the
that the issues identied by Linton have not been resulting volume (Rossi 1985) is still program-
addressed in research, but both focus on old age. matic, but unmet: Hopefully, the next time an
Two decades after the American life studies, editor puts together a volume on these core con-
McMullin (1995) stated that the lack of theoreti- structs of age and gender, the state-of-the art will
cal development concerning the relationship permit a sophisticated integration of new theory
between gender and ageing seems incomprehen- and research that is beyond our contemporary
sible (p. 30). She identied three add on paths ability to provide (p. 17). Since the book was
that have been followed in past work: adding published, numerous overviews devoted to the
gender and/or age to mainstream sociological life course have appeared, for the most part
theorizing, adding gender to sociological theories focusing on age but sidestepping gender. A nota-
of ageing (gendered ageing theory), and adding ble exception is Levy and Widmer (2013), who
age relations to feminist theory (feminist aging challenge scholars to return to sociological work
theory). She argued that if age and gender are published in the mid twentieth century to gain an
organizing dimensions of the social world, then analytical understanding of structural conditions
separating these systems makes no sense. Older that may differentially shape mens and womens
people are not just old, they are either men or life course.
women (p. 37). In the same volume, two In Sects. 5 and 6, we discuss the extent to
European researchers (Ginn and Arber 1995) which societies, through laws and policies, create
remind us that Gender and ageing are inextrica- distinct patterns of interdependence, transitions
bly intertwined in social life; each can only be and life trajectories on the basis of age/sex/gen-
fully understood with reference to the other der. Before we turn to such macro-level structur-
(p. 1). Yet, they go on to say that ageing and gen- ing, we focus on demographic conditions, another
der have rarely been researched in terms of their Rahmenbedingung that creates different patterns
joint inuence. Like McMullin (1995), Ginn and of interdependence and divergence in life phases
Arber (1995) suggest that sociologists con- for men and women. Demographic contexts have,
cerned with ageing and ageism have tended to to a great extent, been neglected in discussions of
add on gender, treating it as a variable rather forces shaping the life course, on both sides of
than integrating it as a fundamental relationship the Atlantic. Mayer (2004) reminds us that
of social organization (p. 2, emphasis added). It is not single individuals but populations that are
In the late 1980s intersectionality was intro- allocated through and streamlined through the
duced as an analytical tool to elucidate the cre- institutional fabric of society across the life time
ation and reproduction of inequalities associated for example, the size of ones cohort, as well as the
preceding and succeeding cohorts, inuences indi-
with salient social categories like gender, sexual- viduals opportunities way beyond individual or
ity, race, age, and class (Crenshaw 1989). Rather situational conditions (p. 165).
136 G.O. Hagestad and P.A. Dykstra

4 Demographic Context Vaupel (2009) commented that in 1840,


of Lives and Relationships Swedish women had the worlds highest life
expectancy at birth 46. Today, Japanese women
4.1 Altered Age Structures hold the world record a bit over 87. While pop-
ulations for most of human history have consisted
The demographic transition (Davis 1945) with its of at least 50 % children and under 5 % old peo-
reduced mortality and fertility rates has created ple, we have witnessed increasing top-
markedly longer lives and altered the balance heaviness; the proportions of young and old are
between young and old in the population. now about equal. According to the 2014
Researchers face many challenges in mapping Population Reference Bureau (PRB) World
how new demographic circumstances have dra- Population Data Sheet, the more developed
matically affected micro-level life course pat- countries now, on a global basis, have 16 % of the
terns and interdependence in intergenerational population under the age of 15; 17 % 65 and
matrices. Such matrices need to be considered on older. These are also the average gures for
a societal level, across cohorts and age groups, as Europe. North America is a bit younger 19 %
well as in families, communities and social net- and 14 %. The oldest population in the world is
works. The paucity of appropriate data is part of the Japanese, where the gures are 13 % and 26
the problem facing researchers interested in % in other words there are twice as many old
social change and relations across generational people as children. Variations in the balance
boundaries. between young and old become more striking
In popular media and policy discussions, the when we consider gender differences in survival
demographic transition is almost uniformly seen to old age.
as a massive increase in the number of old peo- So far, population aging has entailed a femi-
plea big grey wave. The emphasis in schol- nization, due to an increasing gap between life
arly work is also on the older segment of a expectancies of men and women, which results in
changing population. In contrast, a few authors imbalanced sex ratios. In many old populations,
(e.g. Hagestad 2008; Uhlenberg 2009) have there are more than twice as many women as men
argued that demographic changes have altered over the age of 80. The gures for 2013 (PRB
the social worlds of children. Children now rep- 2014) show that in more developed countries, the
resent a much smaller proportion of the popula- average difference in life expectancy at birth is 7
tion than has been the case throughout human years. This is also the average gure for Europe.
history, reecting altered mortality and fertility. In the US, the number is now 5 years. Looking
Victor (2010) comments that in 1901, 40 % of all across Europe, the smallest gender difference is
deaths in England and Wales occurred among in Iceland, where women can expect to live 3
children aged 014, while 44 % were among years longer than men. In sharp contrast, six
individuals aged 65 and older. In 2008, the cor- countries in Eastern and Central Europe have a
responding gures were 1 % and 83 %. Uhlenberg difference of 10 years or more. Belarus, Lithuania
(1980) illustrated how experiences of family and Russia show a difference of 11 years.
deaths have undergone dramatic changes over the Thus, demographic conditions for interdepen-
last century. While losing at least one child was dence between old and young vary widely across
an expectable part of parenthood in the early societies. In 2005, there were 81 women aged 65
1900s, it is now so rare that parents experiencing and older per 100 children under 15 in Italy. The
such loss need national organizations to nd peer corresponding gure for old men was 57 per 100
support. Another change in the patterning of fam- children. The gures for the Russian Federation
ily deaths is that widowhood and the loss of par- were 62 women and 29 men per 100 children. It
ents have developed clearer timing patterns, a is reasonable to conclude that Russian children
trend that may increase potential support from grow up having highly limited contact with old
transition peers. men. Because of changes in life expectancy,
Structuration of the Life Course: Some Neglected Aspects 137

womens friendships and intergenerational ties structures are out of step with the new reality.
are typically characterized by co-longevity and The same argument could also be made using the
long durations. This fact has to a limited degree more traditional concept of cultural lag: shared
been recognized in the life course literature. expectations linked to age status have not kept
The magnitude and complexity of demo- pace with demographic change. As life expectan-
graphic change have not permeated scholarly or cies increased, also in old age, and actuarial pat-
policy discussions of how men and women spend terns became increasingly predictable, scholars
the last decades of life, studies of relationships and policymakers suggested a shift in emphasis
across age groups in society, or research on inter- from time lived (chronological age) to time left in
generational family ties. However, in his impor- dening rights (e.g., Sanderson and Scherbov
tant discussions of the institutionalized life 2007; Vaupel 2009). Interestingly, this distinction
course, Kohli (1986, 2007) emphasized the sig- was made by Neugarten in a 1968 pioneering
nicance of population change in the twentieth paper on awareness of middle age. She argued
century, suggesting that a new demographic sta- that between the ages of 40 and 60, individuals
bility and altered age structures lead to a greater become increasingly aware of nitude and start
emphasis on measurable time, central in the to think in terms of time left, rather than time
development of modern bureaucracies and essen- lived. Ryder (1975), trying to dene a threshold
tial to modern societys division of labor. Kohli for old age, also wrote about the time remaining
(1986) pointed out that during the rst half of the until death as a marker.
1900s, chronological age became increasingly Currently, the world average for the age at
signicant as a basis for marking transitions and which individuals can expect to live 15 more
assigning rights and duties. He stressed that years is 67 (Scherbov et al. 2014). By the end of
demographic stability is a major factor in the the century, the gure is estimated to be the age
emergence of the tripartite life course: a rst of 78. In Denmark, Greece, Italy, the Netherlands,
phase focused on preparation through schooling, Portugal, and Spain, there is now discussion of
a second centered on involvement in work and pension reform that links the statutory retirement
family building, and a third without work, a time age to remaining life expectancy (OECD 2014).
of retirement. These phases were also outlined by Recent analyses using cross-national European
Cain (1964). In addition to age, Kohli pointed to data show that a higher expected length of life
duration as a key dimension of social placement. increases the chances of working close to and
Indeed, some legislation (reected in the German past the ofcial retirement age (Brsch-Supan
concept of Dienstalter), assumed that duration of et al. 2009).
work (reected in the seniority principle) nearly In the US, several recent books take a cultural
coincided with chronological age. Such assump- perspective on the new old age, calling for revised
tions, however, were based on a male life course life maps (see Moen, current volume). They
(Moen 2001; ORand 1988; Srensen 1987). speak of becoming old as entering a new country,
Gender is conspicuously missing in path- starting with Hendricks (1980), Country of the
breaking German work on the institutionalized Old. In 1995, Smith titled her book Old Age is
life course. Another Country. Pipher (1999) used the same
metaphor. Recently, anthropologist Bateson
(2010) has urged older adults to compose a fur-
4.2 Life Maps Out of Step ther life, advocating new educational opportuni-
with Demographic Reality? ties for seniors who try to negotiate unmarked
life trails. In Europe, sociologist Rosenmayr, now
In their discussion of age stratication, Riley and age 90, takes a positive view of missing life
colleagues (1994) argued that in many aging scripts, emphasizing freedom (Rosenmayr 1983).
societies, we observe structural lag: demographic In his latest volume on the issue, published in
shifts have been so rapid and complex that social 2007, he discusses the philosophy of old age,
138 G.O. Hagestad and P.A. Dykstra

with a focus on creativity. Freedom and playful- known about age segregation in residential pat-
ness are also central themes in a rapidly growing terns, with the exception of old age ghettos,
international organization of old women, The such as retirement communities and care institu-
Red Hat Society (Van Bohemen et al. 2014). tions. Spatial segregation is now being addressed
Several authors argue that to create a meaningful by social geographers and city planners (e.g.,
old age, contact with children is essential. As Vanderbeck 2007) but has not been discussed by
Margaret Mead (1970) emphasized, children and a wider social science community. Cultural gaps
old people need to exchange teaching and learn- between age groups are often associated with the
ing, especially in societies undergoing rapid two rst types of segregation and are observed in
change. Such considerations bring us to broader contrasting language, dress, food, and music
perspectives on how demographic shifts, in com- preferences (Hagestad 2008).
bination with other societal change, have affected What are the consequences of age segrega-
relationships between age groups. tion? It makes contact and personal knowledge of
one another difcult for members of different age
groups, thus giving rise to ageism; it blocks
4.3 Altered Opportunities socialization across generational lines, and limits
for Relationships Among Age the spectrum of interpersonal resources in social
Groups: Increasing networks.
Segregation? Cross-age ties are an under-researched topic
(Riley and Riley 2000), perhaps because homo-
Before Kohlis account of the tripartite life geneity is a tenet of friendship research.
course, US social scientists had voiced concerns Homogeneity or homophily (Lazarsfeld and
that the new segmentation of life might lead to Merton 1954) refers to the tendency to form rela-
age segregation. Loand (1968) described col- tionships with others who are similar in some
leges as youth ghettos. Coleman (1961) wrote designated aspect such as age, sex, ethnic back-
about The Adolescent Society, expressing con- ground, and social class (McPherson et al. 2001).
cerns over the exclusion of children and youth As Mollenhorst et al. (2008) show, the age com-
from places of work; adults are away much of position of the contexts in which people meet is
the day in work spaces where there are no chil- an important determinant of the likelihood that
dren and no old people. Later, Coleman (1982) cross-age ties develop. In their network study, the
argued that age segregation deprives the young of mean age difference between condants who met
a proper view of mid-life and old age, and pro- each other at school, in leisure activities, or via
duces adults who have little understanding of the friends was about 9 years, but the mean age dif-
young. A recent US study based on successive ference between those who became acquainted
waves of the General Social Survey shows that via family members was almost 16 years. Few
such concerns are warranted: young adults have authors have addressed the question of whether
become increasingly isolated from older age there are gender differences in the likelihood of
groups outside the family (Smith et al. 2014). having cross-age friendships. A common practice
More than two decades after Colemans warn- is to control for gender in analyses of age
ings, Hagestad and Uhlenberg (2005, 2006) homogeneity in networks rather than to explicitly
raised concerns about three types of age segrega- address gender differences (e.g., Mollenhorst
tion: institutional, spatial and cultural. et al. 2008). Most research shows that women
Institutional segregation is created by policies have larger and more diversely composed net-
that block interdependence between young and works than men (Antonucci 2001), but patterns
old by allocating issues relating to the two age are not always clear cut: gender differences in
groups to separate ministries and administrative personal networks vary by life stage, socioeco-
units. In the spheres of education and leisure, one nomic status and marital history (e.g., Ajrouch
of the age groups is typically excluded. Little is et al. 2005).
Structuration of the Life Course: Some Neglected Aspects 139

Scholars concerned about societal age segre- children with all four biological grandparents
gation and lack of contact between young and old alive. According to Uhlenbergs (1996) estimates,
are often told that such contact is a key function 5 % of 10-year-olds in the US at the start of the
of the family, where members of different age 1900s had four grandparents; by 2005, the gure
groups and cohorts meet, form durable ties and was 40 %. This last gure is very similar to what
engage in reciprocal socialization. emerged from the 2005 Norwegian grandparent-
hood study, a survey anchored in children aged
1012: 41 % had all four (Hagestad 2006).
4.4 Altered Opportunities Microsimulation models reveal an increase in the
for Interdependence Across proportion of 020 year olds with four surviving
Family Generations grandparents in the Netherlands from 10 % in
1950 to 20 % in 1990 (Post et al. 1997).
To describe intergenerational family networks, The availability of grandparents is not limited to
one cannot rely on standard demographic mea- childhood. Estimates for the US (Uhlenberg 1996)
sures such as fertility rates, life expectancy, show that the proportion of 30 year-olds with a
dependency ratios and household composition grandparent alive more than tripled between 1900
(Connidis 2014; Herlofson and Hagestad 2011). and 2000, from 21 % to 75 %. The expectation for
An overview of intergenerational ties requires 2020 is a further increase to 80 %. Uhlenberg pres-
careful attention to anchoring (Herlofson 2013): ents a powerful example of historical change:
whose family network are we describing? 20-year-olds in 2000 were more likely to have a
Uhlenberg (2004) offers an example: decreases grandmother still living (91 %) than 20-year-olds
in adult mortality have little relevance for the in 1900 were to have a mother living (83 %).
availability of grandchildren among mature Studies of intergenerational ties focus on the
adults. Any person who survives to an advanced matrifocal tilt and women as kin-keepers, but
ageregardless of time periodis likely to have such descriptions of womens involvement in
an adult grandchild. However, decreases in adult intergenerational relationships have mostly
mortality are particularly relevant from the per- emphasized cultural constructions of family
spective of grandchildren, because there is an roles, not the demographic basis of gender pat-
increase in the supply of grandparents and terns. An example is repeated reports that mater-
expanded potential for young adults to develop nal grandmothers are the most involved
adult relationships with grandparents. grandparents. Such accounts often neglect the
Many modern societies have myths about the fact that these women typically become grand-
good old days, when children practically grew up parents at the youngest age (reecting timing of
on grandmas lap. Demographers have an alter- parenthood in two generations), are the healthi-
native story about the absence of grandparents in est, and can expect the longest duration of role
the rst part of the twentieth century, especially occupancy. Work on the availability of vertical
for the younger children in large sibships. kin typically uses non-gendered terms, such as
Reduced fertility has also led to clearer sequenc- grandparents and great-grandparents (e.g.,
ing between active parenting and grandparenting, Grundy et al. 1999; Puur et al. 2011), concealing
especially among women, resulting in less com- the fact that women are heavily over-represented
petition between the parent and grandparent in the older generations. Relevant to life course
roles. Uhlenberg (2009) points to one more factor research is the fact that the sequencing of paren-
affecting contact and closeness between grand- tal death and entry into the grandparent role dif-
parents and grandchildren: a reduction in the fers for men and women. Given gender differences
number of grandchild sets. What might be called in age at marriage and life expectancy, men are
a revolution in grandparenthood reects the main more likely to lose fathers before a new generation
drivers of the demographic transition. We have is added, whereas women typically become
seen a dramatic increase in the proportion of grandmothers while their mothers are still alive.
140 G.O. Hagestad and P.A. Dykstra

Cross-national differences in the generational and Komter 2012). For example, research on
structure of families are not easily determined, labor force exit has benetted from looking at the
because the joint effect of demographic trends on retirement decision in a multigenerational per-
family units is not always obvious. An example is spective rather than solely focusing on the retir-
the countervailing effects of increased longevity ing generation. Van Bavel and De Winter (2013)
and postponed childbearing on the generational examined whether grandchild care might encour-
structure of families (Matthews and Sun 2006). age older workers to leave the labor force before
Declining adult mortality means that older family the ofcial retirement age. Their analyses
members are living longer than they did in the revealed that grandparenthood speeds up retire-
past, which increases the likelihood that three, ment, particularly for womenan illustration
four or even ve generations may be alive at the that interdependence plays a stronger role in
same time. In contrast, delayed childbearing womens lives than in mens. The nding is all
means that the age gap between generations is the more compelling, given policy efforts to
widening, which reduces the likelihood that mul- increase labor force participation in the context
tiple generations will be alive at the same time. A of an ageing Europe. Whereas having grandpar-
comparison between Hungary and the ents taking care of grandchildren enables the
Netherlands of the number of descending family middle generation to be gainfully employed, it
generations of people aged 70 and over illustrates suppresses the economic activity of the older
decelerated generational turnover in connection generation, and may constitute an expensive form
with late childbearing (Knipscheer et al. 2000). of childcare.
Since the 1970s, the age at parenthood has been The discovery of grandparents (Segalen
higher in the Netherlands than in Hungary. Not 2010) by fertility researchers is another example
surprisingly, a lower proportion of Dutch (24 %) of new insights gained from considering interde-
than of Hungarian older adults (39 %) reported pendence in the lives of young and old in fami-
four descending family generations, even though lies. A number of studies have shown that the
life expectancy is higher in the Netherlands. decision to have children is taken more readily
when support from grandparents is available
(e.g., Hank and Kreyenfeld 2003). Using Dutch
4.5 Family Generations: longitudinal data, Kaptijn and colleagues (2010)
Recognizing Multiple Links demonstrated that men and women who received
frequent childcare support from their parents
An abundant literature underscores the impor- were more likely to have a second or third child
tance of intergenerational family relationships in than offspring who received no such support.
shaping the life chances of the young, the middle- Based on longitudinal data from 11 European
aged, and the old. Nevertheless, this literature is countries, Aassve et al. (2011) concluded that the
made up of separate foci (Moen et al. 2014): positive effect of grandparental help on child-
parenting tends to be about young families, bearing was much greater in Mediterranean
intergenerational transmission typically countries than in western and northern Europe.
focuses on early and middle adulthood, whereas Research on grandparental care and fertility
informal care is about adult children and age- shows how levels of analysis matter. At the
ing parents. Studies of grandparenting are micro-level of individual lives there is a positive
exceptional in the sense that multiple generations association between grandparental help and
of family members are often considered simulta- childbearing. At the macro-level, however, the
neously: grandchildren, the middle generation, association is negative. Fertility rates tend to be
and grandparents. highest in countries with the most generous
A focus on both the young and the old in fami- public childcare facilities and parental leaves
lies points to forms of life course structuration (Castles 2003; Gornick and Hegewisch 2015;
that have been neglected in earlier work (Dykstra McDonald 2006), that is, in countries where
Structuration of the Life Course: Some Neglected Aspects 141

governments support the combination of parent- supports. These authors argue that the availability
ing duties and employment, thus reducing the or lack of intergenerational family relationships,
necessity of grandparental care. especially with grandparents, is a major factor in
With people living longer and reaching old the widening inequality among young people. In
age in better health (Vaupel 2010), grandparent- other words, cross-generational ties in the family
hood is becoming a more important part of the are a source of economic, cultural and social cap-
life course. In Norway, grandparenthood is being ital. Such concerns have also been raised in dis-
called lifes dessert. cussions of societal age segregation (Hagestad
and Uhlenberg 2005) because it is hard to nd
arenas in which young and old can meet outside
4.6 Limited Vertical Ties: the family realm.
Increased Life Course
Vulnerability?
5 Webs of Interdependent
What happens to the lives of individuals who do Lives: Micro and Macro
not t the picture presented abovethose with Perspectives
no or limited vertical family ties? An issue that is
of particular current interest is rising childless- 5.1 Two Faces of Interdependence
ness rates among men. Some authors, on both
sides of the Atlantic (e.g. Dykstra and Keizer When Elder introduced the concept of interde-
2009; Eggebeen and Uhlenberg 1985), are con- pendence in lives, he focused on family groups:
cerned about mens social integration, support individual lives are inuenced by what happens
through interdependent relationships, and invest- to other family members, whose circumstances
ment in their community, especially in the second are considered when making life course deci-
half of adulthood. North American social psy- sions. Elder et al. (2004) give the following
chological research based on Eriksons concept description: Lives are lived interdependently
of generativity, i.e. investment in younger gener- and socio-historical inuences are expressed
ations, indicates that the concern is warranted. through this network of shared relationships
McAdams and de St. Aubin (1992) found self- (p. 13). Family historians (Hareven 1982; Modell
reported generativity to be associated with par- 1989) argue that with the emergence of the insti-
enting for men but not for women. Compared to tutionalized life course, lives became less contin-
fathers, more childless men felt disconnected gent on conditions in the family realm. In a
from their communities and were not involved home-based economy, the production and repro-
in local organizations. A more recent study duction of the household took precedence over
(McKeering and Pakenham 2000) similarly the interests of its members. The transition to a
found parental generativity (time invested in care wage labor economy, as well as new educational
activities and psychological involvement in par- opportunities, set individuals free from the bonds
enting) more strongly related to societal genera- of the family of origin. Buchmann (1989) speaks
tivity for men than for women. In rural parts of of Freisetzung, a liberation, giving individuals
Europe, social services have difculties organiz- (especially young people) more opportunity to
ing care for old childless men because they are build their own adult lives. The liberation also
severely isolated and often live in remote areas brings uncertainties, as Beck (1992) has argued;
(e.g., Wenger 2009). when individuals are the architects of their own
Other authors (e.g. Esping-Andersen 2002; lives, they run the risk of being left with a sense
Heckman 2006; Srensen 2005) have raised con- of personal failure (see also Furlong and Cartmel
cerns about how children and young people with 1997). To what extent do modern youth actively
limited vertical ties (e.g. with single parents and plan their future? A recent analysis of data from
no available grandparents) nd life course 23 European nations (Hellevik and Settersten
142 G.O. Hagestad and P.A. Dykstra

2013) shows that both micro and macro levels practices (Grandits 2010; Leira 2002; Saraceno
need to be considered. Individuals with greater 2010). In many European societies, laws create
personal security were more likely to plan than or assume interdependence among lives, includ-
those who had fewer personal resources. Yet, ing legal stipulations of age and duration require-
young adults who lived in countries with less ments across family relations. Using Norway as
favorable societal conditions were more likely to an example, laws cross lives in a number of
plan than those in countries with more favorable ways:
conditions.
It is interesting to note that several of the As duties are tied to Bs age. Parents are
authors cited above seem to take a somewhat neg- nancially responsible until the child reaches
ative view of interdependence. Clearly, it is a the age of 18.
multi-faceted phenomenon, in that it represents As rights are tied to Bs age. Parents are enti-
rights, support, continuity and protection against tled to child support (public transfer) until the
risks, as well as obligations, vulnerabilities related child turns 18.
to events and resources of others, and transitions As rights are tied to Bs duration of role occu-
beyond a persons control. Anthropologist David pancy. To qualify for full paid parental leave,
Plath illustrated both faces of interdependence. In the father must have worked for a given num-
an analysis of a Japanese novel, he showed how a ber of months prior to the pregnancy, but he
young woman had her life on hold until her also must have a partner (the mother) who
older sister had made the transition into marriage meets these duration requirements.
(Plath 1980). Plath also wrote of how we need a As duration has negative effects on Bs dura-
convoy (Kahn and Antonucci 1980) of consoci- tion. The law provides a given duration of paid
ates (Schtz 1967), who can serve as co-biogra- care leave (for sick children or other family
phers (Ferrarotti 1981). members), but when the illness period exceeds
the stipulated amount of leave time, the care
provider may lose the continuous duration at
5.2 Interdependence work needed for full pension rights.
as a Policy Issue
In many countries, primogeniture still makes
Europeans often nd it paradoxical that the US rst-borns life progress dependent on their par-
quite possibly the most individualized country in ents dying or giving up the farm or rm. In some
the worldoffers many examples of the power countries, daughters and sons have equal rights;
of family interdependence. Part of this power in other societies, sons have rst rights, even
stems from the lack of state mechanisms for risk when they are not rst-born.
reduction. The rst author, while teaching in US Family responsibility laws dene clear rights
academic institutions, had students crying in her and duties across and within generations. Policies
ofce on numerous occasions because a parent and institutional arrangements may also block
had become unemployed or seriously ill, leaving interdependence, as for example when grandpar-
no money for tuition. ents are not granted the right to raise grandchil-
In all developed societies, the caring and dren when parents cannot provide adequate care,
nancial responsibilities for young and old fam- or when parents have court orders prohibiting
ily members are shared between families and the them from visiting their children after divorce.
state (Kohli et al. 2010), but countries differ How interdependence is shaped on a macro
greatly in their understanding of proper inter- level has not been systematically examined, but
generational family relations (Viazzo 2010). in many modern societies, and in many ways,
Laws dene rights and duties of family members laws and policies create contingent lives. We
towards each other, while policies (or their agree with Esping-Andersen (1997), who states
absence) reward or discourage particular family that lives and relationships must be seen within a
Structuration of the Life Course: Some Neglected Aspects 143

matrix of life-course policies: services, transfers (Saraceno and Keck 2008).1 The Mediterranean
to the old, care for children, support of parenting. countries have the most extensive regulations. In
In other words, we treat interdependence as a Italy, for example, grandparents, siblings, aunts
policy issue, with social psychological conse- and uncles are legally obliged to nancially sup-
quences. This is an underdeveloped domain of port children if their parents are not able to sup-
life course work, for the most part neglected in port them. Many Central European countries
European scholarship on the institutionalized life (e.g., Austria, Latvia) legally obligate grandpar-
course. Can we nd explicit policy efforts to ents to provide nancial support. Western and
shape interdependence by regulating and struc- Northern European countries (e.g., Sweden, the
turing marriage and parenthood, or intergenera- Netherlands, the United Kingdom) typically do
tional ties? To what extent do legal frameworks not legally oblige family members to support
assume, create, and reinforce interdependence children if their parents cannot provide for them.
among lives? Under what circumstances does In a number of countries, adult offspring are
legal regulation create continuity and security under legal obligations to nancially support par-
versus discontinuity and risk for individuals ents. In Italy such rules also hold for grandchil-
whose lives are interconnected? Is As risk Bs dren, as well as for sons-in-law and
security? daughters-in-law, but only if they are legally mar-
As sociology students, we were taught that ried (Saraceno and Keck 2008). The countries
most social roles are reciprocal: Egos rights are that have no legal obligations for adult children to
Alters duty. From this premise, we would con- nancially support their parents tend be in
clude that Egos rights face Alter with prescrip- Northern and Western Europe, but there are
tions. However, we can nd a number of exceptions (Belgium, the Netherlands, France,
examples, especially in relationships between and Germany). The countries that legally oblige
minor children and parents, of how this is not children to provide for their parents tend to be in
always the case. Following divorce that denes Southern, Central and Eastern Europe, but again,
one custodial parent, the non-custodial parent has there are exceptions to this pattern (Hungary,
legal visitation rights. However, these rights typi- Estonia, Bulgaria, and the Czech Republic).
cally take the form of permission. As a conse- Bordone et al. (2012) empirically illustrate
quence, children do not have a right to contact how policy arrangements structure generational
with a non-residential parent. Similar examples interdependence across three generations.
of relational asymmetry are found in states that Combining data from the Survey of Health and
grant grandparents visitation rights. Retirement in Europe (SHARE) with data from
the Multilinks Database on Intergenerational
Policy Indicators,2 they examined the likelihood
5.3 Examples of Laws and Policies that grandparents care for the children of an
Structuring Interdependence employed daughter on a daily basis. Findings
show that grandparents are most likely to be daily
Legal obligations to provide nancial support or caregivers in countries where public childcare
care to family members can be viewed as man- services and parental leaves are least generous
dated interdependence. A power of attorney to (Italy, Greece, Spain, and Poland). They are least
act on behalf of an older person deemed legally likely to care for grandchildren on a daily basis in
unt to make independent decisions, or having to countries that score the best in terms of childcare
accept the authority of parents and guardians, are services (e.g., Belgium), parental leave (e.g., the
other examples of mandated interdependence.
European nations vary widely regarding the 1
See the Multilinks Database on intergenerational Policy
range of family members included in civil Indicators for details. http://multilinks-database.wzb.eu/
laws regulating maintenance responsibilities 2
Ibid
144 G.O. Hagestad and P.A. Dykstra

Czech Republic), or both types of arrangements levels of child involvement by the father after the
(e.g., Denmark). Tobo (2007) argues that grand- introduction of daddy quota, particularly for
parental care in Southern European countries is highly educated men. Herlofson and Ugreninov
part of an effort to improve the life chances of the (2014) report that Norwegian men are more
middle generation. Paradoxically, she notes, involved in childcare after the introduction of the
Spanish grandmothers assume an old-fashioned daddy quota, but not more involved in care for
role to enable their daughters to adopt modern frail parents. Apparently, the policy reform does
gender roles. Grandparental care in Southern not make men generally more caring. Looking
Europe is a clear example of what Leisering after children seems to result in such a depletion
(2004) would call negative life course policy of mens care resources that little is left for the
shaping interdependence between family older generation.
generations.
An example of what Leisering would label
positive life course policy, aiming to shape the 5.4 Intergenerational Care
life course by explicit intervention, can be found Regimes
in parental leave policies, especially leaves for
fathers. Here, the Nordic countries were pio- Rather than focus on individual laws and poli-
neers. In line with Leiserings view, Swedish cies, some scholars have attempted to create
sociologist Therborn (1989) has argued that the models of care regimes, including both care for
Nordic welfare state is based on the assumption the young and the old (e.g., Anttonen and Sipil
that policies can indeed lead to personal change, 1996; Bettio and Plantenga 2004; Daly and Lewis
e.g. create caring fathers and egalitarian partners! 2000; Korpi 2000; Leitner 2003; Sainsbury
Iceland, Norway, and Sweden and, most recently, 1999). An attractive feature of these efforts to
Germany and Portugal (Moss 2014) have intro- map intergenerational care regimes is that they
duced a daddy quota: weeks of parental leave overcome a chopped up view of families by
exclusively reserved for fathers. Arnlaug Leira considering multiple generations. A recent exam-
(2000) highlighted the importance of non- ple is a model developed by Saraceno and Keck
transferable (use or lose) leave entitlements for (2010), who examine how legal and policy frame-
men, describing them as fatherhood by gentle works affect the degree to which country-specic
force. The expanding literature on the gendered institutional frameworks impose reliance on fam-
consequences of leave designs shows increases in ily members and/or support individual autonomy/
mens use of parental leave with the introduction agency. The rst pattern is familialism by default;
of such non-transferable daddy days situations where there are few or no publicly pro-
(Hegewisch and Gornick 2011). vided alternatives to family care and nancial
Has the special quota for fathers made men support. The second is supported familialism,
more caring? Kotsadam and Finseraas (2011) where there are policies, usually in the form of
would say the answer to this question is yes. nancial transfers and leaves, which support
They treated the implementation of the daddy families nancial and caring responsibilities.
quota in Norway as a natural experiment, and The third is defamilialisation, where needs are
compared parents with children born just after partly addressed through public provision (ser-
the reform to parents with children born just vices, income replacement). By identifying and
before the reform. Parents in the treatment measuring actual public provisions rather than
group were less likely to have conicts over the using ideal types of welfare regimes, Saraceno
division of household tasks, and more likely to and Keck capture the nuance that differentiates
share them. In their study of leave policies in countries.
Sweden, Norway, the Netherlands, Canada, An important issue is whether policies
Germany, the United Kingdom, Finland and Italy, involve payments for care, (paid) leaves, or the
Boll and colleagues (2014) found increased provision of care services (Javornik 2014).
Structuration of the Life Course: Some Neglected Aspects 145

When public support is offered in money rather 6 Structuring the Lives of Men
than in kind, families can use it to buy help or to and Women: Gender in Laws
augment the family budget while providing care and Policies
directly. This tradeoff might be different for
families in different socioeconomic circum- Although there is a massive literature on cultural
stances (cf. Gornick and Meyers 2008; Leitner constructions of gender, differential socialization
2003). The strategy of staying at home to pro- and role engagements, we have limited knowl-
vide care is more readily adopted by members edge of how societal laws and policies create dif-
of the working class (in practice: women). This ferent social landscapes and structural maps for
reduces their ability to remain in the labor force life trajectories of men and women.
and contributes to the likelihood of old-age pov- In what follows we explore how gender, in
erty for themselves. combination with age, is a foundation for assign-
Cross-national comparisons reveal that the ing rights and duties. We also examine gender
type of public provision offered has conse- differences in age boundaries for life phases or
quences for gender inequality. Using data from transitions. Our focus is on the EU-28 and OECD
the Survey of Health, Ageing, and Retirement in countries. Where it seems relevant, we expand
Europe (SHARE), Schmid et al. (2012) conrm our focus to other countries.
ndings from many studies that show that
women are more likely to provide intensive care
to aging parents than men. However, the imbal- 6.1 Roles Limited to One Gender
ance in the proportions of men and women pro-
viding such care is higher when aging parents Are there adult roles which have legal rights/
receive public supportin addition to the care duties tied to only one gender? Do we nd coun-
received from adult childrenin the form of tries where citizenship rights, such as suffrage,
cash for care payments than when they receive differ between men and women? Saudi Arabia is
public services (e.g., home help and home nurs- the only United Nations member state in which
ing). Apparently, the public provision of support women do not have the right to vote in national
services helps to keep both men and women elections (Inter-Parliamentary Union 2014).
involved in caring for frail parents, whereas care Nations differ widely in the introduction of wom-
payments are a greater incentive for women than ens suffrage. Early adopters of womens right to
for men. Abendroth and colleagues (2014) dem- vote were New Zealand (1893), Finland (1907),
onstrate the differential effect of cash benets, and Norway (1913). Late adopters are Switzerland
paid leaves, and child care services on womens (1971), Portugal (1976), and Bhutan (2008).3
employment. Using data from the European Military service has long been the domain of
Community Household Panel (ECHP), they men only. Of 33 nations surveyed (EU-28,
show that the motherhood occupational status Norway, Switzerland, the US, Canada and Israel),
penalty is lower in European countries with nine have a system of general conscription:
high expenditures on public childcare. Contrary Austria, Cyprus, Denmark, Estonia, Finland,
to expectations, they did not nd a higher pen- Greece, Israel, Norway and Switzerland (Central
alty in countries with high spending on family Intelligence Agency 2014). In six of these, only
cash benets. The authors argue that paid leaves men are subject to compulsory military service.
and public childcare prevent mothers from being Denmark, Israel and Norway have adopted
sidelined at critical career junctures, whereas conscription for women as well. In Denmark the
cash benets seem to maneuver women into the type of duties might differ, whereas in Israel the
mommy track. These two studies clearly dem- conscript service obligation has a shorter dura-
onstrate how policies (or their absence) shape tion for women.
interdependence within and across family
generations. 3
Source: http://www.ipu.org/wmn-e/suffrage.htm
146 G.O. Hagestad and P.A. Dykstra

Since it has been documented that women, which is a key element in a wide range of legal
across societies, are more likely to provide unpaid regulations (Katz et al. 1973). Examples are the
care than men, whereas men are more often gain- right to sign a contract and being subject to the
fully employed, it is important to ask whether juvenile or the adult criminal system. Utah state
rights to care leaves are differentiated by gender. law provided a clear example: up until 1975, this
Among the OECD countries, Switzerland is the state had a lower age of majority for women
only one with a statutory maternity leave, but no (Goldstein 1988). Parents were mandated to
leave for fathers (Family Database OECD 2014). nancially support men up to the age of 21,
The US is the only OECD member that has no women up to age 18. The assumption was that
statutory entitlement to any kind of parental women would marry earlier and be supported by
leave. Several countries (e.g., Austria, Croatia, their husbands.
Finland, France, Germany, Iceland, Italy, Our search revealed no gender differences in
Norway, Portugal, Sweden) have introduced a the age of majority in OECD countries, but the
daddy quota (a period of leave that is for the actual ages differ across countries. In almost all
exclusive use by fathers on a use-it-or-lose-it OECD countries, the age of majority is 18 years
basis), or a father bonus (a payment, tax break (Family Database OECD 2013). Exceptions are
or additional time away from work) to encourage Canada (19 in certain territories), Japan (20),
fathers to take parental leave (Moss 2014). It is Korea (19), New Zealand (20), and the United
important to note that the design of leave polices States (19 in certain states). Some countries in
differs considerably across countries in terms of the Middle East stand out. Iran and Saudi Arabia
length, level of wage replacement, the exibility have 8 as the age of majority for girls and 14 for
for taking leave, and rules governing fathers boys (OECD/CAWTAR 2014). In Ireland,
access to leave and/or the distribution of leave Portugal and Slovenia, a person can reach the age
between parents (Ray et al. 2010). In Iceland, of majority below the legally dened age if he or
Norway and Sweden, uptake of paternal leave is she marries (Family Database OECD 2013).
mandatory if the full paid parental leave is to be The marriageable age is not necessarily iden-
granted. tical to the age of majority. Again we nd virtu-
An expanding number of developed countries ally no gender differences (Family
offer leave entitlements to care for a wider range Database OECD 2013) in OECD countries. The
of family members (Moss 2014). Conditions for marriageable age is mostly set at 18, and is the
taking leave vary from relatively common sick- same for both men and women. An exception is
ness to critical illness or severe disability. Length, Luxembourg (18 for men and 16 for women). In
payment and other dimensions of leave also vary most member states, persons can marry before
considerably. However, even though descriptions the marriageable age, normally at 16, subject to
of the policies are gender neutral, using terms parental consent or permission from the courts
such as employees and family members, under special circumstances, such as pregnancy.
men are far less likely to make use of such leaves Legislators must strike a fair and reasonable
than are women, particularly if the leaves are balance between adolescents right to be pro-
unpaid (Moss 2014). tected from unwanted sex and the freedom to
engage in self-determined sexual relationships
through age of consent. Denitions of age below
6.2 Gender Differences in Age which all sexual contact was illegal were rst
Borders introduced approximately 200 years ago
(Graupner 2002). At the time the limits were set
Do men and women have different markers around the ages of 1012. By the 1950s, legisla-
between life phases? Do rights/duties tied to age tors started to raise the age of consent. In the ear-
differ for men and women? We start by examin- liest laws, the minimum age limits covered
ing gender differences in the age of majority, vaginal intercourse only. Recently, minimum age
Structuration of the Life Course: Some Neglected Aspects 147

limits have been extended to lesbian and gay rela- time. This logic clearly assumes that women are
tionships. Traditionally, girls were seen as more part of a couple as they near retirement. They
vulnerable, so the age of consent was higher for tend to marry older men, so having an earlier
girls than for boys. A 2002 overview showed that retirement age for her implies that husband and
Estonia, Cyprus and Scotland still had different wife can leave the labor market at approximately
age limits for girls and boys (Graupner 2002). the same time. Ginn and Arber (1995) view the
Current legislation shows no gender difference in earlier retirement age as a means to maintain
the age of consent in OECD countries (Family male nancial dominance, by avoiding the pos-
Database OECD 2013). The minimum age typi- sibility that still-working wives have higher
cally varies between 14 and 16. incomes than retired husbands. Another argu-
Public pensions are a set of policy arrange- ment is that women experience greater difculty
ments that have assumed gendered and partnered nding a new job once they pass a certain age.
life courses, with wives being primarily responsi- Arber and Ginn (1995) note that the lower age of
ble for homemaking and family care, and hus- pension eligibility may have reinforced the wide-
bands primarily responsible for generating income. spread prejudice that women age earlier than
The differential pensionable age ts this bread- men, or that physical signs of aging are more det-
winner model. In 2012, 22 EU and OECD coun- rimental for women (the so-called double stan-
tries had a lower statutory retirement age (age at dard of aging) who become unt for work in
which people gain the right to public old-age pen- their 50s. This might especially apply to work in
sion benets) for women than men, but this num- service industries, or other occupations in which
ber is decreasing (European Commission 2014; physical appearance is important. The lower pen-
OECD 2012, 2014). European countries are cur- sionable age is particularly problematic for
rently in ux about whether the statutory retire- divorced women, because it limits their options
ment age also implies a mandated exit from the to accrue a late-life income of their own.
labor market (Marin 2013). Increasingly, countries Womens permission or prescription for ear-
are enabling employees to work longer. The US is lier labor force exit and public pension benets
rather special with its 1976 Supreme Court ruling might be viewed as a privilege, but as Marin
that mandatory retirement is unconstitutional (2010) points out, an earlier retirement age is
(Abramson 1977). Pension reforms are leading to only an advantage in a pension system that pro-
an increasing equalization of retirement ages of vides women who retire early a higher rate of
men and women. After 2020, only Bulgaria, Chile, return on contributions, so that their pension
Israel, Romania and Turkey will continue to have income matches that of men who retire at a later
different ages for men and women, given what is age. With the move from dened benet pension
known about plans for reforms. Though the equal- systems to dened contribution systems that is
ization of pensionable ages is reasonable, given taking place across all advanced economies,
womens longer life expectancy (Sundn 2010), women are losing the nancial advantages of
women with histories of interrupted employment their privilege to retire early. A shorter qualify-
will have difculty meeting the requirements for ing period is likely to lead to lower pension
full pension benets because they are often forced income and increased poverty risksparticularly
to retire (Marin 2013). It appears that pension for women who cannot rely on a pension shared
reforms are assuming a further masculinization of with a husband. Chlo-Domiczak (2013) points
womens life courses. out that a lower retirement age for women might
A variety of reasons are provided for having lead to discrimination by employers, who run
an earlier retirement age for women (Brocas et al. less risk of losing personnel with a 50 year old
1990). One is that it is a compensation for the man who is probably interested in working at
double shift in many womens lives: combin- least 15 more years than with a 50 year old
ing housework, family care and a paid job. woman who might leave after 10 years. Perhaps
Another is that the couple can retire at the same the years left principle, currently under
148 G.O. Hagestad and P.A. Dykstra

consideration in the context of pension reform, able to provide care to their spouses. Another
will imply that women run less risk of discrimi- explanation is that the men more strongly feel
nation by employers because they must work they are entitled to public support because they
longer. perceive themselves as lacking the necessary car-
As yet, it is unclear what the income conse- ing skills. The gap between de jure and de facto
quences will be of a pensionable age linked to practices represents a major challenge for social
developments in life expectancy. Will women be scientists with an interest in societal structuring
penalized because they live longer? A critical and of her and his adulthood.
gender-sensitive view is needed of the hidden and
implicit assumptions underlying recent reforms.
Do they build on a typical male life course with a 6.3 Gender Differences in Credits
long, uninterrupted work history? Do they build for Role Engagement
on breadwinner assumptions?
Taking the previously described laws and poli- Above, we focused on legislation structuring role
cies together, our conclusion is that there is a con- entry and exit. Gendered life courses also serve
vergence between her and his age boundaries, as the basis for receiving publicly funded benets
rights and duties. Political pressure, leading to through duration requirements. Are there differ-
the Convention on the Elimination of All Forms ences in duration credits for mens and wom-
of Discrimination against Women (CEDAW), has ens role engagement in terms of eligibility for
undoubtedly fuelled this development. The unemployment benets or pensions?
CEDAW is an international treaty adopted in In many European countries, women may
1979 by the United Nations General Assembly. claim pension benets as mothers and as family
Described as an international bill of rights for care providers. They receive credits in recogni-
women, it came into force in 1981 and has been tion of the unpaid work of child rearing and fam-
ratied by 188 of the 193 UN member states. In ily care. Care credits, by acknowledging the time
its 30 articles, the Convention explicitly denes invested in childrearing and looking after depen-
discrimination against women and sets up an dent relatives, are not based on the norm of an
agenda for national action to end such discrimi- uninterrupted work life until retirement. However
nation.4 Our overview has briey touched upon care leaves are the only absences from work
differences between de jure and de facto prac- where xed at rates are sometimes applied (i.e.
tices (e.g., marriageable age, take-up of care predetermined amount)rather than the contrib-
leaves). Societies have not yet bridged the gap utory social insurance principle (i.e. based on job
between legislation aimed at achieving gender history) that prevails in the more male social
equality and established patterns of everyday security arrangements of unemployment, health
lives of men and women. One issue requiring or accident insurance (Marin 2010). Flat-rated
attention is gender-bias in the implementation of benets generally have advantages for less quali-
policies. For example, a recent Dutch study ed and less paid women workers but are detri-
revealed that frail older women living with a part- mental to skilled and well-remunerated women.
ner were more likely to receive publicly funded The more strides women make in the world of
home help than frail older men living with a part- paid work, the greater the gaps between earnings-
nereven though their circumstances were quite related and at-rate pension credits will be.
comparable (Schenk et al. 2014). The authors Countries that have residence-based minimum
suggest that the public servants processing the pensions (e.g., Iceland, the Netherlands and
home help requests perceive older men as less Norway) have a guaranteed retirement
income based on years of residence and has no
4
For more information, visit http://www.ohchr.org/en/ contributory payments or means-testing. Such
hrbodies/cedaw/pages/cedawindex.aspx minimum pensions are favorable to women
Structuration of the Life Course: Some Neglected Aspects 149

because they are not based on employment his- pensions were being paid to surviving ex-wives
tory (Marin 2010). of older colleagues than to retired members who
Care credits are a source of debate between had contributed to the pension schemes.
care feminists, who call for greater recognition
of womens distinct contributions as caregivers
and employment feminists, who feel that many 7 Returning to the Point
women would benet from stronger (not weaker) of Departure
ties to paid work (Ray et al. 2010). The latter
point to disincentives to work and reinforcement In the opening of the chapter, we drew some con-
of traditional assumptions about gender roles, trasts between foundational North American and
particularly when care credits are only awarded European perspectives on the life course, arguing
to women or only to men if women waive their that the former often emphasized the impact of
rights (Expert Group on Gender Equality 2011). micro- and meso-levels of social context, such as
Whereas men receive pensions largely as a families, social networks and communities, on
result of their own employment history, women individual life trajectories. A number of early
as wives, divorcees, and widowsare more North American scholars had a strong connection
likely to be entitled to derived benets, such as to the Chicago school of sociology. This tradi-
survivors benets and benets from pension tion emphasized shared meaning and its creation,
sharing (Lewis 1997). Though benets for the exemplied in Thomas concept, denition of
widowed tend to be couched in gender-neutral the situation. In contrast, European classics
terms, the differences in life expectancy between focused on the macro-level, building especially
men and women (and men marrying younger on the works by Max Weber. They underscored
women) imply that women are the most likely how social institutions shape lives. Interestingly,
recipients. In most European countries, divorcees while the early North American work recognized
are entitled to survivors benets if they have contrasts between her and his lives, Europeans
received maintenance payments from the ex- initially had very little discussion of gender.
spouse and have not remarried (MISSOC 2014). Today, it is very clear that if we want to under-
Systems of old-age assistance based on derived stand contemporary structuring of mens and
rights and marital status make gainful employ- womens lives, we need to build on both macro-
ment less rewarding for women, and lock them and micro perspectives. Silverstein and Giarrusso
into domestic work or into work in the informal (2011) sum it up nicely: Micro-interactions in
sector (Sundn 2010). the family may be shaped by the political econo-
Marin (2010) points out that survivors bene- mies and cultures within which those interactions
ts do not redistribute to women per se, but rather are embedded, specically the way in which wel-
to ever-married couples. Never-married men and fare production is allocated among state, market
never-married women subsidize single breadwin- and family (p. 39). In our chapter, we have
ner families and homemakers in particular. The focused on some structural factors that have not
most conspicuous example of such subsidization been adequately considered in work on gendered
is the granting of generous survivors pension life courses and interdependence among lives.
rights to as many women a man might have
wished to marry over the course of his life
without any cost-sharing on his part. This policy 7.1 Neglected Structural Factors
of what Marin (2010, p. 216) calls state subsi-
dized serial monogamy was apparently quite The rst area of neglect is demographic change:
widespread in the civil service and occupational increasing longevity, combined with reduced fer-
corporatist pension regimes in Central European tility, is creating new late life potential and new
countries. Cut backs were introduced only when opportunities for intergenerational connections.
young women professionals reported that more We argue that crucial insights into life course
150 G.O. Hagestad and P.A. Dykstra

structuring in the family context would be gained underlying mechanisms? Do changing economic,
if scholars adopt a multigenerational focus, mov- political and legal contexts inuence beliefs about
ing beyond the current main foci: couples raising what is possible, desirable, and normal? Do soci-
children, and adult offspring caring for frail older etal conditions take on personal meaning only
parents. Demographic shifts are also increasing when they get translated into family situations
differences between mens and womens lives. and family meaning systems? To what extent are
Since women live longer, they have greater individuals aware of societal forces that have
opportunities for longstanding relationships shaped their choices and behaviors? If they are
across family generations and with age peers. part of a clear cohort pattern, are they aware of it?
The second area of neglect is the role of laws Do we end up telling different stories if we aggre-
and policies in structuring interdependence gate from individual biographical accounts or
among lives and the shaping of gendered lives. examine macro-level societal patterns?
Overall, we nd gender convergence in life struc- In order to begin to address the questions
turing by laws and policies. Yet, we also observe sketched above, we need dialogue and collabora-
strong contrasts between how men and women tion between research communities on two conti-
actually live their lives. nentsone emphasizing culture and shared
Levy, using a concept developed by E. Hughes, meanings on a meso-level; the other stressing
argues that men and women have different mas- macro-level structural conditions. We also need
ter statuses, locating them differently in the to overcome another continental dividethe
worlds of family and work (Levy 2013a, b; chasm between qualitative and quantitative meth-
Krger and Levy 2001). His perspective reects odologies. As Levy (2013a) comments, the dif-
Lintons (1942) and Parsons (1942) discussions ference between a subjectivist biographical
of roles based on age and sex. The master status and a factual life history is often translated into
implies that participation in other roles may be a radical divide.
developed only insofar that it does not interfere To arrive at a better understanding of macro
with the primary responsibility. Thus, mens mesomicro links, we feel that a combination of
involvement in family tasks is secondary to methodological approaches is the route to follow.
breadwinner obligations, womens employment Mixed-methods are often espoused, but not often
is subsidiary to the requirements of their caring practiced in life course research. Scholars prac-
roles. Recently, researchers have shown that such ticing quantitative and qualitative approaches
potential role conict is not limited to mothers of seem to live in separate worlds. It is unusual to
young children, but increasingly also to women nd publications that demonstrate the comple-
in the next generation: grandmothers who strug- mentarity of the two approaches. Rare examples
gle to maintain a work career as well as provide are Melinda Mills research on non-standard
the care for grandchildren (Meyer 2014). work schedules in the Netherlands (Mills and
Tht 2010), Laura Bernardis work (Bernardi
et al. 2007) on fertility in East and West Germany,
7.2 Analytical and Methodological and Helga Krgers study (Krger and Levy
Challenges in Bridging Levels 2001) on the employment careers of German
women. Interestingly, the authors started from
We have discussed Kotsadam and Finseraass substantive puzzles on the supposed inuences of
(2011) comparison of before and after daddy macro-level conditions that their quantitative
quota cohorts, which indeed shows that policy data were not able to solve. Mills, for example,
change results in altered lives. Yet, natural experi- aimed to unravel why Dutch non-standard work
ments like theirs amply illustrate the unanswered schedules did not have negative effects on Dutch
questions in research that attempts to connect familieswhich was completely contrary to
macro- and micro levels. How can we identify previous research based on American families.
Structuration of the Life Course: Some Neglected Aspects 151

She returned to her respondents whose biograph- Bateson, M. C. (2010). Composing a further life: The age
of active wisdom. New York: Vintage Books.
ical accounts revealed that Dutch couples volun-
Beck, U. (1992). Risk society: Towards a new modernity
tarily choose non-standard work schedules so (M. Ritter, Trans.). London: Sage. (Originally pub-
they can spend more time with their families. We lished as Risikogesellschaft: Auf dem Weg in eine
feel that life course researchers should emulate andere Moderne. Frankfurt am Main: Suhrkamp
Verlag, 1986).
this kind of work. The ideal next step is to have a
Bernardi, L., Von der Lippe, H., & Keim, S. (2007). Social
community of researchers who work on recorded inuences on fertility: A comparative mixed methods
and observed life histories, with comparisons study in eastern and western Germany. Journal of
across time and societies. Mixed Method Research, 1(2347), 2007.
doi:10.1177/2345678906292238.
Bettio, F., & Plantenga, J. (2004). Comparing care regimes
Acknowledgments Financial support for work by the in Europe. Feminist Econonomics, 10, 85113. doi:10
second author on this chapter comes from the European .1080/1354570042000198245.
Research Council Advanced Investigator Grant (ERC, Billari, F. C., Goisis, A., Liefbroer, A. C., Settersten,
324211) Families in Context, and from the European R. A., Aassve, A., Hagestad, G., & Spder, Z. (2011).
Union Seventh Framework Large Scale Integrating Social age deadlines for the childbearing of women
Project (EC, 320116) FamiliesAndSocieties. and men. Human Reproduction, 26, 616622.
doi:10.1093/humrep/deq360.
Boll, C., Leppin, J., & Reich, N. (2014). Paternal child-
care and parental leave policies: Evidence from indus-
References trialized countries. Review of Economics of the
Household, 12(1), 129158. doi:10.1007/
Aassve, A., Meroni, E., & Pronzato, C. (2011). s11150-013-9211-z.
Grandparenting and childbearing in the extended Bordone, V., Arpino, B., & Aassve, A. (2012). Policy per-
family (Dondena Working Paper No. 38). Milan: spectives of grandparents in Europe (Dondena
University of Bocconi. Working Paper No. 51). Milan: University of Bocconi.
Abendroth, A. K., Huffman, M. L., & Treas, J. (2014). Brsch-Supan, A., Brugiavini, A., & Croda, E. (2009).
The parity penalty in life course perspective: The role of institutions and health in European pat-
Motherhood and occupational status in 13 European terns of work and retirement. Journal of European
countries. American Sociological Review, 79, 993 Social Policy, 19, 341358.
1014. doi:10.1177/0003122414545986. doi:10.1177/1350506809341515.
Abramson, L. W. (1977). Compulsory retirement, the con- Brocas, A., Cailloux, A., & Oget, V. (1990). Women and
stitution and the Murgia case. Missouri Law Review, social security: Progress towards equality of treat-
42, 2552. ment. Geneva: International Labor Ofce.
Ajrouch, K. J., Blandon, A., & Antonucci, T. C. (2005). Buchmann, M. (1989). The script of life in modern soci-
Social networks among men and women: The effects ety: Entry into adulthood in a changing world.
of age and socioeconomic status. Journal of Chicago: University of Chicago Press.
Gerontology: Social Sciences, 60, 311317. Cain, L. D., Jr. (1964). Life course and social structure. In
doi:10.1093/geronb/60.6.S311. R. E. L. Faris (Ed.), Handbook of modern sociology
Angrist, S. S., & Almquist, A. M. (1975). Careers and (pp. 272309). Chicago: Rand McNally.
contingencies. New York: Dunellen. Cain, L. D., Jr. (1976). Aging and the law. In R. H.
Antonucci, T. C. (2001). Social relations: An examination Binstock & E. Shanahas (Eds.), Handbook of ageing
of social networks, social support, and sense of con- and the social sciences (pp. 342368). New York: Van
trol. In J. E. Birren & K. W. Schaie (Eds.), Handbook Nostrand Reinhold.
of the psychology of aging (pp. 427453). San Diego: Cain, L. D., Jr. (2003). Age-related phenomena; The inter-
Academic Press. play of the ameliorative and the scientic. In R. A.
Anttonen, A., & Sipil, J. (1996). European social care Settersten Jr. (Ed.), Invitation to the life course:
services: Is it possible to identify models? Journal of Toward new understandings of later life (pp. 295
European Social Policy, 6, 87100. 325). Amityville: Baywood Publishing.
doi:10.1177/095892879600600201. Campbell, A., Converse, P., & Rogers, W. (1976). The
Anxo, D., Bosch, G., & Rubery, J. (Eds.). (2010). The quality of American life. New York: Russell Sage
welfare state and life transitions: A European perspec- Foundation.
tive. Cheltenham: Edward Elgar. Castles, F. G. (2003). The world turned upside down:
Arber, S., & Ginn, J. (1995). Connecting gender and age- Below replacement fertility, changing preferences and
ing: A new beginning? In S. Arber & J. Ginn (Eds.), family-friendly public policy in 21 OECD countries.
Connecting gender & ageing (pp. 173178). Journal of European Social Policy, 13, 209227.
Buckingham: Open University Press. doi:10.1177/09589287030133001.
152 G.O. Hagestad and P.A. Dykstra

Central Intelligence Agency. (2014). The world fact book. J. T. Mortimer & M. J. Shanahan (Eds.), Handbook of
https://www.cia.gov/library/publications/the-world- the life course (pp. 319). New York: Springer.
factbook/elds/2024.html. Accessed 4 July 2014. Elder, G. H., Jr., Shanahan, M. J., & Jennings, J. A. (2015).
Cho-Domiczak, A. (2013). Female transition to retire- Human development in time and place. In M. H.
ment (Neujobs working paper no. D16.2B). Retrieved Bornstein & T. Leventhal (Eds.), Handbook of child
from http://ibs.org.pl/projekty/les/neujobs/Female_ psychology and developmental science (Ecological set-
transition _ to_ retirement_Agnieszka_Chlon- tings and processes, Vol. 4, pp. 654). Hoboken: Wiley.
Dominczak.pdf Esping-Andersen, G. (1990). The three worlds of welfare
Cho, S., Crenshaw, K. W., & McCall, L. (2013). capitalism. Princeton: Princeton University Press.
Intersectionality: Theorizing power, empowering the- Esping-Andersen, G. (1997). The comparative political
ory. Signs, 38, 785810. economy dimension. Life course proles within differ-
Coleman, J. S. (1961). The adolescent society: The social ent welfare regimes. Paper prepared for the 1st
life of the teenager and its impact on education. International POLIS project workshop, European
Glencoe: The Free Press. University Institute, Florence.
Coleman, J. S. (1982). The assymetric society. Syracuse: Esping-Andersen, G. (2002). A child-centred social
Syracuse University Press. investment strategy. In G. Esping-Andersen, D. Gallie,
Connidis, I. A. (2014). Age relations and family ties over A. Hemerijck, & J. Myles (Eds.), Why we need a new
the life course: Spanning the macromicro divide. welfare state (pp. 2667). Oxford: Oxford University
Research in Human Development, 11, 291308. doi:1 Press.
0.1080/15427609.2014.967050. European Commission. (2014). Adequacy and sustain-
Crenshaw, K. (1989). Demarginalizing the intersection of ability of pensions. Retrieved from http://ec.europa.eu/
race and sex: A black feminist critique of antidiscrimi- europe2020/pdf/themes/04_pensions_02.pdf
nation doctrine, feminist theory and antiracist politics. Expert Group on Gender Equality and Social Inclusion,
University of Chicago Legal Forum, 4, 139167. Health and Long-Term Care Issues (EGGSI). (2011).
Daly, M., & Lewis, J. (2000). The concept of social care The socio-economic impact of pension systems on the
and the analysis of contemporary welfare states. respective situations of women and men and the effects
British Journal of Sociology, 51, 281298. of recent trends in pension reforms. Retrieved from
Davis, K. (1945). The world demographic transition. http://ec.europa.eu/justice/gender-equality/ les/ equal_
Annals of the American Academy of Political and economic_independence/pensions_report_en.pdf
Social Science, 237, 111. Family Database OECD. (2013). PF 1.8: Legal age
Diewald, M. (2000). Continuities and breaks in occupa- thresholds regarding the transition from child- to
tional careers and subjective control: The case of the adulthood. Retrieved from http://www.oecd.org/els/
East German transformation. In J. Bynner & R. K. family/PF_1_8_Age_threshold_Childhood_to_
Silbereisen (Eds.), Adversity and challenge in life in Adulthood_Dec2013.pdf
the new Germany and in England (pp. 239267). Family Database OECD. (2014). PF 2.1 Key characteris-
London: Macmillan. tics of parental leave systems. Retrieved from http://
Durkheim, E. (1982). Rules of sociological method www.oecd.org/els/soc/PF2_1_Parental_leave_
(S. Lukes, Ed.; W. D. Halls, Trans). New York: Free systems_1May2014.pdf. Accessed 4 July 2014.
Press. (Originally published in 1895 as Les rgles de Ferrarotti, F. (1981). On the autonomy of the biographical
la mthode sociologique). method. In D. Bertaux (Ed.), Biography and society:
Dykstra, P. A., & Keizer, R. (2009). The well-being of The life history approach in the social sciences
childless men and fathers in mid-life. Ageing & (pp. 1927). Beverly Hills: Sage.
Society, 29, 12271242. doi:10.1017/ Fields-Smith, C. (2005). African American parents before
S0144686X08008374. and after Brown. Journal of Curriculum and
Dykstra, P. A., & Komter, A. E. (2012). Generational Supervision, 20, 129135.
interdependencies in families: The MULTILINKS Fry, C. L. (2003). The life course as a cultural construct.
research programme. Demographic Research, 27, In R. A. Settersten Jr. (Ed.), Invitation to the life
487506. doi:10.4054/DemRes.2012.27.18. course: Toward new understanding of later life
Eggebeen, D. J., & Uhlenberg, P. R. (1985). Changes in (pp. 269294). Amityville: Baywood Publishing.
the organization of mens lives: 19601980. Family Furlong, A., & Cartmel, F. (1997). Young people and
Relations, 34, 251257. social change: Individualisation and risk in the age
Elder, G. H., Jr. (1974). Children of the great depression: of high modernity. Buckingham: Open University
Social change in life experience. Chicago: University Press.
of Chicago Press. Ginn, J., & Arber, S. (1995). Only connect: Gender rela-
Elder, G. H., Jr. (1979). Historical change in life patterns tions and ageing. In S. Arber & J. Ginn (Eds.),
and personality. In P. Baltes & O. Brim Jr. (Eds.), Life- Connecting gender & ageing (pp. 114). Buckingham:
span development and behavior (Vol. 2, pp. 117159). Open University Press.
New York: Academic Press. Goldstein, L. F. (1988). The constitutional rights of
Elder, G. H., Jr., Johnson, M. K., & Crosnoe, R. (2004). The women: Cases in law and social change. Madison:
emergence and development of life course theory. In University of Wisconsin Press.
Structuration of the Life Course: Some Neglected Aspects 153

Gornick, J. C., & Hegewisch, A. (2015). Gender, employ- Hendricks, J. (Ed.). (1980). The country of the old.
ment, and parenthood: The consequences of work- Farmingdale: Baywood.
family policies. In D. Keleman (Ed.), Lessons from Herlofson, K. (2013). How gender and generation matter:
Europe? What Americans can learn from European Examples from research on divorced parents and adult
public policies (pp. 1741). Los Angeles: Sage. children. Families, Relationships and Societies, 2,
Gornick, J. C., & Meyers, M. K. (2008). Creating gender 4562. doi:10.1332/204674313X664699.
egalitarian societies: An agenda for reform. Politics & Herlofson, K., & Hagestad, G. O. (2011). Challenges in
Society, 36, 313349. doi:10.1177/00323292083 20562. moving from macro to micro: Population and family
Grandits, H. (2010). Introduction: The reshaping of fam- structures in ageing societies. Demographic Research,
ily and kin relations in European welfare systems. In 25, 337370. doi:10.4054/DemRes.2011.25.10.
H. Grandits (Ed.), Family, kinship and state in contem- Herlofson, K., & Ugreninov, E. (2014). Er omsorgsfulle
porary Europe (The century of welfare: Eight coun- fedre omsorgsfulle snner? Likestilling hjemme og
tries, Vol. 1, pp. 2346). Frankfurt: Campus Verlag. hjelp til eldre foreldre [Are caring fathers caring sons?
Graupner, H. (2002). Sexual consent: The criminal law in Gender equality at home and help to ageing parents].
Europe and overseas. Keynote-Lecture presented at Tidsskrift for Samfunnsforskning, 55, 322346.
the 7th International Conference of the International Hernes, H. M. (1987). Welfare state and women power:
Association for the Treatment of Sexual Offenders, Essays in state feminism. Oslo: Norwegian University
Vienna, Austria. Retrieved from http://members.aon. Press.
at/graupner/documents/Graupner-paper-kn-oFN.pdf Inter-Parliamentary Union. (2014). Women in parliament
Grundy, E., Murphy, M., & Shelton, N. (1999). Looking 2013. The year in review. Retrieved from http://www.
beyond the household: Intergenerational perspectives ipu.org/pdf/publications/WIP2013-e.pdf
on living kin and contacts with kin in Great Britain. Javornik, J. (2014). Measuring state de-familialism:
Population Trends, 97, 1927. Contesting post-socialist exceptionalism. Journal of
Hagestad, G. O. (2006). Transfers between grandparents European Social Policy, 24, 240257.
and grandchildren: The importance of taking a three- doi:10.1177/0958928714525815.
generation perspective. Zeitschrift fr Kahn, R. L., & Antonucci, T. C. (1980). Convoys over the
Familienforschung, 18, 315332. life course: Attachment, roles and social support. In
Hagestad, G. O. (2008). Changes in childrens age and P. B. Baltes & O. G. Brim (Eds.), Life-span develop-
generation mosaics: Challenges to research and policy. ment and behavior (pp. 253286). New York:
In A. Leira & C. Saraceno (Eds.), Childhood: Academic.
Changing contexts (Comparative Social Research Kaptijn, R., Thomse, F., Van Tilburg, T. G., & Liefbroer,
series, Vol. 25, pp. 113132). Bingley: Emerald. A. C. (2010). How grandparents matter: Support for
Hagestad, G. O., & Uhlenberg, P. (2005). The social sepa- the cooperative breeding hypothesis in a contemporary
ration of old and young: A root of ageism. Journal of Dutch population. Human Nature, 21, 393405.
Social Issues, 61, 343360. doi:10.1007/s12110-010-9098-9.
doi:10.1111/j.1540-4560.2005.00409.x. Katz, S. N., Schroeder, W. A., & Sidman, L. R. (1973).
Hagestad, G. O., & Uhlenberg, P. (2006). Should we be Emancipating our children: Coming of legal age in
concerned about age segregation? Some theoretical America. Family Law Quarterly, 7, 211241.
and empirical explorations. Research on Aging, 28, Knipscheer, K., Dykstra, P. A., Utasi, A., & Cxeh-
638653. doi:10.1177/0164027506291872. Szombathy, L. (2000). Ageing and the family. In
Hank, K., & Kreyenfeld, M. (2003). A multilevel analysis G. Beets & K. Miltenyi (Eds.), Population ageing in
of child care and womens fertility decisions in west- the Netherlands and Hungary: A European perspec-
ern Germany. Journal of Marriage and Family, 65, tive (pp. 181199). Amsterdam: Thela Thesis.
584596. Kohli, M. (1985). Die Institutionalisierung des
Hareven, T. K. (1982). Family time and industrial time. Lebenslaufs: Historische Befunde und theoretische
Cambridge, UK: Cambridge University Press. Argumente [The institutionalization of the life course:
Heckman, J. J. (2006). Skill formation and the economics Historical ndings and theoretical arguments]. Klner
of investing in disadvantaged children. Science, Zeitschrift fr Soziologie und Sozialpsychologie, 37,
312(5782), 19001902. doi:10.1126/ 129.
science.1128898. Kohli, M. (1986). The world we forgot: A historical
Hegewisch, A., & Gornick, J. C. (2011). The impact of review of the life course. In V. W. Marshall (Ed.),
work-family policies on womens employment: A Later life: The social psychology of ageing (pp. 271
review of research from OECD countries. Community, 303). Beverly Hills: Sage.
Work & Family, 14, 119138. doi:10.1080/13668803. Kohli, M. (2007). The institutionalization of the life
2011.571395. course: Looking back to look ahead. Research in
Hellevik, T., & Settersten, R. A., Jr. (2013). Life planning Human Development, 4, 253271.
among young adults in 23 European countries: The Kohli, M., Albertini, M., & Knemund, H. (2010).
effects of individual and country security. European Linkages among adult family generations: Evidence
Sociological Review, 29, 923938. doi:10.1093/esr/ from comparative survey research. In P. Heady &
jcs069. M. Kohli (Eds.), Family, kinship and state in
154 G.O. Hagestad and P.A. Dykstra

contemporary Europe (Perspectives on theory and and concealed aspects of parent involvement:
policy, Vol. 3, pp. 225248). Frankfurt: Campus Perspectives from African-American parents in the
Verlag. urban south. Journal of School Public Relations, 28,
Korpi, W. (2000). Faces of inequality: Gender, class, and 350379.
patterns of inequalities in different types of welfare Loand, J. (1968). The youth ghetto: A perspective on the
states. Social Politics, 7, 127191. doi:10.1093/ cities of youth around our large universities. Journal of
sp/7.2.127. Higher Education, 39, 121143.
Kotsadam, A., & Finseraas, H. (2011). The state inter- Marin, B. (2010). Gender equality, neutrality, specicity,
venes in the battle of the sexes: Causal effects of pater- and sensitivity And the ambivalence of benevolent
nity leave. Social Science Research, 40, 16111622. welfare paternalism. In B. Marin & E. Zlyomi (Eds.),
Krger, H., & Levy, R. (2001). Linking life courses, work Womens work and pensions: What is good, what is
and the family: Theorizing a not so visible nexus best? Designing gender-sensitive arrangements
between women and men. Canadian Journal of (pp. 203224). Farnham: Ashgate.
Sociology, 26, 145166. Marin, B. (2013). Welfare in an idle society? Reinventing
Lazarsfeld, P., & Merton, R. K. (1954). Friendship as a retirement, work, wealth, health, and welfare.
social process: A substantive and methodological Farnham: Ashgate.
analysis. In A. Theodore & C. H. Page (Eds.), Freedom Matthews, S. H., & Sun, R. (2006). Incidence of four-
and control in modern society (pp. 1866). New York: generation family lineages: Is timing of fertility or
Van Nostrand. mortality a better explanation? Journal of Gerontology:
Leira, A. (2000). Combining work and family: Nordic Social Sciences, 61, S99S106.
policy reforms in the 1990s. In T. P. Boje & A. Leira Mayer, K. U. (2003). The sociology of the life course and
(Eds.), Gender, welfare state and the market: Towards lifespan psychology: Diverging or converging path-
a new division of labor. New York: Routledge. ways? In U. M. Staudinger & U. Lindenberger (Eds.),
Leira, A. (2002). Working parents and the welfare state: Understanding human development: Dialogues with
Family change and policy reform in Scandinavia. lifespan psychology (pp. 463481). Dordrecht: Kluwer
New York: Cambridge University Press. Academic.
Leisering, L. (2004). Government and the life course. In Mayer, K. U. (2004). Whose lives? How history, society
J. T. Mortimer & M. J. Shanahan (Eds.), Handbook of and institutions dene and shape life courses. Research
the life course (pp. 205225). New York: Springer. in Human Development, 1, 161187. doi:10.1207/
Leiserling, L., & Leibfried, S. (1999). Times of poverty in s15427617rhd0103_3.
western welfare states. Cambridge, UK: Cambridge Mayer, K. U., & Mller, W. (1986). The state and the
University Press. structure of the life course. In A. B. Sorensen,
Leitner, S. (2003). Varieties of familialism: The caring F. Weinert, & L. R. Sherrod (Eds.), Human develop-
function of the family in comparative perspective. ment and the life course (pp. 217245). Hillsdale:
European Societies, 5, 353375. doi:10.1080/1461669 Lawrence Erlbaum.
032000127642. Mayer, K. U., & Schpin, U. (1989). The state and the
Levy, R. (2013a). Analysis of life courses: A theoretical life course. Annual Review of Sociology, 15, 187209.
sketch. In R. Levy & E. D. Widmer (Eds.), Gendered doi:10.1146/annurev.so.15.080189.001155.
life courses: Between standardization and individual- McAdams, D. P., & de St. Aubin, E. (1992). A theory of
ization. A European approach applied to Switzerland generativity and its assessment through self-report,
(pp. 1336). Zurich: LIT Verlag. behavioral acts, and narrative themes in autobiogra-
Levy, R. (2013b). Life course analysis: A eld of intersec- phy. Journal of Personality and Social Psychology, 62,
tions. In R. Levy & E. D. Widmer (Eds.), Gendered 10031015. doi:10.1037/0022-3514.62.6.1003.
life courses: Between standardization and individual- McDonald, P. (2006). Low fertility and the state: The ef-
ization. A European approach applied to Switzerland cacy of policy. Population and Development Review,
(pp. 315337). Zurich: LIT Verlag. 32, 485510.
Levy, R. & Widmer, E. D. (Eds.). (2013). Gendered life McKeering, H., & Pakenham, K. I. (2000). Gender and
courses: Between standardization and individualiza- generativity issues in parenting: Do fathers benet
tion. A European approach applied to Switzerland. more than mothers from involvement in child care
Zurich: LIT Verlag. activities? Sex Roles, 43, 459480.
Lewis, J. (1997). Gender and welfare regimes: Further McMullin, J. (1995). Theorizing age and gender relations.
thoughts. Social Politics, 4, 160177. doi:10.1093/ In S. Arber & J. Ginn (Eds.), Connecting gender and
sp/4.2.160. ageing: A sociological approach (pp. 3041). Bristol:
Lewis, J. (2002). Gender and welfare state change. Open University Press.
European Societies, 4, 331357. doi:10.1080/1461669 McPherson, M., Smith-Lovin, L., & Cook, J. M. (2001).
022000022324. Birds of a feather: Homophily in social networks.
Linton, R. A. (1942). Age and sex categories. American Annual Review of Sociology, 27, 415444. doi:10.1146/
Sociological Review, 7, 589603. annurev.soc.27.1.415.
Loder-Jackson, T. L., McKnight, A. N., Brooks, M., Mead, M. (1970). Culture and commitment: A study of the
McGrew, K., & Voltz, D. (2007). Unmasking subtle generation gap. New York: Doubleday.
Structuration of the Life Course: Some Neglected Aspects 155

Merton, R. K. (1979). The normative structure of science. OECD. (2012). OECD pensions outlook 2012. Paris:
In R. K. Merton (Ed.), The sociology of science: OECD Publishing.
Theoretical and empirical investigations (pp. 267 OECD. (2014). OECD pensions outlook 2014. Paris:
278). Chicago: University of Chicago Press (Originally OECD Publishing.
published as Merton, R. K. (1942). Science and tech- OECD/CAWTAR. (2014). Women in public life: Gender,
nology in a democratic order. Journal of Legal and law and policy in the Middle East and North Africa.
Political Sociology,1, 115126). Paris: OECD Publishing.
Meyer, M. H. (2014). Grandmothers at work: Juggling Orloff, A. S. (1993). Gender and the social rights of citi-
families and jobs. New York: New York University zenship: The comparative analysis of gender relations
Press. and welfare states. American Sociological Review, 58,
Mills, M., & Tht, K. (2010). Nonstandard work sched- 303328.
ules and partnership quality: Quantitative and qualita- Parsons, T. (1942). Age and sex in the social structure of
tive ndings. Journal of Marriage and Family, 72, the United States. American Sociological Review, 7,
860875. doi:10.1111/j.1741-3737.2010.00735.x. 604616.
MISSOC. (2014). European Commission. http://www. Parsons, T., & Bales, R. F. (1955). Family, socialization
missoc.org/MISSOC/INFORMATIONBASE/ and interaction process. Glencoe: Free Press.
COMPARATIVETABLES/MISSOCDATABASE/ Pipher, M. B. (1999). Another country: Navigating the
comparativeTablesSearchResultTree.jsp. Accessed 4 emotional terrain of our elders. New York: Riverhead
July 2014. Books.
Modell, J. (1989). Into ones own: From youth to adult- Plath, D. W. (1980). Long engagements: Maturity
hood in America, 19201975. Berkeley: University of in modern Japan. Stanford: Stanford University
California Press. Press.
Moen, P. (2001). The gendered life course. In R. Binstock Plath, D. W., & Ikeda, K. (1975). After coming of age:
& L. George (Eds.), Handbook of ageing and the Adult awareness of age norms. In T. R. Williams (Ed.),
social sciences (pp. 179190). San Diego: Academic Socialization and communication in primary groups
Press. (pp. 107124). The Hague: Mouton.
Moen, P., Lam, J., Melanie, N. G., & Jackson, M. N. G. Population Reference Bureau. (2014). 2014 world popu-
(2014). Aging families and the gendered life course lation data sheet. http://www.prb.org/pdf14/2014-
2014. In J. Treas, J. Scott, & M. Richards (Eds.), The world-population-data-sheet_eng.pdf. Accessed 12
Wiley Blackwell companion to the sociology of fami- Dec 2014.
lies (pp. 444463). New York: Wiley. Post, W., Van Imhoff, E., Dykstra, P. A., & Van Poppel, F.
Mollenhorst, G., Vlker, B., & Flap, H. (2008). Social (1997). Verwantschaps-netwerken in Nederland:
contexts and core discussion networks: Using a Verleden, heden en toekomst [Kinship networks in the
choice-constraint approach to study similarity in inti- Netherlands: Past, present and future] (NIDI Report
mate relationships. Social Forces, 86, 937965. no. 51). Den Haag: NIDI.
doi:10.1353/sof.0.0010. Puur, A., Sakkeus, L., Pldma, A., & Herm, A. (2011).
Moss, P. (Ed.). (2014). International review of leave poli- Intergenerational family constellations in contempo-
cies and research 2014. Retrieved from http://www. rary Europe: Evidence from the Generations and
leavenetwork.org/lp_and_r_reports/ Gender Survey. Demographic Research, 25, 135172.
Neugarten, B. L. (1968). The awareness of middle age. In doi:10.4054/DemRes.2011.25.4.
B. L. Neugarten (Ed.), Middle age and ageing: A Ray, R., Gornick, J. C., & Schmitt, J. (2010). Who cares?
reader in social psychology (pp. 9398). Chicago: Assessing generosity and gender equality in parental
University of Chicago Press. leave policy designs in 21 countries. Journal of
Neugarten, B. L. (1969). Continuities and discontinuities European Social Policy, 20, 196216.
of psychological issues into adult life. Human doi:10.1177/0958928710364434.
Development, 12, 121130. Riley, M. W., & Riley, J. W., Jr. (2000). Age integration:
Neugarten, B. L., & Hagestad, G. O. (1976). Age and the Conceptual and historical background. The
life course. In R. Binstock & E. Shanas (Eds.), Gerontologist, 40, 266270.
Handbook of ageing and the social sciences Riley, M. W., Johnson, M., & Foner, A. (Eds.). (1972).
(pp. 3555). San Diego: Academic Press. Ageing and society, vol. 3: A sociology of age strati-
Neugarten, B. L., Moore, J. W., & Lowe, J. C. (1965). Age cation. New York: Russel Sage Foundation.
norms, age constraints, and adult socialization. Riley, M. W., Kahn, R. L., & Foner, A. (Eds.). (1994). Age
American Journal of Sociology, 70, 710717. and structural lag: Societys failure to provide mean-
OConnor, J. S. (1996). Citizenship, welfare state regimes ingful opportunities in work, family and leisure.
and gender stratication. Current Sociology, 44, New York: Wiley.
4877. doi:10.1177/001139296044002005. Rosenmayr, L. (1983). Die spte Freiheit: Das Alter, ein
ORand, A. M. (1988). Convergence, institutionalization Stck bewut gelebten Lebens [The late freedom: Old
and bifurcation: Gender and the pension acquisition age, a piece of consciously experienced life]. Berlin:
process. Annual Review of Gerontology and Geriatrics, Severin & Siedler Verlag.
9, 132155.
156 G.O. Hagestad and P.A. Dykstra

Rosenmayr, L. (2007). Schpferisch Altern, eine Settersten, R. A., Jr., & Hagestad, G. O. (1996b). Whats
Philosophie des Lebens [Creative ageing: A philoso- the latest? II. Cultural age deadlines for educational
phy of life]. Mnster: LIT Verlag. and work transitions. The Gerontologist, 36,
Rossi, A. S. (Ed.). (1985). Gender and the life course. 602613.
Hawthorne: Aldine. Silverstein, M., & Giarrusso, R. (2011). Aging individu-
Ryder, N. B. (1975). Notes on stationary populations. als, families, and societies: Micromesomacro link-
Population Index, 41, 328. ages in the life course. In R. A. Settersten Jr. & J. L.
Sainsbury, D. (1999). Gender, policy regimes, and poli- Angel (Eds.), Handbook of sociology of aging (pp.
tics. In D. Sainsbury (Ed.), Gender and welfare state 3549). New York: Springer.
regimes (pp. 245276). Oxford: Oxford University Smith, J. A., McPherson, M., & Smith-Lovin, L. (2014).
Press. Social distance in the United States: Sex, race, reli-
Sanderson, W. C., & Scherbov, S. (2007). A new perspec- gion, age, and education homophily among condants,
tive on population aging. Demographic Research, 16, 1985 to 2004. American Sociological Review, 79, 432
2758. doi:10.4054/DemRes.2007.16.2. 456. doi:10.1177/0003122414531776.
Saraceno, C. (2010). Social inequalities in facing old-age Srensen, A.-M. (1987). Married womens economic
dependency: A bi-generational perspective. Journal of dependency, 19401980. American Journal of
European Social Policy, 20, 3244. Sociology, 93, 659687.
doi:10.1177/0958928709352540. Srensen, A.-M. (2005). Family structure, gender roles
Saraceno, C., & Keck, W. (2008). The institutional frame- and social inequality. In S. Svallfors (Ed.), Analyzing
work of intergenerational family obligations in inequality. Life chances and social mobility in com-
Europe: A conceptual and methodological overview. parative perspective (pp. 108128). Palo Alto:
Deliverable 1.1 for the FP-7 funded project How Stanford University Press.
demographic changes shape intergenerational solidar- Spder, Z., Murink, L., & Settersten, R. A. (2014). Are
ity, well-being, and social integration: A multilinks conceptions of adulthood universal and unisex? Ages
framework. Retrieved from http://www.multilinks- and social markers in 25 European countries. Social
project.eu/wp-content/uploads/2009/ 04/Report _ Forces, 92, 873898. doi:10.1093/sf/sot100.
Saraceno_Keck_Nov081.pdf Sundn, A. (2010). A discussion of retirement income
Saraceno, C., & Keck, W. (2010). Can we identify inter- security for men and women. In B. Marin &
generational policy regimes in Europe? European E. Zlyomi (Eds.), Womens work and pensions: What
Societies, 12, 675696. doi:10.1080/14616696.2010.4 is good, what is best? Designing gender-sensitive
83006. arrangements (pp. 5975). Farnham: Ashgate.
Schenk, N., Dykstra, P., Maas, I., & Van Gaalen, R. Therborn, G. (1989). Social steering and household strat-
(2014). Older adults networks and public care receipt: egies: The macropolitics and the microsociology of
Do partners and adult children substitute for unskilled welfare states. Journal of Public Policy, 9, 371397.
public care? Ageing & Society, 34, 17111729. Tobo, C. (2007). Change and reciprocity in intergenera-
doi:10.1017/S0144686X1300469. tional relationships: The discourse of Spanish working
Scherbov, S., Sanderson, W. C., Samir, K. C., & Lutz, W. mothers. In J. Vron, S. Pennec, & J. Lgar (Eds.),
(2014). Re-measuring twenty rst century population Ages, generations and the social contract: The demo-
ageing. In W. Lutz, W. P. Butz, & S. Kc (Eds.), World graphic challenges facing the welfare state (pp. 191
population & human capital in the twenty-rst century 207). Dordrecht: Springer.
(pp. 563590). Oxford: Oxford University Press. Uhlenberg, P. (1980). Death and the family. Journal of
Schmid, T., Brandt, M., & Haberkern, K. (2012). Family History, 5, 313320.
Gendered support to older parents: Do welfare states Uhlenberg, P. (1996). Mortality decline in the twentieth cen-
matter? European Journal of Ageing, 9, 3950. tury and supply of kin over the life course. Gerontologist,
doi:10.1007/s10433-011-0197-1. 36, 681685. doi:10.1093/geront/36.5.681.
Schtz, A. (1967). The phenomenology of the social Uhlenberg, P. (2004). Historical forces shaping
world. Evanston: Northwestern University Press. grandparent-grandchild relationships: Demography
Segalen, M. (2010). The modern reality of kinship: and beyond. In M. Silverstein & K. W. Schaie (Eds.),
Sources and signicance of new kinship forms in con- Annual review of gerontology and geriatrics (Vol. 24,
temporary Europe. In P. Heady & M. Kohli (Eds.), pp. 7797). New York: Springer.
Family, kinship and state in contemporary Europe Uhlenberg, P. (2009). Children in an aging society.
(Perspectives on theory and policy, Vol. 3, pp. 249 Journal of Gerontology: Social Sciences, 64B, 489
270). Frankfurt: Campus Verlag. 496. doi:10.1093/geronb/gbp001.
Settersten, R., Jr. (2003). Propositions and controversies Van Bavel, J., & De Winter, T. (2013). Becoming a grand-
in life-course scholarship. In R. Settersten (Ed.), parent and early retirement in Europe. European
Invitation to the life course: Toward new understand- Sociological Review, 29, 12951308. doi:10.1093/esr/
ings of later life (pp. 1545). Farmingdale: Baywood. jct005.
Settersten, R. A., Jr., & Hagestad, G. O. (1996a). Whats Van Bohemen, S. R. J. M., Van Zoonen, E. A., & Aupers,
the latest? Cultural age deadlines for family transi- S. (2014). Negotiating gender through fun and play:
tions. The Gerontologist, 36, 178188. Radical femininity and fantasy in the Red Hat Society.
Structuration of the Life Course: Some Neglected Aspects 157

Journal of Contemporary Ethnography, 43, 582600. Viazzo, P. P. (2010). Family, kinship and welfare provi-
doi:10.1177/0891241613505865. sion in Europe, past and present: Commonalities and
Vanderbeck, R. M. (2007). Intergenerational geo- divergences. Continuity and Change, 25, 137159.
graphies: Age relations, segregation and doi:10.1017/S0268416010000020.
re-engagements. Geography Compass, 1(2), 200221. Victor, C. R. (2010). Ageing, health and care. Bristol:
doi:10.1111/j.1749-8198.2007.00012.x. Policy Press.
Vaupel, J. W. (2009). In search of time won: About the Wenger, G. C. (2009). Childlessness at the end of life:
future of aging societies. In Max Planck Society for Evidence from rural Wales. Ageing & Society, 29,
the Advancement of Science (Ed.), Annual report 12431259. doi:10.1017/S0144686X09008381.
2008 (pp. 4750). Munich: Max Planck Society. Young, F. W. (1965). Initiation ceremonies: A cross-
Vaupel, J. W. (2010). Biodemography of human age- cultural study of status dramatization. New York:
ing. Nature, 464, 536542. doi:10.1038/nature08984. Bobs-Merrill.
Part II
Changing Social Contexts
and Life Course Patterns
Family Heterogeneity Over the Life
Course

Sandra Hofferth and Frances Goldscheider

Ongoing debate in the academic study of the fam- experienced dramatic increases in the labor force
ily focuses on the link between family change and participation of women, including married moth-
concomitant economic, ideational, and social ers. When women entered the public sphere for-
change. For example, even though the history of merly dominated by men, this rst half of the
the modern U.S. focuses on the search for indi- gender revolution breached the structure of the
vidual freedom and rights, many political and separate spheres that had characterized the gender
social groups have adhered to traditional views of balance since the mid-nineteenth century, when
the family, reinforcing patriarchal gender and the Industrial Revolution rst took the majority of
intergenerational relationships (e.g., old-order men out of the agricultural household economy.
Amish, Mormon groups). At the same time, we Even before the growth in female labor force
have seen an expansion of family forms, including participation, the life course of women was dra-
a rapid increase in acceptance of equal treatment matically reshaped by declines in fertility and
for, and even marriage between, same gender part- mortality, which meant that a domestic life was no
ners, the development of pioneering technologies longer a life-long job for them. This became par-
for starting families (e.g., assisted reproductive ticularly the case once school came to take so
technologies), and a longer term challenge to tra- much of childrens time and urbanization removed
ditional family structures from changing gender so many of womens farm-based tasks (e.g., main-
roles. Increasing heterogeneity across gender and taining a garden and preserving food together
social class, plus changing views of the role of with caring for small animals) (Stanfors and
biology and even childbearing in family forma- Goldscheider 2015). Women found that they
tion, have led to a proliferation of different family could contribute more to their familys well-being
choices and hence even more potential heteroge- by entering paid employment, particularly once
neity in the life courses of the next generation. the children began school, and increasingly by
Perhaps most profoundly, these changes in continuing in the jobs they had pursued before
gender roles, in particular in the relationships they had a family. This made preparation for
between men and women as couples, are reshap- employment (i.e., more education) increasingly
ing the life course. Since the 1960s, industrialized important for women, and the slowing of mens
countries, including the United States, have wage increases meant that men came increasingly
to depend on an earning spouse. This has begun to
S. Hofferth (*) F. Goldscheider
place increased pressure to share in the domestic
University of Maryland, College Park, MD, USA sphere on men, who, by joining women, will com-
e-mail: Hofferth@umd.edu plete the second half of the gender revolution.

Springer International Publishing Switzerland 2016 161


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_7
162 S. Hofferth and F. Goldscheider

The family prepares the next generation for 1.1 Educational Enrollment/
adopting adult roles, particularly establishing Attainment: Convergence
their own families. The life course principle of by Gender
timing informs our presentation. Experiences as
children in the family of origin set the stage for The big news is that males educational enroll-
later life and generational succession; thus, this ment/attainment superiority has ended, and even
chapter spends considerable time on the social- been reversed, with young women now much
ization of children and youth within the family. more likely than young men to be enrolled in col-
Second, we focus on the principle of linked lives. lege and to have completed college degrees. In
Family linkages may be stable and long-lasting 2012, 44.5 % of women 1824 were enrolled in
or unstable and tenuous. Finally, we emphasize degree-granting institutions compared to 37.6 %
individual agency and change. Behaviors, atti- of men (Federal Interagency Forum on Child and
tudes, and beliefs are not transferred intact. Family Statistics 2014). This is in sharp contrast
Individuals are not conned to repeating the past; with 1970, when only 20 % of young women of
they are inuenced not just by their families but that age were enrolled, compared with 32 % of
also by the social networks, communities, and young men (Federal Interagency Forum on Child
historical period in which they live. The chapter and Family Statistics 2014). Mens increase was
highlights key family behaviors and circum- marginal, womens dramatic. More young
stances that can inuence potential life course women (11.3 %) had completed a BA or more,
trajectories. compared with 8.9 % of young men. This is not
This chapter is organized around the stages of primarily because of differences in high school
the life course: childhood and adolescence, adult- completion, although differences have emerged
hood, and old age. Within each section, to the at this level as well. In 2012 only 9.1 % of 1821
extent research exists, we examine three key year olds were neither high school graduates nor
themes: (1) gender (focusing on the two halves of enrolled in high school; this was somewhat
the gender revolution), (2) biological and social higher for males (10.5 %) than females (7.9 %)
inuences versus individual preference, and (3) (U.S. Census Bureau 2012). The increased edu-
changing norms linked to increased family het- cation of women has implications for families
erogeneity (social class, family structure, immi- through the life course both directly and through
grant/minority status, couples sexual its link to increased labor force participation.
orientation). We focus on describing changing
patterns, describing determinants and summariz-
ing evidence on consequences. Throughout we 1.2 Gender Socialization
will consider the fundamental principles of life
course analysis: historical time and place, timing, Despite the decade-long stall in support for
linked lives, and human agency (Elder 1998), and female labor force participation in the United
in each case, we provide suggestions for pushing States, which ended late in the rst decade of the
the research area forward. twenty-rst century (Cotter and Hermsen 2014),
a number of studies showed that mens and wom-
ens allocation of time to paid and unpaid work
1 Childhood and Adolescence had become more similar (the two halves of the
gender revolution), with the gender gap in the
The changing structures of adult roles by gender unpaid work of cooking, cleaning, and child care
are beginning to reshape childrens preparation in narrowing during the 20002010 decade (Bianchi
childhood. Most dramatically, this is reected in and Milkie 2010). Although a substantial part of
changes in education. However, non-school the narrowing vis--vis housework resulted from
activities are changing rapidly as well, although declines in married womens time, with regard to
not always in directions consistent with the ongo- child care all of the narrowing resulted from
ing changes in adult lives. increases in mens time (Bianchi and Milkie 2010).
Family Heterogeneity Over the Life Course 163

A cross-national study of 20 countries covering legislation, which prohibited discrimination in


the 19652006 period found an increase of an school sports programs. However, the two are not
average of 6 h per week in employed, married at parity because boys also increased their par-
mens household work and child care time (Hook ticipation, but not as dramatically. Although in
2006). 2003 their participation rates were similar (57 %
Given the trend toward gender convergence in for girls vs. 63 % for boys), boys averaged an
economic and domestic roles in young adult- hour more per week than girls playing sports
hood, have there been major differences in social- (4:29 vs. 3:07) (Hofferth 2009). Recent data
ization by gender beyond educational attainment? show that boys 1519 continue to exceed girls in
Lacking information about parental views on time spent in sports (Portereld and Winkler
gender socialization, we document what children 2007). Sports is an area in which convergence
actually do. Because parents are making the deci- between men and women could substantially
sions for their children, examining how children alter attitudes of men towards equality of capa-
spend their time gives us a window on changes in bilities outside the household and increase mens
parental values and beliefs over time. We now willingness to become equal partners within in
have two major studies of American time use in the household. Shared interest in sports might
the past decade. The Panel Study of Income create stronger bonds between men and women.
Dynamics Child Development Supplement In addition, sports participation has been linked
(PSID-CDS) obtained information on the time to greater post-secondary school success for boys
use of children 612, and the American Time Use (Glick and Hohmann-Marriott 2007; Feldman
Survey (ATUS) obtained reports from individuals and Matjasko 2005; McNeal 1995), although it is
15 and older. We use the PSID CDS for childrens not uniformly positive, as it has also been linked
time; for the time of adolescents 1519 we use to greater alcohol use for both boys and girls
the ATUS. (Eccles and Barber 1999).
Studies of both children and adolescents show An important major gender difference also
dramatic convergence by gender in household occurs in the use of electronic media. In 2008
work, albeit to a low level. Housework time boys 1018 spent 14 h per week watching televi-
dropped signicantly for girls between the early sion and girls spent 12. At that time boys play
1980s and 1997, from 4:35 to 3:38 h per week, time consisted more of video game (5.00 vs. 0.94
and again between 1997 and 2003, to 3:09 h, a h) and computer game play (1.86 vs. 0.71 h) than
decline of nearly a third overall (31 %). There girls (Hofferth and Moon 2011). Girls have
was a much smaller decline for boys, who increased their time spent playing video games,
recorded 2 h 52 min, 2 h 42 min, and 2 h 28 min but it is still much less than males. Both males
at these three time points (Hofferth and Sandberg and females have increased their use of the web
2001; Hofferth 2009). Hence, by 2003 girls 612 and email, but girls are more likely to use it (38 %
spent only about a half hour more weekly time in vs. 28 %) and spend more time in email (2.06 vs.
housework than boys of the same age, compared 1.63 h per week). This, again, is an indicator of a
with a difference of nearly 2 h in the early 1980s potentially critical barrier to overcoming gender
(Hofferth 2009). The continued difference differences at home.
between boys and girls in household work is con- Finally, and of some concern, in 20032004
rmed by recent data on 1519 year olds from male teens 1519 spent less time than female
the ATUS (Portereld and Winkler 2007). teens doing homework (Portereld and Winkler
Although gender differences in socialization 2007). Almost 64 % of boys spent fewer than 5 h
have been and remain large, there has been some per week on homework, compared with 58 % of
convergence between boys and girls in one girls. Similarly, only 17 % of boys spent more
important arena. Research has shown that than 10 h on homework, compared with 21 % of
between 1981 and 1997 girls participation in girls. The lower amount of time boys spend in
sports increased (Hofferth and Sandberg 2001). homework is likely linked to their lesser college
This was linked to the passage of Title IX enrollment (Feliciano and Rumbaut 2005).
164 S. Hofferth and F. Goldscheider

The important conclusion is that even though Based upon the time use literature, we see
activities such as household work have declined considerable differences in time use by parental
over time for girls, the gendered nature of their education among children 1519 enrolled in high
time remains strong. Boys spend less time than school (Portereld and Winkler 2007). The larg-
girls both on housework and homework, leaving est distinction is in homework: 4.3 h per week for
them less prepared for the major work and family children whose highest-educated parent did not
roles of adulthood and more prepared to spend nish high school compared with 6 h for children
time in sports and electronic gaming. Gender dif- of parents with a BA and 9 h for children whose
ferences in activities are likely to affect family parent completed a professional or graduate
formation in the early adult life course. degree. Household work goes in the opposite
direction; nearly 4 h per week are spent in house-
hold work by children with a parent who did not
1.3 Heterogeneity in Socialization nish high school compared with 2.6 h by chil-
by Social Class dren with a parent who has a professional or
graduate degree. Differences in sports hours by
In her ethnographic book, Unequal Childhoods, parental education are smaller. Children of highly
Annette Lareau emphasized social class differ- educated parents spend about 5 h per week play-
ences in child rearing, with middle class families ing sports, compared with 6.3 for children of par-
engaging in what she called concerted cultivation ents without a high school degree. The only other
and lower class families engaging in what she major educational difference is that paid work
called natural growth (Lareau 2003). She argued has a curvilinear relationship with education.
that middle class parents make conscious deci- Children of the least educated and the most edu-
sions to promote the growth and development of cated parents spend the least time in paid work
their children, through placing them in organized (Portereld and Winkler 2007). TV viewing and
extracurricular activities, negotiating with insti- video game time are both lower for children of
tutions, and communicating verbally with the more-educated parents, who try to control excess
child. Children of working and lower class par- television and game playing, than children of less
ents, in contrast, place them in fewer organized educated parents (Hofferth 2010). Children of
extracurricular activities; rather, they permit mothers who completed a college degree spent 3
more child-initiated play, promote interaction fewer hours per week watching television and 2
with kin, but spend less time in verbal discus- fewer hours playing video games than children of
sions. Parental goals are primarily to ensure mothers who completed high school or less
safety and discipline, leaving children to develop (Hofferth 2010).
naturally. Although Lareaus social class distinc- Some research examines differences in social-
tion is based upon both parental occupation and ization by immigrant and native families, particu-
education levels, parental education appears to be larly comparing immigrant Hispanic and native
primary and we use it as our indicator of class.1 Hispanic families. Children of immigrants spend
more time studying, less time in activities such as
sports, less time playing video games but more
1
The major criticism of Lareaus conceptualization is that time watching television, and more time in house-
family structure and class are confounded. All of the poor hold work (girls) (Hofferth and Moon 2013).
families in her study were single parent families whereas
all of the middle class families were two-parent families.
Given the large differences in available time in two-parent
and single parent families, natural growth may result more 1.4 Heterogeneity in Socialization
from lesser availability of parental time and resources by Family Structure
than education/class. Research shows that parental goals
and values are similar in both middle and working class
families; what differ are the constraints that income and Family structure, which is linked to social class, has
family structure place on them (Hofferth et al. 2009). a strong effect on childrens share of housework.
Family Heterogeneity Over the Life Course 165

Not only do children take a considerably larger 1.5 Socialization for Family Life:
share if they grow up in a mother-only family, but A Theoretical Approach
boys in particular share much more, especially
adolescent boys, not only replacing the normal As we argue in this chapter, childrens families
fathers tasks (lawn work and home mainte- today are much more heterogeneous than in the
nance), but also more female-typed chores, past. In addition to families with two married,
including grocery shopping, cooking, house- biological parents, children live with two unmar-
cleaning, as well as childcare, dishes, and paper- ried biological parents, one parent and a steppar-
work. There are similar effects on children living ent, single parents, and other relatives such as
in a stepparent family, but they are not as strong grandparents (Hofferth et al. 2013). They may
(Goldscheider and Waite 1991). have two parents of the same gender. Thus social-
Mothers family experiences in childhood and ization for parenthood and family life has become
young adulthood (childhood family structure and complicated. No longer can it be assumed that
nonfamily living in early adulthood) have more children are being reared for participation in two-
mixed effects on their childrens involvement in sex, two-biological parent families. The experi-
household chores. Mothers who themselves grew ence of family structures in early childhood is
up in a mother-only family involve their sons and likely to have long-term effects on socialization
daughters more in household tasks than women for family tasks and behaviors in adulthood.
who grew up in traditional two-parent families, Researchers have studied long term associations
but women who experienced non-family living in for women to some extent, but few have exam-
early adulthood involve them less than women ined long-term associations for men. Parenthood
who did not (even controlling for college educa- is a more heterogeneous experience for men than
tion) (Goldscheider and Waite 1991). women; mothers tend to live with their children,
In sum, the results suggest that girls are still but many fathers do not, and others live with
receiving more training than boys in household stepchildren. The fact that boys raised with a
work, little though both may receive. Boys are stepfather are more likely to become stepfathers
more heavily involved in sports and video/com- themselves implies socialization for different
puter gaming but there is little research that types of fathering roles (Goldscheider and Sassler
informs us what the inuence of the focus on 2006).
such activity in high school has on later family Why would there be different outcomes of
formation and family involvement or on educa- growing up in different family types? Three theo-
tional and occupational success. In terms of retical perspectives on family process in different
social class differences, children of educated family types shape our understandings of these
parents spend less time in household work and likely outcomes: social learning, social control,
less time working for pay during the school and instability.
year than children of less educated parents. Social learning theory predicts that experience
They also spend less time watching television in the family during childhood will inuence
and playing video games and they spend more family attitudes, expectations, and roles in adult-
time in homework and traditional extracurricu- hood. Two-parent families provide important
lar activities other than sports. Finally, very lit- training in the roles of men and women in fami-
tle is known about the socialization of children lies. It cannot be assumed that children who were
of immigrants for later family life. What the raised in other families will learn this lesson;
implications of their different experiences are rather they learn messages from the types of fam-
for boys and girls later life course is known ilies they experience as children. Given how
only from piecing together disparate research important is the emotional context for learning,
ndings. There is clearly room for more sys- the learning that takes place in the home is likely
tematic study. very strong and very salient to children, and
166 S. Hofferth and F. Goldscheider

hence difcult to modify. Growing up with a because of the ambiguity of the role of steppar-
single parent is potentially the most isolating of ent. Even if the parent remarries, a new steppar-
the family types. Although few children spend ent usually has difculty lling the role of the
their entire childhood with only one parent, learn- other parent, which may be why stepfathers have
ing partner skills may be especially difcult for been shown to monitor children less than biologi-
boys with extended time in a single parent family cal fathers (Hofferth and Anderson 2003). The
unless they have role models outside the family. issue of control may be one of relevance to immi-
The proportion of children born to an unmarried grant parents, who are less procient in the lan-
mother doubled between 1980 and 2013, to 41 guage and culture of the host country; children
%; much of the growth was fueled by an increase often serve as translators and guides to their own
in births to cohabiting mothers, now more than parents, reversing family roles (Glick 2010).
half of all unmarried parent births and one- Finally, the third aspect of the family that
quarter of all births (Manning et al. 2014). Black could inuence childrens development of family
mothers are still more likely to be single mothers behaviors is instability. Research suggests that
(as was also the case in 1980), whereas White the disruption of routines may be more detrimen-
and Hispanic unmarried mothers are today more tal to childrens lives than the type of family (Wu
likely to be living with a partner than in the past. and Martinson 1993). Family structure changes
Researchers often include the latter with two- result in more disruption than just changes in par-
parent biological families, though many are ents; changes in parenting, neighborhood, school,
unmarried stepfamilies. They differ from married and friends are also likely to occur. These changes
biological parent families in the lower age, edu- may increase stress and lead to behaviors by chil-
cation, and income of the partners, and the greater dren to reduce conict and increase closeness
fragility of their unions (Sassler 2010). outside the family, such as increased sexual activ-
One area of increased research interest is that ity, precocious adult behavior, and early depar-
of the outcomes for children of growing up with ture from the parental home (Deleire and Kalil
same-sex parents compared with opposite sex 2002; Goldscheider et al. 2014; McLanahan and
parents. France, for example, did not allow gay Sandefur 1994; Wu and Thomson 2001).
and lesbian couples to adopt children, arguing
that heterosexuality was an essential part of cul- 1.5.1 Heterogeneity by Gender
ture that was learned at home (Butler 2004). Research has shown that males responses to dif-
Research on children of lesbian parents has not ferent family processes differ from those of
found differences in the development of children, females. Only a few studies have information
although childrens gender and sexual behavior about family processes and eventual family for-
and preferences have not been extensively stud- mation behavior for both males and females. We
ied (Biblarz and Savci 2010). Even less research focus on ndings from research examining the
has focused on socialization and outcomes for transition for both, as this permits a gender
children living with gay fathers. comparison.
The second theoretical approach argues that Females are very inuenced by their relation-
social control, which also varies across family ship with their mother; those growing up close to
types, is critical to socialization. Research has their mother did not transition to early parent-
shown that, even if the family is warm, children hood as often as those less close. In addition,
who are not monitored do not do as well as those social learning plays an important role; young
who are regularly monitored (Maccoby and women who grew up with a single mother and
Martin 1983). Limit-setting is an important never lived with a father tended to transition to
aspect of appropriate parenting. Of course, moni- motherhood, particularly single motherhood,
toring and keeping track of children is easier sooner than those living with two parents
when there are two parents (Forste and Jarvis (Hofferth and Goldscheider 2010). In this type of
2007). Stepfamilies may be less effective as well household, young women learn how to become a
Family Heterogeneity Over the Life Course 167

single mother and do not learn how to negotiate once having a working wife was problematic for
with a partner. Intergenerational ties in single young men, this is no longer the case; men have
mother families are strong. come to expect a working spouse (Gerson 2010;
Males, in contrast, appear to be inuenced by Goldscheider and Kaufman 2006). As a result,
instability (Hofferth and Goldscheider 2010). although American young adults continue to form
The more family transitions experienced, the conjugal unions and become parents, in most cases
greater the chance of the young man becoming a this occurs much later in young adulthood.
father at a young age. This is probably because
transitions are linked to changes in male father-
gures with whom conict is likely. Both close- 2.2 Union Formation
ness and monitoring are key family processes;
having a close relationship and being monitored, 2.2.1 Delayed Union Formation
but with some say in the rules, appears to be the and Parent-Young Adult
ideal socialization scenario as it delays an early Coresidence
transition to parenthood. Closeness to the mother Age at rst marriage has been increasing over the
was most important to the type of transition to past 30 years; the median age at marriage for
fatherhood. Men who were close to their mother women was 25.8 in 20062010 compared with
growing up became residential fathers even if 21 in 1976. Similarly, the median age at marriage
they were monitored strictly with no say about for men was 28.3 in 20062010 compared with
family rules. However, otherwise similar children 22.9 in 1976. In 20062010, 44 % of women
who were strictly monitored but were not close to compared with 31 % of men were married by age
their mother were more likely to become nonresi- 25 (Copen et al. 2012).
dential fathers (Hofferth and Goldscheider 2010). As a result, young adults gained several years
between completing school and forming a new
1.5.2 Heterogeneity by Social Class family. Although some experienced episodes of
Parental resources are important in family forma- non-family living, many returned or remained in
tion, as their benets accrue to children. Both their parental homes, increasing intergenerational
women and men who grow up with better edu- coresidence (Kahn et al. 2013). Only 8.6 % of
cated parents and in a higher income household young adults aged 2544 were living with their
delay family formation (Hofferth and Goldscheider parents in 1980; in 2010, this was the case for
2010). The theory is that children will have less 16.9 %. Further, although such intergenerational
incentive to leave resource-rich families if they do coresidence often reected the needs of the older
not need to. Parents buy delays in family forma- generation (in 1960, 53 % of Americans 45+
tion, permitting children more preparation time for living with an adult child were nancially depen-
the transition to adulthood (Haveman and Wolfe dent on them), by 2010, this was the case for only
1994; Goldscheider et al. 2014). 27 % (Kahn et al. 2013). This pattern was even
more marked among Black young adults than
non-Blacks (Kahn et al. 2013).
2 Young Adulthood Further, as marriage moves to older ages, the
and Adulthood length of the time in which individuals have to
select a partner is lengthening and the number of
2.1 Transformation options for romantic relationships has expanded
Due to the Gender Revolution from dating to short-term casual sexual relation-
ships (hook-ups) to visiting to cohabitation
Womens life course has been transformed by their (Sassler 2010). This means that the process of
growing participation in the labor force, the rst partner selection has become more complex and
half of the gender revolution. Their earnings have heterogeneous. We do know that cohabitation has
become central to their families well-being. If increasingly become the precursor to marriage,
168 S. Hofferth and F. Goldscheider

with the majority of young adults in their twen- 2.2.2 Social Class Differences
ties having lived with a partner (Sassler 2010). in Union Formation
Although middle class youth increasingly delay What is striking today is the growth in educa-
childbearing until after marrying in their mid- tional homogeneity in mate selection. The pro-
twenties, a substantial minority, particularly portion of partners whose education level is the
those who are not college bound, form their fami- same has increased to 55 % from a low of 45 %
lies early and without marriage. Those who begin in 1960 (Schwartz and Mare 2005). There is sim-
childbearing outside of a marital union are, in ilarly a decline in the proportion of partners who
turn, less likely to later marry. In 2011, 36 % of differ by one or two education categories. For
White mothers, 54 % of Hispanic mothers, and those whose educational levels differ, there has
72 % of African American mothers were unmar- been a sharp decline in those couples in which
ried at their childs birth (McLanahan and Jencks the husbands education exceeds that of the wife;
2015). This has resulted in a bimodal distribution this is a result of the previously described increase
of childrens families, with White children grow- in female schooling levels. Homogamy appears
ing up for the most part with two married parents, to be particularly strong for the college-educated
and Black and increasingly Hispanic children and for those with only a high school degree.
growing up in arrangements, such as cohabitation Cherlin (2010) suggests that there are different
and visiting, that are less stable, with parents who marriage markets for the least educated, those
are economically insecure, younger, less edu- with a high school degree and perhaps some col-
cated, and with potential mental health problems lege, and the college educated.
(Osborne and McLanahan 2007). Barriers to Given the increased time and attention chil-
marriage need to be further explored. It is a dren are spending online, communicating and
widely noted conundrum that same-sex couples gaming, it is only logical that nding a partner is
eagerly anticipate and celebrate marriage, now an objective of complex computerized sys-
whereas many of these young heterosexual cou- tems and is a major business; one in ten American
ples do not see (a good) marriage as an attainable adults reports having used an online dating site,
or even a benecial goal (Edin and Kefalas 2005; 38 % of single Americans who are looking for a
Sassler and Cunningham 2008), particularly as partner have used an online dating site, and 23 %
the economic fortunes of less-educated men have have met a spouse or long-term partner through
tumbled and womens have risen. these services (Smith and Duggan 2013). Even
Individuals choose their partnersmarriages so, it is not a science. The strongest evidence for
are not random eventsbut matching individuals how difcult it is to predict a successful match is
is not an easy task. Given its importance and provided by the Washington Post Date Lab
rapid change in the ways adolescents and young (Antoniades 2011). Success stories are very rare.
adults connect and establish romantic unions, it is Even though participants are chosen by what they
surprising that there is not more research on the share in experiences, preferences and attitudes,
process of partnering and marital/relationship the most common complaint is that the spark is
sorting. Of course, it is only recently that our not there.
large-scale surveys have begun to catalog the pre- In fact, for most young adults, choice is quite
marital relationships of youth, and document constrained by the community they live in, which
dates of cohabitation. In addition, most research is not nearly as broad as the Date Lab area.
has focused on adolescents and youth; given the Individuals who live in low income neighbor-
rise in divorce and remarriage, it is surprising that hoods with poor schools and who do not have
few studies have examined these processes resources to move from them are constrained to
among more mature men and women (Sassler relationships with others in those neighborhoods,
2010). who also have had poor educational opportunities
Family Heterogeneity Over the Life Course 169

and do not know what other relationships are ones stature and position in a society. Rather
possible (Edin and Nelson 2013). Although than contributing economically, children have to
online dating sites expand the range of opportuni- some extent become both a product of the union
ties, still there is a nancial barrier to entry and and a consumption item. Time with children has
such services are unlikely to accept low-income increased whereas another task of the household
men and women. (household work) has declined over the decades
Although a plethora of options are available (Sayer et al. 2004). The growth of intensive par-
for the college-educated youth, and a few ethno- enting means that bearing and rearing a child has
graphic descriptions of family formation among become a major project for middle class families.
those with a high school education or less are in Recognizing various motivations to have a child
print (Edin and Nelson 2013; Edin and Kefalas furthers our understanding of how parenting has
2005), there is little available information about changed and what new motivations have arisen.
the mating and matching process among youth Although some nations (e.g., Italy) have
with some college education. In addition, how extremely low fertility, in the United States child-
the mating process differs for young immigrants bearing has hovered at replacement level, sug-
is very much unknown. It is likely that religious gesting that the cultural context matters. One of
and community organizations serve as the mech- the reasons for a lack of concern about fertility
anism for mate selection for these relatively tra- levels in the U.S. is that the U.S. continues to
ditional ethnic groups. have high levels of immigration; Hispanic
women, the bulk of female immigrants, had a fer-
tility rate of 2.89 compared with the total fertility
2.3 Transition to Parenthood rate of 2.1, the level required to maintain a stable
population (Cherlin 2010).
Although we have pointed to enormous heteroge-
neity in mate selection and family formation, 2.3.2 Biological Relationship
probably the most heterogeneity lies in the con- and Marital Status of Childrens
text of the transition to parenthood and parenting Parents
behavior following childbearing. One of the major stories regarding family change
is the large shift in the structure of families rearing
2.3.1 Meaning of Parenting; children from two married biological parents to a
Motivation for Childbearing variety of other family types, particularly, families
and Later Involvement headed by unmarried mothers. This has resulted
One of the critical questions for modern families from increased out-of-wedlock childbearing;
is: what is the motivation for parenthood and between 1960 and 2013 the proportion of babies
parental (particularly father) involvement? In the born to unmarried mothers increased from 5 % to
past, children were critical to the family econ- 41 % (Child Trends 2014). However, this
omy. Without a source of labor on the farm or in increased heterogeneity is also due to a continued
the business, families could not survive. Even high probability of divorce (Cherlin 2010). Many
after the Industrial Revolution removed labor of these families will eventually be headed by two
from the home, children continued to be social- parents, one of whom is a stepparent. In 2009,
ized in ways that would continue the occupa- 56.7 % of children lived with two married biologi-
tional line from father to child; whether in a cal parents, 3.3 % lived with an unmarried bio-
family business, in a similar factory or industry, logical mother and father, and an additional 1.5 %
or in a similar or related profession. lived with adoptive parents; 27.3 % lived with
Additionally, having a child was important to only one parent; and 4.2 % lived with no parent,
continue the family line or family name. In nearly including grandparents, other relatives, or non-
all societies the family was the important unit; relatives only (Kreider and Ellis 2011). Living
having children and grandchildren increased with a biological parent and a stepparent were
170 S. Hofferth and F. Goldscheider

7.1 % of children. A major implication of these relationship with her children. The same argu-
trends is increased involvement of nonbiological ment applies to stepmothers and their partners
mothers and fathers in parenting children. children. However, because the woman usually
has the option of having her own children and
2.3.3 Social-Normative Expectations often does, there may be less incentive for her to
for Nonbiological Parents invest in the children of her partner unless there is
From a normative standpoint, although the rela- a quid pro quo; he invests in hers as well. The
tionship between husband and wife is institution- proportion of stepfamilies involving a coresiden-
alized through marriage, the relationship of a tial biological father and stepmother is small
stepparent to a stepchild is not (Cherlin 1978). compared to those involving a biological mother
However, there has been substantial research on and stepfather, and less is known about them.
the normative expectations for stepparents and It has been argued that after the relationship
the consensus appears to be that, in the United with the mother, mens attitudes about fathering
States, stepparents are expected not to be parents, are the most powerful psychosocial determinants
although some nancial support is expected. of paternal involvement. Research suggests that
They are expected to be friendly and supportive changes in attitudes are associated with changes
of the biological parent, but not disciplinarians in fathering behavior; of course, we do not know
(Coleman et al. 2000). Research consistently which comes rst. Changes in family structure
shows that stepparents monitor children less than have been occurring over the past 50 years and
do biological parents. The expectation for step- changes in father behavior have been similarly
parent involvement when the parents are unmar- documented. Comparing fathering attitudes over
ried is even more ambiguous than is the case for even the short period between 1997 and 2003
married stepparent couples. Unmarried relation- indicates that fathering attitudes have become
ships have been shown to be shorter in duration more positive over the period (Hofferth et al.
(Seltzer 2000). Another source of ambiguity is 2013). Better educated fathers, those employed
introduced by the regular involvement of the non- part-time, and those caring for a child without
residential biological father of the children, another caregiver had the most positive attitudes
which may interfere with the involvement of a towards fathering. Positive fathering attitudes
stepfather (Hofferth and Anderson 2003). were associated with greater engagement,
Potentially the most ambiguous is the case in warmth, control, and discussion about rules with
which the couple has both biological children their child.
and stepchildren, a blended family. Because dif-
ferential treatment of children in the same family
is obvious and normatively unacceptable, 2.4 Family Processes
research clearly shows less differentiation in the
treatment of biological and stepchildren in such 2.4.1 Gender Differences in Family
families and fewer outcome differences (Hofferth Processes
and Anderson 2003). Does the gender of children matter to family pro-
Even though it makes a lot of sense that bio- cesses? Recent research suggests that a variety of
logical fathers will invest more in children than parental behaviors are linked to the gender of the
nonbiological fathers, there are well-documented child. Is gender becoming less important? The
cases of stepfather care for stepchildren in non- following summary was drawn from a paper by
human species (Hofferth and Anderson 2003). Raley and Bianchi (2006), which summarizes
Investing in the children of his partner (relation- what research tells us about gender differences in
ship or mating effort) increases the chance that family processes. Basically, there is some evi-
she will want to have children with him. In addi- dence for a son preference among fathers. Men
tion, of course, the relationship between the part- were more likely to marry the mother when they
ners is strongly linked to the quality of his had a son, they worked more hours, they spent
Family Heterogeneity Over the Life Course 171

more time with a son, men stayed married longer, partners family. A new partner to the mother of
and they were more likely to gain custody of the the child does not try to replace the childs father
child or at least maintain contact with the custo- in the childs affections.
dial mother and child after divorce. Finally,
fathers were more likely to invest in the college 2.4.3 Differences by Race/Ethnicity
education of sons compared with daughters. The and Immigrant Origin
evidence from time spent with sons versus daugh- The above picture reects primarily research on
ters suggests that this difference is declining; African American and White families. Little is
however, to the extent that men continue to do documented about the Latino family. Immigrant
gender-typed activities themselves, it will not families have a higher birth rate and raise a higher
change very rapidly. number of children in their families. We know
that Latino families tend to be headed by two
2.4.2 Social Class Differences married biological parents; in 2011, 73 % of chil-
in Family Processes dren of immigrant parents were living with two
Resources always matter for the maintenance of parents, compared with 60 % of children of
relationships. The more resources the more ben- native-born American parents (Laughlin 2014).
et to be gained from the relationship. Therefore, However, the fact that in 2011 more than half of
it is important to consider the income/educational Latino mothers were unmarried at birth is a cause
level of the couple when examining family pro- of concern (McLanahan and Jencks 2015). Little
cesses. Recent research argues that relationships is known about pre-birth partnering or post-birth
in middle class families emphasize the bond marriages among Latino groups. Asian youth are
between partnersthe husband and wife, mom even less studied. Today one-quarter of American
and dad. The tie with the mother kept the man children are from immigrant families and in some
with the family and his children. This is the states the fractions are much higher. Research on
package deal that Townsend described how minority and immigrant children are reared
(Townsend 2002). Middle class children reared to become parents and partners is needed as the
in two-parent households expect to have that type population becomes less white, more minority,
of family themselves. Children who move from a and more heterogeneous culturally.
nuclear family of origin are more likely to create
an independent nuclear family unit of two parents
and children. However, if the couple breaks up, 2.5 Intergenerational
the fathers tie with the children is often broken. Consequences
A new stepfather in the family can replace some
of the fathering function of the biological father, Family structure transitions are not random
though usually not all of it. events but are highly responsive to childhood
Kathryn Edin argues that in low income fami- conditions and early life events and circum-
lies, in contrast, the link is primarily between par- stances. Children growing up in the raried
ent and child, not between partners (Edin and atmosphere of high parental resources and educa-
Nelson 2013). The partner relationship is often tion may be able to counter any instability and
brief before a child arrives; it has little time to changes in structures. Further, higher income
develop. Once the relationship dissolves, the families are less likely to disrupt. This leaves
mothers move into relationships with other men those families with the least resources in the
and the fathers with other women. Fathers main- worst conditions upon family disruption.
tain their relationship with the mother of their Research has demonstrated the intergenera-
child solely for the father-child bond. In this situ- tional transmission of parenting processes. A
ation the family is the vertical or extended fam- childs positive engagement with his father con-
ily; fathers have closer ties to their own parents tributes to the formers successful transition to
and family of origin than to their partner and adulthood, and in turn, to greater involvement
172 S. Hofferth and F. Goldscheider

with his own children (Hofferth et al. 2012). Although research on the effects of non-family-
Improved attitudes towards fathering and related experiences in childhood has begun, such
increased involvement of men with their children as how ill health in childhood affects health and
are good signals that mens involvement will con- survival in old age, and how the experience of
tinue to increase. The only caution is that the lit- poverty in childhood has lingering effects, the
erature also suggests that norms and models cumulative effects of family patterns over the full
regarding father involvement are changing rap- life course is one of the greatest open challenges
idly and that in this period it does not take a lot to in the study of the life course and the family
either encourage or to discourage fathers from (Goldscheider 1990).
being involved. From a policy perspective, this In this section, we lay out the challenges sepa-
includes workplace policies that make parental rate spheres couples face at retirement, given the
leave available and supportive employers that great increase in joint survival among couples
make it possible and acceptable to take such who remain married, and provide some evidence
leave. Even today it is still often considered unac- that the challenges might be less for couples who
ceptable for men to take leave even though it is have shared work and family tasks. However, we
acceptable for women (Pleck 1993). have found little research on gender roles in
Again, the reach of early childhood is long. retirement, in part because the pioneers of the
Low-income parents today face a burden of poten- increase in womens labor force participation
tial violence and injury or death in their communi- have only just begun to retire, and few of the
ties through rearms and other forms of violence. male pioneers in sharing household tasks have
How families can better manage their lives to begun to retire. Research that claries how these
reduce this threat and still live relatively free lives challenges might differ by class, race/ethnicity,
has become an important issue. The large and or the presence of non-biological ties, has not yet
increasing gap in life style across economic strata developed. Another important axis of differentia-
and the decreasing willingness of some well-off tion might be couple age differences, given the
groups to support expenditures for the youngest structuring of retirement within the U.S. Social
generations is an increasing problem. Early child- Security system.
hood and school programs will need to be more
and more cognizant of the variety of families rep-
resented in their schools and the needs of families 3.1 Retirement and Sharing
throughout the life course. A recent paper called Housework in Separate
for greater attention by pediatricians to early child- Spheres Marriages:
hood difculties that disadvantaged children and A Theoretical Approach
their families face with the goal of reducing physi-
cal strains or toxic stress that can later become Womens historical specialization in the domestic
serious health impairments absent appropriate sphere has had many repercussionsfor them
intervention (Shonkoff et al. 2012). and for the men they live with. To begin with,
womens specialization in homemaking, together
with increased life expectancy, has created a life
3 Challenges of Aging: course trajectory for them with several distinct
Retirement and Sharing segments and hence a marriage trajectory
Housework marked by built-in radical transformations
(Goldscheider 2000). The male-female relation-
The decisions and experiences made across the ship has become fraught throughout its length
life course accumulate to shape the later phase of with difcult transitions, given the ideology of the
the life course. As childhood activities shape separate spheres and the small families and long
those in adulthood, so both sets of activities lives women achieved due to the demographic
continue to inuence those later in the life course. transition. Men and women who marry in their
Family Heterogeneity Over the Life Course 173

20s or early 30s and have an average of two likely cause considerable suffering for traditional
children can expect their marital relationship to couples. The rst possibility is a problem for
survive until long after the children have grown wives. Husbands often successfully invade their
and are actively involved in their own parental and wives sphere, expecting increased domestic ser-
even grandparent roles. The empty nest stage, vices and supervising their wives activities more
which was once rarely achieved, is now the rule in directly, given greater opportunity and access to
marriages that survive divorce (Glick 1977). do so, leading to anger and resentment among
If one were to describe the ideal life course their wives.
stages of a traditional marriage, given low mor- In the next two possibilities, wives success-
tality and fertility, they would normally include at fully defend their separate sphere. In one, hus-
least four, each with a very different balance in bands enter the family on the wifes termsas
mens and womens relative power and her bumbling assistant. In the other, husbands
responsibility: stay out of the kitchen altogether. In both of
these cases, the husbands were likely to suffer,
both partners childless and employed, a transi- since just when they are experiencing separation
tion for both from full-time student; from their full-time employment, they become
parenthood, with women redening them- servantsor outcastsin their own homes.
selves as childrearers and drastically increased A few families might nd their way to the
male nancial responsibility; fourth possibility, and become egalitarian after
the empty nest, as women retired from daily the children leave home or when husbands have
parenthood and hence reduced household fewer work hours to justify their noninvolvement
responsibility, while men maintain nancial in the home and its tasks. The extent of this
responsibility; and nally change is rarely large, however, because the long
the husbands retirement. years under a different, separate spheres regime
are difcult to undo; this reaction may cause
If one re-negotiation of power and responsi- great suffering, as well. Unlike families used to
bility is difcult for most relationships (consider sharing productive roles inside and outside the
the difculties of renegotiating the parent-child home, separate spheres relationships have little
relationship when the child becomes an adult), experience with re-negotiating the fundamental
dealing with four transformations clearly puts changes needed as the balance of work and lei-
signicant stress on the quality and strength of sure shifts. It is hardly surprising, then, that those
the marital tie. with low levels of marital satisfaction postpone
Further, with women primarily involved with retirement (Kubicek et al. 2010).
childrearing and providing little nancial contri-
bution to the family economy, there was an even
greater need in such families for men to work 3.2 Research on Family History
long hours away from the home to provide sup- and Retirement
port for the children and for their wives who were
raising them. The closeness and warmth devel- How well does this analysis t actual data on
oped during courtship and nurtured in the early couples activities in later life? As we noted, this
years of marriage often withered as their separate is an area that needs much more research. Women
roles as the mothers of children and the bread- continue to do much more housework than men
winners of families separated their lives. after retirement, both in the United States and
Finally, when it came time for him to retire, other industrialized countries (Anxo et al. 2011;
such couples were likely to have a major crisis to Gauthier and Smeeding 2003). There is some
challenge them, since his work life was over but convergence, however, in part because womens
her remaining home tasks were not. There are hours decline and in part because mens hours
four possibilities for this situation, all of which increase. Two studies considered the effect of
174 S. Hofferth and F. Goldscheider

gender role attitudes on this transition, but found 4.2 Adoption and Assisted
either no effect or even counterintuitive effects Reproductive Technologies
(Solomon et al. 2004; Szinovacz 2000). This is
surprising, as research on marital happiness has About 12 % of women 1544 nd it difcult or
found that men aged 5192 with more egalitarian impossible to conceive. The proportion is small
gender role attitudes report greater marital happi- for younger women but reaches one quarter by the
ness (Kaufman 2006). No racial differences were late 30s and early 40s (Chandra et al. 2005). The
found, nor were class differences explored. availability of adoption, assisted reproductive
technology (ART), and surrogacy has improved
the access of couples to childbearing and rearing.
4 Challenges to Ways Adoption has remained at a relatively low
of Thinking About Families level. In 2002, 13 % of women reported rearing a
and Family Formation child who was not related to them, compared
with 11 % in 1995 (Chandra et al. 2005). Women
4.1 Same Sex Couples with less education and lower income are more
likely to be parenting a nonbiological child: 21 %
In 2010, an estimated 5.5 per 1,000 U.S. house- of women who had completed less than high
holds (646,000) were comprised of same-sex school vs. 8 % of college graduates. Similarly, 19
couples (Gates and Cooke 2014). A substantial % of nonHispanic black women compared with
number of these households were rearing chil- 10 % of nonHispanic white and Hispanic women
dren. Data from the U.S. Census Bureaus cared for nonbiological children.
American Community Survey indicate that about Although the proportion of women 2544 who
235,700 minor children were living in same sex used any infertility service was 19 % in 2006
couple households in 2012 (Payne 2014). Of all 2010, the proportion using ART is quite small,
same sex couple households, about one in ve under 1 % (Chandra et al. 2014). About 9 % of
had minor children present. About half of the men 2544 reported using an infertility service.
households had only one child. Not surprisingly, There are basically three forms of ARTs:
more female couple households (28 %) had
minor children than male couple households (13 Intrauterine Insemination (IUI), in which male
%). In comparison, of all family households in sperm are inserted into the female.
the U.S. in 2010, about 45 % had minor children In Vitro Fertilization in which the egg is fertil-
(U.S. Census Bureau 2012). The proportion with ized by the sperm in the laboratory and the
minor children was highest among same-sex cou- embryo placed in the mothers uterus.
ples with the lowest education levels: 40 % of Third Party ART, including Sperm Donation,
female couples and 29 % of male couples with Egg Donation, and Surrogates and Gestational
less than a high school education, compared with Carriers to carry the baby to term. All these
24 % and 9 % for female and male couples, require assistance from third parties.
respectively, with a bachelors degree. It was also
higher among Blacks (44 % and 25 % for female In 20062010, more than one-third of women
and male couples, respectively) than Whites (25 with current fertility problems reported using
% and 11 %, respectively). medical help to get pregnant (36 %) (Chandra
The number of studies of gay and lesbian et al. 2014). Among women aged 2544 in 2006
couples rearing children is still small, and 2010 with current fertility problems, women
because of the difculty in identifying them, still most commonly reported using advice (29 %),
of unknown national representation. One of the infertility testing (27 %), and ovulation drugs (20
leaders in disseminating information about such %). About 7 % reported articial insemination, 3
families is the Williams Institute (Gates and % had ever undergone surgery or treatment for
Cooke 2014). blocked tubes, and 3 % had ever used
Family Heterogeneity Over the Life Course 175

ART. Among women with fertility problems they are also participating in the work force.
there were differences in the use of infertility ser- Much remains to be learned about these powerful
vices by social class; 58 % of women with a changes in gender roles and relationships from
bachelors degree reported their use, compared childhood to old age.
with 33 % of women without a bachelors degree. A number of studies demonstrate the link
Married women were more likely to seek such between early life family structure and process,
assistance (55 %) compared with cohabiting including father involvement, indicating that how
women (21 %) and women who were neither they are reared inuences how young adults rear
married nor cohabiting (21 %). Similarly, seek- their own children. Although future research into
ing such assistance was more common for White linkages between family processes and own par-
than for other women: 44 % for nonHispanic enting is needed, we can conclude that father
Whites, 38 % for Hispanics, and 28 % for Blacks. involvement is likely to continue to increase
Same-sex couples are pursuing these options to because the feedback is positive. Certainly one of
bear children (Biblarz and Savci 2010), and they the most interesting new ndings is that father-
may fall into the other category as described hood is highly motivating for young low-income
here. No national statistics are available for them. males. More needs to be done to capture this
Some research is available on factors associ- motivation and shape it into long-term relation-
ated with seeking to use these technologies to ships with children and with a positive parenting
have a child, but more is needed, particularly for partnership with the childs mother. We do not
same-sex couples. More research is also needed yet know how to do this, unfortunately, as the ini-
on the consequences of their use for parents and tial studies of the Healthy Marriage Initiative
for the children (Biblarz and Savci 2010). (e.g., Building Strong Families) have not demon-
strated overall improvements in outcomes for
fathers or children from the current interventions
5 Conclusions (Wood et al. 2014). We need more tests of inter-
ventions that provide job counseling and jobs as
In this chapter we have pointed out important well as family support and coparenting educa-
convergences in the lives of men and women. tion. Programs need to incorporate economic and
First, we have noted that young women now sur- educational opportunities for young parents.
pass young men in college entry and college Nevertheless, although many young women
graduation. Second, socialization in the home as now prepare through education and early employ-
indicated by household activities has become ment for a sharing relationship predicated on the
more similar, with the household work of boys completion of both halves of the gender revolu-
and girls converging to a low level, whereas girls tion, structural barriers, which lead many young
participation in sports and computing have risen, women (and even more young men) to prepare
though not to the level of boys, and boys do less for a more conventional gender division of labor
homework. Third, womens participation in the (male provider, woman caregiver), remain. This
labor force has risen and mens has declined. is particularly problematic in the United States,
American women still tend to start families ear- which provides families with no paid parental
lier than men and take time off to rear young chil- leave or subsidized, quality child care. This often
dren; however, the participation of men and requires one parent to drop out of the labor
women outside the home in either school or work market, normally the mother. Unlike most other
has been converging such that the difference was industrialized countries, the U.S. also does not
only 9 % in 2013 compared with 21 % in 1980. tax incomes independently, so that the smaller
The major shift is that there is a convergence in income (typically the wifes) is taxed at a higher
involvement with children, with men increasing rate than the larger income. These barriers loom
their involvement with their children and women very high when children are young, when wives
remaining at the same high level even though incomes tend to be lowest and childrens costs in
176 S. Hofferth and F. Goldscheider

time and money tend to be highest, making it dif- Anxo, D., Mencarini, L., Pailhe, A., Solaz, A., Tanturri,
M., & Flood, L. (2011). Gender differences in time use
cult to see the long-term gains in wives earn-
over the life course in France, Italy, Sweden and the
ings from continuous employment relative to the US. Feminist Economics, 179, 159195.
short-term costs they are facing. Further, mothers Bianchi, S., & Milkie, M. (2010). Work and family
face much more of a wage penalty than other research in the rst decade of the 21st century. Journal
of Marriage and Family, 72, 705725.
women (or men), one that is particularly large
Biblarz, T., & Savci, E. (2010). Lesbian, gay, bisexual,
among working class women, who can least and transgender families. Journal of Marriage and
afford reliable childcare. Family, 72, 480497.
Given their frequent use in the U.S., we need Butler, J. (2004). Undoing gender. New York: Routledge.
Blossfeld, H.-P., & Hakim, C. (1997). Between equaliza-
to know more about the outcomes for children
tion and marginalization: Women working part-time in
and families of assistive reproductive technolo- Europe and the United States of America. Oxford:
gies to bear children. We also need more research Oxford University Press.
on families formed by same sex couples. Such Chandra, A., Martinez, G., Mosher, W., Abma, J., &
Jones, J. (2005). Fertility, family planning, and repro-
research is only in its early stages. These families
ductive health of U.S. women: Data from the 2002
permit tests of the source and maintenance of National Survey of Family Growth (Tech. Rep. No.
gender roles and the importance of biology. 23(25)). Hyattsville: National Center for Health
Family formation today is different from the Statistics.
Chandra, A., Copen, C., & Stephen, E. (2014). Infertility
traditional courtship process; the use of complex
service use in the United States: Data from the
statistical algorithms to identify potential matches National Survey of Family Growth, 19822010.
though matchmaking sites may result in a larger National Vital Statistics Report, 73, 119.
set of potential partners than were available in the Cherlin, A. (1978). Remarriage as an incomplete institu-
tion. American Journal of Sociology, 84(3), 634650.
past; this may have advantages for young people
Cherlin, A. (2010). Demographic trends in the United
today. We see no signs in the U.S. of reduction in States: A review of research in the 2000s. Journal of
demand or desire for children. In fact, immigrant Marriage and Family, 72, 403419.
parents tend to have larger families than natives; Child Trends. (2014). Births to unmarried women. http://
w w w. c h i l d t r e n d s . o rg / ? i n d i c a t o r s = b i r t h s - t o -
immigration has been an important source of
unmarried-women
renewal and will continue to be such. Research Coleman, M., Ganong, L., & Fine, M. (2000).
on the needs of immigrant families and their chil- Reinvestigating remarriage: Another decade of prog-
dren should continue to grow and develop as ress. Journal of Marriage and the Family, 62(4),
12881307.
should research on attitudes of whites who are
Copen, C. E., Daniels, K., Vespa, J., & Mosher, W. D.
apparently uneasy at the idea of a majority- (2012). First marriages in the United States: Data from
minority nation (Ingraham 2014). The willing- the 20062010 National Survey of Family Growth.
ness of white Americans to invest in schools and National Health Statistics Reports, 49, 122.
Cotter, D., & Hermsen, J. (2014, July 30). Back on track?
services for minority children will be an impor-
The stall and rebound in support for womens new
tant test of the inclusiveness of our great Melting roles in work and politics, 19772012 (brief report).
Pot and a key factor in U.S. continued world Retrieved October 7, 2014, from https://contemporar-
competitiveness over the coming decades. yfamilies.org/gender-revolution-rebound-brief-
back-on-track/
Deleire, T., & Kalil, A. (2002). Good things come in
threes: Single-parent multigenerational family struc-
References ture and adolescent adjustment. Demography, 39(2),
393413.
Antoniades, C. B. (2011, July 28). Five years of Date Lab: Eccles, J., & Barber, B. (1999). Student council, volun-
Who got hitched, who got ditched and what weve teering, basketball, or marching band: What kind of
learned in our quest for a spark. Washington Post extracurricular involvement matters? Journal of
Magazine. Retrieved from http://www.washington- Adolescent Research, 14, 1043.
post.com/lifestyle/magazine/five-years-of-date-lab- Edin, K., & Kefalas. (2005). Promises I can keep: Why
who-got-hitched-who-got-ditched-and-what-weve- poor women put motherhood before marriage.
learned-in-our-quest-for-a-spark/2011/07/05/ Berkeley: University of California Press.
gIQAILlUeI_story.html
Family Heterogeneity Over the Life Course 177

Edin, K., & Nelson, T. J. (2013). Doing the best I can: Haveman, R., & Wolfe, B. (1994). Succeeding genera-
Fatherhood in the inner city. Berkeley: University of tions: On the effects of investments in children.
California Press. New York: Russell Sage.
Elder, G. H., Jr. (1998). The life course as developmental Hofferth, S. L. (2009). Changes in American childrens
theory. CD, 69, 112. time, 19972003. Electronic International Journal of
Federal Interagency Forum on Child and Family Statistics. Time Use Research, 61, 2647.
(2014). Americas young adults: Special issue, 2014. Hofferth, S. L. (2010). Home media and childrens
Washington, DC: U.S. Government Printing ofce. achievement and behavior. CD, 81, 15981619.
Feldman, A., & Matjasko, J. (2005). The role of school- Hofferth, S. L., & Anderson, K. (2003). Are all dads
based extracurricular activities in adolescent develop- equal? Biology vs. marriage as basis for paternal
ment: A comprehensive review and future directions. investment. Journal of Marriage and Family, 65,
Review of Educational Research, 75, 159210. 213232.
Feliciano, C., & Rumbaut, R. (2005). Educational and Hofferth, S. L., & Goldscheider, F. (2010). Family struc-
occupational expectations and outcomes among adult ture and the transition to parenthood. Demography, 47,
children of immigrants. Ethnic and Racial Studies, 28, 415437.
10871118. Hofferth, S. L., & Moon, U. J. (2011). Electronic play,
Forste, R., & Jarvis, J. (2007). Just like his dad: Family study, communication, and adolescent achievement
background and residency with children among young and behavior, 20032008. Journal of Research on
adult fathers. Fathering, 5(2), 97110. Adolescence, 22, 215224.
Gates, G., & Cooke, A. (2014). United States census snap- Hofferth, S. L., & Moon, U. J. (2013, August). Childhood
shot: 2010. Los Angeles: The Williams Institute. activities, achievement, and immigrant transitions to
Retrieved from http://williamsinstitute.law.ucla.edu/ adulthood. In Paper presented at Annual Meeting of
wp-content/uploads/Census2010Snapshot-US-v2. the American Sociological Association.
pdf?r=1 Hofferth, S. L., & Sandberg, J. F. (2001). Changes in
Gauthier, A., & Smeeding, T. (2003). Time use at older American childrens time, 19811997. In S. Hofferth
ages: Cross-national differences. Research on Aging, & T. Owens (Eds.), Children at the millennium: Where
252, 247274. did we come from, where are we going? (pp. 193
Gerson, K. (2010). The unnished revolution: How a new 229). New York: Elsevier Science.
generation is reshaping family, work, and gender in Hofferth, S. L., Kinney, D., & Dunn, J. (2009). The hur-
America. New York: Oxford University Press. ried child: Myth vs. reality. In K. Matsuka &
Glick, P. (1977). Updating the life cycle of the family. C. Christiansen (Eds.), Life balance: Multidisciplinary
Journal of Marriage and the Family, 39, 513. theories and research (pp. 183206). Bethesda: AOTA
Glick, J. E. (2010). Connecting complex processes: A Press.
decade of research on immigrant families. Journal of Hofferth, S. L., Pleck, J. H., & Vesely, C. K. (2012). The
Marriage and Family, 72, 498515. transmission of parenting from fathers to sons.
Glick, J. E., & Hohmann-Marriott, B. (2007). Academic Parenting: Science and Practice, 12, 282305.
performance of young children in immigrant families: Hofferth, S. L., Pleck, J. H., Goldscheider, F., Curtin, S.,
The signicance of race, ethnicity, and national ori- & Hrapczynski, K. (2013). Family structure and mens
gins. International Migration Review, 41, 371402. motivation for parenthood in the United States. In
Goldscheider, F. (1990). The aging of the gender revolu- N. J. Cabrera & C. S. Tamis-LeMonda (Eds.),
tion. Research on Aging, 12, 531545. Handbook of father involvement (pp. 5780).
Goldscheider, F. (2000). Men, children and the future of New York: Routledge.
the family in the third millennium. Futures, 32, Hook, J. L. (2006). Mens unpaid work in 20 countries:
527538. 19652003. American Sociological Review, 71,
Goldscheider, F., & Kaufman, G. (2006). Willingness to 639660.
stepparent: Attitudes towards partners who already Ingraham, C. (2014, August 13). Whites uneasy at idea of
have children. Journal of Family Issues, 27, mostly minority U.S. (Workblog).
14151436. Kahn, J., Goldscheider, F., & Garcia-Manglano, J. (2013).
Goldscheider, F., & Sassler, S. (2006). Creating stepfami- Growing parental economic power in parent-adult
lies: Integrating children into the study of union for- child households: Coresidence and nancial depen-
mation. Journal of Marriage and the Family, 68, dency in the United States, 19602010. Demography,
275291, erratum 68 (3). 50, 14491475.
Goldscheider, F. K., & Waite, L. J. (1991). New families, Kaufman, G. (2006). Gender roles and marital happi-
no families. Berkeley: University of California Press. ness in later life. Journal of Family Issues, 27,
Goldscheider, F., Hofferth, S. L., & Curtin, S. C . (2014). 735757.
Parenthood and leaving home in young adulthood. Kreider, R. M., & Ellis, R. (2011). Living arrangements of
Population Research and Policy Review. doi:10.1007/ children: 2009. Current Population Reports, P70(126),
s11113-014-9334-9. 125.
178 S. Hofferth and F. Goldscheider

Kubicek, B., Korunka, C., Hoonakker, P., & Raymo, Sassler, S., & Cunningham, A. (2008). How cohabitors
J. (2010). Work and family characteristics as predic- view childbearing. Journal of Marriage and Family,
tors of early retirement in married men and women. 52, 328.
Research on Aging, 32(4), 467498. Sayer, L. C., Bianchi, S. M., & Robinson, J. P. (2004). Are
Lareau, A. (2003). Unequal childhoods: Class, race, parents investing less in children? Trends in mothers
and family life. Berkeley: University of California and fathers time with children. American Journal of
Press. Sociology, 110, 143. doi:10.1086/386270.
Laughlin, L. (2014). A childs day: Living arrangements, Schwartz, C., & Mare, R. (2005). Trends in educational
nativity, and family transition. Current Population assortative marriage form 1940 to 2003. Demography,
Reports, P70(139), 120. 42, 621646.
Maccoby, E., & Martin, J. (1983). Socialization in the Seltzer, J. A. (2000). Families formed outside of marriage.
context of the family: Parent-child interaction. In Journal of Marriage and the Family, 62, 12471268.
E. M. Hetherington (Ed.), Handbook of child psychol- Shonkoff, J., Gerner, A., & the Committee on Psychosocial
ogy (Vol. IV, pp. 1101). New York: Wiley. aspects of Child and Family Health. (2012). The life-
Manning, W., Brown, S., Lamidi, E., & Payne, K. (2014). long effects of early childhood adversity and toxic
Trends in births to single and cohabiting mothers, stress. Pediatrics, 129, e232e246.
19802009 (Tech. Rep. No. FP-14-05). Bowling Smith, A., & Duggan, M. (2013, October 21). Online dat-
Green: National Center for Family and Marriage ing and relationships. Internet and American Life
Research. Project. Retrieved October 4, 2014, from http://
McLanahan, S. S., & Jencks, C. (2015). Was Moynihan Pewinternet.org/reports/21013/online-dating.aspx
right? What happens to children of unmarried moth- Solomon, C., Acock, A., & Walker, A. (2004). Gender ide-
ers. Education Next, 15(2). Retrieved from http://edu- ology and investment in housework: Postretirement
cationnext.org/was-moynihan-right/ change. Journal of Family Issues, 252, 15501571.
McLanahan, S., & Sandefur, G. (1994). Growing up with Stanfors, M., & Goldscheider, F. (2015). The forest and
a single parent: What hurts, what helps. Cambridge, the trees: Industrialization, demographic change, and
MA: Harvard University. the ongoing gender revolution in the United States and
McNeal, R. B., Jr. (1995). Extracurricular activities and Sweden, 18702010 (Tech. Rep. No. SRRD_2014_18).
high school dropouts. Sociology of Education, 68, Stockholm: Stockholm Research Reports in
6281. Demography.
Osborne, C., & McLanahan, S. (2007). Journal of Szinovacz, M. (2000). Changes in housework after retire-
Marriage and Family, 69, 10651083. ment: A panel analysis. Journal of Marriage and
Payne, K. (2014). Demographic prole of same-sex cou- Family 62, 78-92.
ple households with minor children, 2012 (Tech. Rep. Townsend, N. W. (2002). The package deal: Marriage,
No. FP-14-03). Bowling Green: National Center for work and fatherhood in mens lives. Philadelphia:
Family and Marriage Research. Temple University Press.
Pleck, J. H. (1993). Are family-supportive employer U.S. Census Bureau. (2012). Statistical abstract of the
policies relevant to men? In J. C. Hood (Ed.), Men, U.S. 2012. Washington, DC: U.S. Government
work, and family (pp. 217237). Newbury Park: Sage Printing Ofce.
Publications. Wood, R. G., Moore, Q., Clarkwest, A., & Killewald, A.
Portereld, S., & Winkler, A. (2007). Teen time use and (2014). The long-term effects of building strong fami-
parental education: Evidence from the CPS, MTF, lies: A program for unmarried parents. Journal of
and ATUS. Monthly Laboratory Reviews, 130, Marriage and Family, 76, 446463.
3756. Wu, L. L., & Martinson, B. C. (1993). Family structure
Raley, S., & Bianchi, S. (2006). Sons, daughters, and fam- and the risk of a premarital birth. American
ily processes: Does gender of children matter? Annual Sociological Review, 58, 210232.
Review of Sociology, 32, 401421. Wu, L. L., & Thomson, E. (2001). Race differences in
Sassler, S. (2010). Partnering across the life course: Sex, family experience and early sexual initiation: Dynamic
relationships, and mate selection. Journal of Marriage models of family structure and family change. Journal
and Family, 72, 557575. of Marriage and Family, 63, 682696.
Educational Pathways

Robert Crosnoe and Aprile D. Benner

1 Introduction research, and the literature of life course studies


on education is growing. This growth was inu-
The educational career is a transition-rich long- enced by pioneering studies like those of Karl
term trajectory within a highly structured institu- Alexander and Doris Entwisle (1988), who
tional system. On the individual level, this developed a life course model to explain how the
trajectory dictates time use, facilitates the devel- interpersonal, social, and institutional experi-
opment of cognitive, social, and work skills, con- ences of children during the transition into and
nects people to interpersonal relations and social through the early years of schooling serves as a
networks, and paves the way for future socioeco- mechanism in the intergenerational transmission
nomic prospects. On the population level, it of racial/ethnic and socioeconomic inequality.
gauges the extent of a societys investment in its Now, it is being fueled by a new generation of
people and forecasts current and future economic scholars who leverage similar life course models
productivity and social stability. At the connection to organize studies of academic progress and
of these two levels, this trajectory represents the educational inequality and/or use educational
channel by which broad social, cultural, and eco- issues to inform the life course perspective more
nomic forces inuence individual lives and a generally. Such research is especially common in
mechanism in the process by which societal the sociology of education (Bozick and DeLuca
inequality is experienced and reinforced by indi- 2005) and developmental psychology (Benner
vidual people (Kingston et al. 2003; Arum 2000). and Graham 2009), and it is increasingly multi-
As such, education would seem well-aligned with method (Crosnoe 2011). The increasing cross-
the life course perspective, which emphasizes the pollination of life course and education thinking
dynamic and contextually embedded nature of life is, we argue, a positive development for educa-
pathways and how they connect individual devel- tional research, as the life course perspective
opment to population processes (Elder 1998). helps cohere the disparate areas and interests
Historically, however, life course approaches have within that eld. Many educational researchers
been underutilized by educational scholars. are interested in the institutional and contextual
Fortunately, the life course perspective has aspects of educationthe organization of schools
been gaining ground in the eld of educational and their internal curricula, instructional pro-
cesses within classrooms, and the funding and
R. Crosnoe (*) A.D. Benner support of school districts. Many others are
University of Texas at Austin, Austin, TX, USA focused on the developmental experiences of
e-mail: Crosnoe@austin.utexas.edu
those within the educational systemthe
Springer International Publishing Switzerland 2016 179
M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_8
180 R. Crosnoe and A.D. Benner

processes of learning and skill development, the First, we highlight critical educational issues in
interpersonal relationships that form in class- each stage of the life course and how they connect
rooms and schools, psychological orientations to the individual and population levels. Second, we
schooling, and the perceptions of students and focus on specic concepts from the life course
teachers of themselves and others (Arum 2000; perspective and how they are evident in and
Eccles et al. 1993). Because the life course per- informed by educational processes. Reecting
spective focuses attention on the bidirectional our training and the scope of our own research on
interplay of individual people with their contex- education, both stages are grounded in research
tual environments, it provides a conceptual lan- from the U.S., although we will bring in other
guage that allows both of these subelds of countries at times to esh out the discussion.
educational research to speak to each other.
This integrative value of the life course per-
spective in educational research is illustrated by 2 The Educational Life Course
its basic imagery. In this imagery, the life course
is a tapestry of three interwoven threads within On the individual level, the concept of the life
social contexts. The three threads include: (1) course refers to a persons passage through a
developmental trajectories, which tap into ongo- sequence of age-graded roles and contexts with
ing physical, cognitive, and socioemotional associated developmental tasks. On the popula-
growth and maturation; (2) social convoys, which tion level, it refers to the normative or expected
tap into continuity and change in the matrix of ways that people make this passage. We can
social relationships; and (3) social pathways, speak of the life course in general terms or spe-
which tap into sequential movement through cic to a particular domain, such as the educa-
institutional structures and socially dened roles. tional life course that is dened by the
As the transactions among these threads unfold, accumulation of experience within and around an
they shape and are shaped by the contextual envi- educational system, including prior to entry into
ronment, which can be conceptualized on a broad that system, during progress through the system,
spectrum from micro-level (or proximate) eco- and after exit from the system. This educational
logical settings of life like families, meso-level (or life course can be broken down into a few com-
intermediate) organizational settings like school ponents, and we focus on three here. One is
systems, to macro-level (or distal) abstractions learning, which encompasses the mastery of
like stratication systems, the economy, and his- skills and the development of knowledge bases
torical time. In this imagery, what happens to indi- that occur at the intersection of personal capabili-
vidual people over their daily lives, how they ties and instructional strategies (Siegler et al.
actively and passively interact with social struc- 2012). A second component is achievement,
tures, and what is going on in society at large which refers to external evaluations of learning
come together to constitute a complex multi-level (e.g., test scores, grades) and tangible curricular
phenomenon (Crosnoe and Johnson 2011). credentials (e.g., course credits) that bring
Following this imagery and its relevance to rewards or sanctions and that open up or fore-
education in modern society, our goal is to discuss close on future opportunities (Kelly 2008;
contemporary research on education in life course Riegle-Crumb 2006). A third component, attain-
terms. We will not provide an exhaustive review ment, is a summary of persistence, both tempo-
of the literature (see our chapter in the most recent rally (e.g., years of schooling) and in terms of
volume of Handbook on Child Psychology for status (e.g., degrees earned) (Mirowsky and Ross
more) but instead selectively engage with the lit- 2003). These components refer to the formal pro-
erature to capture how education illustrates the cesses of education, or the concrete aspects of
basic components of the life course perspective schooling most closely aligned with the ofcial
and how the life course perspective illuminates mission of modern educational systems to create
educational phenomena. We do so in two stages. a skilled workforce and informed population.
Educational Pathways 181

The educational life course also is formed by elicit socializing responses (Propper et al. 2012;
informal processes, including aspects of psycho- Pianta et al. 2007). The importance of childhood
logical adjustment and interpersonal functioning in the educational life course is that this stage is
that underpin schooling. These informal compo- when the foundation of long-term trajectories is
nents, such as engagement, comfort, and con- built and the roots of disparities in these trajecto-
dence, are additional components of the ries are laid.
educational life course (Crosnoe 2011). Consider one key aspect of learning that
In this section, we cover both formal and infor- unfolds across the transition into formal school-
mal components of the educational life course. In ing: the development of basic mathematical
doing so, we draw on the links between the educa- skills. Even during early childhood, children dif-
tional life course and the developmental stages fer widely in their understanding of mathematical
that help to dene the life course more generally. properties and in their sense of numbers. Because
In other words, what are some critical educational these basic skills then provide the groundwork
components, major forms of stratication, and for all subsequent aspects of math learning, chil-
research grounded policies to reduce such strati- dren who develop more early knowledge of math
cation during childhood, adolescence, and adult- are in a better position to gain more math knowl-
hood? Of course, we recognize that these three edge as they grow up. The process is highly
stages of the life course are somewhat arbitrarily cumulative, as in all academic domains, but it is
dened and demarcated and that the educational particularly apparent in math. As an example,
life courseby denitioncuts across develop- Siegler and colleagues have extensively studied
mental stages. Our purpose for this stage-specic childrens developing ability to approximate
discussion of educational processes, educational numerical magnitudes; in other words, to esti-
inequality, and educational policy is simply to mate which of a set of numbers are closest to
more effectively organize a large literature. Again, some specied value. This ability is typically
we are highlighting a few specic examples gauged with number line tests, in which children
linking individual-level components of education are asked to place some number on a number line
to more population-relevant discussions of educa- that goes from 0 to 100 but without any other
tional stratication and policyas a means of numbers between these two poles shown. Where
illustrating the importance of a more general theo- would a child place 59, close to 0, in the middle,
retical framework for education. or closer to 100? In part, the ability to properly
place the 59 is simply a function of cognitive
maturation. Few young children can correctly
2.1 Educational Processes place a number like this in its correct position on
and Inequality in Childhood the number line, but older children are able to do
it more consistently. Yet, it is also a function of
From early childhood, children start their educa- environmental opportunities to learn. Some chil-
tional life course in earnest, including informal dren have more opportunities to learn and prac-
learning at home and in the community, partici- tice with math, and some children take more
pation in early child care and education pro- advantage of whatever opportunities are available
grams, and their entry into the formal educational to them. As a case in point, children from low-
system. Their experiences during this period income families have slower development of this
largely entail learning key sets of basic skills and mathematical ability because they have less fre-
then building on these skills in increasingly com- quent learning activities with their parents at
plex ways. Such experiences are primarily struc- home, lower exposure to preschool enrichment,
tured by adults within the home, in school, and in and lower-quality math instruction in school.
school-related programs and activities in ways Thus, over-time gains in a key mathematical skill
that both reect and counteract genetically heri- are a product of the interplay of developmental
table traits that affect intellectual capacities and trajectories (maturation, intelligence), social
182 R. Crosnoe and A.D. Benner

convoys (teachers, parents), and social pathways development compounds over time due to the
(enrollment, classes). Importantly, basic math cumulative nature of learning and the large educa-
skills like those tested with the number line then tional system. In the aggregate, this process leads
powerfully predict achievement in math course- to a more unequal population (Alexander et al.
work many years later. Where a child starts in 2014; Kerckhoff 1993). This inequality enables
terms of initial skill levels helps to determine the economic needs of society to be lled by sys-
where they eventually end up in terms of achieve- tematically producing people to take on roles and
ment (Siegler and Lortie-Forgues 2014; Siegler jobs that require different skill levels and come
et al. 2012; Siegler 2009). with different rewards. It also means that some
Now consider how learning during the transi- people are under-placed or over-placed in terms of
tion into formal schooling plays a role in societal the t between the roles and jobs they take and
inequality. The school transmission model for- their skill levels, which could undermine eco-
mulated by Entwisle, Alexander, and colleagues nomic productivity while also being a source of
is a good illustration. The core argument of this societal instability (Arum 2000; Bowles and Gintis
model, which is closely related to the discussion 1976). Of course, children and families do have
of mathematical learning above, is that early agency to counteract this cumulative process, but
socioeconomic and racial/ethnic disparities accu- the unequalizing force of the system is strong.
mulate into long-term disparities in educational Unpacking and understanding this founda-
attainment. In more detail, children from diverse tional role of childhood learning in educational
segments of the population start school with attainment and educational inequality has been a
small but signicant differences in basic math, major activity of life course scholars in educa-
reading, and other academic skills due primarily tion, which, in turn, has been a powerful inu-
to differences in their early childhood environ- ence on educational and social policy. It has
ments (e.g., home, child care, community) before fueled the argumentbest summarized by Nobel
they even enter their kindergarten classrooms but prize-winning economist James Heckman
also to their own characteristics, behaviors, and (2006)that interventions into educational dis-
traits that shape how they interact with these parities bring greater long-term returns to invest-
environments and what they elicit from them. ment when they target early childhood rather
These early skill disparities are then acted upon than later stages of the life course or later levels
by the school system so that they compound from of the educational system. If small but signicant
year to yearone child has better developed ini- gaps in early skills are the foundation of larger
tial skills than another, and so she or he learns inequalities in later education, then closing those
more during the rst year of school, is selected initial gaps is a powerful method for reducing
into a more advanced learning group or class in inequality before it is allowed to grow. This argu-
the second year that further speeds up learning ment underlies the massive expansion of invest-
and leads to more rigorous curricula, and so on. ment in early childhood education programs over
That early skill gap eventually translates into the last decade, as exemplied by the
much larger gaps in educational attainment in implementation of publicly funded preschool in
adulthood through such schooling-related pro- the majority of states in the U.S. by 2014 (Duncan
cesses as well as the summer learning effect in and Magnuson 2013).
which youth from more advantaged backgrounds
experience more educational enrichment when
out of school that then further compounds the 2.2 Educational Processes
learning advantage of these youth (Alexander and Inequality in Adolescence
et al. 2014; Entwisle et al. 2005; Alexander and
Entwisle 1988). Into and through adolescence, young people shift
In this way, the long-term educational diver- into a more formalized, impersonal, and differen-
gence between children that is rooted in early skill tiated curricular environment. Within this
Educational Pathways 183

increasingly complex environment, cognitive and transition into chemistry in tenth grade than a
academic skills are translated into academic cre- similar student who enrolled in life science in
dentials that then become the most visible mark- ninth grade, as life science is typically lower-
ers inside and outside of schools of the preparation status and less challenging than biology and,
and suitability of young people for future educa- therefore, less likely to be viewed by the student
tional and occupational endeavors. During this and others as naturally leading into chemistry.
period, early advantages and disadvantages are Once in a tenth grade chemistry course, the stu-
activated or defused, in part because of how dent will be academically and socially prepared
young people interact with those around them to continue into more challenging science course-
and build social networks, with peers joining work after tenth grade even if the school does not
(and possibly even surpassing) adults as agents of require more science coursework after two sci-
inuence, consultation, and support. ence credits have been fullled. A similar sequen-
One key aspect of achievement that unfolds tial process occurs in math (e.g., taking algebra I
across the transition from elementary to secondary in ninth grade leads to more advanced math cred-
schooling involves the sequential progression of its through high school than taking pre-algebra in
youth through college-preparatory coursework. ninth grade), English (e.g., enrolling in honors
Compared to early stages of schooling, the aca- English in ninth grade leads to more honors and
demic curricula of high schools in the U.S. are Advanced Placement English credits across the
highly differentiated and involve more voluntary high school years than initially enrolling in regu-
choice. Many more courses are available for lar English), and other subjects. Some of these
enrollment, and students technically have the course sequences, upon completion of high
power to select courses (rather than being school, make students more attractive to colleges
assigned), with only some courses being required. than others and enable them to be better prepared
The irony is that amidst all of this choice, sequences for college coursework. Taking challenging sci-
of coursework are highly self-propelling. Once a ence courses and accumulating more science
student enrolls in a course of a certain level in 1 credits is one example of a course sequence that
year, he or she is likely to follow a fairly standard- promotes college matriculation and success in
ized sequence of courses at that same level into college. Thus, where in the academic curriculum
which that the initial course leads. In other words, a student starts high schoolwhich is predicated
once on a specic curricular trajectory, students on the skills that they have developed through
tend to stay on that trajectory, regardless of require- childhood or are merely perceived to have devel-
ments or alternatives, because of inertia, teacher oped, as well as their own choice and agentic
recommendations, parent expectations, skill pro- actionstrongly predicts where they nish in
gression, and other factors. Reecting the chang- that curriculum. The cumulative process emanat-
ing interpersonal orientations that arise during ing from early childhood described in the last
adolescence, peers also play a major role. They section is propelled by a related cumulative
model appropriate behavior, share information, process in adolescence (Crosnoe and Muller
and transmit values to each other, usually in tan- 2014; McFarland 2006; Morgan 2005; Schneider
dem with parental inuences but occasionally in et al. 1999).
opposition to them. Of course, young people are This individual progression through second-
not passive recipients of social inuence. They ary school courseworkand its emanation from
select themselves into social contexts, actively childhood experiencesis fundamentally impli-
work with what they are given, and try to chart cated in socioeconomic, racial/ethnic, and other
their own course. This agency injects variation into disparities in educational attainment at the popu-
general patterns of inertia (Riegle-Crumb and lation level. In a curricular system that is techni-
King 2010; Adelman 2006; Stevenson et al. 1994). cally open and voluntary but that is also
As an example, a high school student enrolled differentiated (i.e., many different kinds of
in biology in ninth grade is more likely to then courses and credits are available), hierarchical
184 R. Crosnoe and A.D. Benner

(i.e., some courses and credits are more valued ferences in informed decision-making related to
than others), and ambiguous (i.e., the norms for SES create disparities in an academic outcome
entering courses and attaining credits and what between high- and low-SES youth even when
they mean for the present and especially the they have the same ability and preparedness. For
future are unclear), some students are able to gain example, a high-SES student enrolls in ninth
a competitive edge over others simply because of grade biology while a low-SES student with the
their family backgrounds. Specically, the family same academic history and skill level enrolls in
and community environments in which they have ninth grade life science because the former
grown up give them access to a broad array of understands why taking biology is more advanta-
social resources that enable informed decision- geous in the long term, because he or she has par-
making that closely reects their actual educa- ents who push for the more challenging course
tional and occupational goals. In other words, and because teachers unconsciously equate
they are able to more strategically and effectively socioeconomic background with skill level.
plan for what they want to achieve. Such inequal- Again, peers will likely play a role during adoles-
ity is evident in educational systems in most cence, either because they directly share informa-
developed countries, but it tends to be more dif- tion and preferences for coursework or because
fusely spread across all levels of the system in the shared coursework doubles as peer time, and
U.S. than in many European countries, which some adolescents may be more or less responsive
often have more formal branch points in which to these peer inuences. This combination of pri-
students are explicitly assigned to academic or mary and secondary effects captures the iterative
vocational tracks (Buchmann and Park 2009; process by which educational disparities are built
Morgan 2005). Although such educational risks over time as students transition from one level of
are experienced by individual students, they have the system to the other. Because of prior opportu-
far-reaching implications for the population, nities, primary effects are at work at each new
given the importance to economic, political, and level, and, because of social resources, secondary
social stability of ensuring that jobs and roles are effects add to those primary effects (Jackson
lled by people with best-suited skills and that et al. 2007; Erikson et al. 2005; Goldthorpe 2000;
social mobility is a feasible reality. Breen and Goldthorpe 1997).
The primary/secondary effects perspective is a Evidence of the ways in which curricular
framework for understanding this stratifying pro- pathways in adolescence can differentiate the
cess. It refers to the effects of something like future prospects of young people has helped to
socioeconomic status (SES) on academic out- qualify the aforementioned shift towards early
comes. Primary SES effects occur when prior intervention in efforts to combat educational
experiences and genetically inuenced capabili- inequality. Even if early skill gaps are the founda-
ties related to SES create disparities in an aca- tion of long-term educational disparities, closing
demic outcome between high- and low-SES these gaps will likely only reduce (rather than
youth. For example, a high-SES student enrolls fully eliminate) disparities later in the educa-
in ninth grade biology while a low-SES student tional career because secondary effects can mean
enrolls in ninth grade life science because the for- that young people diverge even when they have
mer is better prepared for the more advanced the same skill level. In other words, an early
class and better demonstrates the required skill intervention may balance out early childhood
level after enjoying higher-quality instruction skill disparities, but the nature of the system will
from preschool through middle school. In this result in new disparities eventually emerging.
scenario, the SES difference in high school Efforts to do something about those secondary
coursework is an accurate reection of prepared- effects, therefore, can complement early inter-
ness to pursue that coursework, regardless of ventions by acting as boosters (Crosnoe and
whether the preparatory process up to that point Muller 2014). Because a big part of the problem
was equitable. Secondary effects occur when dif- of secondary effects is the availability of aca-
Educational Pathways 185

demic choices in a context in which not all people tive breaks in years of schooling that are demar-
are equipped to make informed choices, many of cated by the awarding of credentials, such as a
these efforts focus on clarifying ambiguity in sec- college degree. Degrees are concrete markers of
ondary school coursework or, even more radi- academic accomplishment that signify to others,
cally, eliminating choices altogether. As one such as potential employers and even prospec-
example, a wide variety of programs link adoles- tive mates, perceived differences in skill, compe-
cents whose parents did not attend college with tence, and value. Even when two people have the
college-educated mentors who can help them and same years of schooling, one with a degree will
their parents navigate the complicated academic have an advantage over one without a degree.
path to college enrollmentwhat classes to take, This importance of degree attainment is often
when to take the SATs, how to build a compelling called a sheepskin effect, and it is a source of
academic resume, how to apply for nancial aid debate about credentialism, or the tendency for
(Gandara 2002). As another example, some people to be rewarded for having degrees more
school districts have simplied and mandated so than having the training and skills those
their math/science coursework, so that all stu- degrees are supposed to signify (Attewell and
dents take the same courses for the same number Domina 2011; Hauser and Koenig 2011; Warren
of years (Attewell and Domina 2008). Overall, et al. 2008).
educational research has shown how the struc- In recent decades, the population of young
tural characteristics of schools may do more to people pursuing college degrees has expanded
shape some dimensions of curricular sequences considerably, and these young people follow
than family background and that the incentives diverse trajectories in this pursuit. For example,
for student behavior created by school reforms although a large portion of American youth enroll
may do more to affect academic decision-making in a 4-year college after leaving high school and
and curricular progress than individual attributes. eventually graduate within several years, others
Collectively, this research suggests that schools enroll but then drop out without nishing, others
can be changed to improve student outcomes go into 2-year colleges and never make it to
(McFarland 2006; Kariya and Rosenbaum 2003). 4-year colleges, others transfer from 2-year to
4-year colleges, and still others go back and forth
between enrolling (at different levels) and drop-
2.3 Educational Processes ping out. To this diversity, we can also add new
and Inequality in Adulthood forms of higher education, such as for-prot
institutions and online enrollment programs
Although formal schooling typically ends fairly (Bonnie et al. 2014; Patrick et al. forthcoming).
early in the life course, the importance of school- The diversity in paths that young people follow to
ing extends across the full span of life. A major college graduation and the substantial numbers
focus of research and theory on the educational who do not make it to college and/or to college
life course, therefore, is on educational attain- graduation is important because of the powerful
mentespecially degree attainmentduring role that college enrollment and graduation play
young adulthood and how its consequences lter in the life course. Because global economic
throughout adulthood and into the later years. restructuring over the last half-century has sig-
The story is about what educational attainment nicantly reduced sectors of the labor market
brings to someone and how what it brings allows (e.g., manufacturing) that historically provided
inequality to reproduce itself across generations. an ample supply of well-paying secure jobs that
To begin, degree attainment is a key aspect of did not require higher education, a college degree
educational attainment. Unlike the years of has become more and more important to success
schooling that a person has reached (a common in the labor market. Indeed, the economic returns
marker of educational attainment in population to a college degreesuch as the increase in life-
studies), degree attainment focuses on qualita- time earnings associated with obtaining a college
186 R. Crosnoe and A.D. Benner

degree versus a high school diploma, or the earn- ment in college in the rst place. Such disadvan-
ings premiumhave increased to historic levels tages include (but are not limited to) less exposure
in the last few decades. Although the costs of to preschool, lower-quality K-12 schooling, the
attaining a college degree (and associated debts) rising costs of higher education, and a lack of
are growing, the returns to such a degree are information about effective paths to college. Yet,
increasing even more rapidly. As a result, a col- disparities in college graduation are found even
lege degree is still a protable investment in the among those who enroll in college, suggesting
long term despite short-term costs, especially in that part of the problem is the higher rates of col-
countries like the U.S. that do not have the kinds lege dropout among the low-income and racial/
of apprenticeship systems supporting transitions ethnic minority students who made it to college
from school to work that have been developed in in the rst place (Bonnie et al. 2014).
some European countries like Germany (Autor A classic theoretical model of the transition
2014; Goldin and Katz 2008; Schoon and into college and its relation to socioeconomic and
Silbereisen 2009; Fischer and Hout 2006). demographic inequality is a good example of
Importantly, the life course returns to a college how life course insights can help shed light on an
degree extend well beyond income to encompass educational issue like college dropout. Tintos
other trajectories such as health, marital stability, (1987) model of institutional departure posits
and civic engagement (Kingston et al. 2003; that students leave college because of a lack of
Mirowsky and Ross 2003). Thus, increasingly, academic and social integrationthey are strug-
college graduation is a dividing line between the gling academically, feel like they do not t in
haves and have-nots in the U.S. and in other socially, generally come to doubt themselves,
industrialized societies. and drift off. Because students from low-income
Moving forward into adulthood, therefore, and racial/ethnic minority families often have
college degree attainment is a source of inequal- lower levels of academic preparation prior to col-
ity. To the extent that college degree attainment is lege entry, enter into segregated college cam-
also inuenced by family background and demo- puses, and are less likely to have the kinds of
graphic factors like gender and race/ethnicity, it social and cultural capital (e.g., contacts with
represents one way that advantages early in life people who can advance their interests, under-
translate into advantages later in life and, on the standing of valued styles of speech or dress) that
population level, one way that inequalities in one bring status in college settings, they might be less
generation translate into inequalities in the next academically and socially integrated at the col-
generation. The evidence of economic and non- leges in which they enroll. Reecting this lower
economic returns to college degrees suggests that degree of integration relative to other students,
graduating from college is a ticket to a more pros- they are at higher risk for dropping out. Because
perous, more stable, and healthier life, but not all they are more likely to drop out, they are less able
youth are equally well equipped to obtain this to enjoy the benets of educational attainment
ticket. Indeed, Whites are more likely to earn a throughout adulthood. They would then need to
degree than African Americans and Latino/as, possess especially high levels of agentic resources
and youth from afuent families are more likely (e.g., condence, persistence) to overcome these
to earn a degree than youth from low-income risks. Their vulnerability during the transition
families (Federal Interagency Forum on Child into college has inspired a number of policies and
and Family Statistics 2014). These disparities in programs to promote their integration at college
college graduation, which forecast disparities in and to protect them from dropout, including
numerous life course domains in the future, are interventions that attempt to build academic and
destabilizing on the societal level. They partially social communities (e.g., study groups, clubs,
result from disadvantages related to socioeco- peer counseling) for students of color and, more
nomic and racial/ethnic stratication throughout generally, for rst generation college students
childhood and adolescence that reduce enroll- (Massey et al. 2011; Steele 1997).
Educational Pathways 187

2.4 Summary of Educational ples that illustrate the theoretical meaning of


Processes and Inequality these life course concepts.

So far, we have highlighted one component of


educational functioning, one conceptual model 3.1 Transitions and Trajectories
of inequality, and associated policy responses for
each of three stages of the life course. We recog- In the life course perspective, a transition is a
nize that this selective discussion may seem change of state or setting that is embedded within
piecemeal to some readers, but we also argue a long-term trajectory. Because trajectories are
that, collectively, these stage-specic discussions often self-propagating and dictated by inertia, the
have helped us to make a larger point. Specically, changes involved in a transition have a height-
we believe that they come together to paint a pic- ened potential to disrupt compared to other peri-
ture of individual educational trajectories that are ods or points along the trajectory. As a result, a
highly cumulative and structured by major transi- transition is when a trajectory is most likely to be
tions in ways that exacerbate socioeconomic and redirected. A negative trajectory may be turned
racial/ethnic inequality but that also suggest criti- around into a more positive direction, or a posi-
cal windows of intervention (e.g., investments in tive trajectory may be interrupted and turned
early childhood education, creation of academic negative. Both kinds of redirection are important
and social learning communities in secondary to understanding inequality, and so transitions are
schools) for addressing these problems. Thus, the likely to be critical periods in the intergenera-
life course perspective offers tools to build a tional transmission of inequality as well as the
more generalized understanding of education that points at which policy intervention may bring the
can then translate research into action. greatest returns (Crosnoe and Johnson 2011;
Schulenberg and Maggs 2002; George 1993).
Transitions play an important role in the educa-
3 Life Course Insights into tional life course as well as in educational dis-
Education parities across the life course, reecting the
staging of the educational system in the U.S. and
In the prior stage of this chapter, we used the in most developed countries. This staging is orga-
stages of the life course to organize our discus- nizational, in terms of levels of schooling (e.g.,
sion of education. In this section, we take a dif- elementary school, secondary school, higher edu-
ferent strategy by using some basic concepts cation) and curricula, and it is physical, in terms
from the life course perspective as the organiza- of multiple classrooms within multiple buildings
tional tool. All of the concepts have been covered and campuses. Students transition through these
or at least alluded to so far. They include life levels and spaces as they attain education as part
course transitions, linked lives, the tapestry of of a long trajectory. Such transitions, therefore,
life course threads, and social context. Again, we are crucial to the cumulative nature of the educa-
provide specic examples of each within the tional career and of educational inequality that
domain of education that tap into different com- has been such a major theme of this chapter
ponents of the educational life course of individ- (Benner 2011).
ual people, aspects of stratication at the The transition into formal schooling and the
population level, and associated policy and inter- transition from high school into college have both
vention responses. Collectively, these examples been discussed so far, and so we now turn to the
show how the life course perspective illuminates transition from elementary school into middle
understanding of the processes of educational school and the transition from middle school into
attainment and educational inequality and, at the high school. These transitions are typically disrup-
same time, demonstrate how educational attain- tive to students academic and socioemotional
ment and inequality can provide concrete exam- functioning in ways that shape future educational
188 R. Crosnoe and A.D. Benner

trajectories. For example, the seminal work of in school, which has implications for their aca-
Simmons and Blyth (1987) documented transition- demic pursuits. As a result of this poor stage-
related declines in both grades and self-esteem in a environment t, early adolescents demonstrate a
sample attending diverse urban middle and high striking drop in academic engagement from
schools. Yet, the degree of such disruptions is in which they never really recover (Eccles 2004;
part dependent on the trajectories that led up to Eccles et al. 1993).
that transition. During these transitions, students Social disruptions are particularly problematic
move into iteratively larger schools with increas- during the transition from middle school into high
ingly higher-stakes academic curricula that are school. In the post-puberty years, aspects of brain
more competitive and more directly geared development and identity development converge
towards preparation for higher education than so that adolescents are more socially oriented,
what they have experienced up to that point. even socially dependent, than they were in child-
Moreover, the act of moving from one school to hood or will be in adulthood. When they move
another upends many relationships, both between between schools, their peer networks are often
students and teachers and among students them- altered in major ways, as students from the same
selves. As a result, transitioning into middle school middle school disperse into different high schools
and then into high school are times of vulnerability and as students from different middle schools
for young people, but that vulnerability is not the meet in the same high school. During this tumult,
same for all students. Differences in vulnerability many adolescents feel lost, and, in their state of
reect differences in individual development (e.g., heightened social orientation, these feelings are
psychological maturity, cognitive capabilities, and especially discomting and have greater potential
social relations) but also group differences in posi- to lter into academic pursuits. Indeed, compared
tion, status, and resources (Benner 2011; to other school transitions, depression, anxiety,
Langenkamp 2010; Barber and Olsen 2004). loneliness, and isolation are more common during
Academic mismatches between teaching and the transition into high school, all of which have
learning are particularly problematic during the implications for academic engagement and
transition from elementary school into middle achievement (Barber and Olsen 2004; Rudolph
school. Basically, middle school organization et al. 2001; Seidman et al. 2003). Although this
and pedagogy are not well-matched with the social disruption is often experienced negatively,
developing needs of the young people that they it does have benets for some students. For exam-
are serving. This school transition co-occurs with ple, ethnographic work has revealed that some
the developmental transition from childhood into students are able to use a school transition to rein-
adolescence, which is a time of dramatic biologi- vent themselves socially, and quantitative work
cal, cognitive, and social changes (Lerner and has demonstrated that this fresh start is more
Galambos 1998; Simmons and Blyth 1987). likely to happen when students who had academic
Early adolescents experience a strong urge to trouble in middle school transition into a high
carve out their own identities and take on more school with few of their middle school peers
autonomy for their lives but still need to maintain (Langenkamp 2010; Kinney 1999).
strong and supportive connections with adults, Just as school transitions have the potential to
but middle schools are not structured in ways to alter educational trajectories of individual stu-
meet these developmental needs. In middle dents, they also have the power to change educa-
school, students typically move from class to tional disparities between student groups. This
class, so that they do not have sustained time with latter change often takes the form of an increase in
any one teacher, but, at the same time, their the magnitude of disparities as the negative aca-
classes maintain top-down teacher-directed demic, social, and other experiences of school
instruction that do not offer them many opportu- transitions are especially pronounced for students
nities to take control of their own learning. In this from racial/ethnic minority groups and/or those
new context, students often struggle emotionally who grew up in low-income families. Their
Educational Pathways 189

heightened vulnerability reects their greater vul- example, past experiences allow Person A to
nerability in the educational system overall. accrue knowledge, skills, and information that
Because they are more often subject to discrimi- rub off on Person B. The relevance of linked lives
nation and segregation and because of their well- to educational research reects the interplay of
documented disadvantages in school quality, individual educational trajectories and school
students of color and/or from low-income fami- pathways with social convoys. In other words,
lies are less able to cope with the pressures and young people move into and through school sur-
hardships that come with school transitions. rounded by other people who contribute to and
Similarly, other students often have more social are affected by their educational experiences.
and institutional resources to ride out the short- Understanding the educational life course, there-
term disruptions associated with school transi- fore, means examining it as part of the series of
tions (Benner and Graham 2009; Crosnoe 2009; linked lives that characterize a persons complex
Barber and Olsen 2004). In these ways, the dis- matrix of social relations (Crosnoe and Johnson
ruptions to the educational career caused by 2011). Here, we focus on two major relationships
school transitions fuel academic stratication, in the social matrix that are highly educationally
translating and magnifying early inequalities into relevant, closely associated with specic devel-
future disparities. For example, analyses of opmental stages, and implicated in educational
nationally representative data on adolescents have inequality.
revealed that the link between middle school First, the parent-child relationship exemplies
math/science skills and high school math/science linked lives during childhood. Beyond issues of
placement is weaker for low-income English lan- genetic heritability of cognitive skills and aca-
guage learners than for other students. They are demically relevant personality traits, parents
more likely to get lost in the system as they transi- powerfully inuence the academic outcomes of
tion from school to school, and, consequently, their children. They do so by organizing aca-
they accrue fewer math/science credits over time demic opportunities for them (e.g., enrolling
(Crosnoe 2009). This process represents a process them in preschool, selecting high-quality schools
of cumulative disadvantage, when an already dis- and curricula, putting them in lessons or other
advantaged group is more affected by some nega- out-of-school activities), actively managing their
tive factor than a more advantaged group and learning and schooling (e.g., home literacy activ-
loses ground as a result (DiPrete and Erlich 2006). ities, participating in school events), emphasizing
the value of schooling (e.g., expressing positive
expectations), and providing home and
3.2 Linked Lives community environments that are conducive to
healthy development (e.g., maintaining a stable
One principle of the life course perspective is family life, living in a safe neighborhood) (Davis-
linked lives, which asserts that the experiences of Kean 2005; Bradley and Corwyn 2002).
people in the same family and network are recip- Importantly, the dynamics of this relationship
rocally related. Any one person is typically con- also allow socioeconomic disparities to magnify
nected with family members, friends, and educational disparities. Most parents want their
important others, and social inuences ow children to succeed academically, regardless of
through these connections (Elder 1998). If Person their SES, but higher-SES parents tend to know
A and Person B have a relationship, anything that how to promote this academic success to trans-
happens to Person A can inuence Person B late their actions into results more than lower-
through the tie between them. That inuence can SES parents. In developmental science, this role
be active, when, for example, Person A has devel- of the parent-child relationship in the stratica-
oped some attitude, orientation, or tendency tion of educational outcomes is best represented
through past experience and explicitly models it by the family socialization model, which is
for Person B. It can also be passive, when, for grounded in Elders (1974) study of children
190 R. Crosnoe and A.D. Benner

coming of age during the Great Depression terproductive academic views and information,
(Conger et al. 2010). It contends that the aca- and discouraging or undermining each other aca-
demic effects of family socioeconomic disadvan- demically. As a result of both sets of processes,
tage on schooling are ltered through intra-family the academic prole of an adolescent typically
dynamics, particularly parenting practices like closely mirrors the proles of his or her friends,
parental involvement in school, cognitive stimu- even when accounting for the initial tendency for
lation, and arrangement of academic activities adolescents to agentically form friendships with
that are more common among higher-SES par- similar others (Crosnoe 2011; Sacerdote 2011;
ents and parenting practices like harsh discipline Ryan 2001).
that are more common among lower-SES parents Some of the most intense controversy in the
(Crosnoe and Cooper 2010; Mistry et al. 2004; eld of educational research has centered on the-
Raver et al. 2007). In sociology, this focus on par- oretical arguments that such peer dynamics factor
enting is exemplied by Lareaus (2003) con- into racial/ethnic disparities in achievement. For
certed cultivation thesis, which contends that the example, the oppositional culture thesis contends
active parental management and involvement of that African-American and Latino/a peer con-
higher-SES parents gives a distinct academic texts equate academic achievement with acting
advantage to their children and teaches them to White, thereby disincentivizing achievement
use the educational system for their own gains. In and leading students in these racial/ethnic groups
both cases, parents are socialized into ideal types to underperform relative to White students of
of parenting, but more disadvantaged parents similar ability levels (Ogbu 1997). Empirical
experience more constraints on their ability to support for this thesis, however, is weak, and
enact their ideal types than more advantaged qualitative work has shown that the evidence of
parents. academic denigration that has been found is
Second, the peer relationship is a primary likely not racialized at all but rather a manifesta-
example of linked lives during adolescence. Like tion of peer denigration of academic achievement
people of all ages, adolescents are highly inu- that pervades the entire adolescent peer culture
enced by those around them. The aforementioned (Harris 2006; Tyson et al. 2005). Newer mixed
peculiarities of adolescent development, how- methods research has also shown that the role of
ever, mean that they are even more easily swayed peer dynamics in racial/ethnic and other dispari-
by peer inuences than children or adults. ties is not so much about negative inuences and
Although conclusions about the magnitude of modeling but instead a reection of the different
peer inuence have been undermined by studies pools of information available to different groups
that failed to account for the bidirectional inu- of students. In other words, both White and
ences among peers and other threats to causal African-American peer groups in a school may
inference (e.g., selection), the general consensus support college-going as a goal for students. Yet,
is that peers model academic behavior for each because of the connection between SES and race/
other (e.g., a friends achievement leads an ado- ethnicity and other aspects of discrimination and
lescent to have a positive association with segregation, the former peer groups may be better
achievement), share academic information with able to help students construct the kinds of aca-
each other (e.g., one friend telling the other about demic resumes that they need to go college than
a new academic program or explaining why tak- the latter (Crosnoe and Muller 2014).
ing advanced math will help college enrollment Immigration is one demographic phenome-
chances), and provide academic support to each non in which the educational role of parent-child
other (e.g., one friend helping the other with relations during childhood and the educational
homework, friends encouraging other during role of peer relations in adolescence come
times of stress). Of course, this inuence can also together to inuence the extent of stratication
go both ways, with friends socializing each other in the educational system. In general, students
into negative academic attitudes, sharing coun- from immigrant families experience more disad-
Educational Pathways 191

vantages than the general U.S. population, 3.3 Intertwined Trajectories


including higher rates of poverty, family-school
language barriers, and lower levels of school At the start of this chapter, we described the
quality. Because they are typically members of basic imagery of the life course as a tapestry of
racial/ethnic minority groups means they also three interwoven threads: developmental trajec-
face more discrimination, in addition to the gen- tories, social pathways, and social convoys. This
eral political and social scapegoating that sur- imagery from the life course perspective is rele-
rounds immigration itself (Garcia-Coll and vant to the contributions of both developmental-
Marks 2009; Crosnoe 2009). Given these disad- ists and demographerstwo main audiences of
vantages, children from immigrant families life course researchto the eld of educational
should be faring poorly in the American educa- research. Scholars from these disciplines have
tional system, and many of them are. Yet, in gen- helped to expand the focus of educational
eral, the norm is for immigrant children to score research away from purely academic and institu-
or rate higher than their peers with native-born tional factors, statuses, and processes to consider
parents in school completion, grade point aver- the ways in which educational experiences are
age, test scores, and other academic factors in entangled with other life experiences (Crosnoe
high school, especially when their generally and Johnson 2011). Here, we discuss three
more disadvantaged socioeconomic circum- examples of the connections among develop-
stances are taken into account. A similar mental trajectories, social pathways, and social
although weaker and less consistent patternholds convoys within the educational life course as a
in elementary school (Hao and Woo 2012; means of blurring the lines between the tradi-
Reardon and Galindo 2009; Pong and Hao 2007; tional domains of educational research and life
Glick and White 2003) and has been reported in course researchembedding education within
other developed countries with large immigrant the general life course and considering individ-
populations from the developing world ual trajectories that aggregate into population
(Washbrook et al. 2012). trends. Doing so is important to understanding
Many explanations are useful to understand- inequality and identifying solutions to social
ing this immigrant paradox, but one of the most problems.
common concerns the strong family ties in immi- First, the connection between education and
grant communities and their implications for peer work represents the interplay of two social path-
dynamics. Immigrant children tend to be more waysthe navigation of the educational system
oriented towards adults than children raised in and the navigation of the labor market, along
the U.S. by U.S.-born parents, and so they are with the sequences of roles and settings associ-
powerfully inuenced by the educational goals ated with both. The general view of the educa-
and aspirations that their parents have for them tional system in the U.S. is that its primary
(and that likely contributed to immigration deci- (although not sole) purpose is to prepare young
sions in the rst place) and less likely to associate people to enter the labor market and contribute to
with (or be inuenced by) peers with attitudes the economic productivity of the nation; how-
and behaviors that are counter to what their par- ever, the weaker mechanisms for coupling voca-
ents value. Thus, social and emotional advan- tional and academic education in the U.S.
tages and protections in the parent-child compared to many European countries, like
relationship blunt other types of disadvantages Germany, mean that many American youth may
and risks, including peer risks. The longer that be more vulnerable during the transition to the
families remain in the U.S. and the more inte- labor market (Bonnie et al. 2014; Schoon and
grated that youth are into Americanized peer Silbereisen 2009). The nature of this preparatory
groups, however, these advantages and protec- role in the U.S. or in other developed countries,
tions fade in power and allow the disadvantages however, shifts depending on timing in the life
and risks to take effect (Kao 2004). course.
192 R. Crosnoe and A.D. Benner

Concurrently, education and work can both 2006; Mirowsky and Ross 2003; Kerckhoff
support and undermine each other. For example, 1993). Yet, even as educational attainment now
most adolescents work for pay at some point predicts work success later, attempting to com-
while they are also attending high school (Apel bine school and work now can reduce educational
et al. 2007). Many view such concurrent school- attainment later. Basically, juggling college
ing and work in positive terms, arguing that work- coursework and a job is related to lower odds of
ing for pay and learning to manage multiple time college graduation, which is the best predictor of
constraints and duties allows young people to ultimate occupational attainment (Bernhardt
develop responsibility, conscientiousness, and et al. 2001). Again, how education and work are
other important skills while also gaining work connected over the life course is important.
experience. The number of adolescents combin- Moreover, socioeconomic differences in the
ing work and schooling has also generated some types of jobs an adolescent can secure, enroll-
concern, fueled by studies showing that adoles- ment in higher education, and the ability to post-
cent workers do worse academically than other pone work to concentrate on higher education
adolescents and are more likely to engage in sub- studies mean that the interplay of school and
stance use. This research is used to characterize work trajectories both reect and can reinforce
work as an academic distraction and an entre socioeconomic inequality over time and across
into age-inappropriate groups with bad inu- generations, which is why reducing educational
ences. Research attempting to reconcile these two disparities across diverse segments of the popula-
sides through more rigorous longitudinal methods tion is one of the widely supported mechanisms
has generally indicated benets (e.g., academic for intervening in the intergenerational transmis-
progress, conventional behavior) from work dur- sion of inequality (Bernhardt et al. 2001; Kane
ing the high school years unless work hours are 1999; Schneider and Stevenson 1999).
quite intensive and/or young people who are Second, the connection between education
working have low social and academic propensi- and health represents the interplay of a social
ties to work (Lee and Staff 2007; Johnson 2004; pathway and developmental trajectory, respec-
McMorris and Uggen 2000; Mortimer 2003). tively, with the navigation of the educational sys-
Prospectively, academic success and educa- tem shaped by and shaping physical and mental
tional attainment in adolescence and young adult- functioning and wellbeing. Although the pur-
hood is one of the most important predictors of poses of the educational system are widely
success in the labor market across adulthood. considered to include the promotion of educa-
Educational experienceespecially in high- tional success on the individual and population
quality schoolscultivates tangible work-related levels, they are not often thought to include the
skills (e.g., math skills needed to be an accoun- promotion of individual and population health.
tant, ne arts training to be an artist), general Yet, health is one of the best-documented out-
cognitive skills that promote success in work comes of education. More educated people are in
(e.g., critical thinking, problem solving), social better health, and more educated societies are
and cultural capital that increase attractiveness to healthier (Mirowsky and Ross 2003).
employers and clients (e.g., knowledge of ne One striking example is the clear link between
arts, good grammar), and it also signals to future educational attainment and mortality. Within any
employers and clients that a person might have given society, more educated people live longer
other valued traits, such as persistence and con- than less educated people. Moreover, societies
scientious, that are thought to go along with edu- with higher rates of educational attainment tend to
cational success. As a result, high school be characterized by lower mortality rates. These
graduates have better labor market outcomes than associations tend to grow over time, both across
high school dropouts and college graduates have the life course and across historical time (Miech
better labor market outcomes than high school et al. 2011; Lynch 2003). Why would educational
graduates (Bonnie et al. 2014; Fischer and Hout attainmentwhich, with the exception of actual
Educational Pathways 193

health education, is not explicitly targeted at less so than health, family statuses and processes
healthbe associated with such a clear marker of are not commonly thought of as outcomes of the
health? Certainly, selection is part of the answer. educational system, but indeed they are. Thus,
Healthier people may be more likely to attain edu- just as family background shapes educational
cation (i.e., a bidirectional interplay between edu- experiences, educational experiences shape
cation and health, not just education affecting future family life (Crosnoe and Cavanagh 2010;
health). Furthermore, the factors that promote edu- Cherlin 2009).
cational attainment might also lead to better health, To begin with some basic family statuses, edu-
such as genetically inuenced cognitive skills and cational attainment predicts numerous aspects of
intelligence (Palloni 2006; Lynch 2003; marriage and partnership across adulthood. In the
Lauderdale 2001). Such selection forces, however, U.S. and many other developed countries, indi-
do not account for all of the links between educa- viduals who have graduated from college are
tion and mortality. Some causal factors are at work more likely to marry than their less educated
as well. Educational attainment comes with mon- counterparts. Once married, college graduates
etary rewards that grant access to higher-quality are also more likely to stay married than others.
health care, safer living and working environ- In other words, education seems to facilitate mar-
ments, and greater health supports (e.g., nutrition, riage and discourage divorce (Musick et al. 2012;
health-related consultations). The benets of edu- Schoen and Cheng 2006). Educational attain-
cation go well beyond money, though, to include ment is also closely related to fertility and associ-
the social, cultural, and cognitive resources that ated parenting statuses and circumstances.
education cultivates (Lynch 2003). For example, College graduates have fewer children on aver-
more educated individuals tend to have broader age than people with less education. Even more
and more diverse networks, so that they have strikingly, they are far less likely to have children
deeper pools of social support. The critical think- outside of marriage. A majority of Americans
ing that they develop through their educational who have not attended college now have children
experiences (e.g., adjudicating between pros and without being married, but less than 10 % of col-
cons, seeing life circumstances as under their own lege graduates do (Cherlin 2009). Importantly,
control) also allows them to make better decisions this interplay of educational attainment and
about health and health-related behaviors family formation has strengthened over time, so
(Mirowsky and Ross 2003; Ross and Wu 1995). that a child with a well-educated mother is likely
On one hand, educational attainment is to enjoy many more advantages (e.g., money,
strongly and inversely related to morality and family stability) today than in the past, while a
many other problematic health conditions and child whose mother has low education faces
outcomes (e.g., depression, functional limita- many more disadvantages. McLanahan (2004)
tions, morbidity). On the other hand, educational has referred to this tighter coupling between
attainment is stratied by SES, race/ethnicity, maternal education and broader family advan-
and other sociodemographic factors. Thus, it is a tages and disadvantages as the diverging desti-
major source of socioeconomic and demographic nies of American children.
disparities in health, which are such a common Turning to basic family processes, educational
focus of life course research and are frequently attainment is also a factor in parenting behavior.
targeted for government intervention (Palloni We have already covered the family socialization
2006; Williams and Collins 1995). model and the concerted cultivation thesis, both
Third, the connection between education and of which contend that family SES differentiates
family formation represents the interplay of a child outcomes through parenting and other fam-
social pathway and a social convoy, with the nav- ily processes, but we want to stress that the litera-
igation of the educational system related to how tures around these theoretical perspectives
individuals enter into and enact family roles and highlight parent education (and particularly
how they participate in family relations. Even maternal education) as a source of that
194 R. Crosnoe and A.D. Benner

socioeconomic differentiation (Conger et al. tional attainment from the micro, meso, and
2010; Lareau 2003). Educational attainment helps macro levels. These levels overlap, although
parents manage their childrens needs and oppor- studies of education often isolate them from each
tunities in developmentally appropriate ways that other (Crosnoe and Benner 2015).
maximize their chances for healthy development First, the micro level consists of proximate
and academic success (Kalil et al. 2012). ecological settings dened by smaller spaces and
Consequently, policy efforts are increasingly tak- numbers of people. The classroom, which is
ing a two generation approach to intervene in embedded in a school, district, and the larger edu-
child health and education disparities. In such an cational system, is an example of a proximate
approach, investments in children (e.g., preschool micro-level context. The classroom is the most
enrichment, health care) are paired with human concrete and identiable setting in which instruc-
capital investments in parents, especially mothers tion and learning occur in interaction with each
(e.g., general equivalency diploma classes, bilin- other, and, as such, it is often a site of intervention
gual education, skill building and job training pro- in efforts to improve learning and reduce educa-
grams). In this way, children benet through tional disparities. For example, consider the
direct services but also indirectly through the debate about whether English language learners
resources that their mothers develop in parallel, have higher academic achievement when enrolled
resources that are in part intended to help them in classrooms in which they are taught primarily
better act on their positive parenting values. This in English or when teachers use a bilingual
approach is relevant to combating socioeconomic approach. Some argue that bilingual instruction
inequality as well as other forms of stratication will interfere with English skill development, but
related to socioeconomic inequality, such as strat- others argue that mastery of the native language is
ication by race/ethnicity and immigration (King a requirement for mastering English. Empirical
et al. 2011; Crosnoe and Kalil 2010; Smith 1995). evidence supports the latter argument. Students
Of course, as this discussion has made clear, do better when instruction geared towards culti-
many of the life course advantages of education vating English language prociency is integrated
help to explain the interplay of educational attain- with content instruction in the native language,
ment with other threads of the life course. with a gradual transition towards content instruc-
Education cultivates resources and skills, opens tion in English as English language prociency
up opportunities, and empowers people to build increases. This strategy works because English
on such resources and skills and capitalize on language learners do not fall behind in academic
such opportunities. Thus, educational attainment content while learning English, and it reects the
stabilizes and supports other forms of attainment. rewards of true bilingualism for cognitive and
Yet, the direction is not solely one way, as those academic skills more generally (Golash-Boza
other forms of attainment comes with advantages 2005; Padilla and Gonzalez 2001; Yeung et al.
and disadvantages that promote or truncate edu- 2000). This immigration-related phenomenon is
cational attainment. an example of the child x instruction x context
approach, which contends that children learn
more readily when their teachers tailor instruction
3.4 Contexts to their specic needs and talents within a gener-
ally supportive environment (Connor et al. 2009).
To return once again to the imagery of the life An English language learner will be more aca-
course as a thread of interwoven tapestry, recall demically successful when a teacher effectively
that this imagery also embeds this tapestry within evaluates what he or she needs both academically
social contexts. We have explained how one con- and in terms of language and develops a pedagog-
tribution of the life course perspective is its ical plan to act on this evaluation, particularly if
emphasis on the multiple levels of contexts. Here, the teacher has the support of parents, other teach-
we offer examples of different contexts of educa- ers, and administrators in this process.
Educational Pathways 195

Second, the meso level consists of larger and exposed to challenging coursework and higher-
more impersonal organizational settings that quality instruction (Carbonaro and Covay 2010).
organize and are inuenced by proximate ecolo- Third, the macro level consists of the diffuse
gies (e.g., peer groups, families) and connect and abstract machinery of society, the societal
them to broader social systems (e.g., stratication institutions and stratication systems that struc-
or political systems). The school is a good exam- ture social life and the micro-level and meso-level
ple. It houses classrooms and is in turn embedded contexts in which social life takes place. Perhaps
in the broader educational system (Arum 2000). the best example of a macro-level institutional
As illustrated by the lengthy and inuential context is the economy, the complex system in
Coleman Report in the 1960s, schools can be which goods and services are traded to balance
characterized by a number of characteristics and supply and demand. The economy is the macro-
processes that inuence student outcomes, both level setting that denes how educational attain-
academic and non-academic. To pick one exam- ment is valued and what it buys people, both
ple, school sector refers to whether a school is individually and as a population. We have already
part of the taxpayer-funded public system that is discussed how long-term changes in the economy
overseen by local, state, and federal governments have changed the nature of educational attain-
or whether it is privately supported by indepen- ment, especially increasing the long-term returns
dent organizations. An important distinction to higher education (Goldin and Katz 2008;
within the general parameter of school sector is Fischer and Hout 2006). We can also think about
Catholic schools, other private schools, and pub- how this long-term economic evolution encom-
lic schools. In general, private schools outperform passes many different shorter-term economic ups
public schools on most indicators of student and downs. Although the long-term trend in the
achievement, reecting a variety of factors, economy is towards higher rewards for college
including greater funding and lower student- education, shorter-term uctuations in the econ-
teacher ratios (Dronkers and Robert 2008; omy (e.g., recessions) have chipped away at these
Lubienski et al. 2009; Coleman and Hoffer 1987). increasing returns. On one hand, the current eco-
This private school advantage is even more pro- nomic climate shapes whether young people pur-
nounced for schools run by the Catholic church, sue higher education. Income effects occur when
especially for children who come from poor fami- hard economic times make people eschew educa-
lies or who are racial/ethnic minorities (Bryk tion in favor of work, as their pressing nancial
et al. 1993). Evidence for this Catholic school needs (and their families needs) mean that they
effect has often been attacked as spurious, the cannot afford to stick with education and instead
argument being that the observed success of need to try to bring in money. Substitution effects
Catholic schools is simply a function of students occur when hard economic times make people
more likely to succeed in the rst place being the concentrate on education over work, as the bad
ones who attend such schools rather than some economy shrinks the number of available jobs,
instructional or curricular advantage of Catholic which, in turn, lowers the opportunity costs of
school education (Hallinan and Kubitschek 2012; enrolling in higher education. In many countries,
Lee and Ready 2009). Efforts to address these evidence points towards income effects as more
concerns have revealed evidence that the Catholic common than substitution effects, but evidence of
school effect is at least in part real, especially for substitution effects (or educational warehousing)
the students who are least likely to select into has been found in the U.S., especially among
Catholic schools in the rst place (Morgan 2001). those from higher-SES backgrounds (Marteleto
In addition to discussions about the strong social et al. 2012; Torche 2010; John 2009; Werum
networks and supports in Catholic schools, this 2001; Shanahan et al. 1997; Felson and Land
causal impact is often attributed to the fact that 1978). On the other, the current economic climate
Catholic schools have more constrained curricula, shapes what happens to those who have pursued
which means that a broader array of students are higher education. Ample evidence from a variety
196 R. Crosnoe and A.D. Benner

of industrialized nations indicates that young peo- Both in terms of research and policy as well as
ple who graduate from college during a recession in the translation between them, the value of the
have lower earnings trajectories over the next two life course perspective for illuminating education
decades than comparable students who graduate and educational inequality is that it tells us when
during better economic times. The weak labor and where to look as well as why we need to look
market means that recession-era graduates are when and where we do. It highlights the impor-
underplaced initially, which then deates their tance of timing; for example, advocating for
upward trajectories (Oreopoulos et al. 2012; Kahn interventions early in the life course supple-
2010). Thus, general economic trends shape moti- mented by boosters later on in the life course, or
vation for and consequences of educational attain- pinpointing critical transition points. It also sug-
ment, and specic economic events qualify this gests that linked lives matter in uid ways; for
economic inuence in particular ways. example, the type of linked lives that require
Classrooms are contained within and orga- attention may change from childhood into ado-
nized by schools, and both are subject to the lescence, and some aspects of linked lives might
inuence of economic forces. At the same time, be more amenable to intervention than others.
what happens in classrooms shapes the academic Finally, it suggests that the rationale for doing
bottom line of schools, and how schools are something about education is that it has the
doing as a whole is a primary factor in economic potential for cascading benets; for example,
stability and expansion. The various social con- improving educational outcomes may enhance
texts in which the life course unfolds, therefore, health, family life, and many other barometers of
are difcult to disentangle. Even when research- a good life or a good society. Thus, the life course
ers attempt to dig down into and unpack one level perspective is a guide for organizing educational
of context, they should do so with the recognition research and pushing it forward, and educational
that it is part of a complex web of environmental research is a good exemplar of the power of the
inuences and processes. life course to promote scientically-based knowl-
edge about a major social issue.
As we have already noted, the integration of
4 Conclusion education and life course research is still rela-
tively new, and so more can be done. Instilling
A point that we have repeatedly made in this some of the merits of life course research within
chapter is that the proliferation of research at the educational research is one need, such as the
intersection of educational studies and life course emphasis on dynamism, contextual embedded-
studies in recent years means that no one review ness, and the cross-pollination between educa-
can fully capture the wide breadth and depth of tion and many other domains. Instilling some of
this integrated eld and, consequently, that we the merits of educational research within life
have to select specic examples to tell a larger course research is another, such as the value of
story. That story is that the educational life course establishing causal effects and of the two-way
is a multi-dimensional construct that, by connect- exchange between science and action. The two
ing the individual and population levels, offers elds must keep engaging with, learning from,
insight into societal inequality and informs efforts and teaching each other.
to reduce it. Put another way, what is going on
inside and around a person is both an effect and
cause of socioeconomic, racial/ethnic, and immi- References
gration-related inequality, and so exploring this
interplay across proximal to distal levels of social Alexander, K. L., & Entwisle, D. R. (1988). Achievement
in the rst two years of school: Patterns and processes.
life is a path towards solving social problems
Monographs of the Society for Research on Child
through evidence-based policy intervention. Development, 53, 1157.
Educational Pathways 197

Alexander, K., Entwisle, D., & Olson, L. (2014). The long Cherlin, A. (2009). The marriage-go-round. The state of
shadow: Family background, disadvantaged urban marriage and the family today. New York: Simon &
youth, and the transition to adulthood. New York: Schuster.
Russell Sage. Coleman, J. S., & Hoffer, T. (1987). Public and private
Apel, R., Bushway, S., Brame, R., Haviland, A. M., high schools: The impact of communities. New York:
Nagin, D. S., & Paternoster, R. (2007). Unpacking the Basic Books.
relationship between adolescent employment and anti- Coleman, J. S., Campbell, E. G., Hobson, C. J.,
social behavior: A matched samples comparison. McPartland, J., Mood, A. M., Weinfeld, F. D., &
Criminology, 45, 6797. York, R. L. (1966). Equality of educational opportu-
Arum, R. (2000). Schools and communities: Ecological nity. Washington, DC: U. S. Government Printing
and institutional dimensions. Annual Review of Ofce.
Sociology, 26, 395418. Conger, R. D., Conger, K. J., & Martin, M. J. (2010).
Attewell, P., & Domina, T. (2008). Raising the bar: Socioeconomic status, family processes, and individ-
Curricular intensity and academic performance. ual development. Journal of Marriage and Family, 72,
Educational Evaluation and Policy Analysis, 30, 5171. 685704.
Attewell, P., & Domina, T. (2011). Educational imposters Connor, C. M., Piasta, S. B., Fishman, B., Glasney, S.,
and fake degrees. Research in Social Stratication, 29, Schatschneider, C., Crowe, E., Underwood, P., &
5769. Morrison, F. J. (2009). Individualizing student instruc-
Autor, D. (2014). Skills, education, and the rise of earn- tion precisely: Effects of child instruction interac-
ings inequality among the other 99 percent. Science, tions on rst graders literacy development. Child
344, 843851. Development, 80, 77100.
Barber, B. K., & Olsen, J. A. (2004). Assessing the transi- Crosnoe, R. (2009). Familyschool connections and the
tions to middle and high school. Journal of Adolescent transitions of low-income youths and English lan-
Research, 19, 330. guage learners from middle school to high school.
Benner, A. D. (2011). The transition to high school: Developmental Psychology, 45, 1061.
Current knowledge, future directions. Educational Crosnoe, R. (2011). Fitting in, standing out: Navigating
Psychology Review, 23, 299328. the social challenges of high school to get an educa-
Benner, A. D., & Graham, S. (2009). The transition to tion. New York: Cambridge.
high school as a developmental process among multi- Crosnoe, R., & Benner, A. (2015). Children at school. In
ethnic urban youth. Child Development, 80, 356376. M. Bornstein & T. Leventhal (Eds.), Handbook of
Bernhardt, A., Morris, M., Handcock, M. S., & Scott, M. A. child psychology and developmental science
(2001). Divergent paths: Economic mobility in the new (Ecological Settings and Processes (series editor:
American labor market. New York: Russell Sage. Richard M. Lerner), Vol. 4). New York: Wiley.
Bonnie, R. J., Stroud, C., & Breiner, H. (Eds.). (2014). Crosnoe, R., & Cavanagh, S. (2010). Families with chil-
The health and well-being of young adults: A critical dren and adolescents: A review, critique, and future
time. Washington, DC: Board on Children, Youth, and agenda. Journal of Marriage and Family, 72, 118.
Families, Institute of Medicine of the National Crosnoe, R., & Cooper, C. (2010). Economically disad-
Academies. vantaged childrens transitions into elementary school:
Bowles, S., & Gintis, H. (1976). Schooling in capitalist Linking family processes, school context, and educa-
America (Vol. 57). New York: Basic Books. tional policy. American Educational Research Journal,
Bozick, R., & DeLuca, S. (2005). Better late than never? 47, 258291.
Delayed enrollment in the high school to college tran- Crosnoe, R., & Johnson, M. K. (2011). Research on ado-
sition. Social Forces, 84, 531554. lescence in the 21st century. Annual Review of
Bradley, R. H., & Corwyn, R. F. (2002). Socioeconomic Sociology, 37, 439460.
status and child development. Annual Review of Crosnoe, R., & Kalil, R. (2010). Educational progress and
Psychology, 53, 371399. parenting among Mexican immigrant mothers of
Breen, R., & Goldthorpe, J. H. (1997). Explaining educa- young children. Journal of Marriage and Family, 72,
tional differentials towards a formal rational action 976990.
theory. Rationality and Society, 9(3), 275305. Crosnoe, R., & Muller, C. (2014). Family socioeconomic
Bryk, A. S., Lee, V. E., & Holland, P. B. (1993). Catholic status, peers, and adolescents path to college. Social
schools and the common good. Cambridge, MA: Problems, 61, 123.
Harvard Press. Davis-Kean, P. (2005). The inuence of parent education
Buchmann, C., & Park, H. (2009). Stratication and the and family income on child achievement: The indirect
formation of expectations in highly-differentiated edu- role of parent expectations and the home environment.
cational systems. Research in Social Stratication and Journal of Family Psychology, 19, 294304.
Mobility, 27, 245267. DiPrete, T. A., & Eirich, G. M. (2006). Cumulative advan-
Carbonaro, W., & Covay, E. (2010). School sector and tage as a mechanism for inequality: A review of theo-
student achievement in the era of standards based retical and empirical developments. Annual Review of
reforms. Sociology of Education, 83, 160182. Sociology, 32, 271297.
198 R. Crosnoe and A.D. Benner

Dronkers, J., & Robert, P. (2008). Differences in scholas- Goldthorpe, J. H. (2000). On sociology: Numbers, narra-
tic achievement of public, private government- tives, and the integration of research and theory.
dependent, and private independent schools a Oxford: Oxford University Press.
cross-national analysis. Educational Policy, 22(4), Hallinan, M., & Kubitschek, W. N. (2012). A comparison
541577. of academic achievement and adherence to the com-
Duncan, G. J., & Magnuson, K. (2013). Investing in pre- mon school ideal in public and catholic schools.
school programs. Journal of Economic Perspectives, Sociology of Education, 85, 122.
27(2), 109132. Hao, L., & Woo, H. S. (2012). Distinct trajectories in the
Eccles, J. (2004). Schools, academic motivation, and transition to adulthood: Are children of immigrants
stage-environment t. In R. Lerner & L. Steinberg advantaged? Child Development, 83, 16231639.
(Eds.), Handbook of adolescent psychology (pp. 125 Harris, A. (2006). I (dont) hate school: Revisiting oppo-
153). Hoboken: Wiley. sitional culture theory of blacks resistance to school-
Eccles, J., Midgely, C., Wigeld, A., Buchanan, C. M., ing. Social Forces, 85, 797834.
Reuman, D., & MacIver, D. (1993). Development dur- Hauser, R. M., & Koenig, J. A. (2011). High school drop-
ing adolescence: The impact of stage-environment t out, graduation, and completion rates: Better data,
on young adolescents experiences in schools and in better measures, better decisions. Washington, DC:
families. American Psychologist, 48, 90101. National Academy Press.
Elder, G. H., Jr. (1974). Children of the great depression: Jackson, M., Erikson, R., Goldthorpe, J. H., & Yaish, M.
Social change in life experience. Chicago: University (2007). Primary and secondary effects in class differ-
of Chicago. entials in educational attainment the transition to
Elder, G. H. (1998). The life course as developmental A-level courses in England and Wales. Acta
theory. Child Development, 69, 112. Sociologica, 50, 211229.
Entwisle, D. R., Alexander, K. L., & Olson, L. S. (2005). John, I. (2009). Economic scarring: The long-term
First grade and educational attainment by age 22: A impacts of the recession. Washington, DC: Economic
new story. American Journal of Sociology, 110, Policy Institute.
14581502. Johnson, M. K. (2004). Further evidence on adolescent
Erikson, R., Goldthorpe, J. H., Jackson, M., Yaish, M., & employment and substance use: Differences by race
Cox, D. R. (2005). On class differentials in educa- and ethnicity. Journal of Health and Social Behavior,
tional attainment. Proceedings of the National 45, 187197.
Academy of Sciences of the United States of America, Kahn, L. B. (2010). The long-term labor market conse-
102, 97309733. quences of graduating from college in a bad economy.
Federal Interagency Forum on Child and Family Statistics. Labour Economics, 17, 303316.
(2014). Americas young adults: Special issue, 2014. Kalil, A., Ryan, R., & Corey, M. (2012). Diverging desti-
Washington, DC: U.S. Government Printing Ofce. nies: Maternal education and the developmental gradi-
Felson, M., & Land, K. (1978). Social, demographic and ent in time with children. Demography, 49, 123.
economic inter-relationships with educational trends Kane, T. J. (1999). The price of admission: Rethinking
in the United States, 19471974. In J. L. Simon (Ed.), how Americans pay for college. Washington, DC:
Research in population economics (pp. 93125). Brookings Institution.
Greenwich: JAI Press. Kao, G. (2004). Parental inuences on the educational
Fischer, C. S., & Hout, M. (2006). Century of difference: outcomes of immigrant youth. International Migration
How America changed in the last one hundred years. Review, 38, 427449.
New York: Russell Sage. Kariya, T., & Rosenbaum, J. E. (2003). Stratied incen-
Gandara, P. (2002). A study of high school Puente: What tives and life course behaviors. In J. Mortimer &
we have learned about preparing Latino youth for M. Shanahan (Eds.), Handbook of the life course.
postsecondary education. Educational Policy, 16, New York: Kluwer/Plenum.
472495. Kelly, S. (2008). What types of students effort are
Garca Coll, C., & Marks, A. K. (2009). Immigrant sto- rewarded with high marks? Sociology of Education,
ries: Ethnicity and academics in middle childhood. 81, 3252.
New York: Oxford University Press. Kerckhoff, A. C. (1993). Diverging pathways: Social
George, L. K. (1993). Sociological perspectives on life structure and career deections. Cambridge, UK:
transitions. Annual Review of Sociology, 19, 353373. Cambridge University Press.
Glick, J. E., & White, M. J. (2003). The academic trajec- King, C., Smith, T., & Glover, R. (2011). Investing in chil-
tories of immigrant youths: Analysis within and across dren and parents: Fostering dual-generation strate-
cohorts. Demography, 40, 589603. gies in the United States. New York: Foundation for
Golash-Boza, T. (2005). Assessing the advantages of Child Development.
bilingualism for the children of immigrants. Kingston, P. W., Hubbard, R., Lapp, B., Schroeder, P., &
International Migration Review, 39(39), 721753. Wilson, J. (2003). Why education matters. Sociology
Goldin, C., & Katz, L. (2008). The race between technol- of Education, 76, 5370.
ogy and education. Cambridge, MA: Harvard Kinney, D. (1999). From headbangers to hippies:
University Press. Delineating adolescents active attempts to form an
Educational Pathways 199

alternative peer culture. New Directions for Child and Morgan, S. L. (2001). Counterfactuals, causal effect het-
Adolescent Development, 84, 2135. erogeneity, and the Catholic school effect on learning.
Langenkamp, A. G. (2010). Academic vulnerability and Sociology of Education, 74, 341374.
resilience during the transition to high school: The role Morgan, S. L. (2005). On the edge of commitment:
of social relationships and district context. Sociology Educational attainment and race in the United States.
of Education, 83, 119. Stanford: Stanford University Press.
Lareau, A. (2003). Uneqal childhoods: Class, race, and Mortimer, J. T. (2003). Working and growing up in
family life. Berkeley: The University of California America. Cambridge: Harvard University Press.
Press. Musick, K., Brand, J. E., & Davis, D. (2012). Variation in
Lauderdale, D. (2001). Education and survival: Birth the relationship between education and marriage:
cohort, period, and age effects. Demography, 38, Marriage market mismatch? Journal of Marriage and
551561. Family, 74, 5369.
Lee, V. E., & Ready, D. D. (2009). US high school cur- Ogbu, J. U. (1997). African American education: A
riculum: Three phases of contemporary research and cultural-ecological perspective. In H. P. McAdoo
reform. Future of Children, 19, 135156. (Ed.), Black families (pp. 234250). Thousand Oaks:
Lee, J. C., & Staff, J. (2007). When work matters: The Sage.
varying impact of work intensity on high school drop- Oreopoulos, P., von Wachter, T., & Heisz, A. (2012). The
out. Sociology of Education, 80, 158178. short-and long-term career effects of graduating in a
Lerner, R., & Galambos, N. L. (1998). Adolescent devel- recession. American Economic Journal: Applied
opment: Challenges and opportunities for research, Economics, 4, 129.
programs, and policies. Annual Review of Psychology, Padilla, A. M., & Gonzalez, R. (2001). Academic perfor-
49, 413446. mance of immigrant and US-born Mexican heritage
Lubienski, C., Weitzel, P., & Lubienski, S. T. (2009). Is students: Effects of schooling in Mexico and bilingual/
there a consensus on school choice and achieve- English language instruction. American Educational
ment? Advocacy research and the emerging political Research Journal, 38, 727742.
economy of knowledge production. Educational Palloni, A. (2006). Reproducing Inequalities: Luck, wal-
Policy, 23, 161193. lets, and the enduring effects of childhood health.
Lynch, S. (2003). Cohort and life course patterns in the Demography, 43, 587615.
relationship between education and health: A hierar- Patrick, M. E., Schulenberg, J. E., & OMalley, P. M.
chical perspective. Demography, 40, 309332. (forthcoming). High school substance use as a predic-
Marteleto, L., Gelber, D., Hubert, C., & Salinas, V. (2012). tor of college attendance, completion, and dropout: A
Educational inequalities among Latin American ado- national multicohort longitudinal study. Youth &
lescents: Continuities and changes over the 1980s, Society.
1990s and 2000s. Research in Social Stratication and Pianta, R. C., Cox, M., & Snow, K. (2007). School readi-
Mobility, 30, 352375. ness and the transition to kindergarten in the era of
Massey, D. S., Charles, C. Z., Lundy, G., & Fischer, M. J. accountability. Baltimore: Brooks.
(2011). The source of the river: The social origins of Pong, S. L., & Hao, L. (2007). Neighborhood and
freshmen at Americas selective colleges and universi- school factors in the school performance of immi-
ties. Princeton: Princeton University Press. grants children. International Migration Review,
McFarland, D. A. (2006). Curricular flows: 41, 206241.
Trajectories, turning points, and assignment crite- Propper, C. B., Shanahan, M. J., Russo, R., & Mills-
ria in high school math careers. Sociology of Koonce, W. R. (2012). Evocative geneparenting cor-
Education, 79, 177205. relations and academic performance at rst grade: An
McLanahan, S. (2004). Children and the second demo- exploratory study. Development and Psychopathology,
graphic transition. Demography, 41, 607628. 24, 12651282.
McMorris, B. J., & Uggen, C. (2000). Alcohol and Raver, C. C., Gershoff, E., & Aber, L. (2007). Testing
employment in the transition to adulthood. Journal of equivalence of mediating models of income, parent-
Health and Social Behavior, 41, 276294. ing, and school readiness for White, Black, and
Miech, R., Pampel, F., Kim, J., & Rogers, R. G. (2011). Hispanic children in a national sample. Child
The enduring association between education and Development, 78, 96115.
mortality the role of widening and narrowing dis- Reardon, S., & Galindo, C. (2009). The Hispanic-White
parities. American Sociological Review, 76, achievement gap in math and reading in the elemen-
913934. tary school grades. American Educational Research
Mirowsky, J., & Ross, C. E. (2003). Education, social sta- Journal, 46, 853891.
tus, and health. New York: Aldine de Gruyter. Riegle-Crumb, C. (2006). The path through math: Course
Mistry, R., Biesanz, J. C., Taylor, L., Burchinal, M., & sequences and academic performance at the intersec-
Cox, M. J. (2004). Family income and its relation to tion of race/ethnicity and gender. American Journal of
pre-school childrens adjustment for families in the Education, 113, 101122.
NICHD Study of Early Child Care. Developmental Riegle-Crumb, C., & King, B. (2010). Questioning a
Psychology, 40, 727745. white male advantage in STEM: Examining dispari-
200 R. Crosnoe and A.D. Benner

ties in college major by gender and race/ethnicity. Siegler, R. S., Duncan, G. J., Davis-Kean, P. E.,
Educational Researcher, 39(9), 656664. Duckworth, K., Claessens, A., Engel, M., & Chen,
Ross, C. E., & Wu, C. L. (1995). The links between educa- M. (2012). Early predictors of high school mathe-
tion and health. American Sociological Review, 60, matics achievement. Psychological Science, 23,
719745. 691697.
Rudolph, K. D., Lambert, S. F., Clark, A. G., & Simmons, R. G., & Blyth, D. A. (1987). Moving into ado-
Kurlakowsky, K. D. (2001). Negotiating the transition lescence: The impact of pubertal change and school
to middle school: The role of self-regulatory pro- context. Hawthorne: Aldine de Gruyter.
cesses. Child Development, 72, 929946. Smith, S. (Ed.). (1995). Two generation programs for
Ryan, A. M. (2001). The peer group as a context for the families in poverty: A new intervention strategy.
development of young adolescent motivation and Norwood: Ablex.
achievement. Child Development, 72, 11351150. Steele, C. M. (1997). A threat in the air: How stereotypes
Sacerdote, B. (2011). Peer effects in education: How shape intellectual identity and performance. American
might they work, how big are they and how much do Psychologist, 52, 613.
we know thus far? Handbook of the Economics of Stevenson, D. L., Schiller, K. S., & Schneider, B. (1994).
Education, 3, 249277. Sequences of opportunities for learning. Sociology of
Schneider, B., & Stevenson, D. (1999). The ambitious Education, 67, 184198.
generation: Americas teenagers, motivated but direc- Tinto, V. (1987). Leaving college: Rethinking the causes
tionless. New Haven: Yale University Press. and cures of student attrition. Chicago: University of
Schneider, B., Swanson, C. B., & Riegle-Crumb, C. Chicago Press.
(1999). Opportunities for learning: Course sequences Torche, F. (2010). Economic crisis and inequality of edu-
and positional advantages. Social Psychology of cational opportunity in Latin America. Sociology of
Education, 2, 2553. Education, 82, 85110.
Schoen, R., & Cheng, A. (2006). Partner choice and the Tyson, K., Darity, W., & Castellino, D. R. (2005). Its not
differential retreat from marriage. Journal of Marriage a black thing: Understanding the burden of acting
and Family, 68, 110. white and other dilemmas of high achievement.
Schoon, I., & Silbereisen, R. K. (Eds.). (2009). Transitions American Sociological Review, 70, 582605.
from school to work: Globalization, individualization, Warren, J. R., Grodsky, E., & Lee, J. C. (2008). State
and patterns of diversity. Cambridge, UK: Cambridge high school exit examinations and postsecondary
University Press. labor market outcomes. Sociology of Education, 81,
Schulenberg, J. E., & Maggs, J. (2002). A developmental 77107.
perspective on alcohol use and heavy drinking during Washbrook, E., Waldfogel, J., Bradbury, B., Corak, M., &
adolescence and the transition to young adulthood. Ghanghro, A. A. (2012). The development of young
Journal of Studies on Alcohol, 63, 5470. children of immigrants in Australia, Canada, the
Seidman, E., Lambert, L. E., Allen, L., & Aber, J. L. United Kingdom, and the United States. Child
(2003). Urban adolescents transition to junior high Development, 83, 15911607.
school and protective family transactions. Journal of Werum, R. (2001). Warehousing the unemployed? Federal
Early Adolescence, 23, 166193. job training programs in the depression-era south.
Shanahan, M. J., Elder, G., & Miech, R. A. (1997). History American Journal of Education, 109, 228265.
and agency in mens lives: Pathways to achievement in Williams, D., & Collins, C. (1995). U.S. socioeconomic
cohort perspective. Sociology of Education, 70, 5467. and racial differences in health: Patterns and explana-
Siegler, R. S. (2009). Improving the numerical under- tions. Annual Review of Sociology, 21, 349386.
standing of children from lowincome families. Child Yeung, A. S., Marsh, H. W., & Suliman, R. (2000). Can
Development Perspectives, 3, 118124. two tongues live in harmony: Analysis of the National
Siegler, R. S., & LortieForgues, H. (2014). An integrative Education Longitudinal Study of 1988 (NELS88) lon-
theory of numerical development. Child Development gitudinal data on the maintenance of home language.
Perspectives, 8, 144150. American Educational Research Journal, 37,
10011026.
College for All: New Institutional
Conicts in the Transition
to Adulthood

Claudia Zapata-Gietl, James E. Rosenbaum,


Caitlin Ahearn, and Kelly Iwanaga Becker

1 Introduction conicts: how the student role conicts with


other roles, in particular family and work roles.
Societal context is crucial for understanding how We also examine simultaneous time conicts:
individuals move through life stages (Elder how college timetables interact with timetables
1998). The current transition to adulthood is for other life events associated with emerging
dened by delays experienced in post-secondary adulthood.
education, career exploration, marriage, and par- Finally, we examine sequential conicts: how
enthood. Less obvious is the interplay between transitions into college, within college, and after
delays in post-secondary education and the tran- college might create conicting expectations, dif-
sitions in the domains of work and family. culties, and delays in college timetables.
Dramatically increased earnings payoffs for Mortimer and Kruger (2000) contend that clear
higher education and a prevailing emphasis on school-work connections may help shape youths
college for all have encouraged more young career plans and outcomes so they can anticipate
adults to plan higher credentials. While these new and prepare for future demands. In contrast, in
changes to the transition to adulthood and the U.S., the vague connections between school-
expanded education can create positive opportu- ing and working, combined with the American
nities for growth and exploration, this also comes ideology of equal, rather unlimited, and ever-
with serious costs and risks, particularly for low- available opportunity, can stimulate quite unreal-
income students who lack the resources for pro- istic thinking about future work roles (Mortimer
longed education and career exploration. and Kruger 2000, p. 485). This analysis suggests
This chapter examines how community col- that poorly structured transitions may create
leges shape the life course of individuals transi- problems for youths planning and progress and
tioning into adulthood, and whether alternative may contribute to the normative long and wind-
institutional structures and better informed ing road described by the concept of emerging
students might reduce some of the unintended adulthood (Arnett 2000, 2004). In other words,
conicts with other life domains. We focus on students life plans may be constrained or aided
several conicts. We examine simultaneous role by their position in and trajectory through educa-
tional systems (Shanahan 2000). We conclude by
suggesting that improved transition structures
C. Zapata-Gietl (*) J.E. Rosenbaum
may reduce sequential, role and time conicts
C. Ahearn K.I. Becker
Northwestern University, Evanston, IL, USA and avoid the problems imposed by extended
e-mail: claudiazapata2015@u.northwestern.edu educational timetables.

Springer International Publishing Switzerland 2016 201


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_9
202 C. Zapata-Gietl et al.

2 Community College Context credential 8 years after high school graduation


and 44 % who enter 4-year colleges with low aca-
The vast expansion of post-secondary education demic achievement do not have a degree after 8
has created extended higher education timelines years (Rosenbaum et al. 2014). Although most
that now characterize the transition to adulthood. high school graduates make it to college, many
In recent decades, U.S. society has made a push do not succeed, and although much research
to give every student the opportunity to get a col- examines the academic challenges they face, rec-
lege degree. In fact, in a recent national cohort onciling the demands of college and other life
(ELS 20022012), virtually all (96 %) high demands is unresolved. We must understand how
school graduates plan to attend college and over this new educational demand ts into the lives of
86 % enroll in college in the 8 years after high youth who previously would not have attended
school (Rosenbaum et al. 2014; cp. Adelman college. In addition, both students and their fami-
2003). This college-for-all policy makes col- lies may have unrealistic expectations of the
leges responsible for shaping the opportunities ways in which college will change the course of
for social mobility of many students who have their lives.
new access to attend college. Todays labor mar- Community colleges, like all higher education
ket increasingly demands higher education cre- institutions, communicate expected degree time-
dentials, making college-for-all, with a broad lines to students, which are based on traditional
denition of college, an appropriate policy. patterns created before higher education became
With their low tuition, convenient locations, widespread. Unfortunately, these expectations
convenient schedules, and open admissions, are frequently hard to meet. Completion of a
community colleges are a common answer to degree now often takes much longer than the con-
these labor market demands. Community col- ventional expectation of a 4-year bachelors
leges enroll nearly half of all college students, degree and a 2-year associates degree (Bound
including many from underserved populations. In et al. 2009). Only 7.1 % of full-time students
2010, community college enrollment was almost complete an associates degree in 2 years and
as high as that of public 4-year college enroll- 31.5 % of full-time students complete a bache-
ment (NCES 2012). Many of these students pur- lors degree in 4 years (Complete College
sue sub-BA credentials such as career certicates America 2011). In fact, the 4-year BA takes an
and associates degrees (AA), which take less average of almost 6 years, while a 2-year associ-
time than a bachelors degree (BA). Since 1969, ates degree takes 34 years. Indeed, a substan-
while BA degrees have doubled, certicates and tial portion of students pursuing 4-year BAs take
AA degrees (largely offered at community col- 8 years or longer, conicting with their initial
leges) have increased fourfold (NCES 2012; nd.). expected timetables (Bound et al. 2009;
About 1.8 million students in 2010 completed a Rosenbaum et al. 2014). This means that while
certicate or associates degree (NCES 2012). students expect to nish their education early in
Because they serve so many youth from disad- their transition to adulthood (by age 22), in real-
vantaged backgrounds (Rosenbaum et al. 2006), ity they may not complete their degrees until they
community colleges are important contexts of are well into their 20s, if they complete at all.
emerging adulthood to consider. This chapter reviews the potential conse-
While critics say college for all is unfeasible quences of these extended degree timetables in
(Murray 2008), and advocates consider it essen- terms of the timing of other adult roles, as well as
tial in the contemporary labor market (Goldin some of the transitional complications that may
and Katz 2008), the reality is that American soci- contribute to longer degree times. Low-income
ety has already persuaded virtually all (96 %) students attending 4-year schools may nd that
high school graduates to have college plans. outside commitments (work and family) as well
Unfortunately, almost half (46 %) of youth who as lack of information may extend timelines
rst enter community colleges do not have any beyond their expectations. We focus on the com-
College for All: New Institutional Conicts in the Transition to Adulthood 203

munity college context because it serves many asserts that the timing of events and experiences
disadvantaged and non-traditional students, who can affect individuals in different ways depend-
might suffer more from extended degree timeta- ing on when they occur in the life course. College
bles. In addition, the exibility community col- attendance has markedly different consequences
leges offer (in an effort to serve the needs of for individuals depending on whether they occur
untraditional students) illustrates how poorly- according to normative timetables, or later in the
structured transitions may contribute to pro- life course. Finally, individuals exist within inter-
longed time lines. However, many of the issues personal networks, and their life course may be
we discuss also pertain to students in 4-year inuenced by other individuals life course
colleges. events, which illustrates the principle of linked
lives. Adjustment to community college life may
be importantly affected by others expectations
3 The Shaping of the Transition and support.
to Adulthood More recently, developmental psychologists
have turned to the concept of emerging adulthood
The transition to adulthood has long received to describe a stage of life that encompasses the
attention from scholars trying to understand how explorations and transitions of 1829 year-olds,
adolescents successfully transition into adult ages when individuals try out and adopt various
roles in society. Life course theory provides a adult roles (like spouse, parent, student, employee
paradigm that takes societal context into account and head of household; Arnett 2000). A primary
when studying individuals lives. Within this par- criticism of the concept of emerging adulthood as
adigm are ve principles (1) life-span develop- a developmental stage lies in questions about
ment, (2) agency, (3) time and place, (4) timing how this stage may be experienced by individuals
and (5) linked lives (Elder 1998). We will outline from varying socioeconomic backgrounds and
each briey within the context of colleges inu- societal positions (Hendry and Kloep 2010; Ct
ence on individuals lives. 2014). In particular, we examine how youth, par-
The principle of life span development states ticularly disadvantaged youth, make choices and
that human development and aging are lifelong actions within the constraints of educational con-
processes, with earlier experiences affecting later texts, which may inuence how they plan their
outcomes. Early educational experiences and lives and enter into the labor market.
accomplishments have life-long consequences Arnett describes emerging adults as entering
for occupational and income attainments and and exiting adult roles at will, and suggests that
adult life styles. Students development may be post-secondary institutions are merely a possible
inuenced at any age by social contexts. The venue for their explorations (Arnett 2000).
principle of agency highlights the active role Instead, we argue that in a historical context
individuals take in constructing their future tra- where college attendance is nearly universal and
jectories through their choices and actions. The most young adults are exposed to college, these
ways educational transitions are currently struc- institutions dene and inuence the roles that
tured place major limits on their exercise of youth adopt. Instead of being merely passive set-
agency because students often lack information tings in which young adults explore their options
about options and their implications. The princi- and interests, we contend that post-secondary
ple of time and place emphasizes that these institutions may actively shape this life transi-
choices and actions occur within the constraint of tion. Furthermore, how they shape this life transi-
a particular historical time and place. The rapid tion may vary widely in relation to students
expansion of community colleges in recent individual place in their life course, their linkages
decades, and the challenges they pose to young with family and social networks, and their previ-
people in process of transition to adulthood, well ous educational experiences and attainments.
illustrates this principle. The principle of timing
204 C. Zapata-Gietl et al.

While Arnett acknowledges the job market Although emerging adulthood delivers the mes-
effects of educational attainment, he portrays sage that extension of the timing of individuals
relatively minor and un-stressful demands from trajectories through adult transitions is penalty-
education where students have lives structured free, we point to the ways in which institutional
around going to classes and doing coursework and societal age norms related to education may
(Arnett 2011, 2006). However, many community create potentially difcult conicts with timeta-
college students do not necessarily follow tradi- bles for other adult role demands.
tional school-to-family timelines and instead Institutions can shape options and perceptions
often marry or have children prior to graduation. during various stages of the life course, and the
Education creates demands and obligations that transition to adulthood is no exception.
may compete with these other roles students Community colleges provide many low-income
adopt. This competition between roles may inu- students with valuable opportunities for higher
ence, and be inuenced by, their college experi- education, which contribute to the life course
ence. The luxury to explore educational through employment opportunities and social
opportunities is especially complicated for low- connections. However, despite their great poten-
income students who must manage competing tial to benet emerging adults, community col-
roles and expectations while enrolled in college. leges present other, potentially harmful
In contrast to Arnett, Pallas (1993) argues that consequences by delaying credential completion.
schooling inuences the sequencing of adult Not only do delays in degree completion impact
roles in other domains, particularly marriage and entry into adult roles, they often result in
parenthood. Pallas envisions education as having increased tuition costs, encroachment on nan-
a strong social impact on the timing, sequencing, cial aid limits, forgone wages, delayed entry into
and even disruption of various major life events. careers and increased demands on and disap-
Pallas suggests that those who remain in school pointments from students support networks.
longer may be unable to fulll both marriage and Further, community colleges fail to offer ways to
parenthood responsibilities and that education diminish these consequences or to explain how
leads to the postponement of marriage and par- students can anticipate and deal with them.
enthood. In other words, men and women delay Indeed, by not announcing the frequency of such
marriage and parenthood not merely because of time delays, students and their family members
the labor market advantages of delay, but because may see delays as indicating individual failure,
school enrollment leaves less time and energy for suggesting low ability and poor prospects.
marriage and children, and less certainty about
future plans. While students expectations may
align with Pallas predictions, students actual 4 Time is the Enemy:
timing for adult transitions may be untraditional. Simultaneous Conicts
In line with Pallas, we suggest that higher edu- in Different Life Domains
cation institutions, where many students spend a
large portion of the transition into adulthood, Life course theorists suggest that societal, histori-
shape expectations for and timing of that transi- cal, and personal contexts shape the sequencing
tion. We argue that the community college con- of emerging adults adult roles, and they make
text in particular imposes constraints that demands and expectations that alter life choices
increasingly lead to conicts with timelines of and plans (Settersten and Hgestad 1996). The
other life decisions made during this life stage. principle of timing is clearly evident in the edu-
While schooling has always been in some con- cational realm. That is, much education literature
ict with the roles of parent, spouse and worker, has focused on how delays may impact education
extending the time spent in post-secondary insti- outcomes. For example, delayed entry to
tutions has prolonged these conicts for an post-secondary education is associated with
increasing number of students in recent years. lower likelihood of degree completion (Bozick
College for All: New Institutional Conicts in the Transition to Adulthood 205

and DeLuca 2005; Complete College America ing of other adult role transitions (marriage,
2011). Other kinds of changes to the expected parenthood and transition out of their parents
sequencing of adult roles are also associated with home). Using multiple regression analysis, they
lower likelihood of degree completion. For found that higher educational expectations were
example, unmarried parents enrolled in college related to delayed expectations for marriage and
are at greater risk of failing to complete their parenthood.
degree (Goldrick-Rab and Sorenson 2010). These The study rst examined adult transition time-
studies are in line with the principle of timing lines without considering degree expectations.
that points to the importance of understanding Consistent with the premise that emerging adult-
how events and experiences affect individuals hood ends at 30 (implying that most have com-
differently at different points during the life pleted adulthood transitions at this age), older
course. The exibility and accessibility of com- students expectations for transitions tended to
munity colleges provide access to education for converge around age 30. For example, while a
individuals at many different stages in the life childless 18 year old expected to wait 8 years
course. (until age 26) for their rst child, a childless 30
Although new educational access has positive year old expected to wait only 4 years (until age
implications for opportunity, it may lead to pro- 34) for their rst child (Table 1). Similar patterns
longed degree timetables that may also conict are evident for age of marriage, but not for mov-
with life events outside of schooling (Settersten ing out of parent household. This suggests that
and Ray 2010; Furstenberg et al. 2004; Settersten these students shared similar norms about when
and Hgestad 1996). Students who spend they personally would transition into marriage
extended time in college may be stuck in associ- and parenting roles.
ated life circumstances that delay self-sufciency, The authors predicted that student expecta-
reduce wages, and increase dependence on fam- tions for the timing of marriage and parenthood
ily support. While these delays can be seen as would reect expected degree timetables. That is,
exibility, expected under emerging adulthood students pursuing AAs should expect to marry
theory, education delays may conict with the earlier than those who planned on pursuing a
implicit and explicit expected timetables of other BA. Multiple regression analysis showed that
institutions, especially for low-income individu- after controls, degree plans for BAs and AAs
als. For example, prolonged degree timetables
may create delayed career entry into adult jobs Table 1 Average expected age for various life
in the primary labor market (Doeringer and Piore transitions
1971), which may prevent young adults from get- Age Marriage First child Full time job Household
ting training and advancement opportunities. 18 25.89 26.72 21.46 21.79
These delays may also affect transitions into 19 25.99 27.12 22.22 22.87
careers, which we will discuss later. 20 26.66 28.3 23.02 23.28
Life course theorists suggest that such con- 21 26.85 28.17 23.74 24.47
texts for transitioning adults may force students 22 27.19 28.64 24.4 25.23
to contend with delays and postponements in 23 27.89 29.76 25.95 26.46
sequencing the adoption of adult roles, including 24 28.8 30.04 26.32 26.88
entering marriage and parenthood. Zapata-Gietl 25 29.42 30.53 27.41 27.56
and Rosenbaum (2013) examined the hypothesis 26 29.58 30.48 28.75 28.75
that educational timetables can conict with 27 30.67 32.07 30.33 29.75
expectations for the adult roles of marriage and 28 31.89 32.42 29.75 31.62
parenthood. The authors used survey data from 29 32.23 33 31.11 33
1,273 students in ten community colleges (ages 30 33.59 34.33 32.25 36.5
1830) to analyze how the degree plans of emerg- Sample: Students attending ten community colleges,
ing adults related to their expectations of the tim- N = 1,273
206 C. Zapata-Gietl et al.

were strongly related to increased duration until are signicantly longer for students with BA and
expected marriage, parenthood, and rst full- AA plans than for students planning certicates,
time job, although they had no impact on timeta- the expectations for these life events do not differ
bles for leaving home. In other words, students very much between those planning AA and BA
who had plans to pursue an AA or BA (as opposed degrees, and the difference is not signicant.
to a certicate) expected to wait longer to get Their expectations do not seem to reect the full
married, have children, or obtain their rst full- degree timetables, with the BAs expecting to wait
time employment than students pursuing certi- only 2 months longer than AAs to marry. This
cates (which typically take about 1 year; See may indicate that although they recognize their
Table 2). These delays suggest that community education plans will extend their timetables, stu-
college students expected degree timetables are dents feel constrained in extending their time-
related to later planned timelines for other life lines further, perhaps by implicit age norms. For
events, compared to students with plans for example, in planning age of rst child, many
quicker education credentials. The nding that women may feel constrained by well-publicized
leaving home is not inuenced by these longer biological timetables. However, the ndings are
educational timelines raise concerns about similar for age of marriage and age of full-time
whether these added household responsibilities job, and similar patterns are seen for men, sug-
may require more time and nancial resources gesting a more general social normative inu-
than when students are living at home. ence, not a biological inuence. Students
However, even if expecting an AA or BA has a planning BAs may not want to delay these life
strong impact compared with students planning events until after they graduate, or students may
certicates, students expected age adjustments not have reconciled actual degree time with age
may not reect the full impact of these longer norms for life events. Alternatively, students may
degrees. While marriage and parenting timetables be optimistic about degree timetables, either

Table 2 Regression estimates for expected duration until life transitions


Establishing
Marriage First child Full time job independent household
Associates degree 0.96** 1.18** 0.67** 0.23
Bachelors degree 1.10** 1.19** 0.99*** 0.60
Certicate (omitted)
None 0.78 0.28 0.59 0.61
Blacka 0.82* 0.56 0.55* 0.12
Latino 1.00** 0.49 0.11 0.64*
Asian or pacic islander 0.49 0.46 0.44 1.31***
Other 0.054 0.06 0.0075 0.28
Female 0.98*** 0.61* 0.22 0.36
Mother has BA 0.19 0.57 0.023 0.018
Explore agreeb 0.56 0.19 0.13 0.24
Explore middlec 0.29 0.14 0.24 0.35
Constant 4.34*** 6.09*** 1.77*** 2.41***
N 736 658 601 657
Sample: Students attending ten community colleges, N = 1,273
a
Omitted categories are Certicate degree, White, Explore- Disagree
b
Explore Agree = Student Agrees or strongly agrees that they want to explore career options even if their degree takes
longer
c
Explore Middle = Student indicated they were neutral about wanting to explore career options even if their degree
takes longer
College for All: New Institutional Conicts in the Transition to Adulthood 207

from lack of information about average time to societal age norms for completing their formal
degree or a belief that they will not be one of the education, however, family support faces unex-
many who take longer to graduate. Remarkably, pected demands; families anticipate providing
students mothers education (our indicator of this support for an institutionally dened amount
socioeconomic status) showed no relationship of time, and may have limited resources or
with students expected age for adult role transi- patience beyond that deadline. We suggest that
tions. This suggests that students expectations one unforeseen consequence of delayed age
are not merely following their parents example. timelines is that students may receive declining
Although the role of institutions, especially edu- family support as they approach and exceed the
cation, in emerging adulthood has been portrayed traditional age for exiting formal education.
in contradictory ways (both inconsequential and Our community college survey asked students
crucial to the prolonged timing of emerging about the extent to which their family supports
adulthood; Arnett 2011), we nd that students their education goals. Although these are cross-
degree plans are, to some extent, related to their sectional analyses, they indicate that older stu-
expected timetables for entering other adult roles. dents perceive declining family support, with a
Future research should examine how these general decline briey suspended at ages 2127
expectations change as the reality of extended for females and 2427 for males. While males
timetables sets in. Students demonstrate agency show a more severe decline through age 50,
in choosing and following the educational trajec- females reported higher support than men until
tories they see represented by specic degrees. age 23, but lower support than men after age 25
While the principal of agency points to the impor- (See Fig. 1). At age 30, men and women diverged
tance of these active choices on the part of stu- again with men reporting much lower support
dents, plans sometimes fail. Research has yet to than women through age 50. These may indicate
explore how unintended delays may inuence the that different age norms operate after age 30, per-
actual decisions students make about entering haps explicitly negotiated within families.
adult roles, what other inuences such as parental Despite being cross-sectional, these ndings sug-
expectations or peers affect their decisions, and gest interesting speculations, for instance, that
how they balance competing inuences and reach extended time in college, which has become
their decisions. Further, research should examine increasingly common (Bound et al. 2009), may
whether these possible delays in important adult lead to declining family support for college par-
life events impact emerging adults perceptions ticularly after age 27. However, multivariate
of their life course progress and decisions as they analyses are needed to consider other confound-
enter adulthood. ing inuences (e.g., college grades, job and fam-
Educational delays may have other implica- ily duties, etc.). We must note that family may
tions. While students may adopt new college refer to family of origin or to spouse, which may
norms in extending their degree timelines, the operate differently. Yet traditional gender role
principle of linked lives draws attention to family expectations that emphasize economic contribu-
members who may continue to hold more tradi- tion for males more than for females may explain
tional age norms (Settersten and Hgestad 1996). the more restricted timetables for males in both
Most community college students live at home kinds of family. However, these conjectures
and commute, meaning that students interact deserve empirical analysis.
with family members (either parents or spouses) Declining family support may in turn have
on a daily basis. Indeed, research has shown that deleterious effects on students ability to persist.
family approval affects persistence (Cabrera Pallas argues that the time spent in college can
et al. 1992). Family support of education may go play an important role in developing plans and
beyond mere approval and extend into additional values in addition to providing job skills.
help, such as facilitating time for studying, pro- According to the socialization theory of
viding transportation to school, and offering schooling school confers knowledge about the
nancial support. As students approach or surpass world for example, the labor market and the
208 C. Zapata-Gietl et al.

My family strongly supports my continued enrollment


.9
.8
.7
.6
.5

18 19 20 21 22 23 24-25 26-27 28-30 31-35 36-40 41-50 51-90


Collapsed Ages
Total Sample Females
Males

Fig. 1 Community college students reports of family support (Sample: students attending ten community colleges,
N = 1,273)

demands of specic jobs [and] can also trans- There is much that we do not understand about
form individuals values or preferences (Pallas the transition to adulthood. Youth who have not
2000). At the same time, the principle of timing completed their education are still in limbo dur-
emphasizes how age serves to sort individuals ing this stage, unable to support themselves with
into positions that allow them access to different the meager earnings their jobs provide. In some
resources and challenges within institutions cases, their parents may be willing to support
(Pallas 2003). That is, individuals of certain ages them, if they are able (but many families are not).
may make decisions that place them on educa- However, education imposes high costs, and par-
tional trajectories that are more or less likely to ents may expect young adults to be nancially
lead to their desired results. For example, younger independent immediately after graduation, even
students with fewer responsibilities may be in a if the individual is still exploring a career, as soci-
better position to take advantage of the enrich- ety encourages them to do. Because of the exi-
ment activities or unpaid internships provided by bility and rapid changes of this time period,
colleges and universities, activities which have parental support is not clearly specied, and it
been shown to increase social integration and may be subject to negotiation or abrupt changes.
degree completion (Tinto 1987). In contrast, we Ultimately, the prolonged exploration of educa-
would expect that students already married or tional opportunities may conict with the needs
raising children may have very different experi- and expectations of many of the new students
ences in these institutions. currently enrolling in community colleges.
College for All: New Institutional Conicts in the Transition to Adulthood 209

5 Three Problematic transition mechanisms designed for students who


Transitions into Adulthood traditionally attend 4-year colleges may fail to
respond to the needs of community colleges new
While the prior section highlights some of the students. Some of these students do not have high
simultaneous conicts among emerging adult achievement expectations or college-educated
roles, we now consider how sequential conicts parents; they do not understand college options
between institutions can be created. The years of and do not have college savings. Additionally,
emerging adulthood, ages 1830, are usually with many older students and students working
accompanied by certain institutional transitions in part time enrolling in community colleges, the
the US. Transitions through educational institu- age at which individuals enter and exit institu-
tions and into the labor force may be poorly aligned, tions of higher education is becoming less stan-
so that a prior institutions preparation does not dardized, which makes structuring transitions
match the later institutions demands. This section particularly difcult. The increased exibility
examines whether poor alignment of these transi- provides new options that allow new students to
tions may account for some of the turmoil and enroll, but this exibility can also pose new chal-
conicts during this life stage, such as delays in lenges to both individuals and institutions.
adult roles or a decline in family support. Analyzing the Educational Longitudinal
Functional sociology provides an image of Study for the class of 2004, researchers
society as a well-designed machine that is coor- (Rosenbaum et al. 2014) nd that, although some
dinated across time and space. Ordinarily, institu- community college students have attained sub-
tions take some responsibility for individuals, BA credentials, almost half (46 %, see Table 3)
providing them with roles, status, responsibility, have no credentials 8 years after high school.
and direction. For example, schools are often Some in this 46 % may still be in college, but if
considered responsible for preparing students to the sample is restricted to students who enrolled
progress within the institution, and for leading to in the rst 2 years after high school, which gives
the next institution. Indeed, students expect that students at least 6 years to complete a credential,
if they take appropriate actions during schooling, 43 % still have no certicate or degree. Similarly,
their careers will unfold as they want and expect. although only 22 % of students who started in
When students enter college and pass their 4-year colleges fail to get any credential, that
courses, they expect to acquire credits, which number doubles to 44 % for those in the bottom
lead to a credential, which in turn leads to better third of test scores who enter 4-year colleges
jobs. In contrast, each of these stages may fail to (Ibid.). Failure to complete any credential can
meet those expectations if transitions are poorly have severe economic and labor market conse-
aligned and unstructured. Higher education insti- quences, making the assumption of adult roles
tutions may lose students at critical transition even more difcult as individuals try to start fam-
points and create conicting expectations when ilies and head households with very small
students arrive at the next institution. Despite the incomes.
importance of educational institutions in shaping We propose that this failure may be in part
the life course, the U.S. poorly synchronizes its attributed to poorly designed and unstructured
many transitions through institutions (Kerckhoff transitions into, though, and out of community
and Bell 1998). American transitions through college. We therefore examine three transitions
school, community colleges, and into the labor where institutions may lose or let down students
market are riddled with difcult and unclear deci- as they pursue college degrees and careers: the
sions that make moving to the next level unneces- high school-to-college transition, the college
sarily hard. enrollment-to-degree transition, and the college-
Community colleges have opened their doors to-work transition. These transitions in commu-
to new kinds of students and must prepare gradu- nity colleges are modeled after the loosely
ates for new labor market demands. However, the structured transitions found in 4-year colleges.
210 C. Zapata-Gietl et al.

Table 3 College attendance, college type, and credential attainment by SES and test score tertiles
Test scorea SESb
All Low test Middle test High test Low SES Middle SES High SES
HS diploma (on 8,537 2,189 2,916 3,395 2,255 2,592 3,348
time, no GED)
Ever attend 86 % 73 % 87 % 95 % 75 % 86 % 95 %
college:
20042012
First college level 2 year 37 % 60 % 42 % 19 % 51 % 41 % 23 %
4 year 59 % 30 % 55 % 80 % 42 % 55 % 76 %
Started at 2 year Some college 46 % 51 % 42 % 44 % 49 % 47 % 41 %
college Certicate 18 % 23 % 17 % 9% 22 % 18 % 11 %
Associates 16 % 15 % 18 % 16 % 16 % 16 % 14 %
degree
Bachelors 20 % 11 % 23 % 30 % 14 % 19 % 33 %
degree plus
Started at 4 year Some college 22 % 44 % 27 % 15 % 36 % 26 % 15 %
college Certicate 5 % 12 % 7% 4% 7% 6% 4%
Associates 5% 9% 5% 4% 8% 7% 4%
degree
Bachelors 67 % 35 % 61 % 77 % 50 % 62 % 76 %
degree Plus
Source: Educational Longitudinal Survey (ELS): 20022012, graduating HS class of 2004
Sample: On-time high school graduates, not enrolled in a post-secondary institution in 2012
a
Test Score is the composite math and language arts standardized test score from 2002
b
SES was created by NCES through ELS, comprised of information on parents occupations and parents education

5.1 High School to College Entry steps they must take to enroll in college credit
courses, which could cause serious delays in their
The transition between high school and college schooling. Indeed, research nds that many high
creates institutional challenges in this country, school seniors who stated in May that they would
largely because of the decentralization of both be attending college in the fall do not actually do
high school and higher education. High schools so, perhaps because of unanticipated obstacles
prepare students prior to the transition, and com- (Stephan and Rosenbaum 2013). Although they
munity colleges may devise procedures that make may ultimately overcome these difculties, and
them responsive to newly-arrived students. evidence shows that many of them do (Rosenbaum
However, if these two consecutive institutions are et al. 2014), this delay may reduce their odds of
not well synchronized, discontinuities arise and success (Bozick and DeLuca 2011). Students
create potential for mistakes, failures, and who do not enroll in the fall may have had dif-
extended degree timetables. High school coun- culties identifying procedures for testing, choos-
selors large caseloads (often over 400 students) ing a program, or procuring funding. These issues
and insufcient information about community can even plague students who do succeed in
colleges increase students difculties in making enrollment, and a lack of know-how or college
the transition to community college. resources can prevent them from gaining traction
First, there is potential for students to get lost in the new institution.
in transition because high schools end their Second, placement tests are major barriers to
responsibility for the student at graduation, but the transition into college. Most community col-
the college may not yet have initiated its respon- leges require students to take this test upon rst
sibility (See Sablan 2013 for a review of summer enrollment. The placement test determines
bridge programs). During the summer after high whether they can enroll in college-level, credit-
school graduation, students might not know the bearing English and math courses, or if they must
College for All: New Institutional Conicts in the Transition to Adulthood 211

take a remedial sequence. Students who exhibit states requires a high-school exit-exam, these
academic deciencies by scoring below the exit-exams require a lower level of academic
selected cut level are often assigned into remedial achievement than college placement exams.
courses, which do not count for college credit and Many seniors pass the high school exit exam, so
which cover topics taught in high school. they expect they are ready for college and are
Unfortunately, most students fail this test and given no information to the contrary. They are
over 60 % of students must take remedial courses then surprised when they fail the college place-
(Bailey et al. 2010; Rosenbaum et al. 2006). ment exam just 3 months later. Such misalign-
Although they are taking classes in college build- ment between educational institutions gives
ings and paying college tuition, students in these students reassuring but mistaken feedback, which
courses do not get college credits, so they often fails to warn them that they are not prepared for
infer that they have not successfully made the college.
transition into college. Unfortunately, the problem of placement tests
This high failure rate is shocking, but where to and remedial courses is exacerbated by colleges
place the blame is less clear. The usual reaction is that do not explain the true nature of these tests.
to blame high schools, but colleges are also at Indeed, instead of using the term, remedial,
fault. In many states, community colleges do not community college staff often use the vague
pose explicit standards, and colleges even use dif- euphemism, developmental education, which
ferent tests and different cutoff points when they students likely do not understand as courses that
use the same test. It is no wonder that high school do not earn college level credits. Students are
graduates are poorly prepared to meet such cha- also told that they cant fail the placement test,
otic standards. Moreover, even when different but no one explains that a low score can have
colleges use the same test, it is very rare that high serious consequences delaying their credential
schools give the college test to their students, to plans by one to four semesters, costing them time
notify them about their skill deciencies. As a and money. These well-meaning stigma-free
result, students usually remain uninformed and euphemisms prevent students from understand-
unable to take actions or make choices that may ing the transition into community college, which
mitigate later delays. can lead to extended timetables and early drop-
Bailey and colleagues (2010) reported that outs, sometimes even before any credits are
most students placed in remedial courses failed to earned.
complete the remedial sequence. Indeed, if they Placement exams are not the only transition
have large achievement shortcomings in math, obstacle. Additional difculties occur during the
nearly all students (83 %) fail to complete the transition out of high school and into college. For
remedial sequence. Despite the predictable risks example, procurement of funding, application
associated with placement in remedial course- timelines, and program selection, all must be
work, students are given no warning about their navigated in order to successfully transition into
poor preparation before they enter college, and college. Society has vastly expanded college
they are not aware of the importance of the place- access, but educational institutions have yet to
ment exam (Rosenbaum et al. 2006). confront the new challenges in the high school to
High schools and colleges are both responsi- college transition.
ble for this lack of student preparation with
regards to the placement test. Although high
schools administer many tests, few of these tests 5.2 Credits Without Credentials
give any information about individuals college
preparation. Since colleges often have different Even after they gain entry to college credit
tests and pass levels, even if high schools gave classes, students face great difculties in actually
college placement tests, their test might not apply completing credentials. Here, the barrier has
to the particular college that a student plans to shifted to college progress, where students
attend. Although high school graduation in many struggle to transition from college attendance to
212 C. Zapata-Gietl et al.

college graduate. Many students accrue enough they end up trying to navigate this complex sys-
credits for a degree, but do not have any creden- tem of programs and courses with no help from
tial. In a national sample of high school gradu- the college, which leads to serious mistakes,
ates, about 8 % had attained Associates degrees wasted time, and dropouts.
10 years later, and another 10 % had enough Many rst-generation college students who
credits for an AA degree but no credentials now populate community college campuses do
(Rosenbaum 2001). not know how to choose realistic degrees that t
Some students report that community college within their timetables or how to choose the right
is like a Kafka-esque nightmare where they run courses to meet degree requirements most ef-
hard but get no traction, where they accrue credit ciently. For instance, one study nds that when
after credit but only get slightly closer to com- deciding to attend college, Latino students get
pleting. Choosing appropriate college courses is helpful information from other Latino students.
unnecessarily complicated. Judging from student However, new students continue to rely on peers
reports, navigating complex college rules for when they get to college even though these peers
choosing courses poses cognitive challenges that possess very little information about college
stymie even the best students. Students discuss requirements (Person and Rosenbaum 2006).
taking an unnecessarily difcult course that they Students have difculties in making program
fail, taking an unnecessarily easy course that choices and degree plans, because they have little
does not count toward degree requirements, tak- information for judging how long their choices
ing the right course offered by the wrong depart- and plans will take and what those choices and
ment, taking the night-school version when only plans require. The principle of agency in the life
the day-version counts, taking courses in the course emphasizes the role of choices in individ-
wrong order, or missing a required course in a uals life trajectories, but their choices and
rigid sequence (Rosenbaum et al. 2006). Even if actions are inuenced by the availability of accu-
students know the courses they need to take, the rate information. When students are unable to
right course may not be offered at a time that ts judge the requirements and timelines for their
their schedule, or it may be full. Credits earned educational plans accurately, their trajectories
can even disappear if students take too long to may be full of unexpected and undesired
complete the degree, as often happens for part- outcomes.
time students.
The problem is both unfathomable complexity
and a lack of warning, and the sheer complexity 5.3 College Without Payoffs
makes warnings prohibitively detailed and hard
to follow. One major issue in community colleges The transition from school to work is rarely easy.
is a dearth of counselors. While high school Even when high schools offered vocational train-
counselors are said to be overburdened with ing, students faced difculties transitioning into
student-to-counselor ratios of 400:1, in commu- the labor force (Stone and Morgan 2012). While
nity colleges, ratios frequently reach 1,500:1 colleges previously provided very little job place-
(Rosenbaum et al. 2015). In the 80 h a counselor ment assistance or specic work preparation to
works during the 2-week registration period, if traditional middle-class students (whose parents
1,500 students were to see an advisor one-on- could help them with advice and job contacts),
one, counselors would have 3 min to advise each college-for-all has enrolled new students who
student. In those 3 min, they would need to assess need additional help for successfully entering the
students prior record and achievement, assist in workforce. The life course trajectories of these
choosing a degree and career goal, determine the new students are constrained, in terms of the
most appropriate program, and choose appropri- principle of time and place, by a workforce that
ate courses. In reality, not all students receive demands college credentials and community
advising, although they probably should. Instead, colleges that have limited success at conferring
College for All: New Institutional Conicts in the Transition to Adulthood 213

credentials and rarely have systematic ways to Table 4 Regression of log income 2011 on educational
attainment
facilitate the transition into work. Today, with
high college dropout rates, many college students Log income 2011
enter the work force without any credential. Graduate degree 0.38***
Failure to complete a college credential has (3.91)
serious consequences. Although having some Bachelors degree 0.36***
college, but no credential led to an economic (7.98)
payoff in the 1970s (Kane and Rouse 1995), Associates degree 0.28***
recent research suggests this is no longer the case (4.93)
(Carnevale et al. 2012; Grubb 2002), and our Certicate 0.22***
analyses conrm that nding. Regression analy- (4.04)
Some college 0.05
sis of the earnings in 2012 of the high school
(1.09)
graduating class of 2004 reveals that each cre-
High school (omitted)a
dential (graduate degree, BA, AA, certicate) has
High school dropout 0.06
substantial earnings payoffs compared to high
(0.75)
school graduates, but college without credentials
SES 2002b 0.04
has none. Compared with high school graduates,
(1.79)
bachelor degrees increased earnings by 36 %,
10th grade test scorec 0.01***
associate degrees by 28 %, and even certicates,
(5.18)
which take about a year, have substantial 22 %
Femaled 0.27***
payoffs (Table 4; see also Jacobson and Mokher
(10.57)
2008). In terms of early earnings payoffs, stu- Black 0.23***
dents who spend time and money on school but (4.28)
who attain no credential or degree are essentially Hispanic 0.03
wasting resources. (This nding applies to youth (0.59)
under age 26, and may not apply to older students Other race 0.01
if they have jobs that reward a few credits and (0.30)
new skills.). Constant 9.80***
The same is true for students who enter 4-year (100.33)
colleges and fail to complete a credential. Almost N 3,677
one-quarter (22 %) of the students who begin in Source: Educational Longitudinal Survey (ELS): 2002
4-year colleges have no credentials (See Table 3), 2012, graduating HS class of 2004
and among those in the bottom third of test Sample: Not enrolled in a post-secondary institution in
2012, employed full time in 2011, at least 2 years since
scores, 44 % get no credentials. Additional analy-
degree completion or college attendance, if applicable
ses indicate that individuals with college but no a
high school category includes all high school gradu-
credential who began in a 4-year college will get ates, including those with a GED
b
no more payoff than those who began in a 2-year SES was created by NCES through ELS, comprised of
information on parents occupations and parents
college. That is to say, 8 years after high school,
education
on average individuals get no payoffs from col- c
Test Score is the composite math and language arts stan-
lege unless they completed a credential. Students dardized test score from 2002
d
with some college do not successfully transi- Omitted categories are male and white
tion into the labor market.
These ndings indicate that students who (2014a, b) ndings using the Youth Development
leave college with no credential get no earnings Study panel that showed that individuals with
benet from having attended college, and would some college had less favorable outcomes than
have been better off accruing work experience those with BA or AA degrees. Signicantly,
than having paid tuition and perhaps entered into while we nd that students with some college
debt. These ndings converge with Vuolo et al.s are not successfully transitioning into the labor
214 C. Zapata-Gietl et al.

market, Vuolo and colleagues (2014b) show that Young adults entering the labor market must
their occupational outcomes are also more also contend with job changes imposed on
adversely affected by economic downturns. younger workers. Having the least seniority, they
We also nd that while students with some are the rst red, and they are subject to high
college but no credentials have the same earnings turnover. While some job changes may have pay-
as high school graduates, almost half (47 %) offs in the future, involuntary turnover can also
acquired student loans with an average debt of slow the acquisition of labor market skills and
nearly $16,000, about the same as those who experiences, which can ultimately reduce long
completed a certicate (Rosenbaum et al. 2014). term wages (Danziger and Ratner 2010).
This lack of earnings coupled with school debt Consistent with the principle of linked lives,
could pose serious consequences for their young adults exibility in absorbing these delays
nancial stability through the period of emerging is linked to the demands, constraints, and sup-
adulthood, which may impact decisions to adopt ports their social networks are able to provide
other adult roles, particularly those requiring a them. Students with family responsibilities or
living wage, such as starting a family. Indeed, with parents who cannot provide support are par-
students cannot escape responsibility for college ticularly affected by job instability.
loans for many yearsit takes 2025 years
(depending on the program) for these debts to be
forgiven. However, debt forgiveness plans only 6 New Structures: Reducing
include loans obtained from the Federal Simultaneous
government, leaving students with private loans and Sequential Conicts
still obligated. Student debt coupled with no
earnings payoffs have long lasting impact (https:// The difculties we have noted in these transitions
studentaid.ed.gov/repay-loans/forgiveness- are not inevitable. Many of them can be improved
cancellation/charts). by creating structured institutional procedures.
For students who do complete a credential, Some transition difculties are related to the
internships and practica often serve as a bridge complexities of community colleges and the
between higher education and work. However, labor market, and students difculty in making
internships are typically unpaid, which presents informed choices. By increasing structural sup-
an obstacle for low-income students, who cannot port and guidance, these educational institutions
afford to work unpaid (Curiale 2009). While may help students acquire the resources, infor-
internship requirements for competitive jobs may mation, and skills necessary for more successful
shut some students out, practica required as part transitions. Indeed, some 2-year colleges have
of curricula may provide valuable work experi- occupational programs that use structured institu-
ence without adding summer or post-graduate tional procedures that shape programs at every
responsibilities, as internships often do. Yet these stage of the process, from college enrollment
are only consistently offered in certain elds, through workforce entry (Rosenbaum et al.
particularly health. 2006). In this section, we consider what struc-
The usual career services ofces in public tured institutional procedures might help students
community colleges have few staff who can only make more orderly progress across transitions,
help a small number of the colleges 30,000 stu- improving students knowledge, plans, and
dents (Rosenbaum et al. 2015). As a result, these careers. In general, we highlight how structured
ofces provide small optional workshops on transition procedures, timely advising, and sys-
making an attractive resume, but they provide tematic information delivery can improve transi-
little practical assistance. They also post jobs on tions. We also suggest supports that may help
the bulletin board, but many of the posted jobs ease conicts for emerging adults who frequently
are for low-skilled work or do not align with the nd themselves occupying multiple roles with
programs offered at the college. conicting demands.
College for All: New Institutional Conicts in the Transition to Adulthood 215

6.1 Institutional Procedures Earlier we noted the shocking 60 % failure


for the High School-to-College rate on college remedial placement tests. Here we
Transition see a reform that could warn students about their
level of preparation, and can even provide instruc-
Structured institutional procedures can reduce tion in high school to address their skill needs.
many of the difculties we have noted in the tran- This reform diagnoses high placement test failure
sition into college. In countries that have a rate as not the fault of high schools or colleges,
national educational system, institutional proce- but the lack of communication between the two.
dures alert younger students to whether they are The Florida reform alerts students to their level of
academically prepared for the next stage of edu- academic preparation for college on the college
cation. For example, in Japanese middle schools, placement test and provides corrective instruc-
students take practice tests in anticipation of an tion. In this way, improved transition structures
exam in ninth grade that determines their high may reduce prolonged degree timetables and thus
school track. These practice tests warn students reduce conicts with other roles adopted during
about their level of preparation, leading some stu- emerging adulthood.
dents to work harder and other students to lower Other institutional mechanisms can help with
their expectations (Kariya and Rosenbaum 1999). transition into college. Counselors are the pri-
Unlike college students in the U.S., who do not mary high school staff responsible for advising
anticipate the remedial placement test and do not students transition from high school to college,
review for it, Japanese students are not surprised but a single counselor cannot advise all students
by the high school entrance test, and have actu- choices. Additionally, high schools may create
ally spent 3 years preparing for it and getting specic institutional procedures that guide this
warnings about their chances of passing it. transition.
Following the Japanese example, Rosenbaum Zapata and Rosenbaum (2013) described
(2001: 276) proposed that instead of the many structured advising procedures in three high
tests that give students test results with no useful schools: a suburban middle-class school which
information about college preparation, U.S. high sent most students to college, an urban low-
schools could administer the community college income school with poor college attendance, and
remedial placement test in junior year of high another urban school serving similarly low-
school. This would inform students about income students but with a new commitment to
whether they are prepared for this test, and allow college attendance. Students need a wide variety
students to prepare during the senior year of high of information and support for choosing colleges,
school. completing applications, seeking nancial aid,
The Florida state legislature recently passed applying to FAFSA (to get Pell grants), and
legislation to implement such a reform. Florida choosing a program. In the suburban high school,
already had a statewide college placement exam this study identied four structured procedures
used in all state community colleges, making it that improved the high school to college
an ideal state to implement this policy. The new transition.
reform mandates that all qualied high school Succinct information sheets: Succinct infor-
juniors take the exam, and students who do not mation sheets can simplify choices. The subur-
meet determined college readiness standards ban high school provided students with clear and
must take a mandatory college preparation course timely information, which documented and fre-
in senior year. This reform is intended to tell high quently reinforced procedures for applying to
school juniors whether they meet the achieve- nancial aid and colleges. When the time came
ment level expected in the state community col- for making certain choices, the school distributed
leges, and it gives them an opportunity to address information sheets that instructed students how
their skill needs (Mokher et al. 2014), thus to make plans and take actions about topics such
enhancing their capacity to exercise agency. as selecting a college.
216 C. Zapata-Gietl et al.

Systematic timetables: Systematic timetables that all aspects of college were explained
can keep students on track. The suburban high because students were often rst-generation
school also created systematic procedures to pro- students and did not have resources at home.
vide students with reminders for the time- This school showed that suburban procedures
sensitive aspects of the college application could also be implemented in the city school
processes such as early-decision and nancial aid context.
deadlines. Timely reminders may assist students In the suburban school and college-focused
in redirecting their goals (when nancial aid lim- urban school, students are supported by system-
its may make shorter degrees more realistic) or atic procedures and advising that deliver infor-
adjusting their current approach. mation in a way that ensures students understand
Software: Software can inform choices and what to do, how to do it, and when. Such infor-
guide students progress. The suburban high mation underlies the successful exercise of
school also used college choice software to agency. They do not assume that parents provide
inform students about which colleges were good key information at home. These schools use insti-
matches, and which were reaches or safety tutional procedures to create a dependable transi-
schools (Becker and Stephan 2011). If students tion from high school to college.
fall behind in the application process, they can be Of course, the best way to improve success
notied or contacted by a school counselor or across the transition is to eliminate the transition
other staff. and merge institutions, which has been done in
Group advising: Group advising can improve some small-scale reforms. Early college high
the dissemination of important information. The schools (ECHS) do just that. ECHS are high
suburban high school used group advising strate- schools that form close connections with a col-
gically, to provide students with expectations and lege and are sometimes even located on a college
strategies for choosing their college. campus (Rosenbaum and Becker 2011). Students
The importance of these systematic features in are prepared in the early years of high school so
the suburban school is evident when contrasted that by 11th grade, they are able to take at least
with the urban school without a commitment to one college course, and by 12th grade they are
college enrollment. That school provided indi- taking several. This structural change gives stu-
vidual meetings with counselors, but very few dents rst-hand experience in college, and seems
systematic procedures, which left students over- to be effective in motivating students to meet col-
whelmed and lacking any strategy for sifting lege standards. In some schools, students com-
through information. Students often lacked the plete enough college credits so they are awarded
basic knowledge necessary to allow them to a high school diploma and an Associates degree
know which questions to ask. Two meetings a by the end of 12th grade.
year with a counselor were not sufcient to guide The kind of fusion created by ECHS requires
them through the complexities of college choice levels of stafng and resources that the typical
and applications. Similarly, the exibility high school is unable to sustain. However, the
provided at community colleges may obscure the close connections formed with colleges can allow
most direct pathways to a college degree. for the kind of alignment in academic require-
In contrast, the urban school with a college ments that can facilitate the transition into post-
commitment had recently initiated new proce- secondary institutions. The transition into college
dures to provide systematic guidance, devised can be improved by creating structured institu-
to resemble suburban school procedures (based tional procedures. Although the U.S. does not
on their own visits to suburban high schools). have a coordinated national educational system,
They differed primarily in that they also pro- institutional procedures can improve coordina-
vided extensive advising on nancial aid pack- tion and alert students about how to prepare for
ages. They also made extra efforts to ensure the next stage of education.
College for All: New Institutional Conicts in the Transition to Adulthood 217

6.2 Institutional Procedures entering college today. In contrast, a study of


to Improve the Credits- 2-year colleges discovered some institutions
Credentials Transition which devised alternative procedures, designed
to reduce the difculties experienced by disad-
Community colleges enroll many students like vantaged students and improve their college suc-
those in the urban high schools we studied. Many cess (Rosenbaum et al. 2015). Community
of these students do not understand college, and colleges can facilitate agency, and respond to
they cannot get their questions answered at home. individuals limited information, by letting stu-
Community colleges could emulate the urban dents choose their end goals, and offering them
high school that focused on college by devising structured curriculum programs that efciently
structured institutional procedures that provide lead to those goals as quickly as possible. In line
systematic information and systematic proce- with an approach that takes the life course into
dures to guide students through the choices they consideration, these colleges create quick suc-
must make in college. In particular, community cesses with short-term credentials, and postpone
colleges could adapt some of the institutional obstacles until after students experience some
procedures from high schools that we described college success. Course choices are structured to
above. Information sheets and website updates reduce mistakes and create efcient progress, and
are not expensive, so community colleges can courses are offered back-to-back in predictable
easily give students information which helps schedules and convenient locations (Rosenbaum
them anticipate and understand their choices et al. 2006). These colleges also create degree
about college majors or remedial coursework. ladders with short-term interim successes along
Systematic timetables can keep students on track the way to ultimate degree goals, and they man-
in colleges, which have many time-sensitive date frequent advising and continual close moni-
requirements such as nancial-aid renewal, major toring, so they detect problems quickly. Although
applications (which may restrict access to needed these colleges have mostly disadvantaged stu-
classes), and other ling deadlines. Software can dents, their structured procedures reduce student
also be used to monitor students progress in get- mistakes and may increase student completion.
ting credits toward their major. Some 4-year col- Systematic analyses indicate that students in
leges (Arizona State University and Austin Peay these colleges have greater condence that they
University) have devised software to notify stu- will make dependable progress in college, get
dents about which courses t their degree goals, relevant curricula, and receive job payoffs
which courses are a good match for their prior (Becker et al. 2014).
achievements, and which will provide serious
challenges. Similarly, Chicagos city colleges are
using software to monitor student attendance, 6.3 Institutional Procedures
grades, and progress, and students are notied to Improve the College-Career
when they fall behind. Finally, group advising Transition
would be less expensive than individual advising,
and requiring it might improve community col- Transitions into the workforce do not need be as
lege students choices and college progress, difcult as they are for American students. For
instead of leaving them to fend for themselves. example, European countries use apprenticeships
More extensive institutional procedures can to provide a structured pathway from school to
also be devised to improve college success for work (Schwartz 2014). Such apprenticeships pro-
disadvantaged students. Although traditional vide clear dependable school-to-work transitions
procedures may work for traditional college stu- by combining education and workplace experi-
dents whose college-educated parents can offer ences in various proportions over time. Although
advice and abundant resources, these procedures this system has challenges, most students see
may not work well for new kinds of students these requirements as manageable and complete
218 C. Zapata-Gietl et al.

the transition smoothly (Hamilton 1990). tacts with certain employers, and they can help
Exposing students gradually to increased students obtain good jobs in the correct eld. The
demands for responsibility teaches them new colleges strive to strengthen these employer con-
capabilities. Hamilton (1990) was impressed that tacts, and they will ask employers how last years
German 18-year-olds in apprenticeships regu- graduates are doing in the job, and what new
larly had more responsibility than Americans skills they should be teaching. Such strong con-
believe 18-year-olds can handle. Kerckhoff and tacts create trust, which enables the college to
Bell (1998: 153) report that European systems communicate trusted signals of students capa-
haveimportant advantages over the US by pro- bilities, which otherwise could not be conveyed.
viding multiple highly visible credentials. These contacts give students condence that their
Apprenticeship systems allow students to be con- college efforts will have labor market payoffs,
dent of dependable labor force outcomes and and students are more motivated as a result
improve the transition by providing clear incen- (Rosenbaum et al. 2006, Chap. 9). Much like
tives and developing student condence in apprenticeships, these college-employer contacts
dependable labor market payoffs. Mortimer and create dependable pathways for the transition
Kruger (2000) also contend that school-work from college to work.
connections may affect youths career plans and Obviously, such extensive job placement ser-
outcomes. While Mortimer and Kruger show the vices are expensive, and most colleges cannot
virtues of a highly structured transition which afford them. Yet many community college occu-
youth can easily understand (as in Germany), the pational faculty already have strong contacts with
connections between schooling and working are employers in their elds (Rosenbaum et al. 2006,
vague in the U.S., and make plans and prepara- p. 199). However, they do not have time to use
tion difcult. Apprenticeship systems have their their contacts to help many graduates because
own challenges (Kupfer 2010), but they do pro- they are tied down with paperwork and adminis-
vide an example of the kind of protections that trative details. It would be relatively inexpensive
can be extended to youth through explicit links to provide clerical assistance for these tasks, to
between school and work. free faculty time to devote to using their contacts
Although American society has very few to help their graduates job search. This would be
apprenticeships, we might ask whether American an important service to graduates and to employ-
colleges could develop institutional structures ers, and it might increase students condence
that would have similar characteristics, wherein that college would have a payoff for them.
students can expect a certain payoff from making
dened progress. American community colleges
may facilitate such transitions by more 7 Conclusion
systematically building practica into the curricu-
lum, which may particularly help disadvantaged Many studies of the life course focus on individu-
youth who have had less exposure to middle class als experiences. Although such research pro-
occupations (Carnevale et al. 2012). vides a good understanding of how individuals
Some 2-year colleges have devised ways to perceive and understand their experiences, most
provide strong support for the college-work tran- life course experiences are also shaped by social
sition at every stage of the job search process context. This chapter has identied some features
(Rosenbaum et al. 2006). In some exemplary col- of the social context that may constrain emerging
leges, students are informed about how to com- adulthood.
municate their job relevant skills in a way that We have identied some unintended negative
employers will recognize them. Schools provide consequences associated with the increased
students with information about job-search strat- access and enrollment in community colleges,
egies and what kinds of jobs are good matches for and we have suggested some of the specic con-
their training. Some colleges have long-term con- icts that need to be considered and improved.
College for All: New Institutional Conicts in the Transition to Adulthood 219

Our analyses show that students may not be fully icts can alert young adults, their parents, and
aware of how their educational plans can inu- educators to notice the ways social contexts cre-
ence other life domains. Ultimately, while emerg- ate conicts, and may suggest alternative ways to
ing adulthood suggests that lengthening reduce these conicts.
timetables are normal during this time period, Clearer expectations of degree timetables
students may be unaware of the extent to which would help students understand the implications
their educational plans may impact their other of their new college opportunities, help them
life domains. Life course theory points to the explain their college time commitments to family
importance of agency in individuals lives, which and employers, and perhaps help them discuss
is constrained by the lack of information and and negotiate changes of age norms for other life
clarity about credentials and timelines. Future events with family members. Colleges create
research should examine how students perceive misleadingly upbeat age expectations that even-
and respond to unmet expectations for transitions tually hurt students who need to develop realistic
into other adult roles. expectations for degree timetables. While some
College-for-all has created a distinctive his- students may become discouraged when they
torical context that simultaneously provides new learn true degree timetables, others may reassess
opportunities and new constraints from educa- their plans and consider alternate, shorter degrees,
tional demands. Unfortunately, the vast increase such as sub-BA credentials, which yield quick
in college access has been undertaken in isola- interim successes, even as some students con-
tion, and has lacked coordination with other insti- tinue to pursue their BA plans. Realistic expecta-
tutions and the demands they make on these new tions about the costs (time, money, labor market
college students. This failure may have vast experience, and family sacrices) of new educa-
implications for creating conicts with other life tional opportunities are especially important for
course events that are expected to occur during low-income and rst generation students who
the transition to adulthood. For example, as stu- may be less likely to have the resources to ride
dents stay in college into their late 20s or beyond, out unexpected delays.
they report less family support for their continua- The payoffs to college only exist for those
tion. Students pursuing extended college timeta- who complete a credential. Students who attend
bles experience simultaneous conicts with other college but get no credentials have successfully
life events, particularly marriage and childbirth. followed societys advice to attend college, but
Prolonged timetables for completing college they get no warning that some college is not
degrees create conicts in other life domains. rewarded, and no advice about how to get quick
Moreover, these conicting pressures may vary credentials. We rightly celebrate high rates of
by social and cultural expectations because indi- college attendance, but we ignore college drop-
vidual trajectories are inuenced by interpersonal outs, offering no advice about alternative creden-
relationships in their linked lives. tials that might be easier and ignoring many of
Causality is always difcult to infer, and cer- the difculties students face after they make it
tainly must be considered very speculative here. into college. While reforms like the National
Yet the studies reviewed here indicate many College Advising Corps have made impressive
kinds of conicts that are likely to inuence efforts to improve college attendance, they need
youths experiences, particularly the kinds of to consider ways to help these new college stu-
experiences associated with their transition to dents to avoid only earning some college
adulthood. If we have identied factors that have instead of a credential.
some inuence, these factors may provide levers Poorly designed transitions can create failures
for social policy to assist youth in dealing with and cause students to believe they are low-
challenges of the transition to adulthood and ability when in fact they might have great poten-
minimizing unintended delays. Merely identify- tial. Smoother transitions into college, through
ing potential simultaneous and sequential con- college, and into the workforce might decrease
220 C. Zapata-Gietl et al.

the conict felt by individuals between their roles Bozick, R., & DeLuca, S. (2005). Better late than never?
Delayed enrollment in the high school to college tran-
as students and their other roles as emerging
sition. Social Forces, 84(1), 531554.
adults. The highly structured transition proce- Bozick, R., & DeLuca, S. (2011). Not making the transi-
dures we outline may help students avoid some tion to college: School, work, and opportunities in the
college but no credential in addition to reducing lives of American youth. Social Science Research,
40(4), 12491262.
extended timetables. These improvements may
Cabrera, A. F., Castaeda, M. B., Nora, A., & Hengstler,
lead to fewer conicts with other roles adopted D. (1992). The Convergence between two theories of
during the transition to adulthood and help ease college persistence. The Journal of Higher Education,
that transition. 63(2), 143164.
Carnevale, A. P., Rose, J. S., & Hanson, A. (2012).
While the transition to adulthood is character-
Certicates: Gateway to gainful employment and col-
ized by prolonged timelines, it is possible that not lege degrees. Washington, DC: Georgetown University
all of these delays are intentional or have positive Center on Education and the Workforce.
effects on individuals life course outcomes. We Complete College America. (2011). Time is the enemy.
Retrieved from www.completecollege.org/docs/
suggest that looking at education inuences in
Time_Is_the_Enemy.pdf
the form of institutional structures is an avenue Ct, J. E. (2014). The Dangerous myth of emerging
for facilitating the lives and goals of young adults. adulthood: An evidence-based critique of a awed
developmental theory. Applied Developmental
Science, 18(4), 177188.
Curiale, J. L. (2009). Americas new glass ceiling: Unpaid
References internships, the fair labor standards act, and the urgent
need for change. Hastings Law Journal, 61, 1531.
Adelman, C. (2003). Principal indicators of student aca- Danziger, S., & Ratner, D. (2010). Labor market out-
demic histories in post-secondary education, 1970 comes and the transition to adulthood. The Future of
2000. Washington, DC: US Department of Education, Children, 20(1), 133158.
Institute of Education Sciences. Doeringer, P., & Piore, M. (1971). Internal labor markets
Arnett, J. J. (2000). Emerging adulthood: A theory of and manpower analysis. Lexington: Lexington Books.
development from the late teens through the twenties. Elder, G. H. (1998). The life course as developmental
American Psychologist, 55(5), 469480. theory. Child Development, 69(1), 112.
Arnett, J. J. (2004). Emerging adulthood: The winding Furstenberg, F. F., Kennedy, S., McLoyd, V. C., Rumbaut,
road from the late teens through the twenties. R. G., & Settersten, R. A. (2004). Growing up is harder
New York: Oxford University Press. to do. Contexts, 3(3), 3341.
Arnett, J. J. (2006). Emerging adulthood in Europe: A Goldin, C., & Katz, L. F. (2008). Transitions: Career and
response to Bynner. Journal of Youth Studies, 9(1), family life cycles of the educational elite. The American
111123. Economic Review, 2008, 363369.
Arnett, J. J. (2011). Emerging adulthood(s): The cultural Goldrick-Rab, S., & Sorensen, K. (2010). Unmarried par-
psychology of a new life stage. In L. A. Jensen (Ed.), ents in college. The Future of Children, 20, 179203.
Bridging cultural and developmental approaches to Grubb, W. N. (2002). Learning and earning in the middle,
psychology: New syntheses in theory, research, and part I: National studies of pre-baccalaureate educa-
policy (pp. 255275). New York: Oxford University tion. Economics of Education Review, 21, 299321.
Press. Hamilton, S. F. (1990). Apprenticeship for adulthood:
Bailey, T., Jeong, D. W., & Cho, S. W. (2010). Referral, Preparing youth for the future. New York: The Free
enrollment, and completion in developmental educa- Press.
tion sequences in community colleges. Economics of Hendry, L. B., & Kloep, M. (2010). How universal is
Education Review, 29(2), 255270. emerging adult-hood? An empirical example. Journal
Becker, K. I., & Stephan, J. L. (2011). Gatekeeping by of Youth Studies, 13, 169179.
software: 21st Century guidance counseling. Jacobson, L., & Mokher, C. G. (2008). Pathways to boost-
Roundtable presentation at American Sociological ing the earnings of low-income students by increasing
Association Conference, Las Vegas. their educational attainment. Report prepared for the
Becker, K. I., Rosenbaum, J. R., Cepa, K., & Ahearn, C. Bill & Melinda Gates Foundation by Hudson Institute
(2014). Turning the question around: Do colleges fail and CNA. Retrieved from http://www.hudson.org/
to meet students expectations? Manuscript submitted les/publications/Pathways%20to%20BBoostin.pdf
for publication. Kane, T. J., & Rouse, C. E. (1995). Labor market returns
Bound, J., Hershbein, B., & Long, B. T. (2009). Playing to two-and four-year college. The American Economic
the admissions game: Reactions to increasing college Review, 85(3), 600614.
competition. Journal of Economic Perspectives, 23(4), Kariya, T., & Rosenbaum, J. E. (1999). Bright ight:
119146. Unintended consequences of detracking policy in
College for All: New Institutional Conicts in the Transition to Adulthood 221

Japan. American Journal of Education, 107(3), Rosenbaum, J. E., Deil-Amen, R., & Person, A. (2006).
210230. After admission: From college access to college suc-
Kerckhoff, A. C., & Bell, L. (1998). Hidden capital: cess. New York: Russell Sage Foundation Press.
Vocational credentials and attainment in the United Rosenbaum, J. E., Ahearn, C., Becker K., & Rosenbaum,
States. Sociology of Education, 71, 152174. J. (2014). The new forgotten half and research direc-
Kupfer, A. (2010). The sociopolitical signicance of tions for supporting them. Presented to the William
changes to the vocational education system in T. Grant Foundation, New York.
Germany. British Journal of Sociology of Education, Rosenbaum, J. E., Rosenbaum, J., Stephan, J., Foran,
31(1), 8597. A. E., & Schuetz, P. (2015). Beyond BA blinders:
Mokher, C. G., Jacobson, L., & Alexander, J. (2014). Cultural impediments to college success. My 4-year
Florida college and career readiness initiative: degree was the longest 8 years of my life. In
Fidelity of implementation in the rst four years. Paper O. Patterson (Ed.), The cultural matrix: Understanding
presented at the Annual Conference of the American black youth. Cambridge, MA: Harvard University
Educational Research Association: AERA, Press.
Philadelphia. Sablan, J. R. (2013). The challenge of summer bridge pro-
Mortimer, J. T., & Krger, H. (2000). Pathways from grams. American Behavioral Scientist,
school to work in Germany and the United States. In 0002764213515234.
M. T. Hallinan (Ed.), Handbook of the sociology of Schwartz, R. (2014). The pursuit of pathways. American
education (pp. 475497). New York: Springer. Educator, 38(3), 2441.
Murray, C. (2008). Real education. New York: Crown Settersten, R. A., & Hgestad, G. O. (1996). Whats the
Press. latest? II. Cultural age deadlines for educational and
National Center for Educational Statistics (NCES). work transitions. The Gerontologist, 36(5), 602613.
(2012). Digest of educational statistics. Washington, Settersten, R., & Ray, B. E. (2010). Not quite adults. On
DC: National Center for Educational Statistics. Tables the Frontier of Adulthood, 1(1), 236.
196197. Shanahan, M. J. (2000). Pathways to adulthood in chang-
Pallas, A. M. (1993). Schooling in the course of human ing societies: Variability and mechanisms in life
lives: The social context of education and the transi- course perspective. Annual Review of Sociology, 26,
tion to adulthood in industrial society. Review of 667692.
Educational Research, 63(4), 409447. Stephan, J. L., & Rosenbaum, J. (2013). Can high schools
Pallas, A. M. (2000). The effects of schooling on individ- reduce college enrollment gaps with a new counseling
ual lives. In M. T. Hallinan (Ed.), Handbook of the model? Educational Evaluation and Policy Analysis,
sociology of education (pp. 499525). New York: 35(2), 200219.
Springer. Stone, J., & Morgan, L. (2012). College and career ready
Pallas, A. M. (2003). Educational transitions, trajectories, in the 21st century. New York: Teachers College Press.
and pathways. In J. T. Mortimer & M. J. Shanahan Tinto, V. (1987). Leaving college: Rethinking the causes
(Eds.), Handbook of the life course (pp. 165184). and cures of student attrition. Chicago: University of
New York: Springer. Chicago Press.
Person, A. E., & Rosenbaum, J. E. (2006). Chain enroll- Vuolo, M., Mortimer, J. T., & Staff, J. (2014a). Adolescent
ment and college enclaves: Benets and drawbacks of precursors of pathways from school to work. Journal
Latino college students enrollment decisions. New of Research on Adolescence, 24(1), 145162.
Directions for Community Colleges, 2006(133), Vuolo, M., Mortimer, J. T., & Staff, J. (2014b). Assessing
5160. the long-term value of college degrees in turbulent
Rosenbaum, J. E. (2001). Beyond college for all: Career economic times. Paper presented at the Annual
paths for the forgotten half. New York: Russell Sage Meeting of the Society for Longitudinal and Life
Foundation. Course Studies, Lausanne.
Rosenbaum, J. E., & Becker, K. I. (2011). The early col- Zapata-Gietl, C., & Rosenbaum, J. R. (2013). Expecting
lege challenge: Navigating disadvantaged students adult roles: Emerging adults in community college
transition to college. American Educator, 35(3), contexts. Paper Presentation at Society of Emerging
1420. Adulthood, Chicago.
Changes in Educational Inequality
in Cross-National Perspective

Pia N. Blossfeld, Gwendolin J. Blossfeld,


and Hans-Peter Blossfeld

1 Introduction Yet not only the secondary school systems


have been transformed, but also the landscapes of
In times of worldwide rapid educational expan- higher education have changed in modern societ-
sion, educational opportunities of successive ies. Up into the 1970s, higher education mostly
birth cohorts vary considerably in modern societ- followed the completion of an academic oriented
ies. Since the mid-twentieth century, all modern secondary school (e.g., a gymnasium, lyce,
societies have experienced an enormous expan- public, or grammar school) and universities,
sion of their secondary educational school and colleges or institutes of technology offered aca-
training systems (Boli et al. 1985; Jackson 2013). demic programs for the brightest and most afu-
Educational reforms have been introduced that ent students. In the following decades, however,
(1) reduced the importance of rigid early selec- not only the participation rate in higher education
tion, (2) extended compulsory education, (3) has been quickly rising (see Blossfeld and
expanded traditional academic tracks, (4) estab- Blossfeld 2014), but also the heterogeneity of
lished more inclusive school types, (5) abolished college students who received some kind of col-
dead-end educational pathways, and (6) created lege degree has been growing.
multiple new alternative routes to higher educa- Given this impressive process of educational
tion rewarding individuals vocational and work expansion, students and their families are, how-
experiences (Benavot and Resnik 2006). ever, not simply passive executors of the
Altogether, these reforms fundamentally changed changes in the macro structures in the educa-
the character of the old sharp divisions between tional system. Family members have agency,
academic and vocational/technical tracks in the make decisions, and weigh the possible future
secondary school systems (Benavot and Resnik costs and benets of educational alternatives
2006) and increased the proportion of young peo- based on the familys position in the system of
ple who have completed at least upper secondary social inequality. The families are therefore the
education (see Blossfeld and Blossfeld 2014). key units of strategic action within the structure
of inequality in modern societies (Erikson and
Goldthorpe 1992). The establishment of more
P.N. Blossfeld (*) G.J. Blossfeld
inclusive school types, the abolition of dead-
Nufeld College, University of Oxford, Oxford, UK
end educational pathways and the creation of
H.-P. Blossfeld
multiple new routes to higher education have
European University Institute Florence,
Florence, Italy opened up many new options for individuals
e-mail: hp.blossfeld@eui.eu and their families to pursue their goals. Here,

Springer International Publishing Switzerland 2016 223


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_10
224 P.N. Blossfeld et al.

we ask the question: How has the educational 2 Conceptual Considerations


attainment of men and women from families and Hypotheses
with different educational resources changed as
a result of the bottom-up universalization of 2.1 Educational Attainment
education in secondary school and by the as a Life Course Process
changing role of elite education? In particular, of Successive Educational
we are interested to nd out which educational Transitions
origin groups could prot most from the addi-
tional opportunities provided by the expanding Given the increasing multitude of educational
upper secondary school and tertiary education pathways in modern secondary and tertiary edu-
systems over time and whether there are simi- cation, comparative research needs a typology to
larities and differences among a broad array of distinguish meaningful and cross-nationally
modern societies. equivalent educational categories (Pfeffer 2006).
Using truly equivalent educational data from In each of the PIAAC societies, OECD country
the new Programme for the International experts have classied respondents diverse edu-
Assessment of Adult Competences (PIAAC), cational degrees based on the International
which was collected in 24 modern countries Standard Classication of Education (ISCED
between August 2011 and March 2012, we com- 1997) (see OECD 1999). Thus, our cross-
pare the changes in the effects of parents educa- national analysis of the PIAAC data provides
tion on their boys and girls educational more comparability than other cross-national
opportunities across birth cohorts. We include 22 comparisons based on independently collected
of the 24 original PIAAC countries: Austria, national surveys. The PIAAC scientic use le
Belgium (Flanders), Canada, the Czech offers three broad highest educational attainment
Republic, Denmark, Estonia, Finland, France, levels: (1) up to lower secondary education, (2)
Germany, Ireland, Italy, Japan, South Korea, the upper secondary educational attainment and (3)
Netherlands, Norway, Poland, the Russian tertiary educational attainment.
Federation, the Slovak Republic, Spain, Sweden, The lower secondary level coincides with the
the United Kingdom (England and Northern end of compulsory education after grade 8 or 9 in
Ireland) and the United States (see PIAAC most nations (UNESCO 2012:33). Lower sec-
2012). We did not include Cyprus and Australia ondary education may prepare students for entry
in our analysis, given specic problems of data into the labor market and/or for upper secondary
access. PIAAC allows us to conduct a highly education (OECD 1999: 33). National boundar-
standardized comparative analysis covering the ies between lower secondary and upper second-
same birth cohorts in a broad range of modern ary education are used to distinguish both levels.
societies from the East to the West, from capital- The completion of lower secondary education is
ist and (former) socialist countries, from coun- therefore possible after 8, 9, or 10 years of
tries with different welfare state regimes and schooling and at the age of 15, 16, or even 17. For
concepts of social equality as well as different example, in the United States, the ISCED classi-
educational institutions. In other words, despite cation denes lower secondary education as
all these cultural, political, economic and struc- grades 7 through 9 and upper secondary as grades
tural differences among countries, have the link- 10 through 12. Thus, a student in the United
ages of parents education to girls and boys States must complete grade 9 in order to attain a
educational opportunities changed in the same lower secondary education and an upper second-
direction and with the same intensity across ary attainment level is roughly equivalent to a
modern societies? high school diploma.
Changes in Educational Inequality in Cross-National Perspective 225

The upper secondary level corresponds to the origin-specic opportunities have been the great-
nal stage of secondary education in most OECD est in the process of educational expansion.
countries. The entrance age to this level is typi- Compared to earlier research on this topic (see
cally 15 or 16 years. There are substantial differ- Breen et al. 2009, 2010), we specically aim to
ences in the typical duration of upper secondary study the different mechanisms at work in the
programs both within and between countries, process of the universalization of upper second-
ranging from 2 to 5 years of schooling (OECD ary education and the expansion of graduation
1999:39). Upper secondary education includes from higher education (see also Erikson and
completing academically oriented secondary Jonsson 1996). We study the transition probabili-
schools such as a gymnasium, lyce, public ties from lower secondary (Y 1) to upper sec-
high school or grammar school and completing ondary or tertiary education (Y 2) and from
vocational schools or apprenticeship programs upper secondary (Y 2) to tertiary education
(e.g. in the dual system in Germany). The cate- (Y = 3):
gory upper secondary level also includes com- Pr(Y 2 | Y 1) = Pr(Y 2) (1)
pleting post-secondary non-tertiary education.
For example, in Germany a skilled craftsman Pr (Y = 3 | Y 2) (2)
(Geselle) can become master craftsman
(Meister). The rst transition represents a choice between
A tertiary educational level is equivalent to leaving the educational system with only lower
graduation from a university of applied sciences, secondary educational attainment and attaining
a professional college or a traditional university an upper secondary or a tertiary level degree
(OECD 1999:51). These qualications are a req- (unconditional probability); and the second tran-
uisite for gaining entry into advanced research sition consists of a choice between leaving the
programs and professions with high skill educational system with only upper secondary
requirements. educational attainment and graduating from
The analysis of an individuals highest educa- higher education, for those who have survived
tional attainment level, however, does not address the rst transition (conditional probability). Thus,
the way persons accumulate formal education educational opportunities are formed in a step-
over the life course, namely, in a sequence of irre- wise process over the life course and can change
versible steps (Mare 1993: 353). Individuals and across birth cohorts (see Elder et al. 2004). Based
families make a sequence of decisions to con- on this life course approach, we are able to study
tinue or to drop out from the educational process. not only the stepwise selection process over the
At each attainment level, one can distinguish life course but also the changes in opportunities
between (1) these decisions, (2) getting admitted across birth cohorts (see Elder et al. 2004).
to the educational institutions, (3) educational
participation in these institutions, and (4) gradua-
tion from these institutions. In our study, we ana- 2.2 Mechanisms of Social Origin
lyze only whether individuals have successfully
completed educational attainments. For each It is well known that the parents educational
cohort this yields transition probabilities from attainment has a strong impact on childrens edu-
lower to higher attainment levels that show when cational success in school. Figure 1 traces the
and at which educational level the greatest implications of a set of causal mechanisms
amount of dropout in the educational career generating the associations between parental
occurs. Across cohorts, these stepwise transition education and childrens opportunities at succes-
probabilities also allow us to study in which sive educational transitions. Given the PIAAC
phase of the educational career the changes in the data, we have to limit our empirical analysis to
226 P.N. Blossfeld et al.

Fig. 1 Causal model of the effects of parental education on educational transitions (Source: Authors presentation)

the total effect of parental education on educa- tions with children, the parental nurturance of
tional transitions (shown as solid lines in Fig. 1) offspring, and the level of cognitive stimulation
rather than its various indirect effects mediated that is provided in the home environment. Better
by different life course processes (shown as educated parents are also more able to support
dashed lines in Fig. 1). In our theoretical exposi- their children in school by parental tutoring and
tion, however, we develop the implications of the homework assistance (Pfeffer 2006). Shanahan
more differentiated model for our analysis of the et al. (2008) introduced the concept of environ-
total effects of parental education. Based on the mental contingency to refer to the congura-
current state of education research, we distin- tions of such contextual factors that jointly
guish four types of indirect effects of parental produce an environmental effect. There is plenty
education on childrens success at the two educa- of empirical evidence that all these factors gen-
tional transitions: the primary effect I mediated erate individual differences in brain develop-
via cognitive skills, the primary effect II medi- ment and cognitive abilities (see Hackman et al.
ated via noncognitive skills, the secondary effect 2010). Since cognitive skills acquired early in
mediated via educational choices, and the ter- life facilitate later learning (learning begets
tiary effect mediated via teachers and schools as learning) and reduce the costs of acquiring fur-
institutions. ther cognitive skills (Carneiro and Heckman
Boudon (1974) called inequalities in cogni- 2005: 90), there is a tendency for family-induced
tive abilities generated in a childs formative inter-individual differences in cognitive achieve-
years by differences in family background the ments to be fairly stable or to even increase sub-
primary effect of social origin on school success sequently in the life course. Empirical results are
(see primary effect I in Fig. 1). The power of generally in line with this argument; it is widely
cognitive skills in predicting educational success accepted that early experiences in the family are
in school and later in adult life is well docu- particularly important for the long-term trajec-
mented in the literature (Almlund et al. 2011). tory of childrens achievements in school and
This primary effect I is strongly determined by later life (see Carneiro and Heckman 2005;
parents educational resources because parents Weinert et al. 2011). In sum, the rst theoretical
are the major agents shaping primary and, to a assumption of our empirical analysis is that the
large extent, also secondary socialization pro- higher the parental education level, the higher
cesses. Their education is assumed to have a the average academic achievement of children in
strong inuence on their everyday life interac- school and the higher the transition probabilities
Changes in Educational Inequality in Cross-National Perspective 227

from one educational level to the next in the edu- skills become increasingly important as students
cational system (see Fig. 1). age and are forced to make their own decisions
Success at the various educational transitions and to regulate their own learning behavior. Thus,
is also determined by noncognitive skills (see pri- the second theoretical assumption of our analysis
mary effect II in Fig. 1). In psychology, noncog- is that the higher the parental education, the more
nitive skills refer to personality indicators, such parents foster noncognitive skills supporting
as metacompetencies, self-concept, self- childrens academic achievement in school and
direction, self-regulation and social competences the higher the transition probabilities from one
(Weinert et al. 2011). These personality traits educational level to the next in the educational
help individuals to regulate their own learning system (see Fig. 1).
and academic achievement. They include interest Childrens educational attainment is also
in learning, taking the initiative to learn, the strongly affected by the so-called secondary
knowledge about ones strengths and weaknesses effects of social origin (see Boudon 1974) (see
in learning, the setting of realistic learning goals, Fig. 1). The secondary effect means that, even if
the planning of learning processes, the motiva- we hold childrens achievement (cognitive and
tion to achieve, etc. Social competencies are peo- noncognitive abilities) constant, families with
ples (reexive) abilities to communicate, higher educational resources make different edu-
interact, cooperate, learn and work with one cational choices at the various educational transi-
another. Social competencies are important in tions. There are several good reasons why better
situations where teamwork is essential. educated parents are in general more inclined to
Educated parents do not only foster childrens push their children to the respective next educa-
cognitive skills but also play an absolutely cen- tional attainment level at almost every ability
tral role in the development of childrens noncog- level (see Bernardi and Cebolla-Boado 2014):
nitive skills by the character of their interactions First, at each educational transition, families
with children and the organization of learning make cost-benet considerations when they
environments at home (Carneiro and Heckman decide whether or not to give their children a bet-
2005; Almlund et al. 2011). Parents can inuence ter educational opportunity. In making this deci-
childrens goals and perceptions as well as their sion, their considerations are shaped by their
self-concepts and opportunities through the structural positions in the system of social
information and experiences they provide. They inequality (Breen and Goldthorpe 1997). The
can also act as role models that shape childrens (subjectively) expected costs of higher educa-
view of how to behave in certain situations tional transitions are higher for families with
(Schoon and Eccles 2014: 49). Children from lower resources and the utilities and (subjective)
more highly educated families will also prot success probabilities are lower for parents with
from a stronger academic climate at home. lower resources (Boudon 1974). Typically, fami-
Behavior genetic studies demonstrate that not lies with fewer resources have a higher time pref-
only cognitive skills but also noncognitive skills erence (which means that they value future higher
are the result of complex geneenvironment inter- educational attainment levels less) and a higher
actions (Johnson et al. 2011; see also Elder and risk aversion (choosing an academically more
Shanahan 2006). The development of noncogni- challenging but economically more rewarding
tive skills is shaped by the family early in life; academic track) (Breen et al. 2014). In a recent
these skills crystallize from childhood through publication, Bukodi and Goldthorpe (2012) esti-
early adulthood. There is empirical evidence that mated the effects of parents class, status and
in this process individuals tend to increasingly education on inequalities of educational
select and manipulate their own environmental opportunities simultaneously. Although the three
circumstances according to their developing per- measures of social origin can be linked to differ-
sonality traits (Johnson et al. 2011; Shanahan ent theoretical mechanisms by which parental
et al. 2008). Over the life course, noncognitive resources impact family decisions, they are also
228 P.N. Blossfeld et al.

highly correlated. Thus, it is generally hard to tainty of educational decision making. In addi-
disentangle and identify the partial effects in sta- tion, the status maintenance mechanism,
tistical terms. We therefore consider these origin described by Breen and Goldthorpe (1997: 283)
measures as complementary rather than compet- for social class, is an important mechanism for
ing. In our analysis of the PIAAC data, we can the intergenerational transmission of educational
only rely on the highest education of father and attainment levels. Parents want their children to
mother as social background information. We attain at least the same level of education that
apply the dominance approach to dene social they have achieved themselves. If children move
origin and take the higher educational attainment down relative to their parents in terms of educa-
level of both parents as an origin measure tional attainment, these losses loom larger (and
(Erikson 1984). In other words, we can only are particularly painful for parents) than similar
model part of the impact of family origin on edu- gains in educational attainment (Kahneman and
cational opportunities and our educational origin Tversky 1979). Thus, differently educated par-
measure will partly reect other family resources, ents have varying aspiration levels right from the
too. However, since the excluded background beginning so that families do not consider chil-
characteristics of parental class, occupational sta- drens absolute educational outcomes but focus
tus and income are causally and temporally sub- on the outcomes relative to their own educational
sequent to parental educational attainment, one degree. Put another way, if we consider our two
can conceive our approach as an analysis of the stage nested sequential transition model, we
total effect of parental education (Pfeffer 2006:2). expect that children from upper secondary and
New studies demonstrate that parental educa- tertiary education backgrounds are more similar
tion is a relatively powerful social origin factor in in their probabilities at the rst transition because
the educational attainment process in modern parents at both levels want their children to obtain
societies (Ermisch and Francesconi 2001; Buis at least an upper secondary degree. We expect a
2013; Bukodi and Goldthorpe 2012; Baker 2014). greater gap in aspirations and transition proba-
This is the case because economic obstacles for bilities between families with tertiary and upper
education, in particular for children from less secondary education at the second transition
advantaged families, have been strongly reduced because parents with a tertiary degree have an
by improving economic living conditions over even higher structural aspiration level. Of course,
the last decades (Breen et al. 2009) and by declin- the increasing saturation of educational attain-
ing family size (Blake 1989). In addition, Breen ments from below is accompanied by a social
et al. (2009) argue that the direct economic costs devaluation of lower educational attainments
of education strongly declined through the aboli- across cohorts (see Blossfeld and Blossfeld
tion of school fees and the lengthening of com- 2014). Consequently, in the process of educa-
pulsory school duration in modern societies. In tional expansion we should expect an upward
summary, parental education seems to have drift in families aspiration levels over time and
become the key origin factor in modern society therefore an increase in educational demand
when it comes to childrens educational opportu- across cohorts. In particular, when the transition
nities (Baker 2014). from lower secondary education to upper second-
In particular, educated parents have better ary education is becoming universal, this will
knowledge about the educational system given structurally push children from parents with
their own experiences with more demanding edu- lower secondary education to increase the edu-
cational pathways (see Lucas 1999). In other cational level of their children across cohorts.
words, they are better able to consult, guide and However, as Blossfeld and Blossfeld (2014) have
maneuver their children through the upper parts argued in an earlier paper these particular educa-
of the school and university system. Their knowl- tional gains are to some extent illusive because
edge about the requirements and academic the normative minimum denitions are changing
demands of school institutions reduces the uncer- as well at the same time across cohorts.
Changes in Educational Inequality in Cross-National Perspective 229

Finally, our theoretical model in Fig. 1 reminds to identify each of the four partial indirect effects
us that parents educational attainment levels are of parental education directly, but this is not pos-
expected to have an impact on the probability of sible with the PIAAC data. To do this, prospec-
educational transitions via the evaluations of stu- tive measures of abilities or retrospectively
dents performance by teachers and educational collected school marks for a long series of suc-
institutions. There is plenty of empirical evidence cessive birth cohorts in a broad array of modern
that children from higher social origins even if societies would be necessary. Such data are not
they have the same academic achievement levels available today.
or the same marks in school as children from less
educated families (see e.g. Dollmann 2011) get
teachers recommendations for demanding edu- 2.3 Competing Theories About
cational institutions more easily or have a higher Changes in Origin-Specic
probability of admission by the next higher edu- Educational Opportunities
cational institution. There are several reasons for Across Cohorts
this advantage of children from higher social ori-
gins (Ditton 2010): (1) Teachers in general attest There are several competing theories about
that children from better educated families have changes in the inequality of educational opportu-
more adequate noncognitive skills. Thus, as nity that occur in the process of educational
described by Bourdieu (1973), it seems that more expansion. First, there is the modernization the-
highly educated parents equip their offspring ory, positing that educational expansion will lead
with a better understanding of the school culture to a general decline in the inequality of educa-
and the ability to act within it (cultural capital) as tional opportunities among all groups across
well as with the corresponding dispositions and cohorts (cf. for example Lenski 1966; Treiman
perceptions (habitus). (2) Teachers assume that 1970). According to this theory the educational
better educated parents are in general more able system mainly expands in response to the func-
to provide support to their children, if necessary. tional requirements of modern societies. In the
So, they ascribe a higher success probability to modernization process, all parents are increas-
these childrens future school careers. And (3) ingly better informed by mass media about the
better educated parents in general exert more educational opportunities of their children so that
pressure on teachers and educational institutions, the secondary effect of social origin is declining.
if this should be necessary. Esser (2014) called In addition, as the educational selection proce-
these mechanisms the tertiary effect of social ori- dures become more rational and less ascriptive,
gin (see Fig. 1). Thus, the fourth theoretical the tertiary effect of social origin declines.
assumption of our empirical analysis is that the Educational opportunities are therefore increas-
higher the education of parents, the more favor- ingly dependent only on students achievements
able are the evaluations of childrens perfor- (primary effect of social origin). However, even
mances by teachers and educational institutions the primary effect is likely to decline in the mod-
and the higher the transition probabilities from ernization process if educational policy increas-
one educational level to the next. ingly invests in early childhood intervention
The four supposed indirect effects of parental programs, all-day schools etc.
education on the educational transition probabili- A competing hypothesis is put forward by
ties in our theoretical model in Fig. 1 all have a the cultural reproduction theorists (see Bowles
positive sign and are complementary. Thus, from and Gintis 1976; Bourdieu 1973; Collins 1971).
this perspective, it is theoretically plausible to They claim that educational certicates help to
estimate a model of the total effect of parental legitimize social inequality in modern societies.
education on the transition probabilities, summa- A major function of educational credentials is to
rizing the various positive indirect effects of par- exclude members of lower social origins from
ents education. Of course, it would be interesting attractive positions in the labor market.
230 P.N. Blossfeld et al.

Reproduction theorists, however, recognize an This 80 % threshold is quite arbitrary and hard
inherent conict between the selection and to justify, since the inequality of educational
socialization functions of education. The role of opportunity might also decline under less extreme
educational institutions is to integrate children conditions. Based on a ceiling effect, increases in
from lower social origins and ethnic minorities transition probabilities exponentially decline
into the dominant value system of the society across cohorts. That is, as the transition probabil-
(see also Meyer et al. 1977). Educational expan- ity for a social group increases, there is less room
sion in upper secondary school is consistent for additional gains. This means that the gains in
with the rising demands of disadvantaged the transition probabilities for children from
groups for more education. In other words, in more highly educated families, who are the fore-
the process of educational expansion from runners, might be smaller than the gains for chil-
below, not only the attainment of primary but dren from less well educated families, who are
also lower and upper secondary schooling will the latecomers. As a result, the inequality of edu-
become universal and increasingly independent cational opportunities declines before the satura-
of social background (Shavit and Blossfeld tion levels of the more privileged groups are
1993). In this process, the better educated fami- reached.
lies are always forerunners and the less educated For the (former) socialist countries in Europe
families are latecomers across cohorts. At the specic changes in the inequality of educational
same time, the privileged groups want to main- opportunities have been observed. For example,
tain their advantages in the upper reaches of Simkus and Andorka (1982) reported for Hungary
social inequality. Hence, they preserve their and Gerber and Hout (1995) for Soviet Russia
privileges in the attainment of higher education. that the effect of social origin on access to univer-
This is possible because of the persistence of sity increased with the establishment of new
primary, secondary and tertiary effects of social elites in the period of state socialism. There is
background across cohorts (see Fig. 1). Thus, evidence that the effect of social origin on educa-
we expect that in the process of expansion of tional attainment rose in formerly socialist
tertiary education the increasing educational Russia, Hungary, Romania and the Slovak
demands of children from more highly educated Republic during the period of transition from
families should always be greater than the socialism to capitalism (Gerber 2000; Iannelli
increasing educational demands of children 2003). It is assumed that the steeply rising com-
from less well educated families. petition among social groups in the transition
Raftery and Hout (1993) suggested an elabo- from a socialist society, committed to social
ration of the reproduction theory regarding edu- equality, to a market society will lead to rising
cational transitions. Their Maximally Maintained inequality of educational opportunity. In our
Inequality (MMI) hypothesis states that the effect empirical analysis, we will test whether these for-
of educational origin on making an educational mer socialist societies follow a specic path in
transition declines across birth cohorts only when the development of origin-specic educational
the privileged groups are more or less saturated opportunities.
at a transition. According to Hout (2006:239) the
saturation threshold is reached, if the propor-
tion of successful students from privileged back- 2.4 Changing Gender
grounds exceeded 80 percent. Put in another Mechanisms Across Cohorts
way, educational expansion only leads to an
increase of educational opportunities for children The basic mechanisms through which parental
from lower educational backgrounds (the late- education inuences the educational
comers), if children from higher educational opportunities of children at various educational
backgrounds (the forerunners) have a transition transitions can be assumed to be at work for both
probability of at least 80 %. men and women (see Fig. 1). However, theory
Changes in Educational Inequality in Cross-National Perspective 231

and empirical research suggest that there are also Girls more school-adequate noncognitive
important differences in the educational oppor- skills also have a strong impact on the evaluation
tunities of men and women by birth cohort and of student performance by teachers. Empirical
social origin. In particular, in most modern soci- studies demonstrate that girls typically get better
eties women of all social origins have achieved grades in their courses in school relative to their
striking gains relative to men in education across performance level on standardized tests (DiPrete
cohorts. Various theories have been offered to and Buchmann 2013). In tracking systems such
explain this dramatic change (see DiPrete and as the German one, where the recommendation
Buchmann 2013; Becker 2014; Schoon and of teachers is often a necessary prerequisite for
Eccles 2014). In the following, we will discuss the transition to upper secondary school, it has
gender-specic mechanisms focusing on the been shown that boys need higher academic skills
four main theoretical dimensions of school suc- than girls to receive a recommendation for the
cess shown in Fig. 1: (1) cognitive skills, (2) gymnasium (Bos et al. 2007). Thus, teachers
noncognitive skills, (3) evaluation of student evaluations of students performance favor wom-
educational performance by teachers and institu- ens higher success rates at educational
tions, and (4) educational decisions of parents transitions.
and individuals. These stylized empirical facts suggest the con-
With regard to cognitive skills, large-scale clusion that girls always have performed better in
assessments, such as TIMSS (Trends in school than boys (DiPrete and Buchmann 2013).
International Mathematics and Science Study), Girls better cognitive skills, their more school-
PIRLS (Progress in International Reading adequate noncognitive skills, and their advantage
Literacy Study) and PISA (Programme for in receiving better grades for the same perfor-
International Student Assessment), suggest that mance in school, are very important factors
women compared to men have an edge in read- underlying the catch-up of females relative to
ing and some weaknesses in mathematics and males in most modern societies and even the
science. However, there is no strong trend in reversal of the gender gap in some countries.
gender-specic academic abilities that could However, they do not explain this dramatic his-
possibly explain the striking gains of women in torical change in gender-specic educational par-
the educational arena in modern societies during ticipation. For example, these advantages cannot
the last decades (DiPrete and Buchmann 2013). explain why young women born in the 1950s or
In terms of noncognitive skills, there seem to be 1960s did not attain higher educational attain-
at least three good reasons for girls greater ment levels than their male contemporaries in
school success (DiPrete and Buchmann most modern societies. It is reasonable to assume
2013:101112): (1) girls on average have an that these women were often barred from going
advantage relative to boys in terms of social to upper secondary schools or university col-
behavior in school: they show higher rates of leges. Thus, a declining educational gender gap
conformity to the school culture and they have and perhaps its reversal in modern societies
lower rates of disruptiveness, aggression, antiso- hinges on an additional theoretical explanation.
cial behavior or attention disorders. (2) Girls on Educational expansion occurred during a his-
average work harder in school and invest greater torical period when gender roles in modern soci-
efforts in doing their homework. And (3) girls on eties were culturally transformed and when
average express a stronger interest in and enjoy- discrimination against women was declining
ment of school. Given these gender differences (DiPrete and Buchmann 2013). One of the most
in noncognitive skills, it is not surprising that important structural developments in contempo-
girls are better able to turn their small cognitive rary societies has been the change in womens life
ability edge at the beginning of school into course patterns (rising employment, declining
higher levels of cognitive growth during the fertility, and the changing use of time) and at
school career (Matthew effect). the same time the relative stability of mens
232 P.N. Blossfeld et al.

roles in the life course (even if there has been a (Steinmetz 2013), a considerable proportion of
slight change in the provider role). This change men still seeks employment in blue-collar jobs
has had a profound impact on the educational and technical occupations, where secondary
decisions of parents. In the 1960s and 1970s, qualications in terms of vocational training are
families in modern societies were still character- often quite sufcient. Thus, for many males the
ized by a male breadwinner-female home- incentive to attain higher formal education seems
maker model and (marital) unions were to be smaller than for women. In summary, we
comparatively stable. At that time, girls educa- expect that the female advantages in academic
tion was less important than boys in terms of performance in school, together with the chang-
families intergenerational status reproduction ing incentives of parents to give their daughters a
(Goldthorpe 1983). This was the time when better education, have increased womens rate of
gender inequality in education meant female dis- educational attainment beyond the rate of their
advantage (DiPrete and Buchmann 2013: 1). male counterparts across cohorts. In other words,
With increasing womens labor force participa- women are increasingly expected to catch up
tion (see Blossfeld and Hakim 1997), however, with men or even outperform men at higher lev-
modern countries are experiencing a fundamental els of educational attainment.
shift from a male breadwinner society towards a In addition, we expect that the rate at which
dual earner society (Blossfeld and Drobni 2001). this change happens is different for women from
In dual earner societies, the status of a family is different educational origins. Since better edu-
not only determined by the economic position of cated parents always tended to hold more gender-
the husband but by the occupational positions of egalitarian attitudes and values and for a long
both (marital or non-marital) partners. Womens time tended to make sure that their sons and
rising gainful employment in gender-segregated daughters received similar educational opportu-
labor markets of modern service societies (see, nities, we expect that the differences between
for example, Steinmetz 2013) means that each men and women with better educated parents are
younger cohort of women can seek employment smaller than for less well educated families
in skilled service and administrative jobs as well among the older birth cohorts (e.g. at the begin-
as in the semi-professions and professions. In ning of the observation window of our study). In
these types of jobs, formal education is a particu- other words, better-educated families have been
larly important requisite for occupational access. forerunners in terms of modern gender-role ori-
In other words, across birth cohorts, the value of entations and educational participation in mod-
womens educational investments in early life ern societies and lesser-educated families are
has sharply increased and parents rising invest- latecomers in this respect. Given the constraints
ments in girls educational attainments can be of the ceiling effect in the educational transitions
viewed as a highly adaptive family strategy in discussed above, we therefore expect that in the
modern societies. These higher educational process of educational expansion women from
investments also make women less vulnerable less well educated families prot more in terms
with regard to union instabilities. In comparison, of their educational success and catch up with
mens adult roles as breadwinners have not men from the same educational origin in succes-
changed so much in modern life courses. After sive cohorts.
entering the labor market, they are still expected
to work full-time until retirement. Based on the
upgrading of skills in the labor market, mens 3 Data and Methods
educational attainment has also become more
important across cohorts. Thus, both womens Using truly equivalent educational data from the
and mens demand for education is rising. PIAAC study, we conduct a cross-national com-
However, based on the high level of occupational parison of changes in the effects of parental edu-
gender-segregation in modern labor markets cation on educational opportunity of men and
Changes in Educational Inequality in Cross-National Perspective 233

women during the process of educational expan- interaction effects are included in our maximum
sion. PIAAC focuses on further education in likelihood estimations:
adult lives and collects cross-sectional informa-
tion on education as well as key cognitive and Parental education We applied the so-called
noncognitive workplace-related skills. The dominance approach (see Erikson 1984) and
PIAAC data set was conducted by the assigned the highest educational level of the
Organisation for Economic Co-operation and parents to the origin family as a whole. We
Development (OECD) and collected in most include dummy variables indicating the high-
countries between August 2011 and March 2012. est educational level of the family: lower sec-
It includes adults between 16 and 65 years old ondary, upper secondary, and tertiary (ref.)
from 24 countries. We restrict our analysis to 22 education.
countries since data access for Cyprus and Cohort trend variable We include a trend vari-
Australia was limited due to political difculties able for the seven birth cohorts and assign the
and data protection restrictions. We include in following values: 1 for cohort 19471952,
our analysis the following countries: Austria 2 for cohort 19531957, 3 for cohort
(AT), Belgium (Flanders) (BE), Canada (CA), 19581962, 4 for cohort 19631967, 5 for
the Czech Republic (CZ), Denmark (DK), cohort 19681972, 6 for cohort 19731977
Estonia (EE), Finland (FI), France (FR), and 7 for cohort 19781982. We tested this
Germany (DE), Ireland (IE), Italy (IT), Japan trend variable in all models against the full set
(JP), the Republic of South Korea (KR), the of birth cohort dummy variables and found
Netherlands (NL), Norway (NO), Poland (PL), that the trend variable excellently represents
the Russian Federation (RU), the Slovak the monotonic changes across cohorts in all
Republic (SK), Spain (ES), Sweden (SE), the countries. Thus we are using the more parsi-
United Kingdom (England and Northern Ireland) monious specication.
(UK), and the United States (US). For a more Gender We include a dummy variable for gen-
detailed description of the PIAAC study design der (1 indicating females and 0 indicating
and the data set we refer the reader to Kirsch and males (ref.)).
Thorn (2013). Migration background In order to control for
PIAAC provides comparable information on migration background we use two dummy
the highest educational levels of the respondents variables indicating that one parent is a
and their parents. As described earlier, PIAAC migrant or both parents are migrants. Natives
was specically designed to establish conceptual are the reference category.
equivalence in the classication of educational Birth cohort * Parental education interactions
attainment levels across countries. The data in the We included interactions between the cohort
scientic use le were coded according to the trend variable and dummy variables for paren-
International Standard Classication of tal education.
Education (ISCED 1997) (OECD 1999). The Birth cohort * Gender interactions. We
randomly selected total sample of adults in the 22 include interactions between the cohort trend
countries analyzed here includes 152,514 indi- variable and the gender dummy variable.
viduals. We restrict our subsample to 100,998 Parental education * Gender interactions We
individuals aged 3065 to make sure that respon- include interactions for the gender dummy
dents have already completed their highest level variable and the dummy variables for parental
of education. education.
We used logistic regression to estimate the
sequential transition model specied in our theo- Since PIAAC applied a complex multistage
retical section (see Buis 2011; Rohwer 2014) and stratied cluster sampling design, we used the
estimated the models separately for each country. PIAAC sampling weights in our Stata estima-
The following explanatory variables and their tion procedures for each of the 22 countries.
234 P.N. Blossfeld et al.

In particular, we applied the Jackknife Repeated expansion of secondary education. This country
Replication (JRR) and Balanced Repeated has a long tradition of universal public education.
Replication (BRR) methods to estimate the sam- Based on the introduction of the comprehensive
pling variance (see Heeringa et al. 2010). high school, enrollment in secondary schools
Based on our logistic probability models, we already steeply increased from 7 % in 1890 to 80
compute for each country and for each of the two % in the 1960s (Benavot and Resnik 2006). Thus,
transitions separate predicted probability plots regarding the rst transition, the United States
(with their condence bands) for the native born was already fairly saturated at the upper second-
men and women from families with the three dif- ary education level at the beginning of the obser-
ferent education levels across the seven birth vation window (birth cohort 19471952). The
cohorts (Hanmer and Kalkan 2013). Netherlands strongly expanded the transition
probability at the rst transition but did not
increase the transition probability at the second
4 Results one. In Austria, the Netherlands, Italy and
Norway the upper secondary level expanded
4.1 The Overall Educational more strongly than the tertiary level. The oppo-
Expansion Dynamics site happened in the Czech Republic, Poland and
in the 22 Countries Russia. Here the tertiary level expanded more
strongly than the upper secondary level. It is
To set the context for our comparative analysis of remarkable that these three countries are former
the changes in origin- and gender-specic educa- socialist countries. At the time of the transition
tional inequalities, we briey describe the overall from socialism to capitalism in the early 1990s,
educational expansion in the 22 countries. Table 1 the educational systems of former socialist coun-
compiles the overall expansion intensities at the tries were characterized by a broad inclusion of
transitions from lower secondary education to students in upper secondary school and a com-
upper secondary completion and from upper sec- paratively small tertiary educational sector. Thus,
ondary education to tertiary graduation between the pathway to tertiary education was strongly
the birth cohorts 19471952 and 19781982 in restricted by the socialist states. Since the fall of
the 22 countries. The overall change in the prob-
ability of making the rst or second transition is
classied as follows: a drop by more than 10 per- Table 1 Comparison of the overall changes in the transi-
centage points is dened as ; a drop by less tion probabilities at the rst and second transition by
country (birth cohorts 19471952 and 19781982
than 10 percentage points is classied as ; no
compared)
change is classied as 0; an increase by less
than 10 percentage points is classied as +; and Change at the transition from
upper secondary to tertiary
an increase of more than 10 percentage points is education
classied as + +. 0 + ++
There are two countries, the United States of Change at the
America and the Netherlands, which stand out in transition from
their dynamics of expansion in the observation lower to upper
0 US
secondary
window. The United States has a constant transi- + EE, DE CZ, PL,
education
tion probability at the rst transition (from lower RU
secondary to upper secondary education) and ++ NL AT, BE, CA,
shows only a moderate further increase of the NO, IT DK, FI,
FR, IE,
probability at the second transition (from upper JP, KR,
secondary education to tertiary graduation). Here SK, ES,
it is important to note that in comparative terms SE, UK
the United States has been a forerunner in the Source: Authors analysis
Changes in Educational Inequality in Cross-National Perspective 235

the Iron Curtain, these so-called transition coun- increase in the inequalities of educational oppor-
tries have been catching up with other modern tunities in the transition from upper secondary to
societies by expanding their higher education tertiary education (which in all three countries
sector in particular. constitutes a historical period in which the over-
In summary, the overall picture in the 22 coun- all transition probability has been rising). As a
tries is one of impressive educational expansion representative example for these three countries,
at both the rst and the second educational transi- we describe the inequality trends for the United
tions within the observation window (see States in more detail.
Table 1). Modern countries typically have univer- Figure 2 shows how the predicted transition
salized their educational systems step by step probabilities to upper secondary and tertiary edu-
from the bottom to the top and turned their ter- cation changed for successive birth cohorts by
tiary educational systems into institutions of parental education and gender in the United
mass education. States. For the transition to upper secondary edu-
cation the transition probabilities for the different
educational origin groups and for men and
4.2 Changes in Origin- women are fairly stable across cohorts (see the
and Gender-Specic two plots on the left panel of Fig. 2 for men and
Educational Opportunities women). Children from parents with upper sec-
ondary and tertiary education have been com-
The question now arises as to what extent men pletely saturated at this transition and did not
and women from different educational back- differ signicantly within the observation win-
grounds beneted from the rising educational dow. In other words, for these two types of origin
opportunities in the 22 countries. In order to families we do not really nd children with less
answer this question, we have estimated logistic than upper secondary education anymore. This is
regression models for the rst and second transi- in agreement with the expectation from the status
tions for each of the 22 countries. The plots of the maintenance hypothesis that children from par-
predicted transition probabilities to upper sec- ents with upper secondary and tertiary education
ondary and tertiary education across the seven are similar in their probabilities at the rst transi-
birth cohorts for men and women from the three tion because parents of both families want their
educational origin families (lower secondary, children to obtain at least an upper secondary
upper secondary and tertiary education) and all degree. In addition, the status maintenance
22 countries is voluminous. These plots are avail- hypothesis (Breen and Goldthorpe 1997) sug-
able from the authors by request. However, the gests a greater gap in transition probabilities
countries can easily be grouped according to between children from these two better educated
their patterns of change in educational inequality families and the children from parents with lower
over time. Given the limited space for this contri- secondary education at the rst transition because
bution, we only describe in detail the develop- parents with lower secondary education have a
ment of educational inequalities based on the much lower structural aspiration level. And
predicted probabilities in three countries. indeed, there is a big and signicant gap between
First, the basic pattern of development of edu- the children from lower secondary education
cational inequalities is very similar in the United families and the other two groups with better
States (US), Germany (DE), and Slovakia (SK). educated parents. It is however surprising that
All three countries have in common that for more there has been a high degree of stability in the
than 30 years: (1) they constantly excluded a sub- transition probability (of about 80 %) for chil-
stantial proportion of children from families with dren from parents with lower secondary educa-
lower secondary education from a successful tion for such a long time. Since the proportion of
transition to upper secondary education; and, at parents with only lower secondary education in
the same time, (2) they have experienced an our data set has been declining from about 30 %
236 P.N. Blossfeld et al.

US: Transition lower secondary - upper secondary education (men)


0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 US: Transition upper secondary - tertiary education (men)

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Pr (1. Transition)

Pr (2. Transition)
2

2
-5

-5

-6

-6

-7

-7

-8

-5

-5

-6

-6

-7

-7

-8
47

53

58

63

68

73

78

47

53

58

63

68

73

78
19

19

19

19

19

19

19

19

19

19

19

19

19

19
Birth cohort Birth cohort
Parental education: Lower secondary Parental education: Lower secondary
Parental education: Upper secondary Parental education: Upper secondary
Parental education: Tertiary Parental education: Tertiary

US: Transition lower secondary - upper secondary education (women) US: Transition upper secondary - tertiary education (women)
0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Pr (1. Transition)

Pr (2. Transition)

2
2

-5

-5

-6

-6

-7

-7

-8
-5

-5

-6

-6

-7

-7

-8

47

53

58

63

68

73

78
47

53

58

63

68

73

78

19

19

19

19

19

19

19
19

19

19

19

19

19

19

Birth cohort Birth cohort


Parental education: Lower secondary Parental education: Lower secondary
Parental education: Upper secondary Parental education: Upper secondary
Parental education: Tertiary Parental education: Tertiary

Fig. 2 Predicted transition probabilities (with 95 % condence intervals) to upper secondary and tertiary education for
successive cohorts in the United States (US) (Source: Authors calculation)

(in cohort 19471952) to about 14 % (in cohort from parents with tertiary education could benet
19781982), this means that a proportion of from the moderate overall educational expansion
about 46 % of children of each birth cohort have of higher education in the United States, while
been excluded in the United States. Of course, the children from families with lower and upper
this is only a small percentage. However, this secondary education clearly lose out at this sec-
exclusion is consequential for these childrens ond transition to tertiary graduation when com-
later life course in a country where the returns to peting with children from tertiary educated
educational attainment have been rising over families. To a large extent this lower graduation
time (DiPrete and Buchman 2013:7). rate might be due to the greater college dropout
If we move to the second transition (from probability of youth from lower social back-
upper secondary to tertiary education) in the grounds. The gains in graduation from college
United States (see the two plots on the right panel are stronger for women than for men, so that
of Fig. 2 for men and women), it is obvious that women from families with tertiary education are
the inequality between the educational opportu- clearly the winners with respect to educational
nities of children from different origin groups is expansion in the United States. The rising educa-
diverging across cohorts. In particular, children tional inequalities in higher education in the
Changes in Educational Inequality in Cross-National Perspective 237

process of educational expansion is consistent gap between these two more privileged origin
with the cultural reproduction theory (Bowles groups and the children from families with lower
and Gintis 1976; Bourdieu 1973; Collins 1971) secondary education has been strongly decreas-
and the MMI hypothesis (Raftery and Hout ing across cohorts. This pattern is in agreement
1993), which claim that privileged families pre- with the MMI hypothesis (Raftery and Hout
serve or even expand their privileges in the attain- 1993) that children from lesser educated families
ment of higher educational degrees at least as can particularly prot from educational expan-
long as they are not yet saturated. sion, if the children from the better educated
In summary, we nd that during the last 30 families are already saturated at the respective
years in the United States inequalities in higher transition. It also conrms the cultural reproduc-
education have increased. These trends coincide tion theory (Bowles and Gintis 1976; Bourdieu
with the more general American experience of 1973; Collins 1971), which recognizes that an
rising societal inequalities during the last important role of educational institutions is to
decades in particular, the decline in the mini- integrate children from lower social origins into
mum wage and rapidly rising wage inequality the dominant value system of the society. And
(Katz and Autor 1999: 1547). The societal model nally, it supports the claim of the modernization
in the United States that minimizes the role of the theory that in the process of educational expan-
state, individualizes risks and promotes market sion there is a general decline in the inequality of
solutions (Esping-Andersen 1999: 75), in order educational opportunities among all groups
to provide for each citizen opportunities accord- because selection and evaluation procedures
ing to achievement regardless of social origin, become more rational and less ascriptive. In other
seems to have been gradually undermined by words, in these six countries children from fami-
strong internal and global competitive forces lies with lower secondary education are less and
since the 1980s (Katz and Autor 1999; DiPrete less likely to be excluded from the process of
et al. 2006). educational expansion and can remarkably
Second, a contrasting development of educa- improve their educational opportunities at the
tional inequalities can be observed in Italy (IT), rst educational hurdle.
Sweden (SE), the United Kingdom (UK) and the However, if we next inspect the second transi-
countries of the former Communist bloc in tion from secondary to tertiary education in
Central and Eastern Europe, such as the Czech Poland (see the two plots on the right panel of
Republic (CZ), Poland (PL), and Estonia (EE). In Fig. 3 for men and women), it becomes apparent
these six countries, educational inequalities are that there is at the same time a trend towards ris-
strongly decreasing at the rst transition and sig- ing inequality, too. In particular, children from
nicantly increasing at the second one. As a rep- parents with tertiary education proted most
resentative example for the pattern of change in from the overall educational expansion of higher
these six countries, we describe in detail the pat- education, while the opportunities of children
tern in Poland (see Fig. 3). from families with lower secondary education
The two plots in the left panel of Fig. 3 show have been quite stable across cohorts, so that they
the changes for Poland in the transition probabili- lose out in relative terms over time. Again, this
ties to upper secondary education for children trend is clearly stronger for women than for men,
from different educational origins, separated for so that women from families with tertiary educa-
males and females. Again, the children from fam- tion gradually catch up and then eventually sur-
ilies with upper secondary and tertiary education pass the males. As mentioned above, trends in
have been fairly saturated at the rst transition educational inequalities in Poland are representa-
since the beginning of the observation window, tive of the trends in educational inequality for the
supporting again the status maintenance hypoth- other two countries from the former Communist
esis (Breen and Goldthorpe 1997). However, in bloc in Central and Eastern Europe, the Czech
contrast to the United States, we now see that the Republic and Estonia, as well as for Western
238 P.N. Blossfeld et al.

PL: Transition lower secondary - upper secondary education (men) PL: Transition upper secondary - tertiary education (men)

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Pr (1. Transition)

Pr (2. Transition)
2

2
-5

-5

-6

-6

-7

-7

-8

-5

-5

-6

-6

-7

-7

-8
47

53

58

63

68

73

78

47

53

58

63

68

73

78
19

19

19

19

19

19

19

19

19

19

19

19

19

19
Birth cohort Birth cohort
Parental education: Lower secondary Parental education: Lower secondary
Parental education: Upper secondary Parental education: Upper secondary
Parental education: Tertiary Parental education: Tertiary

PL: Transition lower secondary - upper secondary education (women) PL: Transition upper secondary - tertiary education (women)
0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Pr (1. Transition)

Pr (2. Transition)
2

2
-5

-5

-6

-6

-7

-7

-8

-5

-5

-6

-6

-7

-7

-8
47

53

58

63

68

73

78

47

53

58

63

68

73

78
19

19

19

19

19

19

19

19

19

19

19

19

19

19
Birth cohort Birth cohort
Parental education: Lower secondary Parental education: Lower secondary
Parental education: Upper secondary Parental education: Upper secondary
Parental education: Tertiary Parental education: Tertiary

Fig. 3 Predicted transition probabilities (with 95 % condence intervals) to upper secondary and tertiary education for
successive cohorts in Poland (PL) (Source: Authors calculation)

European capitalist countries like Italy, Sweden, quite different societies in the process of educa-
and the United Kingdom. Thus, the pattern in tional expansion is the rising competition among
these six countries provides support for the families and individuals with different educa-
hypothesis that privileged families preserve or tional resources. Families try to do the best for
even increase their advantages in the attainment their children. However, they have different loca-
of higher educational qualications. These tions in the structure of inequality and differential
advantages stem partly from their higher aca- access to resources in order to achieve their goals
demic performance and better noncognitive skills as described in our theoretical model (see Fig. 1).
(primary effects I and II) and partly from their The PIAAC data suggest that not only the three
families and teachers educational decision mak- capitalist countries, such as Italy, Sweden and the
ing (the secondary and tertiary effects; see Fig. 1). United Kingdom, are confronted with an increase
Our result here is in agreement with cultural in the competition of families in the transition to
reproduction theory (Bowles and Gintis 1976; higher education in the last three decades, but
Bourdieu 1973; Collins 1971) and the MMI that this is especially true for the countries of the
hypothesis (Raftery and Hout 1993). The factor former Communist bloc in Central and Eastern
that presumably creates similarity among these Europe (Cummings 2003). After the fall of the
Changes in Educational Inequality in Cross-National Perspective 239

Iron Curtain, former socialist countries such as school. However, in terms of growth of students
the Czech Republic, Poland and Estonia have not who have attained at least upper secondary edu-
only strongly expanded their tertiary educational cation, South Korea has experienced an out-
systems (see Table 1), but they have increased the standing expansion in recent decades (see also
competition even more among origin families in Park 2009). Today, only very few South Korean
the transition from socialism to capitalism. The young people enter the labor market after com-
result of this rising general competition seems to pleting junior high school or middle school.
be that even in a context of massive educational Instead, the overwhelming majority of middle
expansion of tertiary education, the effect of school graduates enters high (i.e. upper second-
social origin on higher education has been rising ary) schools. South Koreas academic high
in former socialist countries (see also Gerber schools are in general much less socially differ-
2000; Iannelli 2003; Mateju et al. 2003). entiated (Park 2010). This is because South
Third, there are 13 countries (Austria (AT), Korea introduced an Equalization Policy in
Belgium (Flanders) (BE), France (FR), Canada 1974. Based on this policy, students in public
(CA), Denmark (DK), Finland (FI), Ireland (IE), and private schools are randomly assigned to
Japan (JP), the Republic of South Korea (KR), academic high schools in their districts by a lot-
the Netherlands (NL), Norway (NO), the Russian tery (Park 2010). Figure 4 shows that children
Federation (RU) and Spain (ES)) that manifest from all three educational origin families
clear patterns of declining educational inequali- strongly increased their transition probabilities
ties at both educational transitions. Of course, the in the observation window; this is true in particu-
strength of these declines differs for the rst and lar for women, who started from a lower level of
second transition as well as over the 13 countries. enrollment in birth cohort 19471952. In Fig. 4,
As a rule, the decline in educational inequalities the differences in educational opportunities
has been stronger for women than for men at both among women and men from different educa-
transitions so that the gender gaps decline and tional origins therefore strongly converge and
even tend to reverse in some of the countries. eventually completely disappear at the end of the
Hence, women are again the winners in an era of observation window (birth cohort 19781982).
educational expansion. In order to contrast the Contrary to the MMI hypothesis (Raftery and
changes in educational inequalities in these 13 Hout 1993), Fig. 4 reveals that disadvantaged
countries as much as possible with the changes in children from families with lower secondary
the United States and Poland, we select the education could signicantly increase their edu-
Republic of South Korea (KR). South Korea has cational opportunities in a period when the chil-
not only drastically expanded its overall partici- dren from higher educated families were not yet
pation in upper secondary school and tertiary saturated. Such a catch-up pattern for children
education in the observation window, but experi- from parents with lower secondary school can
enced also an impressive decline in educational also be observed for Ireland (IE), the Netherlands
inequalities at the rst and the second transition. (NL) and Spain (ES). These ndings clearly con-
The two plots on the left panel of Fig. 4 show tradict the MMI theory. They are, however, in
separately for South Korean men and women agreement with the ceiling effect hypothesis that
from different educational origins the changes in predicts declining inequality of educational
the transition probabilities to upper secondary opportunity, if the probabilities of the higher
education (rst transition). It is immediately educational origin groups are below the satura-
clear that South Korea is a latecomer in the pro- tion level and when the probabilities of lower
cess of educational expansion. Thus, there has educational groups are still particularly small.
been lots of room to expand upwards. Hence, in Thus massive educational expansion can indeed
our observation window the three educational reduce inequality of educational opportunity far
origin groups started out from a comparatively before the saturation level for the more privi-
low level of enrollment in upper secondary leged groups is reached.
240 P.N. Blossfeld et al.

KR: Transition lower secondary - upper secondary education (men) KR: Transition upper secondary - tertiary education (men)

1
0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1

0 .1 .2 .3 .4 .5 .6 .7 .8 .9
Pr (1. Transition)

Pr (2. Transition)
2

2
-5

-5

-6

-6

-7

-7

-8

-5

-5

-6

-6

-7

-7

-8
47

53

58

63

68

73

78

47

53

58

63

68

73

78
19

19

19

19

19

19

19

19

19

19

19

19

19

19
Birth cohort Birth cohort
Parental education: Lower secondary Parental education: Lower secondary
Parental education: Upper secondary Parental education: Upper secondary
Parental education: Tertiary Parental education: Tertiary

KR: Transition lower secondary - upper secondary education (women) KR: Transition upper secondary - tertiary education (women)

0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1
Pr (1. Transition)

Pr (2. Transition)
2

2
-5

-5

-6

-6

-7

-7

-8

-5

-5

-6

-6

-7

-7

-8
47

53

58

63

68

73

78

47

53

58

63

68

73

78
19

19

19

19

19

19

19

19

19

19

19

19

19

19
Birth cohort Birth cohort
Parental education: Lower secondary Parental education: Lower secondary
Parental education: Upper secondary Parental education: Upper secondary
Parental education: Tertiary Parental education: Tertiary

Fig. 4 Predicted transition probabilities (with 95 % condence intervals) to upper secondary and tertiary education for
successive cohorts in the Republic of South Korea (KR) (Source: Authors calculation)

In the two plots on the right panel of Fig. 4, the The expansion of the tertiary educational sector in
changes in educational inequalities for the second South Korea took place mainly in the 1960s,
transition (from secondary to tertiary education) 1980s and 1990s. In addition, legislation in 1995
for South Korean men and women are shown. facilitated the creation of many new universities
Clearly, there are great differences between the (Grubb and Lazerson 2009). Based on these
three educational origin groups. Thus, children reforms, all educational origin groups could prot
from more highly educated families have always from the massive and quick expansion of tertiary
had an advantage in completing tertiary educa- education in South Korea; and children from fam-
tion. In South Korea there is severe competition ilies with lower secondary education could catch
on the entrance exam for university (Jang and up in this process. As a result, educational
Kim 2004). In this competition, children from inequalities in the transition to tertiary educa-
families with tertiary education who have not only tional attainment could be effectively reduced.
higher cognitive and noncognitive skills but also The patterns of change in educational inequal-
more favorable resources at home are therefore ity in South Korea (and in the other 12 countries
more successful than children from parents with listed above) contradict the expectations from
upper secondary or lower secondary education. cultural reproduction theory (Bowles and Gintis
Changes in Educational Inequality in Cross-National Perspective 241

1976; Bourdieu 1973; Collins 1971) and the cation. In particular, we sought to understand
MMI hypothesis (Raftery and Hout 1993). They which of the educational origin groups could
rather support the modernization theory, positing prot most from the additional opportunities pro-
that educational expansion will lead to a general vided by expanding upper secondary and tertiary
decline in the inequality of educational opportu- education over time and how these changes dif-
nities among all educational origin groups (cf. for fered among a broad array of modern societies.
example Lenski 1966; Treiman 1970). It is dif- We adopted a life course perspective in order
cult to detect a common mechanism among these to get a better understanding of the processural
13 diverse countries that could explain the pat- nature of educational attainment in the life
tern of declining educational inequality. Only for courses of successive cohorts (Mare 1981).
some of these countries an explanation appears Young people accumulate formal schooling over
quite obvious: In South Korea, Japan, Spain, the life course in a sequence of irreversible steps
Finland, France, and Ireland the expansions in and individuals and families make consecutive
the upper secondary and tertiary educational sys- decisions to continue or to drop out. We looked at
tem have been particularly massive and fast. two important successive transitions in the edu-
Hence, it seems that a very strong educational cational career of children: the transition from
expansion in a short historical period made it lower secondary educational attainment to upper
possible to open up new opportunities for chil- secondary educational attainment and from upper
dren from parents with lower secondary educa- secondary educational attainment to tertiary
tion in the competition among families with graduation.
different resources. Countries such as South Based on our analysis of the rst educational
Korea, Japan and Finland are in addition good transition to upper secondary educational attain-
examples of societies that successfully combine a ment, we can summarize the evidence as follows:
high average level of academic performance in (1) Children from families with tertiary educa-
the large-scale assessments (such as PIRLS and tion were already saturated at the beginning of
PISA) with a high level of equity in educational the observation window (birth cohort 1947
opportunities for all social groups (OECD 2013). 1952). These families are therefore clearly the
forerunners in the process of expansion of the
rst transition in all countries. Thus, even if these
5 Summary and Conclusion families have a high level of educational demand,
they structurally could not further increase their
In the process of educational expansion and transition probabilities to upper secondary school
reforms the old sharp divisions between aca- in the observation window. (2) In most of the 22
demic and vocational/technical tracks in the sec- countries under study, men from families with
ondary school systems of modern societies have upper secondary education were already fairly
greatly declined and the proportion of young saturated at the rst educational transition at the
people who have completed at least upper sec- beginning of the observation window (birth
ondary education has impressively risen. Also cohort 19471952). In comparison, women from
the enrollment in higher education has quickly these families somewhat lagged behind in several
grown in all modern countries and turned their countries in the oldest cohort. This was in par-
tertiary educational systems into institutions of ticular the case in South Korea, the Netherlands,
mass education. Using truly comparative data Spain, Austria, Ireland, and Italy. However, these
from the PIAAC study, we analyzed how the edu- women proted strongly from the expansion of
cational attainment of men and women from the rst transition and caught up with their male
families with different educational backgrounds counterparts within the observation window.
has changed by the bottom-up universalization of Among the youngest cohorts, men and women
education in secondary school and by the declin- from families with upper secondary education
ing role of elite education in modern tertiary edu- were fairly saturated. (3) In almost all countries
242 P.N. Blossfeld et al.

(the exceptions are the United States, Germany hypothesis and the cultural reproduction theory
and Slovakia) there have been impressive gains but also the MMI hypothesis. However, the MMI
for children from lower secondary education hypothesis was not fully conrmed because in
which reduced the inequality of educational some countries men and women from lower sec-
attainment drastically. The United States (with its ondary education backgrounds could catch up
inclusive high school system), Germany (with its even within a context where the more educated
inclusive vocational training system) and most of families had not yet been saturated. Thus, empiri-
the former socialist countries (with their inclu- cal evidence generally is more in agreement with
sive equality programs in secondary schools), the ceiling effect hypothesis than with the MMI
had already comparatively high transition proba- hypothesis. The ceiling effect hypothesis posits
bilities for children from parents with lower sec- that inequality of educational opportunity might
ondary education among the oldest cohort. In the even decline, when the probabilities for the lower
past, these countries clearly were the forerunners educational groups are particularly small and
at the rst transition. But then they did not expand when the probabilities of the higher educational
much further within the observation window and groups are below the saturation level. (2) Women
somehow left children from lower secondary from families with lower and upper secondary
education behind. However, in the great majority education could not only catch up relative to their
of countries there have been gains for children male counterparts from the same educational ori-
from disadvantaged families. These gains have gin level but also with regard to children from
been particularly strong for women in the obser- more privileged families. Hence, one can con-
vation window, so that these women could close clude that women from lower educational origins
the gender gap at the lower end of the inequality are the big winners in the expansion of the rst
pyramid. (4) At the end of the observation win- educational transition in almost all countries.
dow (for cohort 19781982), in most countries These gains of women from families with lower
the transition to upper secondary education was secondary education occurred in a historical
saturated for all origin groups. The exceptions period when gender roles in modern societies
are Spain, Italy, Ireland and the United Kingdom have been culturally transformed and discrimina-
where the transition probabilities for children tion against women has been declining (DiPrete
from parents with lower secondary education and Buchmann 2013). With rising womens labor
were still below Houts 80 % level among the force participation, societies are experiencing a
youngest cohort (born 19781982). In a modern shift from male breadwinner to dual earner fami-
world where upper secondary education has lies (Blossfeld and Drobni 2001). Women
become the minimum standard among the increasingly seek employment in skilled service
younger generation, this lack of education for and administrative jobs. Thus, parents invest
children from disadvantaged families is likely to more in their daughters education. (3) With the
be associated with high unemployment, low universalization of upper secondary education,
career and income opportunities and a high likeli- however, the social value of upper secondary
hood of poverty in the later life course. These attainment is also changing. Reaching at least
children clearly have been left behind by these upper secondary education has become a widely
modern states. shared and common experience for most young
Our empirical ndings for the rst educational people today. When upper secondary education is
transition suggest the following four theoretical becoming nearly universal among the younger
conclusions: (1) In most countries there has been generations, it becomes a kind of necessary con-
a strong decline in educational inequality across dition for every young individual. Reaching
cohorts. Since this decline took place in a context upper secondary educational attainment is noth-
where children from more privileged families ing special anymore. The few who do not (or are
have been fairly saturated, this empirical evi- not able) to make the transition to this new mini-
dence supports not only the modernization mum educational attainment level are, of course,
Changes in Educational Inequality in Cross-National Perspective 243

becoming increasingly disadvantaged and left United Kingdom, Germany, Italy, Sweden, and
behind outsiders. In other words, the impressive the countries of the former Communist bloc in
educational gains of men and women from lower Central and Eastern Europe such as the Czech
secondary education at the transition to upper Republic, Slovakia, Poland, and Estonia) the
secondary education carry an elusive avour. In opportunities to graduate successfully from
comparison to their parents they have clearly higher education for children from lower second-
achieved a nominally higher educational attain- ary education families have become worse. In
ment. But in the process of universalization of three of these nine counties (the United States,
education from below there has also been a Germany and Slovakia), the children from these
decline in the social value of this educational families had quite unchanged opportunities to
attainment level over time. With regard to the reach an upper secondary educational attainment.
younger generations, the relative gain across The additional decline of chances at the second
cohorts seems to disappear. (4) From a life course transition means that these children from disad-
perspective the substantial reduction in origin- vantaged families in these three countries are
specic educational inequalities in reaching the clearly the losers in the process of educational
upper secondary educational level can lead to expansion. In the other six countries (Poland, the
more equality in the attainment of higher educa- Czech Republic, Estonia, Italy, Sweden, and the
tion. But this only occurs if the transition to ter- United Kingdom) children from lower secondary
tiary education does not become more unequal education increased their chances at the rst tran-
with regard to social origin. sition but lost opportunities to graduate from ter-
Our empirical ndings for the second transi- tiary education, if they had successfully
tion to tertiary educational attainment can be completed upper secondary education. Thus, the
summarized as follows: (1) In all countries men gains in the probabilities at the rst transition
and women from parents with tertiary education were at least partly counteracted at the second
had again the highest starting probabilities among transition. This result is of course not consistent
the three different origin groups in the oldest with the expectations of modernization theory
cohort (19471952). The children from these but rather supports cultural reproduction theory.
families did not only have the best opportunities In general, the countries with increasing educa-
to complete upper secondary education but, if tional inequalities at the second transition are
they had attained upper secondary education, those societies that experienced a strongly
they also had the greatest chances to graduate increasing market competition among status and
from higher education. Put in another way, the income groups in society. This is particularly true
very favorable transition probabilities of both for the liberal economies in times of globaliza-
transitions multiply for them over the educational tion (Blossfeld et al. 2005), such as the United
career and produce the highest proportion of ter- States, the United Kingdom, and the former
tiary graduates among all families. The success socialist countries, which strongly liberalized
of children from tertiary families is partly based their economies after the breakdown of the Iron
on their higher academic performance and more Curtain (the Czech Republic, Slovakia, Poland,
school appropriate noncognitive skills and partly and Estonia). From a theoretical point of view,
on the specic educational choices of these stu- the development in Sweden is most interesting.
dents and their parents. (2) The changes in transi- In the past this society was often characterized as
tion probabilities from upper secondary a country with a particularly high level of social
educational attainment to tertiary graduation and gender equality in educational attainment
have not been uniform across the 22 countries. (Erikson and Jonsson 1996). However, in the last
We observe both strongly increasing and decreas- two decades, Sweden also moved away from the
ing educational inequalities between children strict social democratic model by introducing
from different educational origins across cohorts. policies of liberalization, deregulation and priva-
(3) In nine countries (the United States, the tization. In particular, in the educational system
244 P.N. Blossfeld et al.

private schools were introduced and the responsi- career, however, consists of a sequence of partici-
bility for education was shifted from the central pations in often age-graded and institutionally
state to the community level. These reforms seem structured educational contexts. Future cross-
to have turned around the long trend towards national educational research initiatives should
more educational equality. (4) For the remaining therefore collect detailed longitudinal data on the
13 countries the transition probabilities for chil- movement of individuals through the educational
dren from parents with lower secondary educa- systems of a greater number of modern societies.
tion increased relative to the other origin groups Cross-national longitudinal data would also offer
in the transition to tertiary graduation. These the ability to nest individual educational trajecto-
countries are the Republic of South Korea, Japan, ries and educational transitions within changing
Finland, Norway, Denmark, Austria, Belgium institutional and historical contexts. Cross-
(Flanders), the Netherlands, the Russian national research would also prot from better
Federation, France, Ireland, Belgium, Canada data on the long-term relationships between par-
and Spain. In most of these countries there has ents and children and how these relationships
indeed been a substantial reduction of educa- inuence the educational careers of children,
tional inequalities because the declining inequali- adolescents, and adults over the life course in dif-
ties at the rst and second transitions add up over ferent countries. Finally, our theoretical model in
the life course. Rising educational opportunities Fig. 1 suggested that we have to go a step further
for children from lower educational origins are in collecting longitudinal data on (1) the informal
obviously possible. Thus, the question is, What and formal learning environments in the family
explains these declining educational inequali- and within the various institutional settings, (2)
ties? It seems that the massive expansion of the development of cognitive and noncognitive
higher education in a short historical period has competences and (3) the educational decision-
been an important factor. In addition, special edu- making processes over the life course.
cational policy measures, supporting particularly Longitudinal data would shed light on how dif-
disadvantaged children, are important. They ferent competencies are acquired over the life
seem to have been quite effective in reducing span, how they interact over time and across edu-
educational inequalities in several countries. cational stages, and in which way they may con-
tribute to individual and group-specic life
course outcomes. This would allow us to identify
6 Further Research Needs and compare the four types of indirect effects of
for Cross-National parental education on childrens educational
Comparisons attainment (see Fig. 1). In sum, standardized
cross-national panel data are essential for identi-
Our analysis shows that the highly standardized fying trajectories of growth and development
cross-national PIAAC data can provide new over the life course in modern societies and for
insights into the relationships between educa- understanding the causal relationships between
tional expansion and changes in inequalities of competence development, educational decision
educational opportunity across very diverse mod- making and educational participation.
ern societies. However, the PIAAC data also have
severe limitations. The most important drawback
is that they offer only a snapshot of the students References
educational position in a long career at one point
in time. For example, based on the cross-sectional Almlund, M., Duckworth, A. L., Heckmann, J. J., &
Kautz, T. (2011). Personality psychology and econom-
character of the PIAAC study, we were forced to
ics. Discussion Paper No. 5500. Bonn: IZA.
use individuals highest educational attainment Baker, D. (2014). The schooled society. The educational
in order to reconstruct their major educational transformation of global culture. Stanford: Stanford
transitions over the life course. An educational University Press.
Changes in Educational Inequality in Cross-National Perspective 245

Becker, R. (2014). Reversal of gender differences in edu- sion, time discounting preferences, and educational
cational attainment Historical analysis of the west decision-making. European Sociological Review, 30,
German case. Educational Research, 56, 184201. 258270.
Benavot, A., & Resnik, J. (2006). Lessons from the past: Buis, M. L. (2011). The consequences of unobserved het-
A comparative socio-historical analysis of primary erogeneity in a sequential logit model. Research in
and secondary education. In A. Benavot, J. Resnik, & Social Stratication and Mobility, 29, 247262.
J. Corrales (Eds.), Global educational expansion his- Buis, M. L. (2013). The composition of family back-
torical legacies and political obstacles (pp. 189). ground: The inuence of the economic and cultural
Cambridge (MA): American Academy of Arts and resources of both parents on the offspings educational
Sciences. attainment in the Netherlands between 1939 and 1991.
Bernardi, F., & Cebolla-Boado, H. (2014). Previous European Sociological Review, 29, 593602.
school results and social background: Compensation Bukodi, E., & Goldthorpe, J. H. (2012). Decomposing
and imperfect information in educational transitions. social origins: The effects of parents class, status,
European Sociological Review, 30, 207217. and education on the educational attainment of their
Blake, J. (1989). Family size and achievement. Los children. European Sociological Review, 28, 116.
Angeles: University of California Press. Carneiro, P., & Heckman, J. J. (2005). Human capital
Blossfeld, P., & Blossfeld, G. (2014). Educational expansion policy. In J. J. Heckman & A. B. Krueger (Eds.),
from below. The logic of educational expansion in mod- Inequality in America. What role for human capital
ern societies. (Unpublished paper, work in progress). policies? (pp. 77239). Cambridge: MIT Press.
Blossfeld, H.-P., & Hakim, C. (1997). Between equaliza- Collins, R. (1971). Functional and conict theories of
tion and marginalization : women working part-time educational stratication. American Sociological
in Europe and the United States of America. Oxford: Review, 36, 10021019.
Oxford University Press. Cummings, W. (2003). The institutions of education: A
Blossfeld, H.-P., & Dobni, S. (2001). Careers of couples comparative study of educational development in Six
in contemporary societies. A cross-national compari- core nations. Oxford: Symposium Books.
son of the transition from male breadwinner to dual- DiPrete, T., & Buchmann, C. (2013). The rise of women:
earner families. Oxford: University Press. The growing gender gap in education and what it
Blossfeld, H.-P., Mills, M., Klijzing, E., & Kurz, K. means for American schools. New York: Russell Sage.
(2005). Globalization, uncertainty and youth in soci- DiPrete, T., Goux, D., Maurin, E., & Quesnel-Vallee, A.
ety. London: Routledge. (2006). Work and pay in exible and regulated labor
Boli, J., Ramirez, F., & Meyer, J. (1985). Explaining the markets: A generalized perspective on institutional
origins and expansion of mass education. Comparative evolution and inequality trends in Europe and the U.S.
Education Review, 29, 145170. Research in Social Stratication and Mobility, 24,
Bos, W., Hornberg, S., Arnold, K.-H., Faust, G., Fried, L., 311332.
Lankes, E.-M., Schwippert, K., & Valtin, R. (2007). Ditton, H. (2010). Schullaufbahnen und soziale Herkunft
IGLU 2006. Lesekompetenzen von Grundschulkindern eine Frage von Leistung oder Diskriminierung? In
in Deutschland im internationalen Vergleich. Mnster: S. Aufenanger, F. Hamburger, L. Ludwig, & R. Tippelt
Waxmann. (Eds.), Bildung in der Demokratie (pp. 7999).
Boudon, R. (1974). Education, opportunity, and social Opladen: Barbara Budrich.
inquiry. Changing prospects in Western society. Dollmann, J. (2011). Verbindliche und unverbindliche
New York: Wiley. Grundschulempfehlungen und soziale Ungleichheiten
Bourdieu, P. (1973). Kulturelle Reproduktion und soziale am ersten Bildungsbergang. Klner Zeitschrift fr
Reproduktion. In P. Bourdieu (Ed.), Grundlagen einer Soziologie und Sozialpsychologie, 63, 595621.
Theorie der symbolischen Gewalt (pp. 88137). Elder, G. H., & Shanahan, M. J. (2006). The life course
Frankfurt am Main: Suhrkamp. and human development. In R. E. Lerner & W. Damon
Bowles, S., & Gintis, H. (1976). Schooling in capitalist (Eds.), The handbook of child psychology. Theoretical
America (Vol. 57). New York: Basic Books. models of human development (pp. 665715).
Breen, R., & Goldthorpe, J. H. (1997). Explaining educa- New York: Wiley.
tional differentials: Towards a formal rational action Elder, G. H., Johnson, K. M., & Crosnoe, R. (2004). The
theory. Rationality and Society, 9, 275305. emergence and development of life course theory. In
Breen, R., Luijkx, R., Mller, W., & Pollak, R. (2009). J. T. Mortimer & M. J. Shanahan (Eds.), Handbook of
Nonpersistent inequality in educational attainment. the life course (pp. 319). New York: Springer.
Evidence from eight European countries. American Erikson, R. (1984). Social class of men, women and fami-
Journal of Sociology, 5, 14751521. lies. Sociology, 18, 500514.
Breen, R., Luijkx, R., Mller, W., & Pollak, R. (2010). Erikson, R., & Goldthorpe, J. H. (1992). The constant
Long-term trends in educational inequality in Europe: ux: A study of class mobility in industrial societies.
Class inequalities and gender differences. European Oxford: Clarendon Press.
Sociological Review, 26, 3148. Erikson, R., & Jonsson, J. O. (1996). Can education be
Breen, R., van de Werfhorst, H. G., & Meier Jger, M. equalized? The Swedish case in comparative perspec-
(2014). Deciding under doubt: A theory of risk aver- tive. Boulder: Westview Press.
246 P.N. Blossfeld et al.

Ermisch, J., & Francesconi, M. (2001). Family matters: Kahneman, D., & Tversky, A. (1979). Prospect theory: An
Impacts of family background on educational attain- analysis of decision under risk. Econometrica, 47,
ments. Economica, 68, 137156. 263291.
Esping-Andersen, G. (1999). Social foundations of Katz, L. F., & Autor, D. H. (1999). Changes in the wage
postindustrial economies. Oxford: Oxford University structure and earnings inequality. In O. Aschenfelter
Press. & D. Card (Eds.), Handbook of labor economics
Esser, H. (2014). Bildungssysteme und ethnische (pp. 14631555). Amsterdam: North Holland.
Bildungsungleichheiten. Unpublished manuscript. Kirsch, I., & Thorn, W. (2013). Technical report of the
Mannheim Centre for European Social Research. survey of adult skills (PIAAC). http://www.oecd.org/
Mannheim University. site/piaac/_Technical%20Report_17OCT13.pdf
Gerber, T. P. (2000). Educational stratication in contem- Lenski, G. E. (1966). Power and privilege: A theory of
porary Russia: Stability and change in the face of eco- social stratication. Chapel Hill: UNC Press Books.
nomic and institutional crisis. Sociology of Education, Lucas, S. R. (1999). Tracking inequality. Stratication
73, 219246. and mobility in American high schools. New York:
Gerber, T. P., & Hout, M. (1995). Educational stratica- Teachers College Press.
tion in Russia during the Soviet period. American Mare, R. D. (1981). Trends in schooling: Demography,
Journal of Sociology, 3, 611660. performance, and organization. Annals of the
Goldthorpe, J. H. (1983). Women and class analysis: In American Academy of Political and Social Science,
defence of the conventional view. Sociology, 17, 435, 96122.
465488. Mare, R. (1993). Educational stratication on observed
Grubb, W. N., & Lazerson, M. (2009). The education gos- and unobserved components of family background. In
pel. Cambridge: Harvard University Press. Y. Shavit & H.-P. Blossfeld (Eds.), Persistent inequal-
Hackman, D. A., Farah, M. J., & Meaney, M. J. (2010). ity (pp. 351376). Boulder: Westview Press.
Socioeconomic status and the brain: Mechanistic Mateju, P., Rehakova, B., & Simonova, N. (2003).
insights from human and animal research. Nature Transition to university under communism and after
Reviews/Neuroscience, 11, 651659. its demise. The role of socio-economic background in
Hanmer, M. J., & Kalkan, K. O. (2013). Behind the curve: the transition between secondary and tertiary educa-
Clarifying the best approach to calculating predicted tion in the Czech Republic 19481998 (pp. 301324).
probabilities and marginal effects from limited depen- Prague: Institute of Sociology, Academy of Science of
dent variable models. American Journal of Political the Czech Republic.
Science, 57, 263277. Meyer, J. W., Ramirez, F. O., Rubinson, R., & Boli-
Heeringa, S. G., West, B. T., & Berglund, P. A. (2010). Bennett, J. (1977). The world educational revolution,
Applied survey data analysis. Boca Raton: Chapman 19501970. Sociology of Education, 50, 242258.
& Hall/CRC Press. OECD. (1999). Classifying educational programmes.
Hout, M. (2006). Maximally maintained inequality and Manual for ISCED-97 implementation in OECD
essentially maintained inequality: Cross-national countries. Paris: OECD.
comparisons. Sociological Theory and Methods, 21, Park, H. (2009). Inequality of educational opportunity in
253278. Korea by gender, socio-economic background, and
Iannelli, C. (2003). Parental education and young peoples family structure. The International Journal of Human
educational and labour market outcomes: A compari- Rights, 11, 179197.
son across Europe. In I. Kogan & W. Mller (Eds.), Park, H. (2010). Japanese and Korean high schools and
School-to-work transitions in Europe: Analyses of the students in comparative perspective. In J. Dronkers
EU LFS 2000 ad hoc module. Mannheim: Mannheimer (Ed.), Quality and inequality of educdatioan. Cross-
Zentrum fr Europische Sozialforschung. national perspectives (pp. 255273). Berlin: Springer
Jackson, M. (2013). Introduction. How is inequality of Verlag.
educational opportunity generated? The case for pri- Pfeffer, F. (2006). The institutional context of educational
mary and secondary effects. In M. Jackson (Ed.), attainment. Intergenerational educational mobility in
Determined to succeed? Performance versus choice in comparative perspective. Presented at the ECSR
educational attainment (pp. 133). Standford: Prague Conference September 2006.
Standford University Press. Raftery, A. E., & Hout, M. (1993). Maximally maintained
Jang, S., & Kim, N. (2004). Transition from high school to inequality: Expansion, reform, and opportunity in
higher education and work in Korea, from the Irish education, 19211975. Sociology of Education,
competency-based education perspective. 1, 4162.
International Journal of Educational Development, Rohwer, G. (2014). A defence of the sequential transition
24, 691703. model. (Paper is in progress).
Johnson, W., Penke, L., & Spinath, F. (2011). Heritability Schoon, I., & Eccles, J. S. (2014). Gender differences in
in the era of molecular genetics. Some thoughts for aspiration and attainment: A life course perspective.
understanding genetic inuences on behavioural Cambridge: Cambridge University Press.
traits. European Journal of Personality, 25, Shanahan, M. J., Vaisey, S., Erickson, L. D., & Smolen, A.
254266. (2008). Environmental contingencies and genetic pro-
Changes in Educational Inequality in Cross-National Perspective 247

pensities: Social capital, educational continuation, and Treiman, D. J. (1970). Industrialization and social strati-
dopamine receptor gene DRD2. American Journal of cation. Sociological Inquiry, 40(2), 207234.
Sociology, 114, 260286. UNESCO. (2012). International standard classication
Shavit, Y., & Blossfeld, H.-P. (1993). Persistent inequal- of education. UNESCO Institute for Statistics:
ity. Changing educational attainment in thirteen coun- Montreal. http://www.uis.unesco.org/Education/
tries. Boulder: Westview Press. Documents/isced-2011-en.pdf
Simkus, A., & Andorka, R. (1982). Educational attain- Weinert, S., Artelt, C., Prenzel, M., Senkbeil, M. Ehmke, T.,
ment in Hungary. American Sociological Review, 47, & Carstensen, C. H. (2011). Development of
740751. competences across the life span. In H.-P. Blossfeld,
Steinmetz, S. (2013). The contextual challenges of occu- J.-G. Robach, & J. von Maurice (Eds.), Education as a
pational sex segregation. Deciphering cross-national lifelong process. Zeitschrift fr Erziehungswissenschaft,
differences in Europe. Wiesbaden: VS Verlag. 14, 6786.
Work Over the Gendered Life
Course

Phyllis Moen

1 Introduction course. This takes the form of rst, years of prep-


aration (education), then continuous paid
The concept of work is tied to purposive activ- employment throughout adulthood, then the
ity, although the term most often refers to gainful golden years of leisure in retirement (see Kohli
(paid) rather than unpaid labor, such as informal 1986a, b, 2007; Kohli and Meyer 1986; Kohli
family care work or volunteer work for commu- et al. 1991; Moen and Roehling 2005).
nities or non-prot organizations. Paid work is Work by at least one household member is
the principal activity of most prime age adults crucial to supporting individuals and families.
(ages 2555) in advanced economies, institution- But working for the same or several organiza-
alized with the industrial revolution and its corol- tions and in the same or several occupations
lary emphasis on wage employment as distinct throughout the life course not only provides a
from self/family employment in agriculture or source of income (livelihood), it is also key to
small businesses, and distinct as well from the status, skill development, daily routines, infor-
informal family care work taking place within mation acquisition, and social relations. There is
households. (Throughout the chapter, therefore, I a subjective side as well; work shapes identities,
use work to refer to paid work, unless other- values, a sense of mastery, and current as well as
wise designated.) future expectations. In these many ways work
shapes individual and family life quality as well
as life chances at all ages and life stages (Elder
1.1 Work and the Life Course 1994; Glavin 2013; Hitlin and Elder 2007b;
Jahoda 1982; Kohn and Schooler 1983; Schooler
The social and temporal organization of work et al. 1999; Warren et al. 2002). For example,
replete with the rhythms, clocks and calendars of work has consistently been related to both mental
work days, work weeks, work years, and work and physical health. To be sure, healthy people
lives (including entries and retirement exits) have are most apt to work, but work itself promotes
been fundamental to the shaping of the conven- health and well-being (Berkman et al. 2014; Bird
tional mid-twentieth century lock-step life and Rieker 2008; Blustein 2008; Bosma et al.
1998; Carr 1997; Karasek and Theorell 1990;
Landsbergis et al. 1992; Mirowsky and Ross
P. Moen (*) 1989; Pavalko et al. 1993).
Life Course Center and Department of Sociology, Work is also tied to economic safety nets.
University of Minnesota, Minneapolis, MN, USA Unlike European social welfare state provisions,
e-mail: phylmoen@umn.edu

Springer International Publishing Switzerland 2016 249


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_11
250 P. Moen

in the U.S., employing organizations have tradi- sion scheduling reect the work day, and com-
tionally been a major source of social welfare, munity activities are arranged around work
providing (until recently) most health insurance, weeks, weekends, and national holidays from
pensions, and saving plans. Moreover, job tenure work.
(seniority) has historically provided (again, until Employment is clearly the path to economic
recently, and only for white-collar and unionized independence, self-esteem, meaning, and secu-
blue-collar workers) a measure of job and eco- rity, the fundamental public ecology in which the
nomic security accumulating with tenure and adult life course plays out. But jobs remain gen-
translating into retirement security. dered, even though they are no longer ofcially
There is considerable variety and inequality in advertised as for men or for women.
the nature of work and its rewards and costs. Educators, employers, media, and even parents
Employment opportunities, the risks of unem- commonly steer young women (and young men)
ployment, entrance into certain occupations, toward some jobs rather than others, contributing
toxic work environments, job conditions, to the ongoing gendering of occupations
resources, and power are institutionalized and (Fernandez and Sosa 2005). Ridgeway and
stratied in the labor market by policies and tra- Correll (2004) describe how small biasing effects
ditions tied to educational background, job ten- such as these accumulate over careers and life-
ure, and status as well as by (often hidden) times to result in substantially different employ-
stereotypes and discrimination. Taken together, ment experiences and rewards for men and
these policies, traditions and biases serve to fos- women who may have started out with similar
ter and reproduce disparities by social class, race, social backgrounds.
nativity, gender, and age. There are opportunities Both work and the work force have been
for job mobility for some; but both advantages transformed in remarkable ways. The traditional
and disadvantages tied to work trajectories tend social contract rewarding the lock-step of con-
to cumulate over the life course (Dannefer 2003; tinuous full-time employment with job and retire-
Ferraro and Shippee 2009). ment security has disappeared, unlikely to return.
Workers are now more apt to be female, minority,
or older than ever in history. Still, rigid work-
1.2 Work as a Master Life Course time norms, policies, and practices (such as com-
Status pleting education by the 20s or earlier, 9- to 5,
5-day work weeks, hierarchical top-down
Paid work is thus a master status, providing a authority structures, retirement at 65) designed
framework for organizing and integrating liter- for white-collar and unionized blue-collar jobs in
ally every aspect of social existence across the the 1950s (think Mad Men) havent changed.
life course for youth on the precipice of the The fact is, there is a fundamental mismatch
labor market, adults simultaneously building between the social organization of education,
families and careers, and older adults on the edge paid work and retirement (institutionalized in the
of retirement (Barley 1989; Beck 1992; middle of the twentieth century) and twenty-rst
Csikszentmihalyi 1990; Hall 2004; Heinz 2003; century workforce and economic realities.
Hughes 1958; Moen 2003b; Moen and Roehling Adding to this mismatch are outdated gendered
2005). The long arm of the job touches both traditions and expectations regarding paid work
working and non-working adults and retirees, and family care work (such as who is responsible
their partners, their aging parents, their children for breadwinning or family care) that are at odds
and grandchildren. This is the case not only for with increasingly strong societal norms (and pol-
individual and family lives, but also for the vari- icies) endorsing womens employment and gen-
ous institutions, organizations and activities der equality (Blair-Loy 2003; Gerson 2010;
within communities and the broader society. For Moen 1992; Moen and Roehling 2005; Orrange
instance, the rhythms of both trafc and televi- 2007; Reynolds and Aletraris 2010).
Work Over the Gendered Life Course 251

Thus contemporary work and occupational married women war workers to domesticity), a
careers illustrate a key challenge of our times career mystique developed based on middle-class
the need for corporate and public policies pro- white-collar and unionized blue-collar mens
moting innovative organizational structures and occupational trajectories and possibilities (Moen
cultures that better t with twenty-rst century and Roehling 2005). It portrayed continuous,
realities of a changing economy (including the full-time (or more) hard work and commitment
dismantling of traditional social protections), a throughout most of adulthood as the path to suc-
changing workforce (of women as well as men, cess and happiness. The booming Post-World
more older workers), and changing gender and War II economy creating skilled blue-collar and
occupational values (including beliefs that all professional white-collar jobs laid the ground-
adults in the household should work, that family work for the social construction of standardized
and work environments should be egalitarian, good and bad work. Good work was for
and young workers not always buying into tradi- those who could follow the career mystique, that
tional norms and expectations). What are the is, jobs providing high wages, security with
emerging contours of twenty-rst century work, seniority, health care, pensions, sick leave, vaca-
a twenty-rst century life course? tions and other benets (Kalleberg 2011, 2012).
The remainder of this chapter views work Bad work consisted of more marginal engage-
through a life course lens highlighting key life ments part or full time work often intermit-
course themes: that historical context matters, tent with little security and few such benets or
that time and timing are fundamental to aligning supports (Kalleberg 2000). The institutional
individual biographies with macro-level institu- clocks and calendars associated with good work
tions, the dynamics of social change, and the and orderly career paths became key organizers
importance of the constrained agency of individ- of everyday experiences and of the life course as
uals, organizations, public policy makers and an institution. But note that this was based on a
scholars in reshaping the social and temporal white, middle-class and unionized working-class
aspects of work. It further addresses processes male template of work.
inhibiting work redesign and concludes with hid- Consider the ways jobs were allocated by
den assumptions constraining both research and whether or not one was male or female (or preg-
policy innovations. nant), or how workers were allocated social secu-
rity credits, both of which disadvantaged women.
Until the late 1960s, job want ads in newspapers
2 Historical Context: Work specied the gender of the employee, with cleri-
as an Institution cal and secretarial positions specically seeking
women (Pedrianna 2004). Teachers in some
How did we get here? To a place where institu- school districts in the 1960s and 1970s could not
tions feel out of date, no longer tting with continue to teach if pregnant. Before 1978 one
todays intense job demands, chronic job insecu- could not receive a years Social Security credit
rity, uncertain career paths, families without full- except by remaining in the workforce continu-
time homemakers, and transitions into and out of ously throughout the year, something problem-
the work force that take years, not days? A life atic for women moving in and out of paid work
course approach to the study of work necessarily for caregiving or other reasons (Institute for
starts with an historical vantage point (Elder and Womens Policy Research 2005a, b). These poli-
George forthcoming). The development of corpo- cies began to be revised in tandem with both the
rations following the industrial revolution set the womens movement and the corollary unprece-
stage for wage labor, as well as the breadwinner/ dented inux of married women into the work-
homemaker family form (for those who could force (Moen and Roehling 2005).
afford it). After the Second World War and the The career mystique paradigm was never a
return of troops to the civilian labor force (and reality for most minorities, women, or the poorly
252 P. Moen

educated, but it became the widely accepted as men at all points along the life course. This
vision of the ways work and career pathways made life difcult for working women with fam-
should be organized. This false myth traditional- ily responsibilities and for women, immigrants
izing rigid clocks (95, Mondays through and minorities in bad jobs, exacerbating gender
Fridays) and calendars (full-time job entry as a as well as other inequalities (Budig and England
marker of the transition into adulthood, continu- 2001; Correll et al. 2007; Gangl and Ziee 2009;
ous work throughout adulthood; retirement at 62 Staff and Mortimer 2012). For even professional
or 65 as a marker of old age) occurred in tandem women, the reality has become some sort of
with and was only possible because of another career mystique plus (the plus being womens
myth, the feminine mystique (Friedan 1963), conventional family responsibilities).
freeing (male) workers to focus full time on their In tandem with the career mystique, the politi-
jobs because someone else their wives were cal climate of neoliberalism (Acker 2006; Davis
persuaded that their sense of fulllment came 2009; Sennett 1998) over the last several decades
from full-time (and unpaid) care work for their has prioritized paid work over all other activities,
children, their houses, their husbands, their as can be seen in the U.S. in welfare reform leg-
communities. islation passed in 1996 (Personal Responsibility
A life course approach thus underscores the and Work Opportunity Reconciliation Act) and
signicance of social time, the socially con- incentives (delays in eligibility for full Social
structed and institutionalized entry and exit por- Security benets, higher benets for those work-
tals into and out of roles (such as labor market ing until age 70) encouraging older workers to
participation) and relationships (such as marriage extend their years of labor force participation.
and caregiving). The life course regime of paid
work and career paths institutionalized in the
middle of the twentieth century provided (for 3 Time and Timing
many men at least) a lock-step template shaping
much of the workforces biographical experi- A second key life course theme has to do with the
ences, despite the fact that war and economic time and timing of events, related always to age.
downturns generated some sharp dislocations. Life course scholars have sensitized researchers
Institutionalizing age-graded labor market entry to the multiple meanings of age (Brckner and
and exit time tables, occupational ladders, quali- Mayer 2005; Hagestad and Neugarten 1985;
cations for advancement, pensions, and govern- Riley et al. 1994; Settersten and Hagestad 1996a,
ment supports (such as Unemployment Insurance, b; Settersten and Mayer 1997). Age is more than
Social Security, and educational loans) created biological time. It is also about social time, such
what most consider the normal life course. But as the patterning of transitions and trajectories,
this life course regime marginalized those out of the development of timing norms and the concept
step with its specications, those who couldnt of cohort tied to historical context, as well as the
get the educational credentials, job interviews, or cultural and structural clockwork of work.
promotions to jobs with career tracks, as well as Especially consequential is the fact that with
the absence of supports for integrating work and bureaucratization, work came to be about time
family care obligations. rather than piece work, hours at work rather
The womens movement in the 1970s pushed than accomplishments.
to effectively replace the feminine mystique with
the career mystique with women seeking mens
jobs, mens career paths, and mens nancial and 3.1 Scheduling Social Time: Doing
other rewards. Though few men or women could Age and Gender
actually attain it, the false myth (heavy career
commitment as the only path to life quality) The concept of social time captures the tempo-
became legitimated as the way things are, and ral routines, regulations, and rules that dene the
the way things should be for women as well universe of options available for individuals to
Work Over the Gendered Life Course 253

shape their life courses (e.g., Heinz and Marshall Age and aging serve to widen, rather than nar-
2003; Kohli 1986b, 2007; Levy and Widmer row gender differences and inequalities (Moen
2013; Marshall et al. 2001; Moen 2003a; Moen 2001; Moen and Spencer 2006). For example,
and Spencer 2006; Mortimer et al. 2015; contemporary young people in high school and
Shanahan 2000; Shanahan et al. 1997) and incor- college often wear similar styles of clothing, earn
porates gender as well as age expectations. similar (low) wages in retail or service jobs, and
Institutionalized social time clocks dene tradi- take similar classes. It is only as they enter and
tional age-graded pathways to and through paid move through adult roles especially as employ-
work. Age is thus an important determinant of ees, spouses, and parents that womens and
peoples institutionalized social roles, indepen- mens experiences and resources tend to mark-
dent of their capacities and preferences, and is edly diverge (Aronson and Mortimer 2013;
reected in what Riley (1987) refers to as the age Ashby and Schoon 2010). Both advantages and
stratication system. Other meanings of age have disadvantages cumulate over the life course, wid-
to do with the patterning of transitions and tra- ening inequalities within as well as across gender
jectories, the development of timing norms and divides.
the concept of cohort tied to historical context.
Just as there are multiple meanings of age, so
too, are there multiple meanings of gender. 3.2 Timing, Transitions
Gender (or sex) quite obviously refers to the and Trajectories
biological differences between men and women,
but it is equally an ascribed status affecting life Part of the time and timing life course theme
chances and possibilities. There is also a gender relates to the cultural and structural aspects of the
stratication system, sometimes formal but often timing, transitions and trajectories of work and
informal biases allocating some social positions career paths. Rather than think of career as
and opportunities to men and others to women moving up in seniority or climbing an occupa-
throughout the adult course (Bem 1994; Folbre tional ladder, I use the term to capture the
2001; Moen 2001, 2003a; Moen and Chermack dynamic choices and progression of individuals
2005; Moen and Spencer 2006; Risman 1998; West along various role pathways. People follow a
and Zimmerman 1987; Williams 2000; Williams range of careers: educational careers, health
and Dempsey 2014). Note that race and class are careers, family careers, leisure careers, work
also important statuses intersecting with age and careers.
gender to shape the gendered adult course (Acker Following the lock-step path of good jobs or
2006; McCall 2001; Schulz and Mullings 2006). else the more checkered path of bad jobs is
Gender and age are not simply markers; rather, shaped by prior educational training as well as
they play out as biological, social and historical certication, skills, and family background along
forces. This matters because the work scripts with discrimination by race, gender, and social
people follow depend on the combination of their class (Macmillan and Eliason 2003). Good jobs
age and gender. In fact, age and gender constitute offer more than earnings or job security; they also
the backbone of key socialization and allocation provide workers with more autonomy and con-
regimes around paid work (Esping-Andersen trol. Timing also underscores the differences in
2009; Moen 2013), creating distinctive sets of mens and womens experiences of employment.
challenges and limiting the strategic choices and For example, men tend to have longer tenure in
adaptations of women and men as they move to organizations and occupations than women. And
and through adulthood. For example, hiring man- women typically work fewer hours per week than
agers are looking for young adults for entry-level their male colleagues. Thus a gendered life
positions, not someone in their 40s or 50s. What course perspective highlights gender as a key
is key is that both age and gender are intersecting contingency affecting the nature of work and
statuses, setting work expectations, options, and career paths over life biographies and their
resources throughout the adult years. consequences.
254 P. Moen

Womens lives typically underscore the 3.3 Timing Norms


embeddedness of work careers in family careers
and vice versa, the impossibility of isolating one There are other forms of timing as well. Consider
from the other in any articial divide (Blossfeld commonsense notions about the right time for
and Drobni 2002; Blossfeld and Hofmeister doing things. Culturally-grounded norms and
2008; Damaske 2011; Hobson 2014; Sweet policies shape expectations and beliefs about
2014). This has been true of mens lives as well, the right age to be in school, to be in the labor
but was less obvious in earlier times. Recall that force, to move up occupational ladders, to start
white middle-class and unionized blue-collar a family, or to retire from a career (Hagestad
men in the middle of the last century moved up and Neugarten 1985; Neugarten et al. 1965).
occupational or seniority ladders, achieving This reects the institutionalized aspects of
income, status, and job security precisely because work, in terms of expectations and rules around
they had wives who took care of the home front, the time, timing, and duration of jobs (work
leaving them unencumbered, able to focus almost days, work weeks, work years as well as entries
entirely on their jobs throughout their prime and exits at different ages and life stages; see
working years. Kelly and Moen 2007; Moen and Chesley 2008;
Today most workers are married to other Settersten and Hagestad 1996a, b). Work entries
workers, are singles or single parents, and/or or exits experienced as off-time (i.e. earlier or
have children, parents or other aging relatives later than is socially prescribed) may be per-
who require care. Few men or women have ceived as more stressful or disruptive than work
homemaking support, either paid or unpaid transitions that are normatively on-time
(although many husbands still benet from their (George 2003).
working wives doing most of the domestic work). For example, workers in their 50s who are
What some young people increasingly have, unexpectedly laid off or forced into early retire-
however, is parental support as they move into ment due to corporate restructuring may very
the workforce (Mortimer 2011; Mortimer and well experience these off-time transitions as
Fischer forthcoming; Mortimer et al. 2010; disruptive and psychologically stressful. And the
Swartz 2009; Swartz et al. 2011). duration of experiences (such as time unem-
The life course perspective (Elder and George ployed, time out of the workforce, tenure within
forthcoming; Elder et al. 2003; Mortimer and an organization, or cumulative stress) can shape
Shanahan 2003) holds that work transitions (such future opportunities as well as health and well-
as getting or changing jobs, voluntarily leaving being (George 2003, 2014; Pearlin et al. 2005;
the work force, being laid off or retiring) always Schieman et al. 2002).
occur in the context of ongoing trajectories. Another example: men traditionally have been
Understanding the patterning of employment expected to be the family breadwinners, with all
through the life course may be important for the possibilities, power, and pressures associated
understanding both the resources available and with it. Women are now expected to both hold
the meaning of paid work, with men and women down jobs and provide care for inrm and dis-
typically following (or being allocated into) very abled family members as well as for dependent
different patterns. For example, women follow- children. These differing options and imperatives
ing intermittent career paths, moving in and out play out in distinctive trajectories and turning
of the workforce, in and out of part time work in points that shape disparities in mens and wom-
response to shifting family care obligations, nd ens occupational careers, resource accumula-
it easier to get off occupational or organizational tions, and retirement paths (Blossfeld and
tracks than to get back on (Han and Moen 1999a, Drobni 2002; Blossfeld and Hofmeister 2008;
b, 2002; Hobson 2014; Malenfant et al. 2007; George 1993; Moen and Flood 2013; Moen and
Stone 2007). Roehling 2005; Staff and Mortimer 2012).
Work Over the Gendered Life Course 255

3.4 Cohort years. Men in certain cohorts have gone to war


while their younger or older brothers (in different
Age at a given point in time is also an indicator of cohorts) watched these same wars on newsreels
birth-cohort membership and, thus, of life expe- or television. And the cohort of young women
riences shared with others born in the same his- today is the rst participating actively in military
torical context (Riley 1987; Riley et al. 1994; roles previously allocated only to men. It is also
Ryder 1965). People of the same cohort live the rst cohort where young women have more
through the same historical times and are often education than young men.
socialized in similar ways, experiencing the same Cohort differences are especially evident in
events, using the same technologies, and living societal expectations and gender norms around
through the same economies, climates, and pol- paid work and unpaid family work. Recall, the
icy as they move through the life course. The career mystique myth crystallized in tandem with
result is often greater differences across than the feminine mystique (Friedan 1963), a belief in
within cohorts in beliefs and values related to the 1950s and early 1960s that women were
work, gender and adult pathways. For example, socialized to (and should) be exclusively wives
young workers today are far less likely to want and mothers, nding total fulllment in full-time
more job responsibility and more hours on the domesticity. The problem was that many
job compared to young workers in the early working-class and poor women couldnt afford to
1990s (Matos and Galinsky 2012) and fewer say be out of the workforce, often moving in and out
that work is central to their lives (Settersten and of marginal jobs, and middle-class women in col-
Ray 2010). Members of the aging Boomer cohort lege were learning different lessons about using
(born after World War II, from 1946 through their skills in the world of work. Friedans book
1964), Generation X (born 19651974), and (1963) tapped a nerve, revealing many home-
Millenials (19752000) have different beliefs makers longing for a different kind of engage-
about womens and mens roles. Note that family ment and, in doing so, (re) launched the womens
members, neighbors, friends, and coworkers of movement in the 1960s and 1970s.
different ages are also members of different People in different cohorts have different atti-
cohorts, socialized differently and viewing their tudes about work, career paths, and gender
lives past, present and future from vastly dif- (Schuman and Rieger 1992), as well as different
ferent vantage points. options, depending on whether they are women
What gives each cohort its distinctiveness are or men, and whether they have a college educa-
the shared events and experiences of people who tion. Consider the different experiences of the
are roughly the same age when major historical parents of the baby-boom cohort (young adults in
events occur. For Americans moving into adult- the 1940s and 1950s) versus the boomers them-
hood in the early 1960s, a dening moment was selves (young adults in the 1970s and 1980s) ver-
when John F. Kennedy died. For those not yet sus young adults today. Note that what is
born, Kennedys death is the stuff of history taken-for-granted at one point in history (such as
books and documentary lms. For most men as the exclusive family breadwinners in the
Americans who were adults or adolescents at the 1950s) changes. Today, dual-earner households
turn of this century, the transformative experi- are very much the expected norm, and growing
ence of their lives occurred on September 11, numbers of women earn more than their hus-
2001 when planes ew into the World Trade bands. Also new: the transition from school to
Center and the Pentagon. work has lengthened and become more complex
There are other ways that when people are for young people today, often involving a series
born, which birth cohort they belong to, has enor- of unrelated jobs and periods of unemployment
mous implications for the course of their work (Blossfeld and Drobni 2002; Mortimer et al.
256 P. Moen

2015; Vuolo et al. 2012: Mortimer and Moen this immigrants, more older workers, more retirees
volume). along with greater life expectancy). And there is
The boomers (born 19461964) grew up along economic globalization (reducing social protec-
with the womens movement, which, in combina- tions and individualizing risks of job loss and
tion with an expanding economy, resulted in a economic insecurity) as well as high velocity
tremendous growth in womens educational changes in communications and information
attainment and labor-force participation (Moen technologies contributing to blurring boundaries
forthcoming). Boomers were the rst cohort to between work and non-work times and spaces.
experience the mismatch between traditional Another change: prolonged periods of education
career paths and the growing numbers of women and dependency in early adulthood (including
in the workforce. Recall this mismatch reects changes in the odds and timing of obtaining rea-
the changing nature of the labor force. Although sonably secure employment, as well as increases
there are some exceptions (husbands whole wives in cohabitation, delays in marriage and parenting,
are not working and vice versa), todays work- the rise in later life divorce and the growth in
force consists mainly of (1) men whose partners complex families). The conuence of these forces
are also in the labor force, (2) women whose part- are upending conventional life course norms,
ners are also in the labor force, and (3) men and including linear career paths and conventional
women without partners but with family obliga- one-way retirements (Blossfeld et al. 2008; Heinz
tions and/or non-work interests. Still, formal and 2003; Johnson et al. 2012; Moen and Flood 2013;
informal caregiving remains womens work Warner et al. 2010). Medical advances promoting
(Folbre 2001, 2012). This includes many boomer health and longevity, the aging of the large
grandmothers who are both working and caring boomer cohort, declines in the pool (and options)
for their grandchildren (Harrington Meyer 2014). of young job entrants, less physically demanding
Life course scholars (Elder 1996; Moen work, the weakening of unions, and the disman-
2003a; Settersten and Mayer 1997) emphasize tling of traditional pensions and protections for
the life stage principle; this means that the timing older workers and retirees all are part of this
of historical events in peoples lives such as the moving platform of change (Cahill et al. 2005;
delaying of full Social Security benets, new Fullerton and Wallace 2007; Johnson et al. 2011;
communication technologies, or economic down- Kalleberg 2009, 2011; Mortimer et al. 2014;
turns (like the Great Recession felt most strongly Ogums 2012; Sargent et al. 2013). It is not sur-
in 20082010) impacts some cohorts more than prising that the nature of the work life course of
others, given their ages. The life stage principle most people no longer ts the lock-step ideal.
states that the effects of the same historical Fundamental to understanding work in the
change is often different for individuals of differ- twenty-rst century are changes in the nature of
ent ages (different cohorts) and career and life both work and the work force.
course stages, as well as those differing in their
social class backgrounds (Elder 1996, 1998;
Johnson et al. 2012; Moen 2003a, b; Mortimer 4.1 The Changing Nature of Work
et al. 2010).
Work in the early twenty-rst century has
changed in a variety of remarkable ways. This is
4 The Dynamics of Social the consequence of a precarious and polarized
Change global labor market that is producing concerns
over productivity and competitiveness along with
Contemporary employers and employees nd an unsettled economic climate, technological
themselves on a moving platform of multi- advances, the long-standing shift to a service
layered historical, social, and economic changes. economy, and the dismantling of social protec-
Consider rst the changes in workforce and tions as part of a neoliberal agenda emphasiz-
retirement demographics (more women, more ing the self-regulation of markets and the
Work Over the Gendered Life Course 257

individualization of risks (Davis 2009; Glavin supportive of their personal lives, in reducing
2013; Johnson et al. 2012; Kalleberg 2009, 2011; work-family conict and increasing actual and
Silva 2013; Sweet and Meiskins 2013; Vuolo perceived schedule control (Kelly et al. 2011,
et al. 2012). The rise of a contingent workforce 2014), health and health behaviors (Fan and Qian
along with downsizing and the restructuring of 2015; Moen et al. 2011b, 2015) parents per-
job ladders mean that, for many workers, the tra- ceived time adequacy (Hill et al. 2013), and turn-
ditional (secure) organizational career is rapidly over (Moen et al. 2011a, 2014). Such research
becoming obsolete. Lock-step career path tem- designs are important for capturing the changing
plates are thus increasingly irrelevant for contem- nature of work environments and suggest poten-
porary workers. Despite the disappearance of the tial future directions greater exibility and con-
standard employment contract (Stone and Arthurs trol over time, for example. However, employers
2013), many rigid labor market laws, regulations, labor market exibility the ability to reduce
policies, and practices remain in place, continu- their work force at will or offshore jobs to other
ing to provide mid-twentieth century scaffolding countries remains a key component of contem-
shaping (and constraining) twenty-rst century porary work.
clocks, calendars, and rhythms of work. These
rigidities around the expected clockworks of
work and who has control over their jobs and 4.2 The Changing Workforce
their time also promulgate inequalities in work
conditions and resources over the life course. Medical advances promoting health and longev-
Characteristics of jobs, career paths, and the ity, the aging of the large Boomer cohort, increas-
working environment are extremely consequen- ing immigration, different fertility rates, and
tial for the developing individual and tend to dif- womens increased labor force participation are
fer by gender (Moen and Chermack 2005). transforming the demography of the workforce.
Considerable theoretical progress and empirical In fact, the twenty-rst century workforce in
evidence link high demands on the job and low North America and Europe is characterized by a
levels of job latitude with heightened feelings of large number of women (almost equally divided
strain (Elsass and Viega 1997; Karasek and between women and men in the U.S.); growing
Theorell 1990) leading to poor health outcomes numbers of two-earner, single, or single-parent
such as cardiovascular disease and elevated blood employees; increasing numbers of minorities and
pressure. But men are more apt than women to immigrants; and an aging workforce (as boomers
typically occupy jobs with good health benets, move through their 50s, 60s, and 70s). The mis-
those with both high demands and high control match between this heterogeneous workforce
(Bosma et al. 1998; Cheng et al. 2000; de Jonge (increasingly with non-work obligations, values,
et al. 2000; Dwyer and Ganster 1991; Hemingway and interests) and the lock-step expectation of
and Marmot 1998; Kristensen 1995, 1996; continuous full-time employment is another
Landsbergis et al. 1992; Schnall et al. 1994). example of structural lag (Moen 1994; Riley
Moreover, Marshall et al. (1997) nd evidence et al. 1994) or institutional inertia (Moen forth-
suggesting that the demand-control model may coming), in that calcied rules, regulations, cul-
be more applicable to employees in manufactur- tural norms, policies, and practices about work
ing sector jobs (typically men) than to those in time, career paths, and retirement are changing at
service jobs (typically women). They nd that, a glacial pace.
for employees in the service sector, job control As a case in point, increases in life expectancy
does not moderate job demands, while the intrin- and decreases in fertility in Europe and North
sic rewards associated with serving others tend to America are transforming the age distribution of
benet service employees health. the labor pool. The work force is continuing to
Recent natural experiments and eld trials age as boomers grow older, smaller cohorts move
show the importance of employees control over into work, and young people spend more time in
their time, along with having supervisors who are school. The proportion of workers ages 1624 in
258 P. Moen

the U.S. was 16.9 % in 1992, falling to 13.7 % in generation, directly and indirectly by example,
2012 and projected to be 11.3 % by 2022. By what is expected for workers of each gender at
contrast, workers 55 and older constituted only different ages and life stages and from different
11.9 % of the U.S. workforce in 1992; by 2012, socioeconomic backgrounds (see Settersten and
fully 20.9 % of workers in the U.S. were 55 or Owens 2002; Settersten and Ray 2010).
older, projected to be over one in four (25.6 %) Socialization processes thus foster gendered
by 2022 (Toossi 2013). While the retirement work identities as well as different preferences,
transition came to be normatively dened in expectations, values and motivations for women
terms of timing and legitimacy, setting retirement and men at different ages and from the same or
apart from unemployment as a later-life work different backgrounds (Preves and Mortimer
exit that can be planned for, anticipated, and posi- 2013).
tively dened (Graebner 1980; Moen 2012), To understand behavior at any one life stage
comparatively few boomers plan for their retire- requires knowledge of prior processes of social-
ment, and many would prefer a longer transition, ization. Heinz, for instance, describes young
with the possibility of reduced hours spent on peoples biographical action orientation (Heinz
ones career job and the possibility of reemploy- 1999, 2002a, b) that serves as a guide to behavior
ment following retirement from that job (Moen as they move into and through adulthood. As an
et al. 2005). example of the cumulative processes of early
decisions, young women are less apt to take cal-
culus in high school, a choice (reecting earlier
5 Redesigning Work gender socialization) that has enormous implica-
and the Gendered Life tions in terms of their subsequent majors in col-
Course lege. This choice to opt out of high school
calculus considerably reduces the odds that
Attention to three institutionalized processes is women will move into careers in engineering or
central to understanding and developing innova- the natural sciences, given that these occupations
tions to reduce the mismatch between outdated require training in mathematics, widely seen as a
twentieth century policies and traditions around valued but male-typed task (Ceci and Williams
the social and temporal organization of work and 2010; Nosek et al. 2002).
twenty-rst century workforce, technological, Early socialization plays out throughout adult-
and economic realities. These processes also hood. In a 4-year (20002003) Norwegian study
reproduce gender, race and class inequalities by of an entering cohort of undergraduate students
advantaging some early in the career course and in varying professions, Daehlen (2005) found
disadvantaging others, with long-term conse- preferences for work to be deeply rooted in a
quences. These dynamic, interrelated processes persons early socialization, acquired well before
of socialization, allocation, and strategic selec- the choice of a higher educational program and
tion set the clocks and calendars of work and sort sustained through such programs. This study
men and women into different labor markets, dif- offers evidence that gendered preferences are
ferent occupations, different trajectories, and dis- formed well before the transition to adulthood.
parate resources, risks, and rewards. But socialization processes are ongoing
throughout the life course, continuing to shape
adult beliefs, values and expectations (Mortimer
5.1 Socialization and Simmons 1978; Preves and Mortimer 2013).
Daehlen (2005) found that job values change
Liao and Cai (1995) dene socialization as, the with education, more or less in the same direction
process whereby individuals learn and maintain for male and female students (see also Johnson
the morals and values of a society (see also 2001). Goodwin and OConnor (2005) reana-
Kramer 2011). Socialization teaches each new lyzed data from interviews of young men (who
Work Over the Gendered Life Course 259

were ages 1618 in 1962), following up with 94 evaporate (Lippmann 2008). The result? Role
re-interviews in 2001, nding that the career conicts, strains, time pressures and overloads
paths of the respondents they located were for employees in demanding jobs and/or high
extremely gendered. Almost all the men had occupational aspirations who also have family or
found jobs through male relatives, they all personal goals, needs, and obligations (Damaske
stressed making money as important, and 2011; Hobson 2014; Moen et al. 2015; Stone
reported that years ago the Youth Employment 2007; Sweet 2014).
Ofce had steered them into traditionally male Contemporary twenty-rst century cohorts of
jobs. young workers men and women are being
Socialization processes change with changes socialized to simultaneously both embrace and
in the larger opportunity structure of work, but question the career mystique. Recall that social-
remain shaped by existing outdated institutional- ization involves learning from observations.
ized templates. Thus, the cohort of women com- Many young people can no longer see the pay-
ing of age in the 1970s and 1980s rejected the offs of following the lock-step career mystique
feminine mystique, only to learn and embrace the path in light of corporate downsizing, mergers,
career mystique instead. The womens move- and outsourcing, as well as the strains they
ment became an important socializing force observe in the lives of their parents generation.
encouraging women to want and even demand Their beliefs about adulthood, gender equality
mens jobs, mens career aspirations, and and the path to success are often inconsistent,
mens salaries and occupational achieve- ambivalent, and vague (Moen and Orrange 2002;
ments as the path to gender equality. But the Mortimer and Fischer forthcoming; Mortmer
career mystique that a lifetime of working hard et al. 2015; Orrange 2007).
always pays off in occupational and economic Socialization is also subject to changes over
security and success was a false myth, never the life course of an individual and their changing
accessible to most workers. And it was predi- social networks (Liao and Cai 1995). For exam-
cated on having someone else a wife to take ple, consider the transition to parenthood with the
care of all the details of daily living. Accordingly, birth of a child. Becoming a parent cannot be
feminists began to call for men to do their fair understood separately from the taken-for-granted
share of the unpaid family care work, only later norms and scripts that dene what it means to be
recognizing that the structure and culture of paid a working mother or father. Mothers are (still)
work made doing so difcult for all workers, expected to do most of the child care and fathers
regardless of gender (England 2005; Hobson are (still) expected to do most of the breadwin-
2014; Kimmel 2008; Moen and Roehling 2005). ning, even though both are likely to be in the
The feminine mystique is gone, although workforce (Hochschild 1997; Singley and Hynes
women are still socialized and allocated to the 2005).
nations care work for children and youth, com- Adult opportunities still follow informal gen-
munities, the disabled, and the older frail and der norms and expectations even though they are
inrm (Folbre 2001, 2012). The career mystique no longer formally regulated. Consider, as a case
remains dominant in American culture, the epit- in point, educational socialization. On the sur-
ome of independent adulthood, the accepted path face, educational policies are gender-neutral. But
of optimal adult development leading to success, Goodwin and OConnor (2005) nd that schools
fulllment, and gender equality (Moen and reinforce different cultural values, dominant
Roehling 2005), even though young people are masculinity types, vocational preferences, and,
nding it hard to land secure jobs (Mortimer and via the curriculum, link types of knowledge and
Fischer forthcoming; Orrange 2007), midlife skills with masculinities and femininities
adults are stressed managing work and family (p. 453). This accentuates mens traditional
roles (Sweet 2014) and older workers with long expectations and values about male adulthood as
tenure are seeing their job and economic security years of breadwinning and continuous, full-time
260 P. Moen

hard work (see also Moen and Roehling 2005; retirement (Moen 2013). These convoys promote
Orrange 2007; Townsend 2002). and perpetuate disparities by social location (that
As a result of both formal and informal learn- is, by gender, age, social class, marital status,
ing (socialization), gender and age are, literally, immigrant status, race and ethnicity). Social and
in our heads, as scripts guiding behavior and institutional convoys also constitute sources of
expectations of self and others (Fenstermaker and on-going socialization processes: they transmit to
West 2002). We develop age- and gender-related the next generation taken-for-granted norms and
work and family stereotypes and values by hear- expectations about how adulthood should
ing, seeing, and using polarizing language, cate- unfold such as the incidence, timing, durations
gories, expectations, and assessments as and trajectories of education, work career, and
short-hand heuristics and habits (Moen and retirement paths.
Spencer 2006; Ridgeway and Smith-Lovin 1999; Why is there still inequality economic, polit-
West and Zimmerman 1987). Consider, for ical, social and interpersonal between working
example, the ways gender permeates how we women and men? Gender stratication is funda-
think about and refer to work roles (e.g., about mentally about disparities by gender in economic
nurses, where is she? and physicians, is he in power (Andersen 2005, 2009; Blumberg 1984;
yet?), and divisions of labor (e.g., driving Blair-Loy 2009; Esping-Andersen 2009).
trucks is a mans job). Gender and age as ways Serguino (2007) argues that socialization forces
of categorizing and dividing behavior and beliefs, in the form of gender ideology, norms and stereo-
risks and resources combine to shape work roles types reinforce material inequality between
and relationships in ways that seem natural women and men thereby reinforcing the gender
(i.e., second nature). Beliefs about mens greater stratication (allocation) system.
status and competence implicitly shape the These allocation processes institutional
expectations of both men and women about their arrangements that stratify women and men by
own competence and performance compared to their gender are even more important than
others, independently of their actual underlying socialization processes. Kohn and Schooler
abilities (Ridgeway and Correll 2004). (1983) dene stratication as the hierarchical
distribution of power, privilege, and prestige.
Groups and organizations of all types allocate
5.2 Allocation roles, risks, resources and relationships along
with power based on a range of factors. The
Second are allocation processes: the structural allocation of power, privilege and prestige in
arrangements and power differences in groups, society and in organizations depends on ones
organizations and societies that open up some social background, race and ethnicity, educa-
occupational possibilities for men or for women, tional achievement, and occupational level, of
while closing others through processes of gender course, but also on the combination of age and
(combined with race, ethnic, and age) stereotyp- gender (See Andersen 2005, 2009; Hobson 2014;
ing and discrimination (Acker 1992, 2006; Moen Kimmel 2008; Kramer 2011; McCall 2001;
and Spencer 2006; Ridgeway and Correll 2004; Rothman 1979).
Risman 1998; Williams 2000). Just as there are The difference between gender socialization
convoys of relationships (social convoys see and allocation is this: Gender socialization is
Kahn and Antonucci 1980) over the life course, about learning gender, fostering differences in
so too are there convoys of rules, risks, and regu- womens and mens beliefs, values, and identi-
lations (institutional convoys) that shape the dis- ties. Gender allocation is about dividing by gen-
tinctive work experiences of women and men. der, stratifying positions, power and material
Institutional convoys give social relationships resources by whether one is male or female.
structure and meaning, even as they constrain These are two systematic ways that roles,
options around education, work, family and resources and rewards are distributed,
Work Over the Gendered Life Course 261

perpetuating gender inequality. This takes place such as expectations regarding education, parent-
in large part because labor market and business hood, employment, and retirement (see also
rules, routines, and regulations serve to structure Altucher and Williams 2003; Brines and Joyner
the adult course in gendered ways. 1999; Carr 1997; Chesley and Moen 2006;
The notion of career implies an organiza- Clausen 1995; Cooney and Mortimer 1999;
tional ladder to be climbed, but it also stands for Freund 1997; Gerson 1985; Han and Moen
institutionalized life paths involving a series of 1999a, b; Moen et al. 2005, 2006; Pixley and
choice processes (Sweet and Moen 2012). As Moen 2003; Wethington et al. 2003).
Barley (1989; p. 49) points out, the career con- The career concept is useful for understand-
cept can be used as a lens for peering at larger ing both paid and unpaid work over the life
social processes known as institutions. Consider, course, since it incorporates a range of ideas from
for example, the implicit contracts between different perspectives, including transitions and
employers and employees and between family trajectories, individuals and organizations, and
members regarding divisions of labor (at home subjective identities as well as objective paths. It
and at work), job security, and job progression. is also a useful frame for linking the strategic
All of these contracts color individuals views selections by individuals and families with large-
of their experiences both at home and at work, scale institutionalized (and sometimes changing)
affecting their habits, choices, and strategies of policies, practices, and norms in work organiza-
action (Bourdieu 1984; Swidler 1986). tion and society.
What many contemporary workers are experi- Today, especially in light of increasing lon-
encing is the dissolution of these contracts, with gevity, technological advances, shifting gender
no clear-cut alternatives replacing them. The values, a changing workforce, and a globalized
challenge is to develop more exible institutions, economy, much of the existing infrastructure
a range of legitimated paths through work and the around work and the life course is now obsolete.
adulthood, opening up opportunities to custom- In light of this moving platform of social change,
ize education, work and career paths across the how can workers plan for their uncertain futures?
life course. Take the case of younger workers trying to launch
their careers, mid-life workers worried about lay-
offs as well as struggling to manage multiple
5.3 Strategic Selections obligations, or older workers considering leaving
their career jobs or the workforce altogether.
Third are strategic selections: women and men Millenials, genXers and boomers are nding
choose to enter or exit the labor market or par- themselves in a world vastly different from that
ticular occupations and not others at different experienced by their parents at their same ages
points in their lives (Moen and Chermack 2005; (Carr and Manning 2010; Hochschild 2012;
Moen and Chesley 2008; Moen and Wethington Moen 2003a, b; Mortimer and Fischer
1992). But these are constrained choices (Bird forthcoming).
and Rieker 2008: England 2005; Esping- For instance, contemporary older workers
Andersen 2009; Hobson 2014; Kimmel 2008; confront more risk and uncertainty related to
Moen 2013). Individuals learn the institutional- both when they retire from their career jobs and
ized nature of work and the life course through when they exit the labor market altogether. Many
socialization and allocation processes. Then they older workers are developing their expectations
act accordingly. about retirement with the backdrop of corporate
Thus, existing institutionalized arrangements downsizing and increasing job and economic
shape decision making by providing individual insecurity (Moen 2012; Moen et al. 2010; Moen
women and men at different life stages with et al. Forthcoming; Sweet and Moen 2012). In
available lists of reasons, motives, and aspira- 2014, the average age of complete retirement
tions for decision-making (Meyer 1986, p. 205), from the labor force among U.S. workers was
262 P. Moen

around age 62. Exiting their career jobs often coming; Schneider and Stevenson 1999;
occurs even earlier, meaning that contemporary Settersten et al. 2001; Shanahan 2000; Vuolo
workers can expect to spend a signicant portion et al. 2013).
of their life course in encore adulthood (see Moen The twenty-rst century is just such a time of
and Flood 2013; Mortimer and Moen this vol- social transformation. Demographic, cultural,
ume), thinking about or trying to fashion second economic, and technological changes are rewrit-
chapters or encore careers in paid or unpaid ing the nature of work, education/career/retire-
work. Unlike previous generations, for whom ment paths and the gendered life course. Still, the
retirement exits were xed according to public fundamental mismatch between outdated scripts
and private policies, contemporary workers are and new realities perpetuates gender as well as
exercising considerable discretion in the timing class/ race/ethnic and age inequalities. Needed
of their exits from career jobs (or else are laid off are a wider pool of institutionalized options and
or nding themselves encouraged to take early protections from the risks of employment and
buyouts). By contrast, for their parents (mostly economic insecurity for workers of all ages and
fathers) the retirement transition was not so much life stages.
an uncertain or risky event as it was a normative
and routinized inevitability.
Women and men live distinctive work lives 6 Conclusions: Advancing
because they develop different preferences Scholarship and Policy
(Becker 1981; Hakim 2004) as a result of bring-
ing different expectations and values (socializa- Work and careers are what Pearlin (1988; p. 259)
tion) to each fork in the occupational life course. describes as durable arrangements that serve to
Moreover, each fork in the occupational life organize experience over time. In turn, it is pre-
course is stratied by gendered and age-graded cisely this organized experience that is the basis
cultures and biases allocating women and men to for how we see the world around us, how we
different opportunities, risks, and constraints think about it and act toward it. Americans in
(allocation). Even whether or not adults perceive particular equate success and productivity with
that there is in fact a fork in the road (decision paid work. And, as noted throughout, society and
point) sometimes depends on their gender. the economy in all developed nations are ordered
Choosing which occupational path or transition around full-time jobs, full-time workers, and
to take (strategic selections) is guided by both continuous careers, as are organizations, commu-
socialization and allocation processes that persist nities, and families. These rigid structures and
across the life course through ongoing social cultures offer few options, even as work
relationships and institutionalized work, educa- intensication and job insecurity ratchet up
tional, organizational, and occupational regimes uncertainty (Burchell et al. 2002; Smith 2001;
(England 2005; Moen 2013; Person et al. 2005) Uchitelle 2007).
that promote or reduce a sense of agency (Hitlin Historical trends, along with other demo-
and Elder 2007a, b; Hobson 2014). graphic, social and technological transforma-
Large-scale transformations render existing tions, call for new denitions, clocks, and
scripts (learned through socialization and alloca- calendars of work and career paths and new
tion processes) out of date, opening up new or structures promoting productive engagement at
closing down opportunities to women and/or all life-course stages. Recall the concept of career
men. Wars, economic dislocations, technological is most typically dened as a series of positions,
innovations, and social movements, for example, an orderly and hierarchical progression up an
can transform educational, job, career, age, and occupational status ladder. Individuals experi-
gender strategies, as can changes in social poli- encing uneven or downward pathways are seen as
cies and regulatory practices (Elder 1974; Moen deviant, less committed to their jobs, and reaping
and Roehling 2005; Mortimer and Fischer forth- fewer economic or psychic rewards (Kalleberg
Work Over the Gendered Life Course 263

2011; Wilensky 1961). But this formulation con- ory (Dannefer 2003; Ferraro and Shippee 2009),
tains a number of hidden assumptions about the is about existing disparities that cumulate over
nature of the life course, jobs, and social rela- time. But there are also age-based disparities that
tions. These interrelated assumptions constrain accompany growing older, even for those previ-
both innovative scholarship and innovations in ously advantaged. This is a result of past institu-
work redesign and alternative career paths. tionalization of later adulthood as a distinctive
stage of the life course, separated from prime
adulthood through images, meanings and values
6.1 Assumption 1: The Primacy (culture) and social organization (structure) that
of Paid Work as Key serve to create and sustain social group distinc-
to Identity and Fulllment tions (see as senior, the aged, pensioners,
retirees, the young-old, the old-old.
Americans in particular equate success with Whether the age-boundaries around these
occupational status and career mobility in the groupings are becoming more blurred, and
form of rising occupational prestige and high whether they are more or less blurred for indi-
incomes. Employment is not only a central role viduals differentially located in the social struc-
in American society, it is virtually isomorphic ture (such as by gender, class, race and ethnicity)
with contemporary notions of productivity and are important theoretical and empirical questions
achievement. For men, success has been about for future scholarship. Whether later adulthood is
climbing occupational ladders, accruing along being further divided into encore adulthood
the way money, power, and prestige. In the past, and old age stages of the life course is also a
for women, success was cast as marrying a man key topic for investigation (see Mortimer and
with these resources (Clausen and Gilens 1990). Moen this volume). The answers may well turn
Both work and the workforce have changed, and on labor market, income support, and residential
yet achievement and productivity continue to be policies, as well as the health and disability sta-
characterized in terms of occupational trajecto- tuses and family circumstances of individuals
ries. This characterization renders unpaid work within particular subgroups of the population.
at home, in the community discounted, literally, To understand individuals and organizational
in national economic indexes (e.g., the unem- behavior requires investigation of institutional
ployment rate, the gross national product) and in contexts that not only dene appropriate behavior
self-assessments (e.g., I am only a housewife; I but also provide explanations or accounts of that
am just a volunteer). behavior (Friedland and Alford 1991; Sewell
One important contribution of the social sci- 1992). Sometimes there is loose coupling or
ences is in clarifying the relationships between decoupling (deliberate disconnects) between
structured experience and self-concepts (e.g. means and goals (Boxenbaum and Jonsson 2008).
Gecas 2003; Mortimer et al. 2015). Going to col- Attention to meaning and values is also cru-
lege, getting a job, marrying, having a child, cial to the study of work and occupational transi-
serving in the military, being laid off, retiring tions, in order to understand age-graded behavior
all change identities how people see them- (such as workforce exits) as voluntary or involun-
selves, and how others see them. There is tary, expected or unexpected. Thus a retirement
insufcient scholarship on how work exits and exit can be a passage to (well-deserved) leisure,
entrances as well as their duration, timing, and the unwanted result of downsizing, or a desired
expectedness shape self-concepts and feelings of second chance, a transition to a new career.
mastery or control for those at different life For most adults, paid work is a major, if not
stages. the principal, source of purposive activity, social
The Matthew Effect, the idea that the rich get relations, independence, identity, and self-
richer and the poor get poorer, which is the basis respect. The role of worker is the way that we
for the cumulative advantage/disadvantage the- become integrated and acknowledged as adult
264 P. Moen

members of the larger community. But this paths open to women, with the (male) career
ignores the importance and valuing of family mystique the taken-for-granted model. Women
and personal experiences and civic engagement. use it to gauge their own experiences, as do their
It is also based on a mid-twentieth century model colleagues and bosses. But this model does not t
of work, fashioned for those without personal unless women relinquish their traditional family
and family needs, goals, and obligations and no responsibilities as nurturers and caregivers on the
longer has the primacy it once had (Matos and domestic front. Increasingly, the model does not
Galinsky 2012). t mens experiences either (Gerson 2010). Men
Understanding paid work requires attention to have been able to follow the conventional career
the meaning, measurement, and management of model precisely because (1) they did not shoulder
time over the life course of individuals, families, much of the family care work responsibilities,
and organizations. Sensitivity to timing under- and (2) the employer-employee contract provided
scores the need to locate work lives in institu- internal labor markets and avenues for occupa-
tional and organizational as well as historical and tional mobility.
biographical contexts, accentuating age, cohort, The potential signicance of ones career
period, and gender differences and disparities, course for subsequent life quality was rst exam-
but also within-gender, within age, within cohort, ined by Wilensky in the 1950s. In his classic
and within-period inequalities between those in study, Wilensky (1961) investigated the orderli-
low-wage and high-wage work (as well as ness of mens careers, dening disorderly
through the exclusion of some from careers as having a series of functionally and
employment). hierarchically unrelated jobs for at least four
fths of their work history. He found that men
who had orderly careers were more likely (1) to
6.2 Assumption 2: The Male have strong attachments to their communities and
Experience as Template greater participation in community activities, (2)
to integrate work and non-work roles more, and
The use of the career concept is heavily biased (3) to maintain friendships over longer periods of
toward the experience of men, not women (Bem time, as compared to men with disorderly careers.
1994). Thus, career commonly refers to mov- Since social integration has been shown to be
ing through a series of (related) jobs over the life related to later-life health and well-being (Moen
course, the typical experience of men in the sec- et al. 1989, 1992) having had a disorderly work
ond half of the last century. Mens work can be career may also be negatively related to retire-
subdivided into organizational careers moving ment quality.
up internal ladders within a corporation and Unfortunately, Wilensky, as well as other
occupational careers moving up internal lad- scholars who have examined employment history
ders within professions. But when women are the and well-being (e.g., Pavalko et al. 1993), focused
focus, the career concept is typically constrained exclusively on mens career trajectories. But in
to a more narrow denition, referring to their point of fact, women are more apt to have disor-
remaining in, or moving in and out of, the labor derly careers than are men. Given the primacy of
force in tandem with shafting family care their traditional family roles, womens career pat-
obligations. terns are often characterized as erratic,
But from the 1970s on, people have earned uneven, or chaotic, with marriage and chil-
their living in a workforce comprised of both dren often obstacles or at least interruptions in
men and women. Yet we continue to hold men their employment histories, and changes in fam-
and women to a male career trajectory, a path ily responsibilities often requiring women to
that, in reality, is increasingly available to ever move in and out of the labor force and in and out
fewer men or women. Thus far, the gender revo- of part-time work (Moen 1985). Mens tradi-
lution has focused on rendering mens career tional roles as breadwinners has meant that career
Work Over the Gendered Life Course 265

interruptions have been far less common for Fair Labor Standards Act all took as a given
them. Thus, despite changes in womens labor both the lock-step path of (mens) continuous,
force participation in the second half of the twen- full-time employment and the breadwinner-home-
tieth century, mens employment histories con- maker gender divide. These policies, together
tinue to look more orderly than womens. with regulations guiding business practices, con-
In the 1970s, Spillerman (1977) described the stitute an age-graded regime giving structure to
construction of typologies based on empirical reg- the life course: the shared understandings and
ularities in actual career lines (see also Kanter taken-for-granted rules, roles, relationships,
1977). We need empirical accounts of contempo- resources and risks associated with adulthood at
rary twenty-rst century career paths mens and different ages and life stages. Education, employ-
womens, as well as the joint paths of husbands and ment and retirement have served (together with
wives. The need for new career typologies empiri- marriage and parenting) to structure virtually all
cally derived and grounded in community, family, aspects of the adult life course, with retirement
organizational and societal contexts, is becoming frequently a marker of old age. And yet this lock-
increasingly self-evident (Hobson 2014). step sequence of education, employment, and
retirement is obsolete, a cultural relic of a society
that no longer exists.
6.3 Assumption 3: The Inevitability What is required by scholars and leaders is a
of the Lock-Step Life Course thoughtful reappraisal of existing life patterns
and examination of alternatives to the existing
A regime of twentieth century social and organi- lock step. This could lead to a reconguration of
zational policies and practices have created, rein- the life course in ways that create more employ-
forced and perpetuated different life courses for ment and educational options and more variety
men and women. First were efforts to institution- for both men and women in youth, early adult-
alize life into segments predicated on middle-class hood, midlife, and encore adulthood. A key ques-
mens experiences. In fact, the very notion that tion for policy and research is: How do we
adulthood consists of distinctive and identiable respond to institutional inertia? How can we
paths is a product of primarily mid-twentieth cen- change the clocks and calendars of work and the
tury policies and practices developing in Europe lock step life course to better t the realities of
and America around the institutions of education, todays and tomorrows workforces, families,
employment, and retirement (see Hobson 2014; and economies?
Kohli 1986a, b; Marshall et al. 2001; Meyer 1986;
Moen 2013; Shanahan et al. 2002). Recall that in
the middle of the twentieth century, educational, 6.4 Assumption 4: The Individual
employment, and pension legislation and regula- as Focal Point
tion forged a lock-step life course, consisting of
rst full-time public education as preparation for Workers are typically thought of as individuals
adult roles, then an adulthood of continuous, full- without family responsibilities or constraints,
time employment, followed by full-time leisure much less other interests or values, and their
during the golden years of retirement. This adult careers are similarly dened as reecting the
path of continuous, full-time, year-round employ- individuals mobility patterns. But family and
ment, bracketed by schooling at one end and personal considerations and values increasingly
retirement (or death) at the other, became institu- intrude on the work weeks and career paths of
tionalized and taken for granted (Moen 2012, both men and women, as do considerations and
2013; Moen and Roehling 2005). In the U.S., gov- resources associated with a spouses career, chil-
ernment policies Social Security, Unemployment drens schooling, and aging parents needs as
Insurance, Medicare, disability regulations, the well as those of coworkers and supervisors. This
266 P. Moen

conforms to the life course notion of linked lives, tured for breadwinners with homemakers, even
suggesting the importance of couples or house- though neither men nor women are apt to have
holds or work teams as appropriate units of anal- such back up. Family goals and obligations con-
ysis or else key contexts in investigating and tinue to shape womens lives even though most
understanding work over the life course. are entrenched in the workforce (Carr 1997;
And yet existing rhythms, clocks and calen- Garey 1999; Gerson 2010; Han and Moen 1999a,
dars of work and occupational careers remain b, 2002; Ridgeway and Correll 2004). When cou-
based on the presumption that employees have no ples have children, therefore, it is womens wages
family responsibilities, and often that workers are and career progression that suffers (Correll et al.
not embedded in work groups and social net- 2007; Staff and Mortimer 2012). Thus the social
works. But social relationships matter, even organization of the life course typically produces
though this is less blatant than in the 1950s when diverging pathways for men and women
organizations frequently interviewed the wives of (Crompton 1999; Hobson 2014; Moen and
job candidates to ensure that they too were dedi- Roehling 2005) fostering work-family conicts
cated to their husbands upward career mobility. and strains for women and their families (see
As Liljestrm et al. (1978, p. 105) pointed out, Kelly et al. 2008). The life course as an institu-
When industrialization separated homes from tion allocating roles and resources over time is
workplaces the mens contacts with children also replete with informal norms, including those
were diminished. Childhood ended up in wom- held by work teams and social networks, that
ens territory. The child formed more and more shape womens and mens cognitive assessments,
the core of meaning in the married womans their ambitions, stock-taking, and self-image at
life.... No wonder, then that the mother came to various times during their lives (Krger and
be chained to a pedestal of indispensability.... Baldus 1999 p. 356; see also Altucher and
Each parent developed a specialty: the one Williams 2003; Becker and Moen 1999; Blossfeld
became a Mother, the other a Breadwinner. and Huinik 1991; Moen and Orrange 2002;
Middle-class breadwinners had someone at home Townsend 2002).
to do the domestic work of the household as well Contemporary strategies for dealing with the
as the emotional work of family relationships. time crunch frequently involve women cutting
But absent hiring out the nurturing of family rela- back hours on the job and, often, career pros-
tionships (Hochschild 2012), it is hard for con- pects. But the larger issue of integrating work life
temporary workers men or women to manage with home life is occurring in an economic
the competing pressures for time, energy, com- climate of reduced opportunity for mobility for
mitment, and attention between work and family. both men and women, at a time when the male
As boomers move through their 50s, 60s, and 70s template for success ts ever fewer workers,
the care of aging parents and other relatives, regardless of gender. Key topics for future
including ailing spouses, will only accentuate the research: How can the values of individual
disjuncture between the reality of their work lives achievement and success be reconciled with
and their own (as well as others) expectations. familistic values? Can organizational career pat-
Middle-class families in the 1950s, 1960s, and terns accommodate to the shifting personal
1970s could afford one career to a family. needs, goals, and interests of workers over the
Breadwinners who did well in their careers did life course? Are there sustainable and exible
well for their families (Becker 1981; Newman career paths and coworker relationships that pro-
1993). For contemporary workers to attain or mote life quality, organizational effectiveness
maintain their desired standard of living most fre- and gender equality?
quently requires two incomes. The temporal organization of jobs and career
Contemporary gender inequalities are in large paths institutionalized in public and corporate
part the result of the fact that jobs remain struc- policies and practices limits employee control
Work Over the Gendered Life Course 267

and exibility in arranging their daily schedules, represents a set of practices and policies that no
as well as their control and exibility over arrang- longer t with the realities of a changing econ-
ing their career paths and achieving within cou- omy, changing gender roles, blurred lines
ple gender equality. Moreover, wider options in between education, work and retirement, and
the young adult and encore adult years (see cohorts of educated, introspective boomers,
Mortimer and Moen this volume) are difcult if genXers, and millennials. Many of these coexist-
not impossible without the development of ing members of contemporary cohorts value their
greater career and schedule exibilities in the personal and family lives and exible yet mean-
clockworks of workdays, workweeks, work ingful work more than occupational achievement,
years, and work lives (Kelly and Moen 2007). yet they are uncertain as to how to t the pieces of
Needed are work redesigns that provide workers their lives together in an unsettled global climate
with greater exibility and control over their time that individualizes the risks of employment and
(Kelly et al. 2011, 2014; Kossek et al.2014; Moen economic insecurity. Required are new institu-
et al. 2011a). Without such fundamental work tional arrangements for more exible and sus-
redesign, the only way to gender equality seems tainable work and occupational careers for those
to be for young and middle-aged women and men with personal or family interests, goals, and
alike to function as if they have no obligations obligations.
outside of their jobs and for older workers to Employers, policy makers, thought leaders,
make special arrangements with their employers scholars, and individuals have yet to achieve con-
for shifting gears or else fashion their own sec- sensus as to what the nature of twenty-rst cen-
ond chapters. tury work, careers, education, and retirement will
be, or how to restructure the education-work,
work-family, or work-retirement interfaces. Yet
6.5 Whats Next? current assumptions and practices seem increas-
ingly outmoded and inappropriate. The absence
There are two ways to change the gender and age of consensus about new ways to work, learn, and
stratication systems as well as the structural lag retire means that workers continue to confront
around work and career paths, from the top down structural lags in the shape and structure of the
and from the bottom up. Some argue that you life course. The mismatch between outdated rules
have to rst change beliefs, norms and stereo- and updated but ambiguous realities is producing
types and then concrete changes will appear a sense of ambivalence about current and future
(such as equitable income). Others suggest that ways of working (see also Bourdieu 1984;
changing gender- and age-graded systems and Lscher and Pillemer 1998; Mortimer et al. 2015;
structures leads to corresponding changes in gen- Orrange 2007; Suitor and Pillemer 1994).
der and age stereotypes, beliefs and norms The research and policy challenge is to rewrite
(Hobson 2014; Moen 2013; Serguino 2007). In the scripts of the ideal worker and good jobs
reality both are needed, but institutional inertia that reect todays, not yesterdays realities
suggests the importance of building more exible (Hobson 2014; Williams 2000; Williams and
institutions that are more relevant to the changing Dempsey 2014). Both scholars and organiza-
nature of work and the workforce. tional as well as public policy makers need to
DiMaggio and Powell (1991) point to the way rethink and fashion exible and sustainable alter-
institutional arrangements channel social choices. native clockworks, calendars, and pathways of
Organizational and occupational ladders provide work that take into account heightened insecurity
a common understanding of social mobility and and the new individualization of risk, the hetero-
an identiable pattern of progression through the geneous workforce, and the evolving and varie-
work course. But this common understanding gated twenty-rst century life course.
268 P. Moen

References dual earner families (pp. 201231). Oxford: Oxford


University Press.
Blossfeld, H. P., & Hofmeister, H. (2008). Globalization,
Acker, J. (1992). Gendered institutions: From sex role to
uncertainty and womens careers: An international
gendered institutions. Contemporary Sociology, 21,
comparison. Cheltenham: Edward Elgar.
565569.
Blossfeld, H. P., & Huinik, J. (1991). Human capital
Acker, J. (2006). Inequality regimes: Gender, class, and
investments or norms of role transition? How womens
race in organizations. Gender & Society, 20(4),
schooling and career affect the process of family for-
441464.
mation. American Journal of Sociology, 97, 143168.
Altucher, K., & Williams, L. B. (2003). Family clocks:
Blossfeld, H. P., Buchholz, S., Bukodi, E., & Kurz, K.
Timing parenthood. In P. Moen (Ed.), Its about time:
(2008). Young workers, globalization and the labor
Couples and careers (pp. 4959). Ithaca: Cornell
market. London: Edward Elgar.
University Press.
Blumberg, R. L. (1984). A general theory of gender strati-
Andersen, M. L. (2005). Thinking about women: A quar-
cation. Sociological Theory, 2, 23101.
ter centurys view. Gender & Society, 19(4),
Blustein, D. L. (2008). The role of work in psychological
437455.
health and well-being: A conceptual, historical, and
Andersen, M. L. (2009). Thinking about women:
policy perspective. American Psychologist, 63,
Sociological perspectives on sex and gender (8th ed.).
228240.
Boston: Allyn & Bacon.
Bosma, H., Stanseld, S. A., & Marmot, M. G. (1998).
Aronson, P., & Mortimer, J. T. (2013). Subjective prefer-
Job control, personal characteristics, and heart dis-
ences versus objective realities: Voices of full-and
ease. Journal of Occupational Health Psychology, 3,
part-time employed mothers. In P. N. Claster & S. L.
402409.
Blair (Eds.), Visions of the 21st century family:
Bourdieu, P. (1984). Distinction. Cambridge: Harvard
Transforming structures and identities (pp. 185211).
University.
Bingley: Emerald Group Publishing.
Boxenbaum, E., & Jonsson, S. (2008). Isomorphism, dif-
Ashby, J. S., & Schoon, I. (2010). Career success: The
fusion and decoupling. In R. Greenwood, C. Oliver,
role of teenage career aspirations, ambition value and
R. Suddaby, & K. Sahlin (Eds.), Sage handbook of
gender in predicting adult social status and earnings.
organizational institutionalism (pp. 7898). London:
Journal of Vocational Behavior, 77, 350360.
Sage.
Barley, S. R. (1989). Careers, identities, and institutions:
Brines, J., & Joyner, K. (1999). The ties that bind, princi-
The legacy of the Chicago School of Sociology. In
ples of cohesion in cohabitation and marriage.
M. B. Arthur, D. T. Hall, & B. S. Lawrence (Eds.),
American Sociological Review, 64(3), 333355.
Handbook of career theory (pp. 4165). New York:
Brckner, H., & Mayer, K. U. (2005). De-standardization
Cambridge University Press.
of the life course: What it might mean? And if it means
Beck, U. (1992). Risk society. Newbury Park: Sage.
anything, whether it actually took place? Advances in
Becker, G. S. (1981). A treatise on the family. Cambridge,
Life Course Research, 9, 2753.
MA: Harvard University Press.
Budig, M. J., & England, P. (2001). The wage penalty for
Becker, P. E., & Moen, P. (1999). Scaling back: Dual-
motherhood. American Sociological Review, 66,
earner couples work-family strategies. Journal of
204225.
Marriage and the Family, 61(4), 9951007.
Burchell, B., Ladipo, D., & Wilkinson, F. (2002). Job
Becker, H. S., & Straus, A. L. (1956). Careers, personal-
insecurity and work intensication. New York:
ity, and adult socialization. American Journal of
Routledge.
Sociology, LXII(3), 253263.
Cahill, K. E., Giandrea, M. D., & Quinn, J. F. (2005). Are
Bem, S. L. (1994). The lenses of gender: Transforming the
traditional retirements a thing of the past?: New evi-
debate on sexual inequality. New Haven: Yale
dence on retirement patterns and bridge jobs.
University Press.
Washington, DC: U.S. Department of Labor, Bureau
Berkman, L. F., Kawachi, I., & Glymour, M. M. (2014).
of Labor Statistics.
Social epidemiology. New York: Oxford University
Carr, D. (1997). The fulllment of career dreams at
Press.
midlife: Does it matter for womens mental health?
Bird, C. E., & Rieker, P. P. (2008). Gender and health:
Journal of Health and Social Behavior, 38, 331344.
The effects of constrained choices and social policies.
Carr, D. C., & Manning, L. K. (2010). A new paradigm for
New York: Cambridge University Press.
qualitative research in the United States: The era of the
Blair-Loy, M. (2003). Competing devotions: Career and
third age. Qualitative Sociology Review, VI(1), 1633.
family among women executives. Cambridge, MA:
Ceci, S. J., & Williams, W. M. (2010). The mathematics of
Harvard University Press.
sex: How biology and society conspire to limit talented
Blair-Loy, M. (2009). Work without end? Scheduling ex-
women and girls. New York: Oxford University Press.
ibility and work-to-family conict among stockbro-
Cheng, Y., Kawachi, I., Coakley, E. H., Schwartz, J., &
kers. Work and Occupations, 36, 279317.
Colditz, G. (2000). Association between psychosocial
Blossfeld, H. P., & Drobni, S. (2002). Careers of couples
work characteristics and health functioning in
in contemporary societies: From male breadwinner to
Work Over the Gendered Life Course 269

American women: Prospective study. British Medical Elder, G. H., Jr. (1996). Human lives in changing societ-
Journal, 320(7247), 14321436. ies: Life course and developmental insights. In R. B.
Chesley, N., & Moen, P. (2006). When workers care: Cairns, G. H. Elder Jr., & E. J. Costello (Eds.),
Dual-earner couples caregiving strategies, benet Developmental science (pp. 3162). New York:
use, and psychological well-being. American Cambridge University Press.
Behavioral Scientist, 49(9), 12481269. Elder, G. H., Jr. (1998). The life course as development
Clausen, J. (1995). Gender, contexts, and turning points in theory. Child Development, 69(1), 112.
adults lives. In P. Moen, G. H. Elder Jr., & K. Lscher Elder, G. H., Jr., & George, L. (forthcoming). Age, cohort,
(Eds.), Examining lives in context: Perspectives on the and the life course. In M. J. Shanahan, J. T. Mortimer,
ecology of human development (pp. 365392). & M. K. Johnson (Eds.), Handbook of the life course.
Washington, DC: American Psychological New York: Springer.
Association. Elder, G. H., Jr., Johnson, M. K., & Crosnoe, R. (2003).
Clausen, J. A., & Gilens, M. I. (1990). Personality and The emergence and development of the life course. In
labor force participation across the life course: A lon- J. T. Mortimer & M. J. Shanahan (Eds.), Handbook of
gitudinal study of womens careers. Sociological the life course (pp. 319). New York: Kluwer
Forum, 5, 595618. Academic/Plenum Publishers.
Cooney, T. M., & Mortimer, J. T. (1999). Family structure Elsass, P. M., & Veiga, J. F. (1997). Job control and job
differences in the time of leaving home: Exploring strain: A test of three models. Journal of Occupational
mediating factors. Journal of Research on Adolescence, Health Psychology, 2(3), 195211.
9(4), 367376. England, P. (2005). Gender inequality in labor markets:
Correll, S. J., Benard, S., & Paik, I. (2007). Getting a job: The role of motherhood and segregation. Social
Is there a motherhood penalty? American Journal of Politics, 12, 264268.
Sociology, 112(5), 12971338. Esping-Andersen, G. (2009). The incomplete revolution:
Crompton, R. (1999). Restructuring gender relations and Adapting to womens new roles. Cambridge: Polity
employment: The decline of the male breadwinner. Press.
Oxford: Oxford University Press. Fan, W., & Qian, Y. (2015). Long-term health and socio-
Csikszentmihalyi, M. (1990). Flow: The psychology of economic consequences of early-life exposure to the
optimal experience. New York: Harper Collins 19591961 Chinese famine. Social Science Research,
Publishers. 49, 5369.
Daehlen, M. (2005). Change in job values during educa- Fenstermaker, S., & West, C. (2002). Doing gender, doing
tion. Journal of Education and Work, 18(4), different: Inequality, power, and institutional change.
385400. New York: Routledge.
Damaske, S. (2011). For the family? How class and gen- Fernandez, R. M., & Sosa, M. L. (2005). Gendering the
der shape womens work. New York: Oxford job: Networks and recruitment at a call center.
University Press. American Journal of Sociology, 111(3), 859904.
Dannefer, D. (2003). Cumulative advantage/disadvantage Ferraro, K. F., & Shippee, T. P. (2009). Aging and cumula-
and the life course: Cross-fertilizing age and social tive inequality: How does inequality get under the
science theory. Journal of Gerontology, 58(6), skin? The Gerontologist, 49(3), 333343.
327337. Folbre, N. (2001). The invisible heart: Economics and
Davis, G. F. (2009). Managed by the markets: How family values. New York: New Press.
nance re-shaped America. Oxford: Oxford University Folbre, N. (2012). For love and money: Care provision in
Press. the U.S.. New York: Russell Sage.
de Jonge, J., Bosma, H., Peter, R., & Siegrist, J. (2000). Freund, A. M. (1997). Individuating age salience: A psy-
Job strain, effort-reward imbalance and employee chological perspective on the salience of age in the life
well-being: A large-scale cross-sectional study. Social course. Human Development, 40, 287292.
Science & Medicine, 50(9), 13171327. Friedan, B. (1963). The feminine mystique. New York:
DiMaggio, P. J., & Powell, W. W. (1991). Introduction. In Bantam Doubleday Dell.
W. W. Powell & P. J. DiMaggio (Eds.), The new insti- Friedland, R., & Alford, R. A. (1991). Bringing society
tutionalism in organizational analysis. Chicago: back in: Symbols, practices, and institutional contra-
University of Chicago Press. dictions. In W. W. Powell & P. J. DiMaggio (Eds.), The
Dwyer, D. J., & Ganster, D. C. (1991). The effects of job new institutionalism in organizational analysis
demands and control on employee attendance and sat- (pp. 232263). Chicago: University of Chicago Press.
isfaction. Journal of Organizational Behavior, 12(7), Fullerton, A. S., & Wallace, M. (2007). Traversing the
595608. exible turn: U.S. workers perceptions of job secu-
Elder, G. H., Jr. (1974). Children of the great depression. rity, 19772002. Social Science Research, 35(1),
Chicago: University of Chicago Press. 201221.
Elder, G. H., Jr. (1994). Time, human agency, and social Gangl, M., & Ziee, A. (2009). Motherhood, labor force
change: Perspectives on the life course. Social behavior and womens careers: An empirical assess-
Psychology Quarterly, 57, 415. ment of the wage penalty for motherhood in Britain,
270 P. Moen

Germany and the United States. Demography, 46(2), Heinz, W. R. (2002a). Transitional discontinuities and the
341369. biographical shaping of early work careers. Journal of
Garey, A. I. (1999). Weaving work and motherhood. Vocational Behavior, 60, 220240.
Philadelphia: Temple University Press. Heinz, W. R. (2002b). Self-socialization and post-
Gecas, V. (2003). Self-agency and the life course. In J. T. transitional society. Advances in Life Course Research,
Mortimer & M. Shanahan (Eds.), Handbook of the life 7, 4164.
course (pp. 369388). New York: Kluwer Academic/ Heinz, W. R. (2003). From work trajectories to negotiated
Plenum Publishers. careers: The contingent work life course. In J. T.
George, L. K. (1993). Sociological perspectives on life Mortimer & M. J. Shanahan (Eds.), Handbook of the
transitions. Annual Review of Sociology, 19, 353373. life course (pp. 185204). New York: Kluwer
George, L. K. (2003). Life course research: Achievements Academic/Plenum Publishers.
and potential. In J. T. Mortimer & M. J. Shanahan Heinz, W. R., & Marshall, V. W. (2003). Social dynamics
(Eds.), Handbook of the life course (pp. 671680). of the life course: Transitions, institutions, and inter-
New York: Kluwer Academic/Plenum Publishers. relations. Weinheim: Deutscher Studien Verlag.
George, L. K. (2014). Taking time seriously: A call to Hemingway, H., & Marmot, M. (1998). Psychosocial fac-
action in mental health research. Journal of Health tors in the primary and secondary prevention of coro-
and Social Behavior, 55(3), 251264. nary heart disease: A systematic review. In S. Yusuf,
Gerson, K. (1985). Hard choices: How women decide J. Cairns, J. Camm, E. Fallen, & B. Gersch (Eds.),
about work, career, and motherhood. Berkeley: Evidence based cardiology (pp. 269285). London:
University of California Press. BMJ Publishing Group.
Gerson, K. (2010). The unnished revolution: How a new Hill, R., Tranby, E., Kelly, E. L., & Moen, P. (2013).
generation is reshaping family, work, and gender in Relieving the time squeeze? Effects of a white-collar
America. New York: Oxford University Press. workplace change on parents. Journal of Marriage
Glavin, P. (2013). The impact of job insecurity and job and Family, 75, 10141029.
degradation on the sense of personal control. Work and Hitlin, S., & Elder, G. H., Jr. (2007a). Agency: An empiri-
Occupations, 40, 115142. cal model of an abstract concept. Advances in Life
Goodwin, J., & OConnor, H. (2005). Engineer, mechanic Course Research, 11, 3367.
or carpenter? Boys transitions to work in the 1960s. Hitlin, S., & Elder, G. H., Jr. (2007b). Time, self, and the
Journal of Education and Work, 18(4), 451471. curiously abstract concept of agency. Sociological
Graebner, W. (1980). A history of retirement. New Haven: Theory, 25, 170191.
Yale University Press. Hobson, B. (2014). Worklife balance: The agency and
Hagestad, G. O., & Neugarten, B. L. (1985). Age and the capabilities gap. New York: Oxford University Press.
life course. In R. H. Binstock & E. Shanas (Eds.), Hochschild, A. R. (1997). The time bind: When work
Handbook of aging and the social sciences (pp. 3561). becomes home and home becomes work. New York:
New York: Van Nostrand Reinhold. Metropolitan Books.
Hakim, C. (2004). Key issues in womens work: Female Hochschild, A. R. (2012). The outsourced self: Intimate
diversity and the polarization of womens employ- life in market times. New York: Henry Holt and
ment. London: TheGlassHouse. Company, LLC.
Hall, D. T. (2004). The protean career: A quarter-century Hughes, E. C. (1958). Men and their work. Glencoe: Free
journey. Journal of Vocational Behavior, 53, 113. Press.
Han, S. K., & Moen, P. (1999a). Work and family over Institute for Womens Policy research. (2005a, June).
time: A life course approach. The Annals of the Women and social security: Benet types and eligibil-
American Academy of Political and Social Sciences, ity (IWPR Publication No. D463). Washington, DC.
562, 98110. Institute for Womens Policy research. (2005b, August).
Han, S. K., & Moen, P. (1999b). Clocking out: Temporal The gender wage ratio: Womens and mens earnings
patterning of retirement. American Journal of (IWPR Publication No. C350). Washington, DC.
Sociology, 105, 191236. Jahoda, M. (1982). Employment and unemployment: A
Han, S. K., & Moen, P. (2002). Coupled careers: Pathways social-psychological analysis. Cambridge: Cambridge
through work and marriage in the United States. In University Press.
H. Blossfeld & S. Drobni (Eds.), Careers of couples Johnson, M. K. (2001). Changes in job values during the
in contemporary societies: From male breadwinner to transition to adulthood. Work and Occupations, 28,
dual earner families (pp. 201231). Oxford: Oxford 351354.
University Press. Johnson, R. W., Mermin, G. B., & Resseger, M. (2011).
Harrington Meyer, M. (2014). Grandmas at work: Job demands and work ability at older ages. Journal of
Juggling families and jobs. New York: New York Aging & Social Policy, 23(2), 101118.
University Press. Johnson, M. K., Sage, R. A., & Mortimer, J. T. (2012).
Heinz, W. R. (1999). From education to work: Cross- Work values, early career difculties, and the U.S.
national perspectives. New York: Cambridge economic recession. Social Psychology Quarterly,
University Press. 75(3), 242267.
Work Over the Gendered Life Course 271

Kahn, R. L., & Antonucci, T. C. (1980). Convoys over the Kohli, M., & Meyer, J. W. (1986). Social structure and
life course: Attachment, roles, and social support. In social construction of the life stages. Human
P. B. Baltes & O. C. Brim (Eds.), Life-span, develop- Development, 29, 145149.
ment, and behavior (pp. 254283). New York: Kohli, M., Rein, M., Guillemard, A.-M., & van Gunsteren,
Academic. H. (1991). Time for retirement: Comparative studies of
Kalleberg, A. L. (2000). Nonstandard employment rela- early exit from the labor force. Cambridge: Cambridge
tions: Part-time, temporary, and contract work. Annual University Press.
Review of Sociology, 26, 341365. Kohn, M. L., & Schooler, C. (1983). Work and personal-
Kalleberg, A. L. (2009). Precarious work, insecure work- ity: An inquiry into the impact of social stratication.
ers: Employment relations in transition. American Norwood: Ablex.
Sociological Review, 74(1), 122. Kossek, E. E., Hammer, L. B., Kelly, E. L., & Moen, P.
Kalleberg, A. L. (2011). Good jobs, bad jobs: The rise of (2014). Designing work, family & health organiza-
polarized and precarious employment systems in the tional change initiatives. Organizational Dynamics,
United States, 1970s2000s. New York: Russell Sage. 43(1), 5363.
Kalleberg, A. L. (2012). Job quality and precarious work: Kramer, L. (2011). The sociology of gender: A brief intro-
Clarications, controversies, and challenges. Work duction (2nd ed.). New York: Oxford University Press.
and Occupations, 39(4), 427448. Kristensen, T. S. (1995). The demand-control-support
Kanter, R. M. (1977). Work and family in the United model: Methodological challenges for future research.
States: A critical review and agenda for research and Stress Medicine, 11, 1726.
policy. New York: Russell Sage. Kristensen, T. S. (1996). Job stress and cardiovascular dis-
Karasek, R. A., & Theorell, T. (1990). Healthy work: ease: A theoretic critical review. Journal of
Stress, productivity and the reconstruction of working Occupational Health Psychology, 1, 246260.
life. New York: Basic Books. Krger, H., & Baldus, B. (1999). Work, gender and the
Kelly, E., & Moen, P. (2007). Rethinking the clockwork of life course: Social construction and individual experi-
work: Why schedule control may pay off at home and ence. Canadian Journal of Sociology, 24(3),
at work. Advances in Developing Human Resources, 355379.
9(4), 487506. Landsbergis, P. A., Schnall, P. L., Dietz, D., Friedman, R.,
Kelly, E. L., Kossek, E. E., Hammer, L., Durham, M., & Pickering, T. (1992). The patterning of psychologi-
Bray, J., Chermack, K., & Murphy, L. (2008). Getting cal attributes and distress by job strain and social sup-
there from here: Research on the effects of work- port in a sample of working men. Journal of Behavioral
family initiatives on work-family conict and business Medicine, 15(4), 379405.
outcomes. Academy of Management Annals, 2(1), Levy, R., & Widmer, E. (2013). Gendered life courses
305340. between individualization and standardization. A
Kelly, E. L., Moen, P., & Tranby, E. (2011). Changing European approach applied to Switzerland. Wien: Lit
workplaces to reduce work-family conict: Schedule Verlag.
control in a white-collar organization. American Liao, T. F., & Cai, Y. (1995). Socialization, life situations,
Sociological Review, 76(2), 265290. and gender-role attitudes regarding the family among
Kelly, E. L., Moen, P., Fan, W., Oakes, J. M., Okechukwu, white American women. Sociological Perspectives,
C., Davis, K. D., Hammer, L., Kossek, E., King, R. B., 38(2), 241260.
Hanson, G., Mierzwa, F., & Casper, L. (2014). Liljestrm, R., Svensson, G. L., & Mellstrm, G. F.
Changing work and work-family conict in an infor- (1978). Roles in transition: Report of an investigation
mation technology workplace: Evidence from a group- made for the advisory council on equality between
randomized trial. American Sociological Review, men and women. Stockholm: Liber Forlag.
79(3), 485516. Lippmann, S. (2008). Rethinking risk in the new econ-
Kimmel, M. S. (2008). The gendered society. New York: omy: Age and cohort effects on unemployment and
Oxford University Press. re-employment. Human Relations, 61(9), 12591292.
Kohli, M. (1986a). The world we forget: A historical Lscher, K., & Pillemer, K. (1998). Intergenerational
review of the life course. In V. W. Marshall (Ed.), ambivalence: A new approach to the study of parent-
Later life: The social psychology of aging (pp. 271 child relations in later life. Journal of Marriage and
303). Beverly Hills: Sage. the Family, 60, 413425.
Kohli, M. (1986b). Social organization and subjective Macmillan, R., & Eliason, S. R. (2003). Characterizing
construction of the life course. In A. B. Srensen, F. E. the life course as role congurations and pathways: A
Weinert, & L. R. Sherrod (Eds.), Human development latent structure approach. In J. T. Mortimer & M. J.
and the life course: Multidisciplinary perspectives Shanahan (Eds.), Handbook of the life course
(pp. 271292). Hillsdale: Erlbaum. (pp. 529554). New York: Kluwer Academic/Plenum
Kohli, M. (2007). The institutionalization of the life Publishers.
course: Looking back to look ahead. Research in Malenfant, R., LaRue, A., & Venzina, M. (2007).
Human Development, 4, 253271. Intermittent work and well-being: One foot in the
272 P. Moen

door, one foot out. Current Sociology, 55(6), Handbook of work family integration: Research, the-
814835. ory, and best practices (pp. 95118). New York:
Marshall, N. L., Barnett, R. C., & Sayer, A. (1997). The Elsevier.
changing workforce, job stress and psychological dis- Moen, P., & Flood, S. (2013). Limited engagements?
tress. Journal of Occupational Health Psychology, 2, Womens and mens work/volunteer time in the encore
99107. life course stage. Social Problems, 60(2), 206233.
Marshall, V. W., Heinz, W., Krger, H., & Verma, A. Moen, P., & Orrange, R. (2002). Careers and lives:
(2001). Restructuring work and the life course. Socialization, structural lag, and gendered ambiva-
Toronto: University of Toronto Press. lence. Advances in Life Course Research, 7, 231260.
Matos, K., & Galinsky, E. (2012). National study of Moen, P., & Roehling, P. (2005). The career mystique:
employers. New York: Families and Work Institute. Cracks in the American dream. Boulder: Rowman &
McCall, L. (2001). Complex inequality: Gender, class, Littleeld.
and race in the new economy. New York: Routledge. Moen, P., & Spencer, D. (2006). Converging divergences
Meyer, J. W. (1986). The institutionalization of the life in age, gender, health, and well-being: Strategic selec-
course and its effects on the self. In F. E. Weinert & tion in the third age. In R. Binstock & L. George
L. R. Sherrod (Eds.), Human development and the life (Eds.), Handbook of aging and the social sciences
course: Multidisciplinary perspectives (pp. 119216). (pp. 127144). Burlington: Elsevier Academic Press.
Hillsdale: Erlbaum. Moen, P., & Wethington, E. (1992). The concept of family
Mirowsky, J., & Ross, C. E. (1989). Social causes of psy- adaptive strategies. Annual Review of Sociology, 18,
chological distress. New York: Aldine de Gruyter. 233251.
Moen, P. (1985). Continuities and discontinuities in wom- Moen, P., Dempster-McClain, D., & Williams, R. M., Jr.
ens labor force participation. In G. H. Elder Jr. (Ed.), (1989). Social integration and longevity: An event his-
Life course dynamics: 1960s to 1980s (pp. 113155). tory analysis of womens roles and resilience.
Ithaca: Cornell University Press. American Sociological Review, 54, 635647.
Moen, P. (1992). Womens two roles: A contemporary Moen, P., Dempster-McClain, D., & Williams, R. M., Jr.
dilemma. Westport: Greenwood. (1992). Successful aging: A life course perspective on
Moen, P. (1994). Women, work and family: A sociological womens roles and health. American Journal of
perspective on changing roles. In M. W. Riley, R. L. Sociology, 97, 16121638.
Kahn, & A. Foner (Eds.), Age and structural lag: The Moen, P., Sweet, S., & Swisher, R. (2005). Embedded
mismatch between peoples lives and opportunities in career clocks: The case of retirement planning.
work, family, and leisure. New York: Wiley. Advances in Life Course Research, 9, 237265.
Moen, P. (2001). The gendered life course. In L. George & Moen, P., Huang, Q., Plassmann, V., & Dentinger, E.
R. H. Binstock (Eds.), Handbook of aging and the (2006). Deciding the future: Do dual-earner couples
social sciences (pp. 179196). San Diego: Academic. plan together for retirement? American Behavioral
Moen, P. (2003a). Midcourse: Navigating retirement and Scientist, 49(9), 122.
a new life stage. In J. Mortimer & M. J. Shanahan Moen, P., Sweet, S., & Hill, R. (2010). Risk, resilience,
(Eds.), Handbook of the life course (pp. 269291). and life-course t: Older couples encores following
New York: Kluwer Academic/Plenum Publishers. job loss. In P. S. Fry & C. L. M. Keyes (Eds.), New
Moen, P. (2003b). Its about time: Couples and careers. frontiers in resilient aging (pp. 283309). New York:
Ithaca: Cornell University Press. Cambridge University Press.
Moen, P. (2012). Retirement dilemmas and decisions. In Moen, P., Kelly, E. L., & Hill, R. (2011a). Does enhancing
J. Hedge & W. Borman (Eds.), The Oxford handbook work-time control and exibility reduce turnover? A
of work and aging (pp. 549569). New York: Oxford naturally-occurring experiment. Social Problems,
University Press. 58(1), 6998.
Moen, P. (2013). Constrained choices: The shifting insti- Moen, P., Kelly, E. L., Tranby, E., & Huang, Q. (2011b).
tutional contexts of aging and the life course. In Changing work, changing health: Can real work-time
L. Waite (Ed.), Perspectives on the future of the sociol- exibility promote health behaviors and well-being?
ogy of aging (pp. 175216). Washington, DC: National Journal of Health and Social Behavior, 52, 404429.
Research Council. Moen, P., Kaduk, A., Kossek, E. E., Hammer, L., Buxton,
Moen, P. (forthcoming). Encore Adulthood: Boomers on O. M., ODonnell, E., Almeida, D., Fox, L., Tranby,
the Edge of Risk, Renewal, and Purpose. New York: E., Oaks, J. M., & Casper, L. (2015). Is work-family
Oxford University Press. conict a multi-level stressor linking job conditions to
Moen, P., & Chermack, K. (2005). Gender disparities in mental health? Evidence from the work, family and
health: Strategic selection, careers, and cycles of con- health network. Research in the Sociology of Work, 26:
trol. Journal of Gerontology, 60B, 99108. 177-217.
Moen, P., & Chesley, N. (2008). Toxic job ecologies, lag- Moen, P., Kelly, E. L., Fan, W., Lee, S-R., Almeida, D.,
ging time convoys, and work-family conict: Can Kossek, E. E., & Buxton, O. (Forthcoming). Does a
families (re)gain control and life course t? In D. S. exibility/support organizational initiative improve
Lero, K. Korabik, & D. L. Whitehead (Eds.), high tech employees subjective wellbeing? Evidence
Work Over the Gendered Life Course 273

from the Work, Family, and Health Network. American Journal of Personality and Social Psychology, 83,
Sociological Review. 4459.
Moen, P., Kelly, E. L., Oakes, J. M., Lee, S-R., Bray, J., Ogums, R. U. (2012). Living and working in retirement, a
Almeida, D. M., Hammer, L., Hurtado, D., & Buxton, new paradigm in the U.S. exploring retirees attitudes
O. (2014, August). Can a exibility/support initiative and beliefs toward working. Journal of Economic &
reduce turnover? Results from the work, family and Social Studies (JECOSS), 2(1), 2150.
health study. Presentation at the Annual Meeting of Orrange, R. M. (2007). Work, family, and leisure:
the American Sociological Association, held in San Uncertainty in a risk society. Boulder: Rowman &
Francisco. Littleeld.
Mortimer, J. T. (2011). Transition to adulthood, parental Pavalko, E. K., Elder, G. H., Jr., & Clipp, E. C. (1993).
support, and early adult well-being: Recent ndings Worklives and longevity: Insights from a life course
from the Youth Development Study. In A. Booth, perspective. Journal of Health and Social Behavior,
S. Brown, N. Lansdale, W. Manning, & S. McHale 34(4), 363380.
(Eds.), Early adulthood in a family context (pp. 2734). Pearlin, L. (1988). Social structure and social values: The
New York: Springer. regulation of structural effects. In H. J. OGorman
Mortimer, J. T., & Fischer, L. (Forthcoming). Career (Ed.), Surveying social life (pp. 252264). Middletown:
beginnings American style: Agency and oundering in Wesleyan University.
subjective perspective. In T. T. Swartz, D. Hartmann, Pearlin, L. I., Schieman, S., Fazio, E. M., & Meersman, S.
& R. Rumbaut (Eds.), Crossings to adulthood: How (2005). Stress, health, and the life course: Some con-
young Americans understand and navigate their lives. ceptual perspectives. Journal of Health and Social
Leiden: Brill Publishers. Behavior, 46, 205219.
Mortimer, J. T., & Moen, P. (this volume). The changing Pedrianna, N. (2004). Help wanted NOW: Legal opportu-
social construction of age: Precarious identity and nities, the womens movement, and the battle over sex-
enactment of early and encore stages of adult- segregated job advertisements. Social Problems, 51,
hood. In M. Shanahan, J. T. Mortimer & M. K. 182201.
Johnson (Eds.), The handbook of the life course (Vol. Person, A. E., Rosenbaum, J. E., & Deil-Amen, R. (2005).
2). New York: Springer. Colleges, careers, and the institutional structure of the
Mortimer, J. T., & Shanahan, M. J. (2003). Handbook of transition to adulthood. Advances in Life Course
the life course. New York: Kluwer/Plenum Publishers. Research, 9, 125149.
Mortimer, J., & Simmons, R. G. (1978). Adult socializa- Pixley, J. E., & Moen, P. (2003). Prioritizing careers. In
tion. Annual Review Sociology, 4, 421454. P. Moen (Ed.), Its about time: Couples and careers
Mortimer, J. T., Kim, M. & Swartz, T. T. (2010). (pp. 183200). Ithaca: Cornell University Press.
Unemployment, parental help, and self efcacy during Preves, S., & Mortimer, J. T. (2013). Socialization for pri-
the transition to adulthood. Paper presented at the mary, intimate, and work relationships in the adult life
IUSSP Seminar on Intergenerational Ties and course. In J. DeLamater (Ed.), Handbook of social
Transitions to Adulthood, sponsored by Bocconi psychology (2nd ed., pp. 151190). New York:
University, the MacArthur Network on Transitions to Springer.
Adulthood, and the IUSSP Scientic Panel on Reynolds, J., & Aletraris, L. (2010). Mostly mismatched
Transitions to Adulthood in Developed Countries. with a chance of settling: Tracking work hour mis-
Milan, 89 Nov 2010. matches in the United States. Work and Occupations,
Mortimer, J. T., Vuolo, M., & Staff, J. (2014). Agentic 37(4), 471511.
pathways toward fulllment in work. In A. C. Keller, Ridgeway, C. L., & Correll, S. J. (2004). Unpacking the
R. Samuel, M. M. Bergman, & N. K. Semmer (Eds.), gender system: A theoretical perspective on gender
Psychological, educational, and sociological perspec- beliefs and social relations. Gender & Society, 18(4),
tives on success and well-being in career development 510531.
(pp. 99126). New York: Springer. Ridgeway, C. L., & Smith-Lovin, L. (1999). The gender
Mortimer, J. T., Lam, J., & Lee, S.-R. (2015). system and interaction. Annual Review of Sociology,
Transformation, erosion, or disparity in work identity? 25, 191216.
Challenges during the contemporary transition to Riley, M. W. (1987). On the signicance of age in sociol-
adulthood. In K. McLean & M. Syed (Eds.), Oxford ogy. American Sociological Review, 52(1), 114.
handbook of identity development (pp. 319336). Riley, M. W., Kahn, R. L., & Foner, A. (1994). Age and
New York: Oxford University Press. structural lag: The mismatch between peoples lives
Neugarten, B. L., Moore, J. W., & Lowe, J. C. (1965). Age and opportunities in work, family, and leisure.
norms, age constraints, and adult socialization. New York: Wiley.
American Journal of Sociology, 70, 710717. Risman, B. J. (1998). Gender vertigo: American families
Newman, K. S. (1993). Declining fortunes: The withering in transition. New Haven: Yale University Press.
of the American dream. New York: Basic Books. Rothman, R. A. (1979). Occupational roles: Power and
Nosek, B. A., Banaji, M. R., & Greenwald, A. G. (2002). negotiation in the division of labor. The Sociological
Math = male, me = female, therefore math me. Quarterly, 20, 495515.
274 P. Moen

Ryder, N. B. (1965). The cohort as a concept in the study Shanahan, M. J., Elder, G. H., Jr., & Miech, R. A. (1997).
of social change. American Sociological Review, History and agency in mens lives: Pathways to
30(6), 843861. achievement in cohort perspective. Sociology of
Sargent, L. D., Lee, M. D., Martin, B., & Zikic, J. Education, 70, 5467.
(2013). Reinventing retirement: New pathways, new Shanahan, M. J., Mortimer, J. T., & Krger, H. (2002).
arrangements, new meanings. Human Relations, Adolescence and adult work in the twenty-rst cen-
66(1), 321. tury. Journal of Research on Adolescence, 12,
Schieman, S., Van Gundy, K., & Taylor, J. (2002). The 99120.
relationship between age and depressive symptoms: A Silva, J. M. (2013). Coming up short: Working-class
test of competing explanatory and suppression inu- adulthood in an age of uncertainty. New York: Oxford
ences. Journal of Aging and Health, 14, 260285. University Press.
Schnall, P. L., Landsbergis, P. A., & Baker, D. (1994). Job Singley, S. G., & Hynes, K. (2005). Transitions to parent-
strain and cardiovascular disease. Annual Review of hood: Work-family policies, gender, and the couple
Public Health, 15, 381411. context. Gender & Society, 19(3), 378397.
Schneider, B., & Stevenson, D. (1999). The ambitious Smith, V. (2001). Crossing the great divide: Worker risk
generation: Americas teenagers, motivated but direc- and opportunity in the new economy. Ithaca: Cornell
tionless. New Haven: Yale University Press. University/ILR Press.
Schooler, C., Mulatu, M. S., & Oates, G. (1999). The con- Spillerman, S. (1977). Careers, labor market structure,
tinuing effects of substantively complex work on the and socioeconomic attainment. American Journal of
intellectual function of older workers. Psychology and Sociology, 83, 551593.
Aging, 14, 483506. Staff, J., & Mortimer, J. T. (2012). Explaining the mother-
Schulz, A. S., & Mullings, L. (2006). Gender, race, class hood wage penalty. Demography, 49, 121.
and health: Intersectional approaches. San Francisco: Stone, P. (2007). Opting out? Why women really quit
Jossey-Bass. careers and head home. Berkeley: University of
Schuman, H., & Rieger, C. (1992). Historical analogies, California Press.
generational effects, and attitudes toward war. Stone, K., & Arthurs, H. (2013). Rethinking workplace
American Sociological Review, 57, 315325. regulation: Beyond the standard contract of employ-
Sennett, R. (1998). The corrosion of character: Personal ment. New York: Russell Sage Foundation
consequences of work in the new capitalism. Publications.
New York: Norton. Suitor, J., & Pillemer, K. (1994). Family caregiving and
Serguino, S. (2007). Plus a change? Evidence on global marital satisfaction. Journal of Marriage and the
trends in gender norms and stereotypes. Feminist Family, 56, 681690.
Economics, 13(2), 118. Swartz, T. T. (2009). Intergenerational family relations in
Settersten, R. A., Jr., & Hagestad, G. O. (1996a). Whats adulthood: Patterns, variations, and implications in the
the latest? Cultural age deadlines for educational and contemporary United States. Annual Review of
work transitions. The Gerontologist, 36(2), 178188. Sociology, 35, 191212.
Settersten, R. A., Jr., & Hagestad, G. O. (1996b). Whats Swartz, T. T., Uno, M., Mortimer, J. T., Kim, M., &
the latest? II. Cultural age deadlines for educational OBrien, K. B. (2011). Safety nets and scaffolds:
and work transitions. The Gerontologist, 36(5), Parental support in the transition to adulthood. Journal
602613. of Marriage and Family, 73, 414429.
Settersten, R. A., Jr., & Mayer, K. U. (1997). The mea- Sweet, S. (2014). The work-family interface: An introduc-
surement of age, age structuring, and the life course. tion. Thousand Oaks: Sage.
Annual Review of Sociology, 23, 233261. Sweet, S., & Meiksins, P. (2013). Changing contours of
Settersten, R. A., Jr., & Owens, T. (2002). New frontiers in work: Jobs and opportunities in the new economy (2nd
socialization. London: Elsevier. ed.). Thousand Oaks: Pine Forge Press.
Settersten, R. A., Jr., & Ray, B. (2010). Not quite adults: Sweet, S., & Moen, P. (2012). Dual earners preparing for
Why 20-somethings are choosing a slower path to job loss: Agency, linked lives, and resilience. Work
adulthood and why its good for everyone. New York: and Occupations, 39(1), 3570.
Bantam/Random House. Swidler, A. (1986). Culture in action: Symbols and
Settersten, R. A., Jr., Furstenberg, F. F., Jr., & Rumbaut, strategies. American Sociological Review, 51,
R. G. (2001). On the frontier of adulthood: Theory, 273286.
research and public policy. Chicago: University of Toossi, M. (2013). Labor force projections to 2022: The
Chicago Press. labor force participation rate continues to fall. Monthly
Sewell, W. H., Jr. (1992). A theory of structure: Duality, Labor Review (Dec): 128.
agency, and transformations. American Journal of Townsend, N. W. (2002). The package deal: Marriage,
Sociology, 98, 129. work, and fatherhood in mens lives. Philadelphia:
Shanahan, M. J. (2000). Pathways to adulthood in chang- Temple University Press.
ing societies: Variability and mechanisms in life course Uchitelle, L. (2007). The disposable American: Layoffs
perspective. Annual Review of Sociology, 25, 667692. and their consequences. New York: Vintage Books.
Work Over the Gendered Life Course 275

Vuolo, M., Staff, J., & Mortimer, J. T. (2012). Weathering West, C., & Zimmerman, D. H. (1987). Doing gender.
the great recession: Psychological and behavioral Gender & Society, 1(2), 125151.
trajectories in the transition from school to work. Wethington, E., Pixley, J. E., & Kavey, A. (2003). Turning
Developmental Psychology, 48, 17591773. points in work careers. In P. Moen (Ed.), Its about
Vuolo, M., Mortimer, J. T., & Staff, J. (2013). Adolescent time: Couples and careers (pp. 168182). Ithaca:
precursors of pathways from school to work. Journal Cornell University Press.
of Research on Adolescence, 24(1), 145162. Wilensky, H. (1961). Orderly careers and social
Warner, D. F., Hayward, M. D., & Hardy, M. A. (2010). participation. American Sociological Review, 26,
The retirement life course in America at the dawn of 521539.
the twenty-rst century. Population Research and Williams, J. C. (2000). Unbending gender: Why family
Policy Review, 29(6), 893919. and work conict and what to do about it. Oxford:
Warren, J. R., Hauser, R. M., & Sheridan, J. T. (2002). Oxford University Press.
Occupational stratication across the life course: Williams, J. C., & Dempsey, R. (2014). What works for
Evidence from the Wisconsin Longitudinal Study. women at work: Four patterns working women need to
American Sociological Review, 67, 432455. know. New York: New York University Press.
Military Service in Lives: Where Do
We Go From Here?

Andrew S. London and Janet M. Wilmoth

1 Introduction Society of America (GSA)-sponsored journals


(The Journals of Gerontology: Series A, The
The military is a complex, dynamic, and poten- Journals of Gerontology: Series B, The
tially transformative early life-course institution Gerontologist, The Journal of Gerontology,
(Bennett and McDonald 2013; Kelty and Segal Public Policy & Aging Report), Research on
2013) about which we know less than we should. Aging, and Journal of Aging and Health pub-
It is increasingly recognized that military service lished a total of 101 military-related articles, or
is a somewhat hidden or missing variable in three per year, on average (Wilmoth and London
the literature on the life course and aging (Spiro forthcoming). Perhaps marking a turning point in
et al. 1997). In the lead article to a special issue of the eld, the nine articles published in the peak
Research on Aging on Military Service, the Life year2006include the introduction and six
Course, and Aging, which he guest-edited, research articles that were published in
Richard A. Settersten, Jr. (2006: 12) writes: Setterstens special issue of Research on Aging.
Most scholarship on aging is based on cohorts However, the vast majority of the military-related
born early in the 20th century, and these cohorts publications in these seven journals during this
have had signicant experience with war. time period were based on cross-sectional
Wartime experiences may therefore be critical research designs and non-representative samples,
but largely hidden variables underlying current and focused exclusively on veterans or Veterans
scientic knowledge about aging. We concur Administration health care facilities and pro-
with this claim, although we contend that mili- grams that serve older veterans. Although these
tary service, more generally, has been understud- studies address important, policy-relevant ques-
ied as an early-adulthood institutional inuence tions, the degree to which prior service in the
on the life course and aging (Wilmoth and armed forces affects variation in long-term life-
London 2013). course outcomes remains virtually unexamined
During the 34-year period from January 1980 because almost none of these studies directly
through December 2013, the ve Gerontology compares veterans to non-veterans over time in
well-specied models.
The sample of articles noted above clearly
A.S. London (*) J.M. Wilmoth does not constitute or represent the body of
Department of Sociology, Aging Studies Institute,
research published on military service and the
and Institute for Veterans and Military Families,
Syracuse University, Syracuse, NY, USA life course. In fact, there is a rapidly emerging
e-mail: anlondon@maxwell.syr.edu body of quantitative, longitudinal and qualitative

Springer International Publishing Switzerland 2016 277


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_12
278 A.S. London and J.M. Wilmoth

research on military service in lives that builds Our intent in this chapter is to layer, weave,
upon strong foundations laid over several decades and integrate each of these approaches, with a
(Wilmoth and London 2013; see also, Camacho focus on life-course stages as the primary scaf-
and Atwood 2007; Card 1983; Carlson and folding for organizing our arguments, reviews of
Andress 2009; Hogan 1981; London forthcom- evidence, methodological challenges, and gaps in
ing; London and Wilmoth 2008; MacLean and the literature, and discussions of directions for the
Elder 2007; Mettler 2005; Modell and Haggerty future. Unless otherwise stated, our focus is on
1991; Wilmoth and London forthcoming). Some the U.S. military. By adopting this approach, we
of the earliest foundations were laid by Glen can frame a focused set of issues that are germane
H. Elder, Jr. (Elder 1986, 1987). As he made sem- to a specic stage of the life course. Then, with
inal conceptual contributions to the development those issues in mind, we can review exemplary
of the life-course perspective, he and a group of conceptual and empirical research, pay attention
colleagues also elaborated a set of studies that to historical variations, draw attention to princi-
empirically demonstrated the importance of ples of the life course that are particularly salient,
examining military service in lives (Clipp and and describe processes that are related to entrance
Elder 1996; Dechter and Elder 2004; Elder 1986, into the military, military service experiences, and
1987; Elder and Bailey 1988; Elder and Clipp exit from the military, as well as the consequences
1988a, b, 1989; Elder et al. 1991, 1994, 1997, of military service for those who served and those
2010; Pavalko and Elder 1990). Elders inuence whose lives are linked to them.
extended beyond his own direct scholarly contri- Military service could have an effect on virtu-
butions to the eld to the training and support of ally everything life-course researchers study, but
scholars from various disciplines who have pur- data constraints limit what we know. We address
sued their own investigations of military service this issue most directly toward the end of the
and the life course (Gade and Ruise 2013). He chapter. Moreover, because it is a hidden or
remains a stalwart and engaged advocate for missing variable in most domains of investiga-
studying military service in lives (Elder 2013). tionwith notable exceptions in some, as the
One can imagine approaching a discussion of chapter aims to describethere are many ques-
military service and the life course in various tions that are relatively unexplored. Identifying
ways. One might think about the military life causal relationships remains a daunting challenge
course as encompassing three periodspre- in most domains of investigation. In this chapter,
service, active-duty, and veteranand focus on our review of such questions, knowledge, meth-
describing experiences within and across these odological challenges, and gaps in knowledge
stages exclusively among people who serve. One will by necessity be selective and relatively
might extend that framework to think about the broad; cited sources provide more detailed, criti-
ways that those who serve are similar to and dif- cal reviews of theory and evidence related to spe-
ferent from age-similar peers. One might focus cic domains of investigation, and point to gaps
on life-course stages, such as childhood and ado- in the literature. There is a need for additional
lescence, the transition to adulthood, midlife, theoretical delineation of the characteristics of
retirement, and old age, and consider how trajec- military service that matter and empirical investi-
tories through, and outcomes within, each stage gation of the mechanisms through which military
are associated with military service. One might service shapes the life course. Some of the best
focus on the principles of the life courselife- life-course research links theory and method-
long development, location in time and place, ologically rigorous empirical investigation to
timing and sequencing, human agency, and provide compelling evidence of the how and why
linked lives (Elder and Johnson 2002; Elder et al. military service matters. Such studies contribute
2003; Giele and Elder 1998; Settersten 2003; to our understanding of military service in lives,
Mortimer and Shanahan 2003)in relation to as well as the role of social institutions in shaping
military service. the life course more generally.
Military Service in Lives: Where Do We Go From Here? 279

2 Life-Course Perspectives activation of Reserve and National Guard


on Military Service components, inclusion of women, minorities, and
gays and lesbians) change over time. Because of
Each of the ve commonly recognized principles data constraints, one of the great challenges of
of the life-course perspectivelife-long develop- life-course research on military service is speci-
ment, location in time and place, timing and fying the ways in which service in particular
sequencing, human agency, and linked lives times and places affects specic outcomes.
(Elder and Johnson 2002; Elder et al. 2003; Giele The principle of human agency directs atten-
and Elder 1998; Settersten 2003; Mortimer and tion to the role that individuals play in their own
Shanahan 2003)is germane to understanding lives as they actively choose among various
the consequences of military service in lives. opportunities and deal with the constraints that
Human development and aging must be under- they encounter as they age in time and place.
stood as processes that unfold over time. What Because choices are structured by historical and
happens earlier conditions what happens in sub- social circumstances, as well as personality, indi-
sequent life stages; goal-setting for an imagined, viduals with apparently similar social locations
desired future might also do the same. The prin- may make very different choices. Although the
ciple of life-long development highlights the military is often considered a total institution
critical importance of having comprehensive data (Burland and Lundquist 2013) that strips a sub-
on military service experiences, as well as longi- stantial amount of agency from individuals while
tudinal data that enable researchers to address they are on active-duty and can produce identity
selection and disentangle aging from period and conict during the transition to civilian life
cohort effects. Change is fundamental to this (Smith and True 2014), individuals always act
principle of the life course, and can be seen in upon their circumstances. The choices service
studies of the consequences of military service members make, including the initial choice many
that draw on long-term longitudinal data. make to volunteer for service, re-enlistment deci-
The principle of location in time and place sions, seeking entrance into a military academy,
emphasizes how lives are contextualized by his- or choosing a military career, can have signicant
torical period and specic geographies. It points consequences over the life course for them and
to the need to consider cohort differences for those whose lives are linked to them. How
(Wilmoth and London forthcoming), as well as military service affects post-service choices and
where service happens and how service affects actions is also a critical issue. The relatively high
mobility and (re)settlement patterns (Bailey rate of suicide among veterans, as well as active-
2011, 2013). Active-duty military personnel duty personnel, underscores this point
move more frequently than civilians, and they (LeardMann et al. 2013).
move longer distances when they relocate. They The principle of timing and sequencing
are also disproportionately recruited from south- focuses on when events and transitions occur in
ern and rural parts of the country. The military is relation to age and other events and transitions. In
a complex system of inter-related places, and, the literature on military service and the life
like all places, these sites are socially organized course, this principle has received the most atten-
and have cultures that vary over time. Wars hap- tion in relation to the transition to adulthood
pen in particular theaters of operation, but are because the active-duty period is generally initi-
geographically dispersed such that not all service ated and ended in early adulthood. Questions
members who are active-duty during a war are related to the timing and sequencing of service in
exposed to combat. The historical period of ser- individual lives, and how military service
vice also matters, as the circumstances of service inuences the acquisition of marital, parenting,
(e.g., length of overseas duty, environmental educational, and occupational roles have received
exposures, ratio of wounded to died), and military the lions share of attention in the literature. The
policies (e.g., stringency of screening policies, consequences of transitions, including their
280 A.S. London and J.M. Wilmoth

disruptiveness to already established life-course active duty and received a better-than-


trajectories, depend in part on whether they are dishonorable discharge. While there is opportu-
expected or unexpected, as well as if they are nity for advancement in the armed forces,
considered on-time or off-time. Culturally especially among those pursuing military careers,
based, age-graded norms allow for some plan- each rank initially selects individuals with par-
ning and determine whether a transition is ticular characteristics and provides them with
deemed to be on-time or off-time. Off-time tran- access to different types of assignments, risks,
sitions do not coincide with the expected timeta- and rewards. Military occupational specialties
ble and are often, but not always, associated with vary by branch and rank, and provide different
negative consequences. kinds of education and training, which variably
The principle of linked lives focuses attention translate into the civilian labor market (Kleykamp
on the importance of social relationships and the 2013). Exposure to combat is not a prerequisite
interdependence of lives. The intergenerational for either active-duty or veteran status, or benets
transmission of military service within families is eligibility, although specic benets are available
one topic that has received considerable attention to veterans who served in specic wars or experi-
in the literature, with the U.S. Department of enced a service-connected disability (Wilmoth
Defense estimating that 57 % of active-duty and London 2011; Wilmoth et al. 2015a). The
troops serving in 2011 were the children of cur- availability of specic veterans benets varies
rent or former active-duty or Reserve service by historical period, duration of service, rank,
members (Human Resources Strategic and disability status, and the take-up of benets
Assessment Program 2011). Addressing how varies across the life course (Bennett and
participation in the military, whether for a short McDonald 2013; Street and Hoffman 2013;
time or a 20-year career, affects service mem- Wilmoth and London 2011). Family members
bers parents, spouses, and children is a critically may also receive benets as a result of a spouse
important issue that warrants substantially more or parents active-duty service or service-related
attention than it has received. death.
The military is a potentially transformative
institution that tends to engage young people,
3 Military Service in Lives: often for a relatively short period of time. For
A Conceptual Model most of U.S. history, less than 1 % of the popula-
tion served in the armed forces at any given time,
The military is a particular kind of institution that with the exception of brief periods when the
reects that social context in which it is orga- country was at war (see Figure 1 in Segal and
nized (Kelty and Segal 2013; Segal and Segal Segal 2004). Cumulatively, as of September
2004). In the United States, the armed forces 2014, it is estimated that there were almost 22
include the Army, Navy, Air Force, Marine million veterans living in the U.S., of whom
Corps, and Coast Guard. These branches are about 75 % served during wartime (National
organized hierarchically by rank, with a clearly Center for Veteran Statistics and Analysis 2014).
specied and enforced chain of command and Based on data from the 2009 to 2013 American
ne-grained distinctions drawn, respectively, Community Survey, veterans represent about 9 %
within the enlisted and ofcer ranks. Reserve and of the civilian population aged 18 and above
National Guard components are variably con- (U.S. Census Bureau 2015). Many more individ-
nected to, but distinct from, the various branches uals live their lives linked to those who have
of the armed forces. Individuals serving full-time served, which underscores the prevalence of mili-
in one of the ve branches of the armed forces are tary service experiences. Although the size of the
considered active-duty service members; veteran active-duty force has declined over the course of
status is conferred upon those who served on the late-twentieth and early-twenty-rst centuries
Military Service in Lives: Where Do We Go From Here? 281

(see Figure 1.1 in Wilmoth and London 2013), in military service in relation to other life course
2011, 1,411,424 Americans were on active duty transitions and events matters (Elder 1986, 1987;
in the armed forces (Department of Defense Hogan 1981). Moreover, as Teachman (2013:
2012). Women comprised 14.5 % of active-duty 282283) notes: the effects of military service
personnel. The largest share served in the Army may not be proportionate across the life course
(561,437), followed by the Air Force (328,821), military service may produce results that vary
Navy (320,141), Marine Corps (201,026), and according to stage in the life courseWhat may
Coast Guard (42,011). Overall, the ratio of appear to be a null effect of military service at
enlisted personnel to ofcers was 4.91. The one point in the life course may be very different
average ages of enlisted personnel and ofcers at earlier or later points in time. Overall, this
were 27.4 and 34.7 years, respectively. conceptual model of the life-course consequences
Serving in the military variably offers young of military service suggests that it is important to
adults an array of risks and benets, and the examine the potential effects of military service
effects of military service may be mixed and during different stages of the life course.
countervailing. For any particular life-course out-
come, the extent to which military service is ben-
ecial or has negative consequencesif it has 4 Childhood and Adolescence
any consequences at alldepends on the particu-
larities of historical circumstance, the prevailing Life-course and military researchers who address
organization and policies of the armed forces, childhood and adolescencethe pre-military
individual characteristics, service experiences, period of the life coursehave paid particular
and access to benets. We advocate using a attention to identifying early life-course factors
cumulative exposure model of the life course for that encourage or discourage voluntary military
studying military service in lives (see Figure 1.2 in service. This body of research on selection into
Wilmoth and London 2013). Specically, we military service addresses intergenerational
propose that exposure to the military in a particu- issues, as well as economic and non-economic
lar historical period by individuals with particu- inuences on entry into the military (Teachman
lar characteristics inuences processes of and Tedrow 2014). Other questions that are ger-
cumulative inequality that produce variation in mane to life-course researchers with interests in
outcomes over the life course. Early-life charac- children and youth pertain to the linked-lives
teristics and circumstances shape exposure to aspect of the life-course perspective. In part moti-
military service, and military service can medi- vated by the past decade of war and recognition
ate, moderate, or have no inuence on the asso- that an estimated two million children of active-
ciations between such factors and later-life duty personnel have variably been subjected to
outcomes. Additionally, participation in the mili- the exigencies of parental war-related deploy-
tary has the potential to directly (re)shape mid- ment, deployment-related stress, and combat-
and later-life educational, occupational, wealth related injury and death (Cozza et al. 2013),
accumulation, marriage/family, civic engage- researchers are increasingly investigating how
ment, and health trajectories and outcomes. exposure to the military early in life shapes child
Whether and how military service affects the life development. A recent issue of The Future of
course is individually, socially, and historically Children (Cozza and Lerner 2013) on Military
contingent, rather than automatic, and can best be Children and Families epitomizes this trend, and
evaluated by comparing those with military ser- serves as an excellent starting point for research-
vice histories to those without them, controlling ers with such interests. Whether, how, and why
for selection and taking into account the hetero- such childrens childhood, adolescent, and adult
geneity in military service experiences to the life-course trajectories and outcomes will differ
extent possible. The timing and sequencing of from those of children who have not had similar
282 A.S. London and J.M. Wilmoth

experiences is an important question for future given the substantial intergenerational inuence
research. Data collection must begin now and be on military service that has operated in the past
sustained over time in order to rigorously address (see Figure 3.16 in Sackett and Mavor 2003).
the questions life-course researchers will likely But, understanding selective inuences on ser-
ask about these children in the future. Finally, vice is also important for assessing the distribu-
Burland and Lundquist (2013) have recently tion of the risks that wartime service entails
argued that children in veteran families may also (Gimbel and Booth 1996), determining the racial
retain variable and complex relationships to the and socioeconomic status equity of draft-era con-
military that are largely unexplored in life-course scription, and obtaining adequate control of
research. selection in life-course research (Wolf et al.
2013). In the absence of near-universal participa-
tion or random assignment to military service,
4.1 Selection into the Military which is closely approximated in the context of a
draft lottery, accounting for voluntary selection
From 1941 to 1973, American men were subject into military service is especially challenging, yet
to a draft, and during World War II, the Korean critically important, for determining the extent to
War, and the Vietnam War, many men were in which observed associations represent causal
fact conscripted into service (Flynn 1993). For relationships (Wolf et al. 2013).
example, about 10 of the 16 million men who Demographic characteristics, such as age,
served during World War II were conscripted race/ethnicity, nativity status, gender, and sexual
(Segal and Segal 2004). It is worth noting, how- orientation have variably constrained entry into
ever, that even while the draft was active, many the military, as well as military service experi-
men and all women who served in the U.S. mili- ences and access to benets, in various historical
tary did so voluntarily (Card 1983; Segal and periods (Brown 2013; Campbell 2013; Kelty and
Segal 2004). With the end of the Vietnam-era Segal 2013; Lutz 2013). Historically, however,
draft and the advent of the All-Volunteer Force in rates of participation in the military over 50 %
1973, all individuals who have entered the mili- were observed among each single-year cohort of
tary have done so voluntarily. men born between 1915 and 1935, with the rate
In the context of a draft, or in national contexts topping 70 % among men in the 19191927 birth
where military service is a compulsory duty of cohorts (Hogan 1981). Such high levels of par-
citizenship (e.g., Denmark, Israel, Singapore), ticipation were achieved through a mix of con-
selective inuences on service and subsequent scription and voluntary enlistment. Interestingly,
life-course trajectories and outcomes are less in the context of the Vietnam-era draft, and per-
pronounced, although they may still be present haps the World War II-era draft as well, the dis-
depending on the stringency of screening pro- tinction between voluntary and compelled service
cesses. In such contexts, non-service may have blurred. As documented by Card (1983), the
both material and symbolic consequences. In extent to which men felt coerced into serving var-
contexts where service is voluntary, understand- ied among those who were drafted and those who
ing variation in who serves is important for help- were not, as well as by socioeconomic and human
ing guide military personnel policy, recruitment, capital characteristics. For example, approxi-
and planning (Bachman et al. 2000; National mately the same percentages of men with higher-
Research Council 2006; Sackett and Mavor and lower-than-average academic aptitude,
2003). In the U.S., ensuring an adequate number respectively, indicated that they were drafted but
of active-duty personnel has been a particular served willingly (21.6 % and 28.2 %) or were
policy concern in the era of the All-Volunteer drafted and served unwillingly (2.1 % and 2.0
Force. The rapid decline in the number of indi- %). However, a signicantly lower percentage of
viduals born to at least one veteranfrom 40 % men with higher-than-average academic aptitude
in 1970 to 8 % in 2000is a particular concern enlisted willingly (40.4 % versus 56.4 %), while a
Military Service in Lives: Where Do We Go From Here? 283

signicantly higher percentage enlisted unwill- Bill benets, as well as non-economic factors,
ingly (i.e., to avoid the draft) (25.9 % versus such as having a parent or other family member
8.2%). In the contemporary U.S. context, indi- who previously served in the military, living near
viduals can no longer be compelled to serve by a military base, being patriotic, having a thirst for
the state; however, they may feel compelled to adventure, or needing to belong. While there are
some degree by other factors, such as family many factors that shape entrance into the mili-
expectation, lack of alternate opportunity, or the tary, as is the case with disqualifying conditions,
need to escape abuse within the family. The the factors that lead an individual to volunteer in
extent to which these kinds of felt social pres- one historical period may be different from those
sures contextualize what on the surface are vol- that inuence voluntary entry in another.
untary enlistment decisions, the experience of Given that the factors that inuence entrance
serving in the military, or post-service outcomes into military service change over time, and selec-
has received little attention in the literature. tion into military service is an issue of high pol-
In the current All-Volunteer Force era, indi- icy relevance in the era of the All-Volunteer
viduals must volunteer for service. This consti- Force, there is a need for continued, expanded
tutes a rst-stage selection process. However, the data collection, theoretical development, and
desire to enter the armed forces voluntarily in this research into early-life inuences on military ser-
or any other era does not necessarily translate vice. For researchers, being able to account for
into entry because the military is a highly selec- early-life factors that inuence entry into the
tive institution that screens and excludes individ- military, combat assignments during service, and
uals deemed unqualied for service. The post-service outcomes of interest is critical for
characteristics that disqualify an individual from determining the causal effects of military
service vary across historical periods and con- service.
texts, such as periods of war and peace, but have In a recent article, Teachman and Tedrow
included physical and mental health problems, (2014: 4647) write that our understanding of
obesity, disability, low IQ, lack of a high school the factors that lead young men and women to
diploma or equivalent, heavy drug use, and hav- choose military service in the all-volunteer era
ing a felony conviction (National Research remains skeletal, and that there is no commonly
Council 2006). Just on the basis of the pre- accepted model for military enlistment. In this
induction physical exam, from 1950 to 1971, article, they focus on delinquency, and deftly
rates of rejection were always over 30 %, but theorize and empirically demonstrate a nonlinear
mostly at or over 40 %, and varied substantially relationship between delinquency and military
from year to year (Wolf et al. 2013). The life- service. Youth who demonstrate some propensity
course consequences of rejection from the mili- toward delinquent behavior are positively
tary for those who desired, and perhaps expected selected into the military because it provides a
or were expected, to serve is also a topic that has legitimate institutional context for the transition
received little attention in the literature. to adulthood. The military offers delinquent
A substantial body of research documents that youth a range of perceived benets, such as
a broad range of factors can inuence voluntary adventure, comradery, income, an exit from the
decisions about enlistment (Elder et al. 2010; circumstances that contribute to delinquent and
Fligstein 1980; Johnson and Kaplan 1991; criminal behavior, and an opportunity for a fresh
National Research Council 2006; Sackett and start. Some may be encouraged to enter the mili-
Mavor 2003; Teachman et al. 1993a, b). As dis- tary by family members, counselors, or social or
cussed by Teachman and Tedrow (2014), these juvenile justice workers, which may contribute to
include economic factors, such as relatively low a positive association between delinquency and
family socioeconomic status, constrained oppor- service; such potential inuences raise questions
tunity in the labor market, or the desire to obtain about the extent to which some delinquent youth
training or higher education through access to GI feel coerced into service. Their analysis of data
284 A.S. London and J.M. Wilmoth

from the 1997 National Longitudinal Survey of science/medicine, public service, literary/
Youth demonstrates that lower levels of delin- linguistic, social service, business management,
quency increased the likelihood of serving in the computation, and labor). Compared to those who
military, while higher levels of delinquency and did not serve, Card (1983) reported that those
criminality discouraged entrance, possibly who served were found to: be less calm; receive
because the military deemed such individuals to less guidance outside of high school; participate
be morally unt for service. less in extracurricular activities; and be less inter-
Due to cohort replacement, the veterans who ested in all of the vocations for which signicant
served during World War II and the Korean War differences emerged. Although men from upper-
are aging out of the population at present, while class backgrounds were signicantly less likely
those who served during the Vietnam War are to have served than men from other socioeco-
entering the retirement and post-retirement age nomic class backgrounds, and lower-class and
ranges (Wilmoth and London forthcoming). As Black men were more likely to serve in the Army
such, Vietnam-era veterans and their contempo- and Marine Corps where combat exposure was
raries will be of increasing interest to life-course more common, her ndings led Card (1983: 24)
scholars who study older adults (London forth- to conclude: The groups of veterans and non-
coming). Given this and the fact that few long- veterans are not as different from one another as
term studies that include Vietnam-era veterans popular belief might suggest. Not one of the
have included the prospective measurement of group averages lies more than 0.2 standard devia-
pre-service characteristics that might have inu- tions away from the overall class average. To
enced voluntary entrance into the military, it is some extent, she concluded, the draft helped level
instructive to note that some research from the the playing eld. However, it is likely that mili-
Vietnam era suggests that those who did serve tary screening and the differential rejection of
may have been more similar than dissimilar to both draftees and enlistees was also a factor. For
those who did not serve. example, with respect to race, countervailing
Josena Card (1983) drew on extensive, pre- forces may have worked to equalize participa-
service, baseline Project Talent data, which were tion. Blacks were more likely than whites to be
collected when the men she studied were 14 or 15 drafted because they were less able to obtain
years old, to identify factors that differentiated deferments, but they were four times more likely
those who served from those who did not. She to be rejected because they failed pre-induction
also distinguished service members who did not physical and mental health exams.
serve in the Vietnam theater of operations (i.e., While unobserved heterogeneity related to
non-Vietnam veterans) from those who did serve selection into military service, as well as mortal-
in Vietnam and were, therefore, potentially ity selection (Wolf et al. 2013), will remain issues
exposed to combat (i.e., Vietnam veterans). in life-course research that focuses on older
Comparing those who did not serve to non- Vietnam-era veterans, Cards (1983) results pro-
Vietnam veterans and Vietnam veterans, respec- vide some of the most comprehensive evidence
tively, she found no statistically signicant regarding what might and might not matter with
differences by: race; size of town in which the respect to selection into the military during the
individual grew up; religion; scores on eight dif- Vietnam War era. This should help guide inter-
ferent cognitive tests; self-ratings of study habits pretive efforts in future research.
and skills; education plans; high school course
and grades; 9 out of 10 personality characteristics
(except calmness); 7 out of 11 high school expe- 4.2 Children and Youth in Military
riences (except amount of guidance received Families
elsewhere, amount of extracurricular reading,
variety of extracurricular activities, and degree of Although questions related to selection into mili-
participation in extracurricular activities); and 10 tary service have received the lions share of
of 17 vocational interests (except biological attention in the literature related to the pre-service
Military Service in Lives: Where Do We Go From Here? 285

period of the life course, another issue that life are questions that lend themselves well to
warrants mention in relation to the contemporary future investigation. Similar questions should be
historical circumstances of children and youth, asked specically in relation to the large number
and future life-course studies, pertains to the of Reservists who were mobilized as part of the
experiences of military children. Increasingly, recent war efforts and their children. Reserve
the post-9/11 military is comprised of active-duty families are less accustomed to deployment.
service members with families; of the two mil- Some research suggests that all members of acti-
lion who have served in the active-duty military vated Reserve families experience boundary
since 9/11, nearly 45 % had children (Cozza et al. ambiguity, which is a state in which family
2013). In 2011, the ratio of spouses and children members are inconsistent in their reports of who
to active-duty service members was 1.41 (Clever is in or out of the family and who is performing
and Segal 2013). Overall, in 2011, there were 1.2 what roles within the family (Faber et al. 2008).
million children of active-duty personnel, and Such ambiguity, if prolonged, may have signi-
another 743,736 children in the families of cant consequences for children.
Reserve and National Guard members. It is beyond the scope of this chapter to pro-
While they may share some experiences with vide an extensive review of the ndings contained
so-called military brats from earlier eras, who in the recently published issue of The Future of
were fewer and often raised overseas during Children (Cozza and Lerner 2013) and related
peacetime (Ender 1996, 2002; Wertsch 1991), it literature. However, we believe it is important for
is likely that the experiences of this generation of life-course researchers with interests in children
military children will be different in many ways. and youth, and how experiences in childhood
As described in detail in a recent issue of The shape adult life-course trajectories and outcomes,
Future of Children on Military Children and to recognize these contemporary circumstances.
Families (Cozza and Lerner 2013), children in The research community needs to think carefully
military families face a particular set of circum- about the kinds of data that will be necessary to
stances that have the potential to impact develop- investigate whether and how military-related
mental trajectories and outcomes. Post-9/11 childhoods affect lives as they unfold over time
military service has been particularly consequen- (Chandra and London 2013; Teachman 2013). A
tial for service members, their spouses, and chil- failure to act now risks leaving the contemporary
dren because it has occurred during the wars in experiences of a large number of military chil-
Iraq and Afghanistan. Military parents and chil- dren hidden in future life-course research.
dren during this period faced lengthy, and often
multiple, deployments with concomitant
deployment-related separation and substantial 4.3 Children and Youth in Veteran
deployment-related stress. While military fami- Families
lies and children exhibit substantial resilience,
and military programs aim to assist families dur- A third and nal issue that is pertinent to life-
ing the deployment of a member, there is evi- course researchers with interests in the military
dence of negative consequences for children. and children and youth pertains to children in
Among military children, longer parental deploy- veteran families. As noted recently by Burland
ments were associated with increased problems and Lundquist (2013: 172): The literature on
in school, in the family, and with peers (Chandra military and family focuses almost exclusively
et al. 2010; Lester et al. 2010). Whether such on current families in the military, not on veteran
problems persist and represent seeds of longer- families, whereas the existing veteran life-course
term problematic trajectories and outcomes, rep- literature is focused mostly on individuals, not on
resent relatively transient difculties in families. Veteran families are those in which at
challenging circumstances, or generate stress- least one member served on active-duty in the
related growth that benets these children later in armed forces. Some, but not all children in
286 A.S. London and J.M. Wilmoth

veteran families were military children. 2013). While age at entry and exit from the
Sometimes, family formation begins after the military dene duration of service, they also
period of active-duty service ends, such that chil- mark transitions out of and into civilian adult-
dren of veterans never directly experience the hood. Prior to service or in its absence, civilian
military institution directly. However, they may status signies non-veteran; after service, civilian
experience it indirectly, as a result of the effects status signies veteran. Since the military is one
of military service on their parent or parents, on of the largest single employers of young men,
family life, and in relation to the material and increasingly young women, the period of
resources to which they have access. In the con- active-duty service coincides with, and to a cer-
text of providing a comprehensive review of the tain degree constitutes, the transition to adult-
literature on veteran families, Burland and hood for many Americans. Given extensive
Lundquist (2013) argue that the military often heterogeneity in military service experiences and
maintains complex ties to veteran families long outcomes, as well as the transformative potential
after the service members active-duty service of the military as a quasi-total institution, veteran
has ended. These ties may operate through GI status can signify an enormous range of personal
Bill benets for education or home loan guaran- circumstances in early adulthood. While condi-
tees. In the context of service-connected disabil- tioned by period-specic and dynamic economic,
ity or the death of a service member, the ties labor market, social, political, and familial
between the military and the veteran family may factors, mid- and later-life effects of military
actually be intensied as pensions are paid out service are emergent in or anchored to active-
over the long-term. The extent to which these duty service in early adulthood. For example,
complex and variable ties to the military continue combat-related injuries, accidents, or military
to shape the lives of children and youth whose sexual trauma that occur during the active-duty,
lives are linked to veterans are important ques- young-adult period of the life course point to
tions that to date remain understudied. direct mechanisms by which military service can
negatively impact young lives (MacLean 2013).
On the positive side, education and training in the
5 The Transition to Adulthood military can in some circumstances transfer and
provide a competitive edge in the civilian labor
Life-course scholars have demonstrated that the market (Kleykamp 2013).
pathways taken during the demographically How military service affects the transition to
dense period of young adulthood (Rindfuss 1991) adulthood, and thereby, subsequent life-course
can leave an indelible mark on the course of trajectories and outcomes, has been the subject of
human lives (Shanahan 2000). Military service is a substantial amount of theoretical consideration
a salient pathway to adulthood for many young and empirical investigation (Wilmoth and London
Americans (Bennett and McDonald 2013; Kelty 2013). Yet, there is a need for additional specica-
et al. 2010; Kelty and Segal 2013). Of the tion and testing of the mechanisms by which serv-
1,411,424 active-duty personnel in 2011, 13.3 % ing in the military service has effects on particular
were 25 years old or younger and another 22.5 % outcomes. Some of the most inuential research
were 2630 years old (Department of Defense on military service and the transition to adulthood
2012). Considering only enlisted personnel, 49.3 has focused on discontinuities in the life course
% were 25 years old or younger. turning points and life course disruption, and on
Most active-duty service begins and ends dur- how military service may benet individuals from
ing the stage of the life course often identied as disadvantaged backgrounds (Bennett and
the transition to adulthood. For the 1940, 1945, McDonald 2013; Browning et al. 1973; Elder
1950, 1955, 1960, and 1965 single-year birth 1986, 1987; Hogan 1981; Kelty et al. 2010;
cohorts, less than 5 % of men remained on active Sampson and Laub 1996). Other research focuses
duty by age 30 years (Figure 13.2 in Wolf et al. on the timing and sequencing of life-course
Military Service in Lives: Where Do We Go From Here? 287

transitions into marital, parenting, and occupa- an unprecedented opportunity to better their lives
tional roles in relation to military service. Recent though on-the-job training and further education
work takes up military service in relation to the (see also Laub and Sampson 2003). Mettler
emergent new transition to adulthood paradigm (2005) has made similar arguments.
(Berlin et al. 2010; Settersten and Ray 2010), Glen H. Elders (1986, 1987) early theorizing
which we believe is a useful framework for guid- about how military service could produce discon-
ing future research on the transition to adulthood. tinuities in the life course focused attention on
age at entry into military service, and the relation
of military service to psychosocial maturation
5.1 Military Service and the timing of other life-course transitions. By
and Discontinuities in the Life focusing on early versus late entry into the
Course military, he further elaborated the military-as-
turning-point hypothesis and a corollary life-
The potential for military service to produce dis- course-disruption hypothesis. Focusing on
continuities in the life course has been a long- service during World War II, Elder (1987) argued
term focus in the literature. Early theorizing that early entry constituted a social and psycho-
focused on the potential for the military to benet logical moratorium that delayed the transition to
individuals from disadvantaged backgrounds. adulthood. By delaying the assumption of mari-
One conceptual modelcommonly referenced tal, parenting, and civilian occupational roles,
as the military-as-turning-point hypothesis individual service members could maximize their
suggests that the military knifes off negative utilization of service benets and the resources
early-life inuences and provides a bridging made available by the institution. Early entrants
environment in which service members can were likely the most disadvantaged, who had
obtain resources that reset their life course in fewer alternatives and were, perhaps, proactively
benecial ways (Browning et al. 1973; Lopreato seeking a route out of difcult early-life circum-
and Poston 1977). Settersten and Ray (2010: 35) stances. Such individuals were most likely to
write that the military is: designed to shape the benet from the health and educational benets
futures of young adults by providing a setting in available to them through military service. In
which they can successfully live, work, and learn. contrast, relatively late age at entry into the mili-
By coupling expectations and demands with tary had the potential to disrupt established mari-
guidance, mentoring, and other resources, mili- tal, parenting, and occupational trajectories.
tary service helps young adults acquire skills and Later entrants likely came from more-advantaged
fosters a sense of competenceit also provides a socioeconomic backgrounds, and were more
bridge from school to higher education or the likely to have completed their educations, be
labor force by providing tuition credits, loan for- married, have children, and have launched their
giveness, nancial stipends, access to jobs, or careers prior to entering the military. As a result,
health insurance and other benets. While such military service later in the transition to adult-
benets and bridging functions are not auto- hood could disrupt established life-course pat-
matic, and military service during war time (and terns. Moreover, late entrants would likely have
to a lesser extent peace time) carries a substantial less motivation and opportunity to take advantage
risk of injury (MacLean 2013) that can produce a of the educational benets available to veterans.
negative turning point in the life course, there is Thus, through various mechanisms, the gains that
evidence that military service may be particularly would accrue to more-disadvantaged, earlier
advantageous for individuals from disadvantaged entrants might not materialize to the same degree
backgrounds. For example, Sampson and Laub or with the same effects in their lives, and they
(1996: 364) conclude: Military service in the might actually be disadvantaged due to service-
World War II era provided American men from connected injury or the disruption that serving
economically disadvantaged backgrounds with engendered in their family and work lives.
288 A.S. London and J.M. Wilmoth

Empirical examination of these theoretical entry was benecial; most of the evidence was
propositions yielded substantial support. Elder consistent with life course disruption and conti-
(1986) reported that those who entered military nuity of disadvantage interpretations.
service at younger ages did in fact come from Emerging research pays attention to other
more-disadvantaged family backgrounds, had mechanisms by which military service can pro-
poorer grades, and reported lower feelings of duce discontinuities in the early-adult life course.
self-adequacy. However, relative to non-veterans, Miech et al. (2013) theorized that the stringent
they were able to achieve equal occupational out- antidrug policies that the military successfully
comes, to have more stable marriages, and to instituted in the mid-1980s, and expanded and
experience larger gains in psychological strength rened subsequently, could substantially reduce
through mid-life. The positive turning point they illicit drug use, even among those who used drugs
experienced was in part due to their later transi- prior to entering the military. Knowledge of when
tion to adult roles and responsibilities. This inter- the policies were instituted and what they entailed
pretation is consistent with the beliefs of the allowed for the design of a natural experiment to
majority of early entrants, who were much more evaluate the impact of military service on illegal
likely than later entrants to report retrospectively drug use in young adulthood. These investigators
that their lives had followed a different and more conducted an age-period-cohort analysis of data
rewarding course as a result of their military ser- from the 1985, 1988, and 19902010 National
vice (Elder 1987). In addition to the corollary Survey of Drug Use and Health in order to exam-
evidence regarding late entry that is reported by ine the impact of military service on long-term
Elder (1986, 1987), other studies provide direct trends in hallucinogen use. Results were consis-
evidence in support of the life-course-disruption tent with the hypothesis that a mandatory anti-
hypothesis. For example, Elder et al. (1994) drug policy enforced within the context of a
report, based on data from the relatively advan- powerful total institution can lead to a life-long
taged men in the Stanford-Terman longitudinal reduction in hallucinogen use. Among birth
study who experienced World War II, that each cohorts who were young adults immediately
year that entrance into military service was before the implementation of the antidrug poli-
delayed reduced the economic and job benets cies, the odds of past-year hallucinogen use were
associated with military service, and increased twice as high for veterans than for nonveterans
the risk of life-course disruption and related over the life course. This difference disappeared
costs. Moreover, partly as a result of the work- among birth cohorts that passed through young
life disadvantages they experienced, late- adulthood after the antidrug policies were imple-
mobilized men were at greatest risk of negative mented even though veterans had signicantly
physical health trajectories over the life course. higher overall rates of illegal drug use in adoles-
London and Wilmoth (2006) also examined the cence. Further analysis indicated that, after the
turning-point and life-course-disruption hypoth- drug-testing policies were implemented, veterans
eses in their study of the inuence of race and had signicantly lower prevalence of past-year
early-life socioeconomic status disadvantage on hallucinogen use than nonveterans among the
mens later-life mortality. Based on analyses of subgroup who reported a history of illegal drug
data from men in the Health and Retirement use before age 18 years. These trends were not
Study (HRS) and the Study of Assets and Health explained by trends in recruits tendencies to use
Dynamics among the Oldest-Old (AHEAD), illegal drugs prior to entry into the military,
which include a nationally-representative sample although it is conceivable that knowledge of the
of cohorts who experienced World War II, the militarys drug policy dampened pre-service drug
Korean War, and the Vietnam War, they found use among those who aspire to enter the military.
little evidence that military service or early age at Overall, the results of this study point to a specic
Military Service in Lives: Where Do We Go From Here? 289

mechanism by which service in the armed forces and/or began their rst full-time jobs while
produced a turning point that led to a lasting, life- enrolled in school. To the extent that the veterans
long behavior change. attending college under GI Bill benets were
from poor socioeconomic backgrounds, interco-
hort changes in transition behavior was assured.
5.2 Timing and Sequencing Hogan (1981) also discusses how military ser-
in the Transition to Adulthood vice affected marriage patterns. Faced with sud-
den, sometimes unexpected, induction into the
The timing and sequencing of life-course transi- armed services during the draft, some men chose
tions and events is a central concern in life-course to marry at once, while others postponed mar-
studies generally (Elder and Johnson 2002; Elder riage until after they completed their wartime ser-
et al. 2003; Giele and Elder 1998; Settersten vice. Unmarried men who took up GI Bill
2003; Shanahan 2000). Concerns about timing education benets after their service was com-
and sequencing are embedded within the military- pleted would have to wait a long time to marry if
as-turning-point and life-course-disruption they waited until after they completed school and
hypotheses that were discussed in the previous found a rst full-time civilian job. Thus, many
section. Questions related to how military service veterans married on-time, but prior to complet-
affects the timing and sequencing of transitions ing school or beginning work. Hogan (1981: 214)
and events in the early-adult life course was a notes: For most men, therefore, military service
central focus of Hogans (1981) book Transitions represented an obstacle to the smooth completion
and Social Change: The Early Lives of American of the transition from adolescent to adulthood.
Men. In this book, he used data from the In addition to Hogans (1981) early, seminal
Occupational Changes in a Generation II survey, work on the topic of how military service shaped
which was conducted in conjunction with the inter-cohort variation in timing and sequencing in
March 1973 demographic supplement to the the transition to adulthood, other studies have
Current Population Survey. Men from early- to investigated the effects of military service on the
mid-twentieth century birth cohorts, between the timing and sequencing of role acquisition in
ages of 20 and 65 years were included. These are early-adulthood. Much, but not all, of this work
cohorts in which participation in the military was focuses on marriage and childbearing. For exam-
particularly high. ple, Card (1983) found that Vietnam-era veterans
Hogan (1981) documented various effects of nished school, got married, and became fathers
military service on timing and sequencing. For at signicantly older ages than their non-veteran
example, with respect to the timing of school classmates. Call and Teachman (1996) examined
completion in relation to military service, he marital stability in relation to the timing of rst
notes that men from lower socioeconomic status marriage and Vietnam-era military service (i.e.,
backgrounds are less likely to attend college, and, before, during, or after service). They found that
if they complete college, they take longer and Vietnam combat and non-combat veterans mar-
more variable amounts of time to do so. The GI ried at the same rate as non-veterans, and mar-
Bill allowed many veterans, and some veterans riages contracted before or during service were
from lower socioeconomic backgrounds, to delay not negatively impacted with respect to marital
entry into the labor force and attend college stability. However, marriages contracted after
(Bound and Turner 2002; Stanley 2003). This service experienced increased stability. Focusing
accentuated inter-cohort variation. As Hogan on those serving during the All-Volunteer Force
(1981: 212) writes in relation to the availability era, Lundquist (2004) reported that marriage was
of GI Bill education benets: It created a group more common and occurred earlier in the lives of
of military-experienced birth cohorts among military personnel than in the lives of civilians,
whom schooling was completed relatively late, at while Lundquist and Smith (2005) reported that
widely varying ages, and many persons married military personnel also started childbearing
290 A.S. London and J.M. Wilmoth

earlier. Teachman (2007a) examined the effect of (2010) argue that military service during the era
military service and race on the timing of a rst of the All-Volunteer Force is one of many path-
marriage, and found that military service ways to adulthood, rather than a delay or a detour
increased the likelihood of marriage for both in the process of transitioning to adulthood (see
whites and blacks. After taking selectivity, also, Kelty and Segal 2013; Bennett and
income, and economic stability into account, McDonald 2013). Kelty et al. (2010: 181) further
observed effects were particularly strong for argue that serving in the military results in transi-
black men. Usdansky et al. (2009) used data from tions to adulthood that are more stable and
the Fragile Families and Child Well-Being Study orderly than is the case among those who have
and found that military service prior to a non- not served in the military. While it is beyond the
marital birth signicantly increased the odds of scope of this chapter to delve substantially into
marriage within 5 years for blacks only. the emergent evidence regarding the changes in
Although there is some research that addresses the transition to adulthood that constitute the new
timing and sequencing issues in relation to mili- paradigm, we believe that the new transition to
tary service, the research in this area is thin and adulthood paradigm frames a set of important
many gaps in the research remain. As Burland issues that can and should be taken up in relation
and Lundquist (2013: 173174) note in relation to military service. For example, building on
to the life-course-disruption hypothesis and the Hogans (1981) work, it would be interesting to
timing of family formation: By implication, it know whether and for whom the wars in Iraq and
may also be that families created after service are Afghanistan, in conjunction with the Great
more positively beneted than families that form Recession of 2008, have complicated the stability
before or during service. This family formation and orderliness of transitions to adulthood among
timing comparison has scarcely been studied and recent cohorts of veterans.
is something that merits further research.
Beyond marriage and childbearing, many other
questions about the timing and sequencing of 6 Middle Age and Later Life
transitions and events in the young-adult life
course warrant additional attention. Consistent Life-course researchers often focus on the mid-
with the life-course principle of life-long devel- and later-life consequences of military service
opment and the cumulative inequality perspec- via the lens of veteran status since most individu-
tive, the patterning of events in early-adulthood is als who serve have rejoined the civilian popula-
likely to be an important determinant of what tion by age 30 (Wolf et al. 2013) and sampling
comes next. frames for household-based surveys often
exclude active-duty military personnel. The con-
sequences of early-adult military service on vet-
5.3 The New Transition erans mid- and later-life outcomes may be
to Adulthood positive or negative, and transient, long-lasting,
or permanent. Not all will be apparent in young
The new transition to adulthood generally adulthood; some may emerge in later life for the
entails a lengthening of the time it takes for youth rst time or as a recurrence. Although a substan-
to leave home, complete school, enter the work- tial body of research on the consequences of mili-
force, marry, and have children, as well as sub- tary service exists (Wilmoth and London 2013),
stantial variability in the sequencing of the determining the causal inuence of military ser-
various transitions that occur during this vice on a range mid- and later-life outcomes has
demographically-dense period of the life course been hampered by: the lack of long-term,
(Berlin et al. 2010; Settersten and Ray 2010). In comparative, longitudinal data; inadequate con-
their contribution to an issue of The Future of trol of selection into military service; and insuf-
Children on Transition to Adulthood, Kelty et al. cient measurement of the timing of military
Military Service in Lives: Where Do We Go From Here? 291

service in relation to other life course events. The successfully to a narrow range of birth cohorts
absence of data on the heterogeneity of military and selected outcomesprimarily education and
service experiences is also a critical issue. Some earnings. Although combat-exposed veterans
long-term studies are based on a relatively nar- demonstrate considerable resiliency over the
row range of birth cohorts, which contributes to course of their lives (Aldwin et al. 1994; Ardelt
uncertainty about the generalizability of ndings. et al. 2010; Elder and Clipp 1989), combat expo-
Other studies, such as those that use the HRS, are sure has been a particular focus of attention
only able to observe individuals who survive to because it signies the sacrice soldiers make
midlife, which raises a concern about mortality and it clearly reects an observable causal effect
selection (i.e., left censoring) (Wolf et al. 2013). of early-adulthood military service. Combat
Despite these limitations, extant research pro- exposure can have life-long consequences for
vides some insight into the long-term, potentially disability (Elder et al. 1997; MacLean 2010,
causal effects of military service on lives. 2013), physical and mental health problems
Contributors to a recent edited volume entitled (Elder et al. 2009; Vogt et al. 2004), inadequate
Life-Course Perspectives on Military Service sleep (London et al. 2014), and mortality (Bedard
(Wilmoth and London 2013) provide comprehen- and Deschnes 2006; Dobkin and Shabani 2009;
sive reviews of what we know about the mid- and Elder et al. 2009), and it may also affect many
later-life consequences of military service for other health-related processes and outcomes that
labor market outcomes among veterans and mili- unfold over the life course.
tary spouses (Kleykamp 2013); family relation- For example, in her study Lives after Vietnam,
ships (Burland and Lundquist 2013); spatial Card (1983) collected data that allowed her to
mobility (Bailey 2013); health (MacLean 2013); describe the extensive combat-related experi-
and later-life health and nancial security (Street ences of Vietnam veterans (i.e., those who actu-
and Hoffman 2013). A review of the ndings ally served in Vietnam and were potentially
from those chapters is beyond the scope of this exposed to combat), and how those varied by
chapter. Here, we focus critical attention on a branch, race, socioeconomic background, and
range of issues that are germane to understanding academic aptitude. She was also able to draw on
the long-term consequences of military service, the extensive pre-service baseline data collected
and provide selected, illustrative examples from by Project Talent in order to control admirably
the literature that point to potential directions for for selection into the military. Among Vietnam
future research. veterans, those who served in the Army and
Marine Corps were signicantly more likely to
have combat experiences than those who served
6.1 Identifying Causal Effects in the Navy, Air Force, and Coast Guard.
Specically, they were more likely to: receive re
In the literature on military service and the life from the enemy (91.9 % versus 65.4 %); re own
course, identifying causal effects has been a per- weapon at enemy (68.8 % versus 34.7 %); kill
sistent challenge. Some researchers have used enemy (45.7 % versus 20.8 %); see someone get
experimental audit studies to examine the impact killed (68.9 % versus 47.9 %); see enemy
of veteran status on employment (Kleykamp wounded (70.2 % versus 37.9 %); see American
2009, 2013) and experimental vignette studies to wounded (88.6 % versus 67.1 %); see enemy
examine the inuence of veteran status on public dead (77.3 % versus 38.9 %); see American dead
attitudes (MacLean and Kleykamp 2014). While (77.6 % versus 59.9 %); and be in a combat situ-
the World War II- and Vietnam-era draft lotteries, ation where survival was in jeopardy (74.0 % ver-
and other variables, have been used to instrument sus 45.3 %). Non-whites and men with
military service (Angrist 1990, 1993; Angrist and lower-than-average socioeconomic backgrounds
Chen 2011; Angrist and Krueger 1994), instru- were more likely to serve in the Army and Marine
mental variable methods have only been applied Corps than the other branches, although there
292 A.S. London and J.M. Wilmoth

were no signicant race or socioeconomic status outcomes dissipate, as suggested above, while at
differences in the specic combat experiences other times they emerge in specic contexts or at
that were reported by Vietnam veterans. However, particular points in the life course. In some cir-
Vietnam veterans with low academic aptitude cumstances, military service has positive effects
were more likely than those with high academic that emerge and dissipate over time.
aptitude to see American wounded, see enemy For example, Whyman et al. (2011) recently
and American dead, and be in a combat situation focused on peacetime service in order to docu-
where survival was in jeopardy. ment whether non-war military service may have
Partly as a result of these experiences, veter- benecial effects on health and well-being
ans at the rst follow-up had signicantly more through midlife. They used data from the National
symptoms of post-traumatic stress disorder Longitudinal Survey of Youth and reported that
(PTSD)nightmares, loss of control over behav- men who served on active duty but did not see
ior, emotional numbing, withdrawal, hyper- combat were less likely to experience depressive
alertness, anxiety, depressionthan non-Vietnam symptoms than non-veterans and men who served
veterans (i.e., those who served but not in the in the Reserves. To help establish a causal rela-
Vietnam theater of operations) and non-veterans. tionship, the Reserves were chosen as the com-
They also had higher conviction rates for misde- parison group because they share many
meanors and felonies, and lower life satisfaction unobserved characteristics with veterans, but
in general. However, demonstrating change over have not served on active duty. Results also indi-
time and life-long development, at age 36, their cated that this benecial effect of non-combat
ratings of various aspects of their lives military service dissipated after discharge. Their
relationship with wife or girlfriend, having and analysis provides an important counterpoint to
raising children, occupation or job, nancial the prevailing emphasis in the social science lit-
security and material well-being, health, relation- erature and media on the negative mental health
ship with other relatives, relationship with consequences of military service, and also sug-
friends, intellectual development, personal gests that the mental health consequences of mil-
understanding and awareness, socializing and itary service can change over the life course.
entertaining, and recreational activitieswere Even if negative, positive, or neutral conse-
consistently higher than similar ratings taken just quences of military service are evident at one
7 years prior. While this was true of their class- point in the life course, it does not necessarily
mates ratings too, the increases reported by the mean that the same will be the case at other points
Vietnam veterans tended to be greater than those in the life course. For example, one of the con-
reported by their classmates (Card 1983: 149). tributors to the special issue of Research on Aging
Whether and how these mens lives continued to on Military Service, Aging, and the Life Course
change through mid- and later life is unknown as that Richard A. Settersten, Jr. edited focused on
there has not been a broad-scale long-term fol- late-onset stress symptomatology among aging
low-up of the Project Talent sample to date. combat veterans (Davison et al. 2006). London
et al. (2012) used data from the HRS to estimate
growth curve models of mens depressive symp-
6.2 Dynamic Outcomes tom trajectories in later life by veteran status. Net
of a broad range of other factors, they found that
The negative and positive outcomes of military veterans had signicantly fewer depressive
service are dynamic and can change over the life symptoms than non-veterans around retirement
course. However, the long-term data sets that age. Moreover, they found evidence of a positive
life-course researchers have available for analy- turning point; military service offset the effect of
sis often do not contain an adequate number of early-life socioeconomic disadvantage on mens
points of observation to track trajectories of sta- later-life depressive symptoms. Only non-
bility and change with age. Sometimes, negative veterans from socioeconomically disadvantaged
Military Service in Lives: Where Do We Go From Here? 293

backgrounds had elevated depressive symptoms environment to the extent that it encourages
in later life. The depressive symptoms trajecto- physical activity during and after the period of
ries of the other three groupsadvantaged non- active duty, but is a risk environment to the extent
veterans and veterans, respectively, and that it encourages eating habits that can lead to
disadvantaged veteranswere signicantly weight gain. Results from well-specied models
lower and remarkably similar around the mean support the conclusion that military service is
age (66.2 years), although all groups converged associated with higher BMI.
at the oldest ages. Recent research using data from the HRS
examined inter-cohort variation in mens later-
life BMI trajectories and came to a similar con-
6.3 Inter- and Intra-Cohort clusion (Wilmoth et al. forthcoming). Without
Variation any controls in the models, veterans exhibit lower
BMI, on average, than non-veterans. However,
Secondary data analysts depend on large-scale with controls for birth cohort, early-life charac-
data collection efforts to support their research teristics that occur prior to military service,
agendas. Generally, they take what they can get potentially mediating mid- to late-life character-
in terms of measured outcomes and the birth istics, and methodological controls for proxy
cohorts that are represented in the data sets they report, attrition, and death during the study
use. Researchers who are interested in studying period, the effect of veteran status was found to
the same processes among different cohorts to be small, marginally signicant, and positive
determine if the effect of military service has (i.e., veterans were heavier than non-veterans).
changed must try to align different data sets, Overall, the results demonstrated the large effect
which introduces an element of uncertainty of the secular trend in increased weight across the
rooted in indirect comparisons. Nevertheless, the populationyounger cohorts were substantially
admirable research program elaborated by Jay heavier than older cohorts regardless of their vet-
Teachman and Lucky Tedrow over the last decade eran statusand a substantively small but con-
provides important insights into inuences of sistent, positive intra-cohort effect of veteran
military service on education, income, and mar- status.
riage/family outcomes across cohorts (Teachman There also appears to be intra-cohort variation
2004, 2005, 2007a, b; Teachman and Tedrow in mid- to later-life health on the basis of veteran
2004, 2007, 2008). status. Wilmoth et al. (2010) used data from the
Some data sets do allow for direct inter- and 19922006 waves of the HRS to estimate growth
intra-cohort comparison of dynamic outcomes. curve models that examined differences in health
For example, two recent studies that focus on trajectories between non-veterans and veterans,
military service in relation to body mass index veterans with and without wartime service, and
(BMI) are illustrative. In the era of the All- war service veterans who served during World
Volunteer Force, Teachman and Tedrow (2013) War II, Korea, Vietnam, and multiple wars. They
argue that eating and exercise patterns are more focused on mens health conditions, activities of
in balance during the active-duty period than they daily living limitations, and self-rated health.
are during the more-sedentary veteran period. Veterans had better health at the mean age of 66.2
Eating habits established while in the military, years, but experienced greater age-related
when physical activity and caloric demands are changes in health than non-veterans. Similarly,
relatively high, carryover into civilian life, but men who served during wartime had better health
activity levels and exercise habits do not carry at the mean age, but more age-related changes in
over to the same extent. As a result, during the health than men who did not serve during
transition to civilian life, veterans gain weight wartime. Among war veterans, Vietnam veterans
and then carry that weight with them as they age. were in poorer health than veterans of World War
From this perspective, the military is a protective II, the Korean War, and multiple wars at the mean
294 A.S. London and J.M. Wilmoth

age, but they experienced less substantial age- women. Drawing on the life-course perspective,
related health changes than men who served dur- they theorized that the likelihood of engaging in
ing those earlier wars. extramarital relationships might be increased by:
factors that select individuals into military ser-
vice; the opportunities that deployment-related
6.4 Timing and Sequencing separations present; the male-dominated culture
in the Life Course of the military; the sex industries that thrive
around overseas bases and in the locations where
Data on timing and sequencing of life course deployed service members take rest and relax-
events with respect to military service are often ation; and, possibly, the post-service careers cho-
not collected systematically, which makes it sen by veterans, if they disproportionately
impossible to isolate the long-term effects of mil- increase business-related travel and separation
itary service in relation to other events during from a spouse. Using data from the National
young adulthood. When such data on timing and Health and Social Life Study (NHSLS), they
sequencing are available, they can help explain found that 32 % of veterans had engaged in an
the sources of variation in observed mid- and extramarital relationship, which was about twice
later-life outcomes. For example, with respect to the rate of non-veterans. Multivariate analyses
timing in the life course, the relatively earlier age indicated that veterans, particularly male veter-
at marriage and childbearing among veterans ans, were signicantly more likely than non-
may increase the odds of divorce and shape pat- veterans to ever engage in an extramarital
terns of remarriage over the life course (Adler- relationship, and that both veteran status and hav-
Baeder et al. 2006; Lundquist and Xu 2014). ing had an extramarital relationship increased the
With respect to sequencing, Hogan (1981) found odds of ever experiencing divorce.
that veterans of World War II, the Korean War, Using the same pre-service, active-duty, post-
and the Cold War did not experience the mid- service life-course theoretical framework, in a
career earnings decits that are usually associ- subsequent study, London and Wilmoth (2015)
ated with marrying before school completion. used data from three national surveys to investi-
However, men who served in the armed forces gate variation in mens participation in commer-
during the Vietnam War and completed college cial sexual relationships by veteran status. Taken
after marriage did experience the socioeconomic together, the samples included men who turned
attainment decits that are usually associated 18 years old between 1922 and 2010 and thus
with a non-normative transition to adulthood. served in the military from the early part of the
These mid-life socioeconomic attainment out- twentieth century through the early part of the
comes are important because they lay the founda- twenty-rst century. They found that male veter-
tion for nancial security in later-life and ans were signicantly more likely than non-
retirement (Street and Hoffman 2013), while also veterans to report ever paying for sex, with rates
shaping the capacity for intergenerational trans- across the three sub-studies ranging from 10.9 to
fers to children and, sometimes, aging parents. 14.6 % among non-veterans and from 25.3 to
In order to break new ground with an under- 33.9 % among veterans. In multivariate models
analyzed outcome and establish foundations for that controlled for demographic and early-life
future investigation, life-course researchers factors to the extent possible with available data,
sometimes proceed with studies that only allow the odds of ever paying for sex were 2.253.10
for the theoretical description of the processes times higher among veterans than among non-
before, during, or after military service that may veterans. In a supplemental analysis using data
produce an observable association. For example, from the General Social Survey (GSS), longer
London et al. (2013) examined the association duration of service was associated with an
between veteran status and ever having an extra- increased odds of ever paying for sex. While the
marital relationship among ever-married men and data they had available did not allow them to
Military Service in Lives: Where Do We Go From Here? 295

determine the timing of the commercial sexual connected disability has important implications
relations in the mens lives or in relation to mili- for injured service members lives, but also for
tary service, or demonstrate a causal relationship the individuals whose lives are linked to them.
between serving in the military and ever paying Sometimes, service-connected injuries generate
for sex, the strength and consistency of the nd- short-terms demands for family care; at other
ings provided compelling evidence of an associa- times, caregiving demands may extend over the
tion that they deemed worthy of further theorizing long-term, with substantial consequences for
and empirical investigation. caregivers careers and well-being. It is notewor-
thy that recent evidence indicates that households
that include disabled service members are at sig-
6.5 Linked Lives nicantly greater risk of poverty and material
hardship, which affect all household members
Adequate data to examine how military service (Hein et al. 2012; London et al. 2011; Wilmoth
affects the lives of those who are linked to veter- et al. 2015b).
ans as they agegrandparents, parents, siblings,
spouses, and childrenare generally not avail-
able in single-respondent surveys, and are under- 7 Conclusion
analyzed when they are available. The literature
on military service has paid some attention to In their Annual Review of Sociology essay from
spouses outcomes. For example, military over two decades ago, Modell and Haggerty
spouses are often conceptualized as tied (1991: 205) wrote that studies of military service
migrants because active-duty service entails and the life course aim to connect the micro-
substantial mobilityrelocations every 34 and macro-levels of analysis, thus connecting the
years (Bailey 2011, 2013). Frequent moves soldiers story to that of his [or her] changing
decrease military spouses attachment to the society. In setting the agenda for future life-
labor force (Kleykamp 2013), which can reduce course studies of military service, Teachman
opportunity costs for childbearing. The increased (2013: 275) recently wrote: Research on mili-
mobility-decreased labor force attachment- tary service has a long and rich history. Indeed,
increased fertility nexus among military wives the roots of quantitative Sociology can be traced
has been examined to some extent in the litera- to research sponsored by the military during
ture (Gill et al. 1994; Gill and Haurin 1998). This World War IIYet despite this research heritage,
is arguably a causal sequence that is driven by the fty years after the publication of The American
tempo of service-related moves to which service Soldier, we know surprisingly little about the
members and their families are subjected. The ways military service is linked to the lives of the
military places substantial demands on woman as men and women who have served their country.
wives and mothers, particularly ofcers wives In our view, military service is an emerging
(Harrell 2000, 2001; Lundquist and Xu 2014). variable in life-course studies. A small group of
Given the very large cohorts of older men who life-course researchers have pursued programs of
served in the military, most of whom were mar- research over many decades, and several younger
ried, how linkage to men who serve affected mili- scholars, many of whom are referenced in this
tary spouses lives and womens aging is a topic chapter, are poised to do so. Others have made
that warrants considerably more attention among singular, or less sustained but none-the-less
life-course scholars than it has received. To date, important, contributions to the eld. There is
few studies consider the long-term consequences much to be gained by continued investigation.
for spouses later-life outcomes. The recent, longest-ever, post-9/11 wars in Iraq
Serving in the military is risky, and many indi- and Afghanistan have generated a groundswell of
viduals are harmed as a result of serving their interest and attention in the research community,
country (MacLean 2010, 2013). Service- the media, and the public. Many special issues of
296 A.S. London and J.M. Wilmoth

journals that focus on military service in some susceptibilities; and international, comparative
way that is relevant for life-course researchers research questions that focus on how different
have recently been published or are forthcoming state policies structure the life course in relation
(e.g., The Future of Children, Substance Use & to military service.
Misuse, The Gerontologist, The Journal of The time to tackle these methodological chal-
Gerontological Social Work). There is renewed lenges and pursue these substantive topics is now,
energy in the eld, and much to be gained by the as aging veteran cohorts from World War II and
careful theoretical specication and empirical the Korean War, who are currently over the age of
testing of the mechanisms by which participation 75, are replaced by the retirement-aged Vietnam
in the military produces effects on the life course. War cohort, the middle-aged Gulf War cohort,
In order to maximize the potential for study- and the young-adult aged post-9/11 cohort that
ing the military service in lives, new and enhanced served in Iraq and Afghanistan. In some respects,
data collection will likely be necessary (Chandra we have missed the opportunity to fully elucidate
and London 2013; Teachman 2013). It is critical the mechanisms by which military service affects
that these data collections take into account the lives based on the experiences of the cohorts who
previously mentioned methodological challenges served in the military during the middle of the
that hamper our current understanding of the twentieth century. New data collection on these
long-term consequences of military service. In cohorts could still yield important insights if it
particular, we need nationally representative data carefully attends to the previously discussed
that: measures selection into military service and methodological issues. But, the most promising
specic military service experiences in order to research is on the All-Volunteer Force-era cohorts
identify the causal aspects of service that initiate because they can be prospectively tracked
enduring life-course effects; considers a range of through middle age and later-life. Building on
positive and negative life course outcomes to existing studies and pursuing new data collection
reveal the heterogeneous effects of military ser- with these younger cohorts can lay a foundation
vice within and between cohorts; collects data at that will lead to better understanding of the long-
regular intervals throughout mid- to late-life to term impact of military service on lives.
track the trajectory of emerging and dissipating
effects; contains information on the timing of
events to clarify how military service relates to References
other life course events like completion of educa-
tion, union formation and dissolution, childbear- Adler-Baeder, F., Pittman, J. F., & Taylor, L. (2006). The
prevalence of marital transitions in military families.
ing, and emergence of health conditions; and
Journal of Divorce and Remarriage, 44(12), 91106.
links the data of primary respondents to signi- Aldwin, C. M., Levenson, M. R., & Spiro, A. (1994).
cant others, such as parents, spouses, children, Vulnerability and resilience to combat exposure: Can
and grandchildren, to understand how the effects stress have lifelong effects? Psychology and Aging, 9(1),
3444.
of military service reverberate through lives.
Angrist, J. D. (1990). Lifetime earnings and the Vietnam
Addressing these methodological concerns will era draft lotteryEvidence from Social Security
enable researchers to understand the enduring administrative records. American Economic Review,
effects of military service and to pursue core and 80(3), 313336.
Angrist, J. D. (1993). The effect of veterans benets on
emerging substantive topics within life-course
education and earnings. Industrial & Labor Relations
studies, such as many that have been mentioned Review, 46(4), 637652.
throughout this chapter. Future research might Angrist, J. D., & Chen, S. H. (2011). Schooling and the
also examine variation on the basis of gender, Vietnam-era GI Bill: Evidence from the draft lottery.
American Economic Journal: Applied Economics, 3,
race/ethnicity, and sexual identity; gene-
96118.
environment interactions and the extent to which Angrist, J. D., & Krueger, A. B. (1994). Why do World
military service is a risky or protective early-life War II veterans earn more than nonveterans. Journal
environment for individuals with specic genetic of Labor Economics, 12(1), 7497.
Military Service in Lives: Where Do We Go From Here? 297

Ardelt, M., Landes, S. D., & Vaillant, G. E. (2010). The Card, J. J. (1983). Lives after Vietnam: The personal
long-term effects of World War II combat exposure on impact of military service. Lexington: Lexington
later life well-being moderated by generativity. Books.
Research in Human Development, 7(3), 202220. Carlson, E., & Andress, J. (2009). Military service by
Bachman, J. G., Segal, D. R., Freedman-Doan, P., & twentieth-century generations of American men.
OMalley, P. M. (2000). Who chooses military ser- Armed Forces & Society, 35(2), 385400.
vice? Correlates of propensity and enlistment in the Chandra, A., & London, A. S. (2013). Unlocking insights
US armed forces. Military Psychology, 12(1), 130. about military children and families. The Future of
Bailey, A. K. (2011). Race, place and veteran status: Children, 23(2), 187198.
Migration among black and white men, 19402000. Chandra, A., Lara-Cinisomo, S., Jaycox, L. H., Tanielian,
Population Research and Policy Review, 30(5), T., Han, B., Burns, R. M., & Ruder, T. (2010). Views
701728. from the homefront: The experiences of youth and
Bailey, A. K. (2013). Military employment and spatial spouses from military families. Santa Monica: RAND
mobility across the life course. In J. M. Wilmoth & Corporation. http://www.rand.org/content/dam/rand/
A. S. London (Eds.), Life-course perspectives on mili- pubs/technical_reports/2011/RAND_TR913.pdf
tary service (pp. 185199). New York: Routledge. Clever, M., & Segal, D. R. (2013). The demographics of
Bedard, K., & Deschnes, O. (2006). The long-term military children and families. The Future of Children,
impact of military service on health: Evidence from 23(2), 1339.
World War II and Korean War veterans. American Clipp, E. C., & Elder, G. H., Jr. (1996). The aging veteran
Economic Review, 96(1), 176194. of World War II: Psychiatric and life course insights.
Bennett, P. R., & McDonald, K. B. (2013). Military ser- In P. E. Rusking & J. A. Talbott (Eds.), Aging and
vice as a pathway to early socioeconomic achievement post-traumatic stress disorder (pp. 1951).
for disadvantaged groups. In J. M. Wilmoth & A. S. Washington, DC: American Psychiatric Press.
London (Eds.), Life-course perspectives on military Cozza, S. J., & Lerner, R. M. (2013). Military children
service (pp. 119143). New York: Routledge. and families: Introducing the issue. The Future of
Berlin, G., Furstenberg, F. F., Jr., & Waters, M. C. (2010). Children, 23(2), 311.
Introducing the issue. The Future of Children, 20(1), Cozza, S. J., Haskins, R., & Lerner, R. M. (2013). Keeping
318. the promise: Maintaining the health of military fami-
Bound, J., & Turner, S. (2002). Going to war and going to lies and children. The Future of Children Policy Brief.
college: Did World War II and the G.I. Bill increase http://futureofchildren.org/futureofchildren/publica-
educational attainment for returning veterans? Journal tions/docs/23_02_PolicyBrief.pdf
of Labor Economics, 20(4), 784815. Davison, E. H., Pless, A. P., Gugliucci, M. R., King, L. A.,
Brown, M. T. (2013). Military service and lesbian, gay, King, D. W., Salgado, D. M., Spiro, A., & Bachrach, P.
bisexual, and transgender lives. In J. M. Wilmoth & (2006). Late-life emergence of early-life traumaThe
A. S. London (Eds.), Life-course perspectives on mili- phenomenon of late-onset stress symptomology
tary service (pp. 97118). New York: Routledge. among aging combat veterans. Research on Aging,
Browning, H. L., Lopreato, S. C., & Poston, D. L. (1973). 28(1), 84114.
Income and veteran status Variations among Dechter, A. R., & Elder, G. H., Jr. (2004). World War II
Mexican Americans, blacks and anglos. American mobilization in mens work lives: Continuity or dis-
Sociological Review, 38(1), 7485. ruption for the middle class? American Journal of
Burland, D., & Lundquist, J. H. (2013). The best years of Sociology, 110(3), 761793.
our lives: Military service and family relationships Department of Defense. (2012). 2011 Demographics: A
A life-course perspective. In J. M. Wilmoth & A. S. prole of the military community. Washington: Ofce
London (Eds.), Life course perspectives on military of the Deputy Under Secretary of Defense. http://
service (pp. 165184). New York: Routledge. w w w. m i l i t a r y o n e s o u r c e . m i l / 1 2 0 3 8 / M O S /
Call, V. R. A., & Teachman, J. D. (1996). Life-course tim- Reports/2011_Demographics_Report.pdf
ing and sequencing of marriage and military service Dobkin, C., & Shabani, R. (2009). The health effects of
and their effects on marital stability. Journal of military service: Evidence from the Vietnam draft.
Marriage and the Family, 58(1), 219226. Economic Inquiry, 47(1), 6980.
Camacho, P. R., & Atwood, P. L. (2007). A review of the Elder, G. H., Jr. (1986). Military times and turning points
literature on veterans published in Armed Forces & in mens lives. Developmental Psychology, 22(2),
Society, 19742006. Armed Forces & Society, 33(3), 233245.
351381. Elder, G. H., Jr. (1987). War mobilization and the life
Campbell, D. (2013). Womens lives in wartime: The course: A cohort of World War II veterans. Sociological
American Civil War and World War II. In J. M. Forum, 2(3), 449472.
Wilmoth & A. S. London (Eds.), Life-course perspec- Elder, G. H., Jr. (2013). Foreword. In J. M. Wilmoth &
tives on military service (pp. 4867). New York: A. S. London (Eds.), Life-course perspectives on mili-
Routledge. tary service (pp. xixxxiii). New York: Routledge.
298 A.S. London and J.M. Wilmoth

Elder, G. H., Jr., & Bailey, S. L. (1988). The timing of Flynn, G. Q. (1993). The draft 19401973. Lawrence:
military service in mens lives. In J. Aldous & D. M. University Press of Kansas.
Klein (Eds.), Social stress and family development Gade, P. A., & Ruise, B. (2013). United States military
(pp. 157174). New York: Guilford. services sponsorship of life-course research: Past,
Elder, G. H., Jr., & Clipp, E. C. (1988a). Wartime losses present, and future. In J. M. Wilmoth & A. S. London
and social bonding Inuences across 40 years in (Eds.), Life-course perspectives on military service
mens lives. Psychiatry Interpersonal and Biological (pp. 243253). New York: Routledge.
Processes, 51(2), 177198. Giele, J. Z., & Elder, G. H., Jr. (1998). Methods of life
Elder, G. H., Jr., & Clipp, E. C. (1988b). War experiences course research: Qualitative and quantitative
and social ties: Inuences across 40 years in mens approaches. Thousand Oaks: Sage Publications.
lives. In M. W. Riley (Ed.), Social change and the life Gill, H. L., & Haurin, D. R. (1998). Wherever he may go:
course (Social structures and human lives, Vol. 1, How wives affect their husbands career decisions.
pp. 306327). Beverly Hills: Sage Publications. Social Science Research, 27(3), 264279.
Elder, G. H., Jr., & Clipp, E. C. (1989). Combat experi- Gill, H. L., Haurin, D. R., & Phillips, J. (1994). Mobility
ence and emotional health Impairment and resilience and fertility in the military. Social Science Quarterly,
in later life. Journal of Personality, 57(2), 311341. 75, 340353.
Elder, G. H., Jr., & Johnson, M. K. (2002). Perspectives Gimbel, C., & Booth, A. (1996). Who fought in Vietnam?
on human development in context. In C. von Hofsten Social Forces, 74(4), 11371157.
& L. Bergman (Eds.), Psychology at the turn of the Harrell, M. C. (2000). Invisible women: Junior enlisted
millennium (Social, developmental, and clinical per- Army wives. Santa Monica: RAND.
spectives, Vol. 2, pp. 153175). East Sussex: Harrell, M. C. (2001). Army ofcers spouses: Have the
Psychology Press Ltd. white gloves been mothballed? Armed Forces &
Elder, G. H., Jr., Gimbel, C., & Ivie, R. L. (1991). Turning Society, 29(1), 5575.
points in life: The case of military service and war. Hein, C., London, A. S., & Wilmoth, J. M. (2012).
Military Psychology, 3(4), 215231. Veteran status and material hardship: The moderating
Elder, G. H., Jr., Shanahan, M. J., & Clipp, E. C. (1994). inuence of disability. Social Service Review, 86(1),
When war comes to mens lives: Life-course patterns 119142.
in family, work, and health. Psychology and Aging, Hogan, D. P. (1981). Transitions and social change: The
9(1), 516. early lives of American man. New York: Academic.
Elder, G. H., Jr., Shanahan, M. J., & Clipp, E. C. (1997). Human Resources Strategic Assessment Program. (2011).
Linking combat and physical health: The legacy of Status of forces survey of active duty members:
World War II in mens lives. American Journal of Leading indicators. Arlington: Defense Manpower
Psychiatry, 154(3), 330336. Data Center.
Elder, G. H., Jr., Johnson, M. K., & Crosnoe, R. (2003). Johnson, R. J., & Kaplan, H. B. (1991). Psychosocial pre-
The emergence and development of life course theory. dictors of enlistment in the all-voluntary armed forces:
In J. T. Mortimer & M. J. Shanahan (Eds.), Handbook A life-event-history analysis. Youth and Society, 22(3),
of the life course (pp. 319). New York: Kluwer 291317.
Academic/Plenum. Kelty, R., & Segal, D. R. (2013). The military as a trans-
Elder, G. H., Clipp, E. C., Brown, J. S., Martin, L. R., & forming inuence: Integration into or isolation from
Friedman, H. W. (2009). The lifelong mortality risks of normal adult roles? In J. M. Wilmoth & A. S. London
World War II experiences. Research on Aging, 31(4), (Eds.), Life-course perspectives on military service
391412. (pp. 1947). New York: Routledge.
Elder, G. H., Jr., Wang, L., Spence, N. J., Adkins, D. E., & Kelty, R., Kleykamp, M. A., & Segal, D. R. (2010). The
Brown, T. H. (2010). Pathways to the all-volunteer military and the transition to adulthood. Future of
military. Social Science Quarterly, 9(2), 455475. Children, 20(1), 181207.
Ender, M. G. (1996). Growing up in the military. In C. D. Kleykamp, M. (2009). A great place to start? The effect of
Smith (Ed.), Strangers at home: Essays on the effects prior military service on hiring. Armed Forces &
of living overseas and coming home to a strange land Society, 35(2), 266285.
(pp. 125150). Bayside: Aletheia Publications. Kleykamp, M. (2013). Labor market outcomes among
Ender, M. G. (2002). Military brats and other global veterans and military spouses. In J. M. Wilmoth &
nomads: Growing up in organization families. A. S. London (Eds.), Life-course perspectives on mili-
Westport: Praeger. tary service (pp. 144164). New York: Routledge.
Faber, A. J., Willerton, E., Clymer, S. R., MacDermid, Laub, J. H., & Sampson, R. J. (2003). Shared beginnings,
S. M., & Weiss, H. M. (2008). Ambiguous absence, divergent lives: Delinquent boys to age 70. Cambridge,
ambiguous presence: A qualitative study of military MA: Harvard University Press.
reserve families in wartime. Journal of Family LeardMann, C. A., Powell, T. M., Smith, T. C., Bell,
Psychology, 22(2), 222230. M. R., Smith, B., Boyko, E. J., Tomoko, I. H.,
Fligstein, N. D. (1980). Who served in the military, 1940 Gackstetter, G. D., Ghamsary, M., & Hoge, C. W.
73. Armed Forces & Society, 6(2), 297312. (2013). Risk factors associated with suicide in current
Military Service in Lives: Where Do We Go From Here? 299

and former US military personnel. Journal of the MacLean, A. (2013). A matter of life and death: Military
American Medical Association, 310(5), 496506. service and health. In J. M. Wilmoth & A. S. London
Lester, P., Mogil, C., Saltzman, C., Woodward, K., Nash, (Eds.), Life course perspectives on military service
W., Leskin, G., Bursch, B., Green, S., Pynoos, R., & (pp. 200220). New York: Routledge.
Beardslee, W. (2010). Families overcoming under MacLean, A., & Elder, G. H., Jr. (2007). Military service
stress: Implementing family-centered prevention for in the life course. Annual Review of Sociology, 33,
military families facing wartime deployments and 175196.
combat operational stress. Military Medicine, 176(1), MacLean, A., & Kleykamp, M. (2014). Coming home:
1925. Attitudes toward U.S. veterans returning from Iraq.
London, A. S., & Wilmoth, J. M. (2006). Military service Social Problems, 61(2), 131154.
and (dis)continuity in the life course Evidence on Mettler, S. (2005). Soldiers to citizens: The G.I. Bill and
disadvantage and mortality from the Health and the making of the greatest generation. New York:
Retirement Study and the Study of Assets and Health Oxford University Press.
Dynamics among the Oldest-Old. Research on Aging, Miech, R. A., London, A. S., Wilmoth, J. M., & Koester,
28(1), 135159. S. (2013). The effects of the militarys antidrug poli-
London, A. S., & Wilmoth, J. M. (2008). Military service. cies over the life course: The case of past-year halluci-
In D. Carr, R. Crosnoe, M. E. Hughes, & A. Pienta nogen use. Substance Use & Misuse, 48, 837853.
(Eds.), Encyclopedia of the life course and human Modell, J., & Haggerty, T. (1991). The social impact of
development (pp. 292299). Farmington Hills: Gale. war. Annual Review of Sociology, 17, 205224.
London, A. S., & Wilmoth, J. M. (2015). Veteran status Mortimer, J. T., & Shanahan, M. J. (2003). Handbook of
and paid sex among American men: Findings from the life course. New York: Kluwer Academic/Plenum
three national surveys. Archives of Sexual Behavior, Publishers.
44(4), 9971009. National Center for Veterans Analysis and Statistics.
London, A. S., Hein, C. M., & Wilmoth, J. M. (2011). (2014). Ofce of the Actuary, Veteran Population
Work-related disability, veteran status, and poverty: Projections Model (VetPop2014), Table 2 L. http://
Implications for family well-being. Journal of Poverty, www.va.gov/vetdata/Veteran_Population.asp
15, 330349. National Research Council. (2006). Assessing tness for
London, A. S., Brown, M. T., & Wilmoth, J. M. (2012). military enlistment: Physical, medical, and mental
Long-term consequences and turning points: Veteran health standards. Washington, DC: National
status and mens later-life depression. Presented at the Academies Press.
107th annual meeting of the American Sociological Pavalko, E. K., & Elder, G. H., Jr. (1990). World War II
Association, Denver, 1720 August. and divorce: A life-course perspective. American
London, A. S., Allen, E., & Wilmoth, J. M. (2013). Journal of Sociology, 95(5), 12131234.
Veteran status, extramarital sex, and divorce: Findings Rindfuss, R. (1991). The young adult years: Diversity,
from the 1992 National Health and Social Life Survey. structural change, and fertility. Demography, 28(4),
Journal of Family Issues, 34(11), 14521473. 493512.
London, A. S., Burgard, S. A., & Wilmoth, J. M. (2014). Sackett, P., & Mavor, A. (2003). Attitudes, aptitudes, and
The inuence of veteran status, psychiatric diagnosis, aspirations of American youth: Implications for mili-
and traumatic brain injury on inadequate sleep. tary recruitment. Washington, DC: National
Journal of Sociology & Social Welfare, XLI(4), 4967. Academies Press.
Lopreato, S. C., & Poston, D. L. (1977). Differences in Sampson, R. J., & Laub, J. H. (1996). Socioeconomic
earnings and earnings ability between black veterans achievement in the life course of disadvantaged men:
and non-veterans in the United States. Social Science Military service as a turning point, circa 19401965.
Quarterly, 57(4), 750766. American Sociological Review, 61(3), 347367.
Lundquist, J. H. (2004). When race makes no difference: Segal, D. R., & Segal, M. W. (2004). Americas military
Marriage and the military. Social Forces, 83(2), population. Population Bulletin, 59(4), 140.
731757. Settersten, R. A., Jr. (2003). Invitation to the life course:
Lundquist, J. H., & Smith, H. L. (2005). Family formation Toward new understandings of later life. Amityville:
among women in the US Military: Evidence from the Baywood.
NLSY. Journal of Marriage and Family, 67(1), 113. Settersten, R. A., Jr. (2006). When nations call: How war-
Lundquist, J. H., & Xu, Z. (2014). Reinstitutionalizing the time military service matters for the life course and
family: Lifecourse policy and marriage in the military. aging. Research on Aging, 28(1), 1236.
Journal of Marriage and Family, 76(5), 10631081. Settersten, R. A., Jr., & Ray, B. (2010). Whats going on
Lutz, A. C. (2013). Race-ethnicity and immigration in the with young people today? The long and twisting
U.S. military. In J. M. Wilmoth & A. S. London (Eds.), path to adulthood. The Future of Children, 20(1),
Life-course perspectives on military service 1941.
(pp. 6896). New York: Routledge. Shanahan, M. J. (2000). Pathways to adulthood in chang-
MacLean, A. (2010). The things they carry: Combat, dis- ing societies: Variability and mechanisms in life
ability, and unemployment among U.S. men. American course perspective. Annual Review of Sociology, 26,
Sociological Review, 75(4), 563585. 667692.
300 A.S. London and J.M. Wilmoth

Smith, R. T., & True, G. (2014). Warring identities: U.S. Census Bureau. (2015). http://factnder.census.gov/
Identity conict and the mental distress of American f a c e s / t a b l e s e r v i c e s / j s f / p a g e s / p r o d u c t v i ew.
veterans of the wars in Iraq and Afghanistan. Society xhtml?src=CF
and Mental Health, 4(2), 147161. Usdansky, M. L., London, A. S., & Wilmoth, J. M. (2009).
Spiro, A., Schnurr, P. P., & Aldwin, C. M. (1997). A life- Veteran status, race-ethnicity, and marriage among
span perspective on the effects of military service. fragile families. Journal of Marriage and Family,
Journal of Geriatric Psychiatry, 30(1), 91128. 71(3), 768786.
Stanley, M. (2003). College education and the midcentury Vogt, D. S., King, D. W., King, L. A., Savarese, V. W., &
GI Bills. Quarterly Journal of Economics, 118(2), Suvak, M. K. (2004). War-zone exposure and long-
671708. term general life adjustment among Vietnam veterans:
Street, D., & Hoffman, J. (2013). Military service, social Findings from two perspectives. Journal of Applied
policy, and later-life nancial and health security. In Social Psychology, 34(9), 17971824.
J. M. Wilmoth & A. S. London (Eds.), Life-course per- Wertsch, M. E. (1991). Military brats: Legacies of
spectives on military service (pp. 221242). New York: childhood inside the fortress. New York: Harmony
Routledge. Books.
Teachman, J. D. (2004). Military service during the Whyman, M., Lemmon, M., & Teachman, J. (2011). Non-
Vietnam era: Were there consequences for subsequent combat military service in the United States and its
civilian earnings? Social Forces, 83(2), 709730. effects on depressive symptoms among men. Social
Teachman, J. D. (2005). Military service in the Vietnam Science Research, 40, 695703.
era and educational attainment. Sociology of Wilmoth, J. M., & London, A. S. (2011). Aging veterans:
Education, 78(1), 5068. Needs and provisions. In J. L. Angel & R. Settersten
Teachman, J. D. (2007a). Race, military service, and mar- (Eds.), Handbook of the sociology of aging (pp. 445
ital timing: Evidence from the NLSY-79. Demography, 461). New York: Springer.
44(2), 389404. Wilmoth, J. M., & London, A. S. (2013). Life-course per-
Teachman, J. D. (2007b). Military service and educational spectives on military service. New York: Routledge.
attainment in the all-volunteer era. Sociology of Wilmoth, J. M., &, London, A. S. (forthcoming). The
Education, 80(4), 359374. inuence of military service on aging. In K. F. Ferraro,
Teachman, J. D. (2013). Setting an agenda for future L. K. George, & D. Carr (Eds.), Handbook of aging
research on military service and the life course. In and the social sciences (8th edition). New York:
J. M. Wilmoth & A. S. London (Eds.), Life course per- Springer.
spectives on military service (pp. 254274). New York: Wilmoth, J. M., London, A. S., & Parker, W. M. (2010).
Routledge. Military service and mens health trajectories in later
Teachman, J. D., & Tedrow, L. M. (2004). Wages, earnings, life. Journal of Gerontology: Social Science, 56(6),
and occupational status: Did World War II veterans 744755.
receive a premium? Social Science Research, 33(4), Wilmoth, J. M., London, A. S., & Hein, C. M. (2015a).
581605. The use of VA disability compensation and Social
Teachman, J. D., & Tedrow, L. M. (2008). Divorce, race, Security disability insurance among working-aged
and military service: More than equal pay and equal veterans. Disability and Health Journal, 8(3),
opportunity. Journal of Marriage and Family, 70(4), 388396.
10301044. Wilmoth, J. M., London, A. S., & Hein, C. M. (2015b).
Teachman, J., & Tedrow, L. (2013). Veteran status and Economic well-being among older adult households:
body weight: A longitudinal xed-effects approach. Variation by veteran and disability status. Journal of
Population Research and Policy Review, 32(2), Gerontological Social Work, 58(4), 399419.
199220. Wilmoth, J. M., London, A. S., & Himes, C. L. (forthcom-
Teachman, J. D., & Tedrow, L. M. (2014). Delinquent ing). Inter-cohort variation in the consequences of
behavior, the transition to adulthood, and the likeli- U.S. military service on mens body mass index trajec-
hood of military enlistment. Social Science Research, tories in mid- to late-life. In C. Burton-Jeangros,
45, 4655. S. Cullati, A. Sacker, & D. Blane (Eds.), A life course
Teachman, J. D., Call, V. R. A., & Segal, M. W. (1993a). perspective on health trajectories and transitions.
Family, work, and school inuences on the decision to London: Springer.
enter the military. Journal of Family Issues, 14(2), Wolf, D. A., Wing, C., & Lopoo, L. M. (2013).
291313. Methodological problems in determining the conse-
Teachman, J. D., Call, V. R. A., & Segal, M. W. (1993b). quences of military service. In J. M. Wilmoth & A. S.
The selectivity of military enlistment. Journal of London (Eds.), Life-course perspectives on military
Political & Military Sociology, 21(2), 287309. service (pp. 254274). New York: Routledge.
Criminal Justice and the Life
Course

Sara Wakeeld and Robert Apel

1 Introduction interdependency, and agencyand we use these


to structure the remainder of the review.
Few institutions in the United States have under-
gone such radical change as the criminal justice
system over the last four decadeschanges that 2 Context: History,
brought it to the fore as an institution powerfully Demography, and the Prison
structuring the lives of those who come into con- Environment
tact with it. We focus our analysis on the United
States for two reasons. First, historical develop- Context refers to the idea that the life course of
ments in the United States have transformed the individuals is embedded in and shaped by the his-
criminal justice system and, especially, the prison torical times and places they experience over
from an institution that few experience to one that their lifetime (Elder 1998: 3). With respect to
has profound inuences on the life course. the criminal justice system, there are three major
Second, because a central premise of life-course contexts that must be considered: historical,
theory is to link the institutional and historical sociodemographic, and geographic. A fourth
context to the unfolding of individual lives, it is contextual facet that cannot be overlooked is the
virtually impossible to do justice to the perspec- prison environment itself.
tive without focusing the discussion on a specic
place and time. We do, however, compare devel-
opments in the United States with other nations 2.1 Historical Context
throughout. We take advantage of the conceptual
tools of life-course analysis in order to integrate Since the early 1970s, the American criminal jus-
much of what we know from existing research tice system has grown in size, scope, and inu-
about the criminal justice system and its conse- ence. These shifts were historically unprecedented
quences. Elder (1998) denes four key principles in scale and, importantly, largely unanticipated
of the life-course perspectivecontext, timing, by criminologists. Writing in the 1970s, for
example, prominent criminologists Alfred
Blumstein and Jacqueline Cohen (1973) mar-
S. Wakeeld (*) R. Apel veled at the remarkable stability of punishment
School of Criminal Justice, Rutgers University, between 1930 and 1970. At the time of their writ-
New Brunswick, NJ, USA ing, the prison incarceration rate in the United
e-mail: sara.wakeeld@rutgers.edu; States had achieved an apparent steady state of
robert.apel@rutgers.edu

Springer International Publishing Switzerland 2016 301


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_13
302 S. Wakeeld and R. Apel

about 110 per 100,000, and only rarely strayed Still, while there is substantial cross-state
outside of the interval 100120 per 100,000. This variation, the mass incarceration of large propor-
stability persisted through the crime rate increases tions of the population is unique to the United
of the 1960s, compelling them to propose a con- States in comparison to the rest of the world.
servation theory whereby the threshold response Figure 3 compares incarceration rates in the
to criminal behavior routinely adapts to maintain United States relative to others across the globe,
a constant level of punishment. An analysis of showing a rate in the United States that is 510
state imprisonment rates provided further evi- times the norm internationally.1 Moreover, while
dence of this homeostasis (Blumstein and Moitra the United States knows no peer in its rate of
1979). However, as if to mock this conservation incarceration, the countries that come closest
theory, in the 1970s policymakers began laying have little else in common with the United States
the foundation for the grand social experiment and are typically characterized by repressive or
(Frost and Clear 2009) that is now commonly unstable regimes (for example, Russia, Cuba, and
referred to as mass incarceration. Rwanda). Countries that more closely approxi-
Sociologist David Garland is usually credited mate the United States on other metrics (social,
with coining the phrase mass imprisonment political, or economic) have also increased their
and highlights two primary characteristics of the use of incarcerationfor example, the UK,
practice: (1) a rate of incarceration that is far Denmark, or the Netherlandsbut no other
beyond any historical or comparative standard Western democracy approaches the sheer scale of
(Garland 2001b: 5) and (2) incarceration rates imprisonment evident in the United States over
that result in the systematic incarceration of the last four decades. Unsurprisingly, the
whole groups in the population (Garland 2001b: American prison boom coincided with massive
6). Figures 1, 2, 3, and 4 demonstrate just how expansion in spending on the criminal justice
much has changed with respect to crime and system overall, including police, courts, and cor-
criminal justice since the 1970s in the United rections (see Fig. 4).
States (Federal Bureau of Investigation 1960 The punitive turn is generally explained
2012; Sourcebook of Criminal Justice Statistics through a conuence of factors including the
2014). As Garlands characterization suggests emergence of law-and-order politics and the
and Fig. 1 displays, the imprisonment rate has politicization of crime in the 1970s and 1980s
grown steadily, whether or not the crime rate is (Beckett 1997; Garland 2001a; Jacobs and Helms
moving upward or downward. 1996), economic realignment (Western 2006),
While most observers refer to mass incarcera- and increasingly deterministic and punitive sen-
tion as a national trend, there is substantial varia- tencing policies (National Research Council
tion in the practice across states. Figure 2 2014; Simon 2007). Examples of such practices
compares incarceration rates across states in include mandatory minimum sentencing regimes
2013. Some states, like Maine or Minnesota, (that require a prison term be served), so-called
never dabbled in the mass incarceration of its three strikes and youre out laws (that dramati-
population while others, most notably in the cally lengthen prison sentences for recidivists),
South, have maintained the highest mass incar- and the expansion of legal, civil, and social liabil-
ceration rates since beginning the practice in the ities that accompany a felony conviction (that
late 1970s. Still others, best exemplied in entrench former inmates in a criminal class)
California and Florida, were initial leaders in the (Olivares et al. 1996; Shannon et al. 2014;
high reliance on incarceration as a response to Wakeeld and Uggen 2010).
crime but have since scaled back because of
crushing budget decits, bloated prison popula- 1
Note that Fig. 1 depicts imprisonment rates (incarcera-
tions, or, in the case of California, a federal court
tion in prisons) while Fig. 3 compares incarceration rates
order to reduce their populations because of (incarceration rates in prisons and jails) hence the differ-
inhumane conditions. ence in overall rates across the gures.
Criminal Justice and the Life Course 303

Fig. 1 U.S. Violent Crime Rate and Incarceration Rate per 100,000 (19602012)

Fig. 2 Incarceration Rate per 100,000, by U.S. State (2013)

For our purposes of examining the effects of incarceration era experienced rapidly increasing
shifts in crime and punishment on the life course, levels of formal criminal punishment and surveil-
trends in crime and punishment suggest that lance and widely varying levels of exposure to
cohorts who came of age during the mass crime. The average young adult in the United
304 S. Wakeeld and R. Apel

Fig. 3 Incarceration Rate per 100,000, by Country (2013)

Fig. 4 Criminal Justice System Expenditures in the United States, 19822007

States today now has the novel historical experi- 2.2 Sociodemographic Context
ence of living in a low-crime society with the
highest rates of criminal punishment in the world. Recall that Garlands mass incarceration crite-
In contrast, among the most disadvantaged and ria include both high (relative to a historical or
socially marginalized, crime, police surveillance, comparative norm) and demographically concen-
and criminal punishment are common experi- trated rates of incarceration. The graphs in the
ences. It is to this group that we now turn. preceding section show how far out of any
Criminal Justice and the Life Course 305

historical or comparative mean the United States cohort shifts in the likelihood of arrest and incar-
is with respect to its reliance on the prison. Yet ceration provide another salient example. A
mass incarceration is consequential for our recent study found that about 30 % of Black
understanding of the life course as much because males had been arrested at least once by age 18
of Garlands second criterion as the rst. and almost half (49 %) had been arrested at least
Incarceration in the United States is increasingly once by the age of 23; the corresponding arrest
dened as the mass imprisonment of marginal- risk for White males was 22 % and 28 % respec-
ized menmostly Black and Hispanic men with tively (Brame et al. 2014). The race gap in cumu-
very little education and very poor job prospects. lative risk of incarceration by age 30 for men who
Contact with the criminal justice system is thus came of age during the prison boom is much
driven by the intersection of group memberships largerabout 26.8 % of all Black males had been
dened by race (Black and Hispanic), sex (male), incarcerated compared to 5.4 % for White males.
socioeconomic status (poorly educated, unem- Among cohorts born 30 years earlier, prior to the
ployed or low-wage workers), and neighborhood prison boom, the comparable risks are 14.7 %
(highly concentrated disadvantage). (Black) and 3.8 % (White) (Western and Pettit
As with the causes of the prison boom, the 2010: 11). Similarly large race gaps in the risk of
causes of such stark racial disparities in the prison experiencing the incarceration of a parent are evi-
population are a point of signicant debate dent as wellraising the possibility of an inter-
among scholarsranging from narratives of generational transmission of inequality that
largely unintended consequences of race-neutral occurs via contact with the criminal justice sys-
policies (Neal and Rick 2014) to mass incarcera- tem, criminal conviction, and prison experience
tion as racial subjugation by another name (Pettit et al. 2009; Wakeeld and Wildeman
(Alexander 2010). Indeed, others view the prison 2013; Wildeman 2009).
boom as extrapenological, or unrelated to The race-sex-class nexus is the most striking
crime rates, but instead driven by the same racial- characteristic of the modern prison population,
ized forces that resulted in slavery or Jim Crow but other characteristics dene the group as well.
segregation (Wacquant 2000, 2001). Several The expansion of the criminal justice system in
studies do suggest that the proportion of sentenc- the United States has also increased the rate of
ing disparities explained by racial differences in women serving timefrom about 5 per
arrest has declined since the dawn of the prison 100,000 in 1970 to 65 per 100,000 in 2013and
boom (Baumer 2013) and racial disproportional- prisons are also characterized by an aging popu-
ity is surely the most striking feature of the mod- lation with signicant mental and physical health
ern criminal justice system in the United States problems (Beck and Maruschak 2001; Bureau of
(Wakeeld and Uggen 2010). This debate is not Justice Statistics 2007; Myers and Wakeeld
yet settled, howeverwe would argue that 2014; National Commission on Correctional
although racial disparities are evident in all stages Health Care 2002). Indeed, spending time in the
of the justice system, the causal mechanisms that nations prisons and, especially, jails are a com-
produce them are not clear. mon experience for those with severe mental ill-
While race gaps in criminal involvement exist, ness (National Commission on Correctional
they do not neatly explain race gaps in who ends Health Care 2002; Teplin 1984).
up in prison (National Research Council 2014;
but see Neal and Rick 2014). Black-white gaps in
arrest and incarceration are more disparate with 2.3 Geographic Context
respect to drug crimes, yet numerous self-report
studies show that Whites report comparable or Contact with the criminal justice system tends to
higher levels of substance use relative to Blacks be concentrated geographically as wellseveral
(Bachman et al. 1991; Beckett et al. 2006; studies show that a large proportion of inmates
National Research Council 2014). Race gaps and come from a relatively small number of neigh-
306 S. Wakeeld and R. Apel

borhoods (Clear 2007; Gonnerman 2004; 2.4 The Prison as Context


Sampson and Loefer 2010). The spatial con-
centration of criminal justice experience is par- Finally, the relevance of context for life-course
tially explained by developments in law analysis leads to the additional realization that
enforcement that resulted in the concentrated the prison or jail environment is a potentially
surveillance of whole groups and neighbor- salient context in its own right. The effects of
hoods. Serious criminal behavior is highly spa- doing time, prison culture, and living in a total
tially concentrated, as is the law enforcement institution was a source of sociological and crim-
response to it (see Smith 1986). The very exis- inological attention long before the rapid increase
tence of the crime hot spot in the modern crim- in the incarceration rate (E. Goffman 1961; Irwin
inological lexicon reects the non-random 1970; Jacobs 1977; Sykes 1958/2007). Classic
distribution of crime along with other indicators studies of institutionalization and prison culture
of physical and social disorder (see Sherman generally pre-date the prison boom in the United
et al. 1989). Wilson and Kellings (1982) meta- Stateswith relatively few exceptions (e.g.,
phor of a broken window to characterize the Irwin 2005; Kruttschnitt and Gartner 2005;
relationship between disorder and serious preda- Skarbeck 2014) sociologists and criminologists
tory crime became the dominant rationale for increasingly turned their attention to life after
enforcement of public order violations (so-called prison in the form of collateral consequences
incivilities). Consequently, many urban police research (see Wakeeld and Uggen 2010 for a
departments have embraced the aggressive use review) or focused on the control and mainte-
of eld interrogations of suspicious persons nance of growing inmate populations (e.g.,
especially the so-called stop-and-frisk or Terry DiIulio 1987; Steiner and Wooldredge 2009) dur-
stop (Terry v. Ohio, 392 U.S. 1 [1968]). Whatever ing the prison boom. That the prison environment
its law enforcement benets, however, it has as a consequential context for later life outcomes
been heavily criticized on the basis that it received less research attention during the height
harasses poor and minority residents while hav- of the prison boom represents a large gap in the
ing questionable impact on crime and disorder literature (see Kreager et al. 2015; Simon 2000
(Gelman et al. 2007; Goffman 2014; Tyler 2006; for further discussion).
but see Zimring 2012). There are reasons to think that the prison con-
We noted in the historical section of our essay text is important for understanding life-course
that even at such high levels of incarceration, the outcomes, whatever the rate of incarceration, but
average person in the United States has little con- this question surely takes on more relevance as
tact with the criminal justice system and simply the prison population grows larger. Incarceration
benets from coming of age during a period with could promote the accumulation of a perverse
very little crime. Moreover, many scholars of the form of knowledge known as criminal capital.
collateral consequences of the prison boom, for Criminal capital is the accumulation of skills and
example, routinely describe them as hidden, knowledge that, while useful in the criminal
invisible, (e.g., Manza and Uggen 2006; Pettit underworld and in prison, are of little practical
2012; Western and Beckett 1999) or otherwise utility in more conventional settings (Hagan
opaque to the general publicthe growth in 1993). The most obvious way that correctional
incarceration and overwhelming reach of law institutions could provide criminal capital is
enforcement has only recently become a source through peer inuence. Inmates might become
of public discussion and debate. The demo- more deeply embedded in crime because they
graphic and spatial concentration of crime, law spend time in the company of fellow captives
enforcement attention, and prison experience is who strengthen their orientation to unlawful
perhaps one explanation for the relative lack of behavior. For example, Bayer et al. (2009)
public discussion of the deleterious consequences reported higher recidivism probabilities among
of the prison boom until very recently. juveniles who were exposed to other youthful
Criminal Justice and the Life Course 307

offenders remanded to the same Florida correc- make a bad situation less bad by some measures,
tional facility, and who had committed the same provided they are matched to the appropriate
type of offense. In other words, they found rein- services.
forcing effects of inmate peers on those offenses We end this section by reiterating a curiosity:
in which youth had prior experience, but no such much of the research on the social structure and
reinforcing effects on offenses in which youth context of the prison took place before the mas-
had no prior experience. sive increases in incarceration in the late 1970s
Criminal behavior might also be maintained and early 1980s. Research on the modern prison
through prison socialization, or what has been boom is thus largely de-contextualized today. Yet
called prisonization or institutionalization we know that the modern prison presents a num-
(Clemmer 1940; Goffman 1961; Sykes ber of difculties for inmate adjustment while
1958/2007). Adaptations to the habits and customs incarcerated and upon release. Modern prisons
that prevail in correctional institutions may not be are often overcrowded places, mixing less experi-
easily shed upon release and could therefore create enced offenders with a criminal core, high rates
problems of adjustment in conventional society. of mental illness prevail, especially in jails,
Correctional institutions are tightly controlled and the provision of vocational training and treat-
environments wherein inmates are stripped of their ment services is highly variable. Often as a result
identities and cut off from social supports. Like of scarce data sources, few studies are able to
Brooks Hatlen in The Shawshank Redemption link variation in institutional experience to later
the elderly prison librarian who was paroled after outcomes, thus much current research on impris-
serving 50 years in prison but who took his own onment and the life course ignores the enormous
life shortly after releasesome returning ex-pris- black box of institutional context.
oners might experience psychological difculty
making the transition back into the free world after
a lengthy prison sentence. Alternatively, in prisons
that house the most dangerous offenders (e.g., 3 Timing: Age-Graded Effects
maximum security prisons), James Jacobs (1977) of Criminal Justice Contact
suggests that socialization into a violent code of
conduct could worsen recidivism risks. Timing is the principle that the developmental
It is not immediately obvious, however, that impact of a succession of life transitions or events
the incarceration experience is altogether nega- is contingent on when they occur in a persons
tive for inmates, as the foregoing discussion life (Elder 1998: 3). Age is a central feature of
about criminal capital and prison socialization life-course analysis of the criminal justice sys-
might suggest. Although the rehabilitative ideal tem, and encompasses such diverse themes as the
was never truly realized even in its supposed hey- age distribution of crime and criminal justice
day of the 1970s (Phelps 2011), there is compel- involvement, the age-graded consequences of
ling evidence that some common prison-based involvement, the aging of the inmate population,
programs lower the recidivism of participants. and the very conception of criminal culpability
For example, in a meta-analysis of prison-based (and the historical justication for a juvenile jus-
programs, Wilson et al. (2000) concluded that tice system distinct from the criminal justice
programs which provide adult basic education system).
and post-secondary education produce consistent
reductions in recidivism, as well as improvement
in post-release employment. Furthermore, these 3.1 Age and the Evolution
and other programs generate net benets with of Punishment
respect to their cost effectiveness (Aos et al.
2006). This suggests that, at least for some incar- Age is inextricably linked to crimeself-
cerated offenders, the prison experience can reported criminal involvement tends to peak in
308 S. Wakeeld and R. Apel

middle adolescence and fall off sharply by the 3.2 Aging in Prison
late 20s. Indeed, the aggregate age-crime curve
varies so little that criminologists Michael Just as the proportion of juveniles serving time
Gottfredson and Travis Hirschi (1986) deemed it increased, so too did the proportion of aging and
unworthy of social explanation, though crimi- elderly inmates. Lengthier sentences resulted in
nologists continue to examine its shape across signicant graying of the prison population.
place, space, and demographic groups (e.g., While the changes in the juvenile justice system
Lauritsen 1998; Liu 2015), as well as to develop brought immediate criticism, the problems asso-
life-course theories for individual-level trajecto- ciated with an aging prison population took lon-
ries of criminal involvement, persistence, and ger to reveal. Beginning in the 2000s, however,
desistance (Moftt 1993; Sampson and Laub bloated state prisons began to buckle at the
1994). For our purposes, the punitive turn weight and cost of such large prison populations
described above also dramatically altered the (Auerhahn 2006; Mitka 2004). The costs imposed
consequences of criminal involvement for by older inmates are considerable; the group is
juveniles. characterized by the signicant health problems
While age is inextricably linked to criminal that often accompany lives led in instability, vio-
involvement, youthful offenders were historically lence, and substance abuse. The most notorious
buffered from the most serious forms of criminal example is Californiaa leader in punitive sen-
punishment. This is much less true today. Many tencing laws like Three Strikes and the reliance
of the social, cultural, and political changes that on incarceration as a response to crimeas over-
drove punitive policy adoption in the adult crimi- crowding and lack of adequate medical care in its
nal justice system inuenced the juvenile justice prisons were deemed tantamount to cruel and
system as wellin addition, real increases in unusual punishment by the U.S. Supreme Court
youth crime and gang violence in the 1980s and in 2011 (Simon 2014).
1990s gave the impression that the countrys
youth were out of control (Welch et al. 2002).
Prominent commentators warned of remorseless 3.3 Age-Graded Consequences
and violent juvenile super-predators and a of Criminal Justice Contact
wave of violent crime (DiIulio 1995). This never
happenedas Fig. 1, presented earlier, clearly Moving beyond the age prole of prison popula-
showsbut the commentary did support a host tions and the relative unimportance of age for
of policy shifts, to dramatic effect, in the juvenile punishment decisions, we can also think of the
justice system (Roberts 2004; but see Nagin et al. age-graded consequences of contact with the
2006). Among them, the transfer of juveniles to criminal justice system. First, custodial conne-
adult court became relatively common and juve- ment during the late teens and early twenties cor-
niles, some only pre-teens, were increasingly responds to a critical period in the life coursethe
subject to adult punishmentlong custodial sen- transition to adulthood. The expansion of the
tences, mandatory life without parole, and even criminal justice system corresponds to the same
death (Bishop 2000; Fagan and Zimring 2000; historical moment that life-course sociologists
Feld 1999). More subtle changes that reduced began to describe the transition to adulthood as
exibility, encouraged mandatory sentences, and more variable, more precarious, and, for some,
reduced legal protections were also evident lengthier than during earlier eras (Arnett 2000;
throughout the early stages of the prison boom Shanahan 2000). While we can offer no direct test
taken together, all of these trends made it difcult of this hypothesis at the aggregate level, locking
to discern substantive differences between juve- up large numbers of adolescents and young adults
nile and criminal justice by the early 1990s (Feld presents numerous challenges for their ability to
1990, 1997). achieve common markers of adult status.
Criminal Justice and the Life Course 309

In life-course parlance, serving time puts lengthening of sentences for violent offenders,
youthful inmates off-time with their same age most serve relatively short sentences prior to
peers (Neugarten et al. 1965; Uggen and Wakeeld returning to their communities (Petersilia 2003).
2005). This happens in a variety of waysincar- These characteristics of the prison population,
ceration removes inmates from pro-social peers coupled with the sheer size of the criminal
that may aid subjective desistance processes class in the United States (Shannon et al. 2014),
(Massoglia and Uggen 2010) and reduces their results in enormous potential for spillover of
attainment of adult status markers like obtaining a incarceration effects to those socially-connected
degree (Page 2004), stable employment (Apel and to inmates. Among the most studied are the
Sweeten 2010; Pager 2003), family formation effects of the prison population for social inequal-
(Apel et al. 2010; Massoglia et al. 2011), and civic ity, labor markets, neighborhoods, families, and
engagement (Manza and Uggen 2006). Mass schools.
incarceration thus targets a critical population for
inclusionyoung people on the cusp of adult-
hoodand dramatically inuences their ability to 4.1 Inequality in Labor Market
move forward once they leave the criminal justice and Family Outcomes
system. Importantly, these effects are not limited
to convicted felons and former inmates. The era The prison boom has shaped the labor market in
of mass incarceration is notable not just for the the aggregate and employment prospects at the
breadth of effects observed for inmates but the individual level. Former inmates have great dif-
degree to which the experience of incarceration culty obtaining employment and incarceration
cascades to other groups socially connected to experiences, even those that are relatively short,
inmates (Comfort 2007). It is to these groups that are associated with signicant reductions in earn-
we turn in the next section. ings over the life course (Apel and Sweeten 2010;
Pager 2003; Ramakers et al. 2014; Western
2006). In the aggregate, the size of the prison
4 Interdependency: Linked population obscures indicators of economic
Lives, Social Ties, health by removing large numbers of the unem-
and Spillover ployed to prisons where they are uncounted in
household employment surveys (Pettit 2012;
Interdependency touches on the notion that lives Western and Beckett 1999); this is most problem-
are lived interdependently, and social and histori- atic because it leads to the false conclusion that
cal inuences are expressed through this network racial inequality in wages/income is improving
of shared relationships (Elder 1998: 4). when it is in fact worsening (Neal and Rick 2014;
Individuals who become entangled in the crimi- Western 2006).
nal justice system are embedded within familial A large research literature also details the near
and social relationships with others who might universal harmful effects of imprisoning large
experience vicarious consequences because of numbers of fathers for the wellbeing of their chil-
that entanglement, not to mention that the con- dren and partners. The effects of paternal incar-
centration of criminal justice involvement in cer- ceration are global and nearly always in the
tain neighborhoods potentially erodes the ability direction of harmworsening everything from
of communities to solve their own problems. mental health and behavioral problems (Geller
Many of us have a tendency to think of inmates et al. 2009; Wakeeld and Wildeman 2013;
as social isolates even before they entered the Wildeman 2010) and nancial security
criminal justice system. Yet the majority of (Schwartz-Soicher et al. 2011) to school perfor-
inmates were employed prior to entering prison mance (Turney and Haskins 2014) and infant
(Bureau of Justice Statistics 2007), most have mortality rates (Wakeeld and Wildeman 2013).
children (Mumola 2000), and, despite signicant The effects of maternal incarceration are more
310 S. Wakeeld and R. Apel

heterogeneous (Turney and Wildeman 2015; tem. The effects of incarceration for adult mens
Wildeman and Turney 2014), reecting both life course trajectories are limited because they
greater seriousness and the smaller number of have already accumulated a host of disadvan-
incarcerated mothers, but the chaos imposed on tages by the time they enter prison. This neces-
children by their mothers involvement in the sarily limits the ability of prison to make a bad
criminal justice system is clear (E. Johnson and situation worse or drastically improve their cir-
Waldfogel 2004; Siegel 2011). cumstance. Contrast this with the experiences of
Similar effects are observed with respect to their children. For all children relative to adults,
the partners of inmates; incarceration is associ- there is more room for an important change in
ated with divorce (Apel et al. 2010; Massoglia trajectory (in both positive and negative direc-
et al. 2011), maternal depression (Wildeman tions) because their lives are much less xed and
et al. 2012), harsh parenting (Turney 2014) and consequential childhood events are often more
abuse (Wakeeld 2005), and nancial insecurity important than events that occur later in life. As a
(Braman 2004; Schwartz-Soicher et al. 2011). result, the effects of paternal incarceration for
Taken together, the consequences of incarcera- children are much more global (inuencing men-
tion for families are overwhelming. In perhaps tal and physical health, educational outcomes,
the most apt description we have read, Megan occupational attainment, and increasing the risk
Comfort (2008) refers to the instability, hassles, of homelessness), tend to be larger in magnitude,
and stigma imposed on the families of inmates as and have much larger potential effects for social
a form of secondary prisonization. inequality (Wakeeld and Wildeman 2013,
The rapid increase in incarceration in the 2011).
United States has had important implications for
social inequality as well. Shifts in inequality
depend on two factors: rst, that some experience 4.2 Institutional Spillover
or event do measureable harm and second, that and Prison-Like Contexts
there are large race and class differences in the
likelihood of experiencing it. Both of these are Finally, just as incarceration effects spill over to
clearly characteristic of the prison boom in the socially-connected others, trends in the criminal
United States. An abundance of research justice system structure other social institutions
describes the social disabilities and racial dispro- as a result of co-occurring cultural and political
portionality associated with incarceration trends. The same social undercurrents that have
declines in mental and physical health, wages, given rise to stark changes in the criminal justice
and family wellbeing are well-documented, system can be seen in other institutions that have
among a host of others (see Wakeeld and Uggen taken on prison-like conditions. The same sorts
2010; Wildeman and Muller 2012 for broad of practices that now dominate the criminal jus-
reviews)and suggest that the prison boom is a tice system can be seen in the practices of public
consequential institution for the maintenance and schools, for example. Indeed, the overlap between
durability of social inequality. punitive practices in the criminal justice system
The effects of paternal incarceration for chil- and punitive school discipline is so large that
dren and social inequality are often larger in many commonly refer to the reciprocal institu-
magnitude than the effects of incarceration for tional linkages between the two as a school-to-
inmates themselves (Wakeeld and Wildeman prison pipeline (Kim et al. 2010), and one can
2013; Wildeman and Muller 2012). While most think of schools as sites of preparation for later
children of incarcerated parents will not serve carceral contact (Hirscheld 2008). The compar-
time themselves, the negative effects of parental ison between schools and prisons is most notable
imprisonment in childhood nonetheless reect a in two respects: rst, racial disparity in school
form of intergenerational transmission of disad- discipline, suspension, and expulsion is col-
vantage via contact with the criminal justice sys- ored in much the same way as incarceration risk
Criminal Justice and the Life Course 311

(Skiba et al. 2002), and second, innovations to are an emerging area of research and offer yet
disciplinary policy parallel innovations to sen- another example of the long reach of the prison
tencing policies in effect and tone. It is hard, for into all facets of social life. It is nearly impossible
example, to view Zero Tolerance policies in for one to imagine how to establish a stable life-
schools as anything other than directly analogous course trajectory without interacting with
to Mandatory Minimum or Three Strikes laws in employers, schools, banks, or families. We sug-
the criminal justice system (Schwartz and Rieser gest that all of these trendsmass incarceration,
2001). Both innovations rely on harsh punish- mass surveillance, and institutional spillover
ment for rst (or second or third) offenses, reduce add up to a population that is hidden, locked out,
discretion by allowing little to no to deviation and living on the margins of society (Pettit 2012).
from prescribed punishments, and result in enor- Having fully described the structural conditions
mous racial disparities in outcomes. and constraints that characterize the mass incar-
More subtle forms of criminal justice system ceration era, we complete our essay by assessing
spillover are found with respect to broader insti- the role of individual agency within the criminal
tutional avoidance following contact with the justice system.
criminal justice system. Mass surveillance and
police attention is focused on a small number of
neighborhoods and demographic groupsas a 5 Agency: Crime Decision
result, high warrant rates, police raids, and cor- Making, Disrepute,
rectional supervision characterize some neigh- and Individual Constraints
borhoods and demographic groups. In her
ethnography of one Philadelphia neighborhood, Agency is the acknowledgement that individuals
for example, Alice Goffman (2014) describes a construct their own life course through the
process of institutional avoidance and a life led choices and actions they take within the opportu-
constantly on the run following initial criminal nities and constraints of history and social cir-
justice system contact. The men in Goffmans cumstances (Elder 1998: 4). However one thinks
study spent some time in prison but contact also about choices in the presence of constrained
structures their lives (and the lives of their fami- options and extreme disadvantage, it is neverthe-
lies and friends) once releasedmuch of their less the case that crime pays for many offend-
time was spent avoiding institutions in an effort ers. For other offenders, embedded in criminal
to avoid entanglements with police or parole of- subculture and in search of respect, criminal jus-
cers. Living on the run also precludes any form of tice entanglement might actually confer a badge
stable employment or family life because to of honor. For yet other offenders, crime might
avoid the police is to avoid establishing a stable very well provide a means to express their con-
residence or employment, seeking medical care tempt for conventional society. These possibili-
at hospitals (or being present for the birth of ones ties point to the fact that much criminal behavior
child), or engaging any other institution that may arises from a series of choices made under a vari-
keep a record of stable attendance. Using quanti- ety of institutional and social constraints, and that
tative data, sociologist Sarah Brayne (2014) has it can often serve a purpose that is quite useful
conrmed this resultthose who have had con- when viewed from the perspective of the offender.
tact with the criminal justice system are signi- For individuals heavily involved in street
cantly less likely to report contact with criminal activity, criminal justice involvement is
surveilling institutions that might lead to crimi- an ever-present liability. To use a labor market
nal justice system entanglements (e.g., hospitals, analogyand at the risk of implying that crimi-
banks, employers, or schools). nal behavior is a kind of career when in fact it is
Subtle forms of institutional avoidance as a highly intermittent and resembles no such
result of contact with the criminal justice system thingarrest is an occupational hazard that
(such those described by Goffman and Brayne) accompanies high-risk behavior. The economic
312 S. Wakeeld and R. Apel

benets of crime are most obvious in the case of an intense, pressing need for cash bordering on
drug trafcking. Early studies of the pecuniary desperation. Jacobs and Wright (1999) observed
benets of drug selling indicated that it yielded that many offenders are immersed in a street-
mean earnings of about $1,000 per year (Vicusi based culture that rewards fast living, referring
1986; Wilson and Abrahamse 1992). What is to fetishized consumption of status-enhancing
notable about these estimates is their timing luxury goods and cash-intensive, every night is
the late 1970s, prior to the introduction of crack a Saturday night illicit activities (e.g., gambling,
cocaine to urban drug markets. The earnings heavy drinking, hard drug use). Financial des-
opportunities from drug selling would change peration arises when participants need a quick
considerably in the ensuing decade. For exam- infusion of cash in order to keep the party going,
ple, Reuter and colleagues (MacCoun and as it were. Crime often provides the most ef-
Reuter 1992; Reuter et al. 1990) conducted cient means available to resolve an acute nan-
interviews with admitted drug offenders on pro- cial crisis. While this suggests that much criminal
bation in the District of Columbia in 1988. activity has an economic motivation, it is not,
These offenders reported net income (gross strictly speaking, in the sense that offenders face
income less expenses) from drug selling of over nancial hardship and are pushed by circum-
$700 per month at the median ($1,800 at the stances completely beyond their control.
mean; $2,500 at the 75th percentile). (Note that In an empirical demonstration of the occupa-
the 1988 estimates should be doubled to convert tional facet of some criminal activity, Matsueda
them to 2014 dollars.) Among those who sold and colleagues (1992) analyzed data from the
drugs daily, they estimated a wage of $30 per h National Supported Work Demonstration Project,
for only 3 h of work each day, which was an evaluation of a transitional work program
about 4 times the wage from legitimate employ- from the 1970s. The sample was composed of
ment for this sample. men released from jail or prison in the 6 months
Some offenders are also immersed in a street prior to entry into the study, as well as men
culture, and might feel compelled to adopt the involved in a drug treatment program in the year
attributes that are believed to characterize that prior to study entry. They demonstrated that ex-
culture, such as aggressiveness, a confrontational inmates rated some legitimate occupations sig-
style, and an exaggerated sense of masculinity nicantly lower than ex-addicts, despite the fact
(Anderson 1999). Coupled with the notion of that both groups were chronically unemployed.
deviance avowal, this draws attention to the Moreover, for both groups, the number of prior
fact that some individuals may actively seek out arrests was positively correlated with the prestige
criminal labels resulting from conviction or accorded to criminal occupations. While a lack of
incarceration (Turner 1972). For example, crimi- legitimate work opportunities for the ex-inmates
nal justice involvement may offer prestige and might have accounted for this effect, it is also
confer a badge of honor upon individuals from possible that it stemmed, in part, from outright
certain segments of the population where such deance toward conventional society and its
experiences are a normal part of the life course. institutions (Sherman 1993).
Thus, far from being a social stigma, criminal Putting aside the numerous advantages to be
behavior and its legal consequences justice may acquired by risking involvement with the criminal
serve as a status symbol that legitimates an justice system, the mark of a felony conviction
offenders criminal accomplishments in the eyes or a prison sentence has the ironic effect of con-
of his or her peers. straining opportunities and therefore channeling
The two foregoing possibilitiescriminal jus- offenders into long-term criminality. While a
tice involvement as an occupational hazard and a felony conviction or prison sentence would be a
badge of honorare not necessarily mutually consequential event in the life course in any era,
exclusive. A theme that has emerged from inter- the modern era is characterized by an
views with active offenders is that, in their pur- overwhelming number of collateral consequences
suit of a party lifestyle, they commonly experience from which it is nearly impossible to escape
Criminal Justice and the Life Course 313

(American Bar Foundation 2015; Olivares et al. reduction benets, there are a growing number of
1996). Formal legal consequences bar felons empirical studies that document these diminish-
from employment, educational opportunities, ing returns to prison growth, indicating that incar-
family reintegration, public housing, various ceration is far less effective at preventing crime
forms of civic engagement, among others, and today than it was in the not-too-distant past
effectively restrict the number of potential path- (R. Johnson and Raphael 2012; Liedka et al.
ways to redemption. Informal legal consequences 2006; Spelman 2000). Relatedly, many studies of
of a felony conviction (stigma, collateral conse- the deterrent impact of incarceration relative to
quences, and exposure to widespread surveil- non-custodial sanctions indicate that, at best,
lance, among others) also limit the ability of incarceration has no relationship with criminal
former inmates to rebuild their lives away from recidivism and might under certain circumstances
the criminal justice system. Agency is thus more make recidivism more likely (Nagin et al. 2009).
limited and operates in countervailing ways for The potential to scale back our reliance on
those involved in crimeit rst brings people imprisonment as a response to crime control makes
into the system through individual-level choices it all the more critical that sociologists and crimi-
but the contours, breadth, and scope of criminal nologists do more to describe how and for whom
punishment today limits the ability of individuals imprisonment is most consequential. Estimates of
to fully move beyond that criminal past. Even the relationship between crime and incarceration
long after a criminal identity has been shrugged vary widely, suggesting that increases in the incar-
off, ex-offenders nd that their legitimate options ceration rate may have reduced crime by as little as
are severely limited. For convicted felons and zero and as much as 25 % but are hotly debated
former inmates in the current era, contact with (see, for example, reviews, research, and commen-
the criminal justice system carries with it a per- tary in Durlauf and Nagin 2011; Levitt 1996;
manent mark of dishonorable status. Wakeeld and Uggen 2010; Western 2006). At rst
glance, these estimates are surprisingly low given
the sheer scale of imprisonment in the United
5.1 Conclusion and Future States. One explanation for this is that while lock-
Challenges ing up large numbers of crime-prone young men
reduces crime, locking up so many young men is a
In this essay, we have used core concepts from strategy with rapidly diminishing returns and has
life-course sociology to examine the role of the the potential to increase crime through its collateral
criminal justice system in modern life. We end consequences.
our discussion by suggesting fruitful pathways Mass incarceration brought fundamental
for research in the area to move forward. Most changes in the types of law violators that are sent
importantly, we note a number of signs that sug- to prison. This is no doubt due in large part to the
gest the prison boom era is receding. The increase war on drugs that commenced in the early 1970s
in the incarceration rate has slowed (and even during the Nixon administration and accelerated
decreased in some states and at the federal level), during the Reagan administration of the 1980s.
signicant legal reforms in the last 10 years have Since that time, growth in the drug incarceration
reduced the scope and punitiveness of criminal rate has far outpaced growth in other crimes, con-
punishment, and the public is currently engaged tributing to 33 % of total prison growth (Blumstein
in a sustained discussion surrounding police and Beck 1999; Caulkins and Chandler 2006).
practices and issues of social justice. Put simply, once you have locked up the most
It is becoming clear that our experiment in violent, troublesome, and high-rate offenders,
mass incarceration is a failure, imposing too many further increases to the incarceration rate tend to
unintended consequences for social life and concentrate on less serious offenders, yielding
inequality for too small reduction in crime (Clear smaller crime benets.
and Frost 2013; National Research Council 2014; Yet because of the phenomenon of diminish-
Simon 2014). With respect to ever smaller crime ing returns, discussed above, these estimates are
314 S. Wakeeld and R. Apel

invalid for contemporary deliberation. We do not are substantial interventions, of course. Yet many
engage this debate in any great detail but, follow- of the reforms aimed at reducing the size and
ing Sampson (2011), suggest that even if one scope of the criminal justice system are limited to
takes the mean crime reduction estimate (~10 the low-hanging fruit of non-violent and low-
15 %) as fact, this is an enormously important level offenders. This is a worthy goal but it is
social intervention and likely to have profound worth noting that even if every drug offender in
effects on peoples lives. Given voluminous Americas prisons were released tomorrow, the
research demonstrating the harms associated United States would still have the highest incar-
with exposure to violence (e.g., Bingenheimer ceration rate in the world. Similarly, Californias
et al. 2005; Sharkey 2010; Singer et al. 1995), a prison population will remain a crushing problem
reduction in the violent crime rate on the order of for the state unless it allows for some discussion
15 % is not to be taken lightly. These potential of the release of elderly inmates who were con-
benets, however, must be weighed against the victed of violent crimes.
potential costs to social life that accompany such Today, a full third of all youth are arrested by
a high rate of incarceration. That mass incarcera- their early 20s (Brame et al. 2011), often for less
tion is associated with a host of damages to for- serious crimes. African American men who
mer inmates, their families, and their communities dropped out of high school are more likely to
arises from the fact that mass incarceration swept spend time in prison than to eventually obtain a
up too many citizens who might better have been college degree or spend time in the military
left in the community. (Western 2006). Paternal incarceration is a con-
The lack of research on heterogeneity in crimi- sequential cause of the intergenerational trans-
nal justice contact and its consequences is a glar- mission of disadvantage (Wakeeld and
ing weakness in the areawithout providing Wildeman 2013). Differences in risks across
information on who should go to prison, research- cohorts and demographic groups like these sug-
ers are out of conversation with policy makers in gest the potential for aggregate-level, cohort-,
how to most effectively reduce the prison popula- and period-specic effectsand evoke classic
tion while maintaining low crime rates. Similarly, life-course analyses on the long-term effects of
while it is important to call attention to the massive other institutional, economic, and political shifts
increases in social inequality brought about by the (e.g., Elder 1974, 1999; McAdam 1988;
prison boom (Goffman 2014; Wakeeld and McLanahan and Percheski 2008). While the
Uggen 2010; Western 2006), appeals to social jus- criminal justice system or incarceration experi-
tice have little chance for success if reductions in ences may have occupied a niche position in life-
the prison population or changes in policing prac- course sociology in the past, our review suggests
tices results in large increases to the crime rate. that it ought to occupy a central place going
Again, following Sampson (2011), we suggest that forwardalongside the family, the labor market,
research that better adjudicates between the conse- or social movementsas a force that powerfully
quences of incarceration for social life while also alters the lives of those who come into direct con-
accounting for its crime reduction impact is an tact with it, as well as those who do so only indi-
important consideration for future research. rectly because of who their family members are
We also see reason for tempering optimism or the communities in which they reside.
about recent sentencing reforms and [small]
reductions in the incarceration rate in some states.
The abolition of the death penalty and mandatory References
life without parole sentences for juveniles by the
Supreme Court (in 2005 and 2011, respectively), Alexander, M. (2010). The new Jim Crow: Mass incar-
ceration in the age of color blindness. New York: The
California Realignment (a program aimed to
New Press.
reduce state prison populations by 40,000) and its American Bar Foundation. (2015). National inventory of
reform of Three Strikes in the last election, or the collateral consequences of conviction. http://www.
recent reforms to federal drug sentencing laws abacollateralconsequences.org. Accessed 31 Jan 2015.
Criminal Justice and the Life Course 315

Anderson, E. (1999). Code of the street: Decency, vio- ment hypothesis. Journal of Criminal Law and
lence, and the moral life of the inner city. New York: Criminology, 70, 376390.
W. W. Norton. Braman, D. (2004). Doing time on the outside:
Aos, S., Miller, M., & Drake, E. (2006). Evidence-based Incarceration and family life in urban America. Ann
public policy options to reduce future prison construc- Arbor: University of Michigan Press.
tion, criminal justice costs, and crime rates. Olympia: Brame, R., Turner, M. G., Paternoster, R., & Bushway,
Washington State Institute for Public Policy. S. D. (2011). Cumulative prevalence of arrest from
Apel, R., & Sweeten, G. (2010). The impact of incarcera- ages 8 to 23 in a national sample. Pediatrics, 129(1),
tion on employment during the transition to adulthood. 2127.
Social Problems, 57(3), 448479. Brame, R., Bushway, S., Paternoster, R., & Turner, M. G.
Apel, R., Blokland, A. A. J., Nieubeerta, P., & van (2014). Demographic patterns of cumulative arrest
Schellen, M. (2010). The impact of imprisonment on prevalence by ages 18 and 23. Crime and Delinquency,
marriage and divorce: A risk set matching approach. 60, 471486.
Journal of Quantitative Criminology, 26, 269300. Brayne, S. (2014). Surveillance and system avoidance:
Arnett, J. J. (2000). Emerging adulthood: A theory of Criminal justice contact and institutional attachment.
development from the late teens through the twenties. American Sociological Review, 79(3), 367391.
American Psychologist, 55(5), 469480. Bureau of Justice Statistics. (2007). Survey of inmates of
Auerhahn, K. (2006). Selective incapacitation, three state and federal correctional facilities, 2004.
strikes, and the problem of aging prison populations: U.S. Department of Commerce, Bureau of the Census.
Using simulation modeling to see the future. Ann Arbor, MI: Inter-university Consortium for
Criminology & Public Policy, 1(3), 353388. Political and Social Research (producer and
Bachman, J. G., Wallace, J. M., OMalley, P. M., Johnston, distributor).
L. D., Kurth, C. L., & Neighbors, H. W. (1991). Caulkins, J. P., & Chandler, S. (2006). Long-run trends in
Racial/ethnic differences in smoking, drinking, and incarceration of drug offenders in the United States.
illicit drug use among American high school seniors, Crime and Delinquency, 52, 619641.
19761989. Journal of Public Health, 81(3), Clear, T. R. (2007). Imprisoning communities: How mass
372377. incarceration makes disadvantaged neighborhoods
Baumer, E. P. (2013). Reassessing and redirecting worse. New York: Oxford University Press.
research on race and sentencing. Justice Quarterly, 30, Clear, T. R., & Frost, N. A. (2013). The punishment imper-
231260. ative: The rise and failure of mass incarceration.
Bayer, P., Hialmarsson, R., & Pozen, D. (2009). Building New York: NYU Press.
criminal capital behind bars: Peer effects in juvenile Clemmer, D. (1940). The prison community. New York:
corrections. Quarterly Journal of Economics, 124, Rhinehart.
105147. Comfort, M. (2007). Punishment beyond the legal
Beck, A. J., & Maruschak, L. (2001). Mental health treat- offender. Annual Review of Law and Social Science,
ment in state prisons. Washington, DC: 3(1), 271284.
U.S. Government Printing Ofce. Comfort, M. (2008). Doing time together: Love and fam-
Beckett, K. (1997). Making crime pay: Law and order in ily in the shadow of the prison. Chicago: University of
contemporary American politics. New York: Oxford Chicago Press.
University Press. DiIulio, J. (1987). Governing prisons: A comparative
Beckett, K., Nyrop, K., & Pngst, L. (2006). Race, drugs, study of prison management. New York: The Free
and policing: Understanding disparities in drug deliv- Press.
ery arrests. Criminology, 44(1), 105137. DiIulio, J. (1995). The coming of the super-predator. The
Bingenheimer, J. B., Brennan, R. T., & Earls, F. J. (2005). Weekly Standard, pp 2328.
Firearm violence exposure and serious violent behav- Durlauf, S. N., & Nagin, D. S. (2011). Imprisonment and
ior. Science, 308(5726), 13231326. crime: Can both be reduced? Criminology and Public
Bishop, D. M. (2000). Juvenile offenders in the adult Policy, 10(1), 1354.
criminal justice system. Crime and Justice: A Review Elder, G. H. (1974, 1999). Children of the great depres-
of Research, 27, 81167. sion: Social change in life experience (25th anniver-
Blumstein, A., & Beck, A. J. (1999). Population growth in sary ed.). Boulder: Westview Press.
US prisons, 19801996. In M. Tonry & J. Petersilia Elder, G. H. (1998). The life course as developmental
(Eds.), Crime and justice: A review of research (Vol. theory. Child Development, 69(1), 112.
26, pp. 1761). Chicago: University of Chicago Press. Fagan, J., & Zimring, F. E. (Eds.). (2000). The changing
Blumstein, A., & Cohen, J. (1973). A theory of the stabil- borders of juvenile justice: Transfer of adolescents to
ity of punishment. Journal of Criminal Law, the criminal court. Chicago: University of Chicago
Criminology, and Police Science, 63, 198207. Press.
Blumstein, A., & Moitra, S. (1979). An analysis of the Federal Bureau of Investigation. (19602012). Crime in
time series of imprisonment rate in the states of the the United States. Uniform Crime Reports Series.
United States: A further test of stability of the punish- Washington, DC.
316 S. Wakeeld and R. Apel

Feld, B. C. (1990). The punitive juvenile court and the Jacobs, B. A., & Wright, R. (1999). Stick-up, street cul-
quality of procedural justice: Disjunctions between ture, and offender motivation. Criminology, 37(1),
rhetoric and reality. Crime and Delinquency, 36, 149174.
443466. Johnson, R., & Raphael, S. (2012). How much crime
Feld, B. C. (1997). Abolish the juvenile court: reduction does the marginal prisoner buy? Journal of
Youthfulness, criminal responsibility, and sentencing Law and Economics, 55(2), 275310.
policy. Journal of Criminal Law and Criminology, Johnson, E., & Waldfogel, J. (2004). Children of incarcer-
88(1), 68136. ated parents: Multiple risks and childrens living
Feld, B. C. (1999). Bad kids: Race and the transformation arrangements. In M. E. Patillo, D. F. Weiman, &
of the juvenile court. New York: Oxford University B. Western (Eds.), Imprisoning America: The social
Press. effects of mass incarceration (pp. 97131). New York:
Frost, N. A., & Clear, T. R. (2009). Understanding mass Russell Sage.
incarceration as a grand social experiment. In A. Sarat Kim, C. Y., Losen, D. J., & Hewitt, D. T. (2010). The
(Ed.), Studies in law, politics, and society (Vol. 47, school-to-prison pipeline: Structuring legal reform.
pp. 151191). Bingley: JAI Press. New York: NYU Press.
Garland, D. (2001a). The culture of control: Crime and Kreager, D. A., Schaefer, D. R., Bouchard, M., Haynie,
social order in contemporary society. New York: D. L., Wakeeld, S., Young, J., et al. (2015). Toward a
Oxford University Press. criminology of inmate social networks. Justice
Garland, D. (2001b). The meaning of mass imprisonment. Quarterly. 129
Punishment and Society, 3(1), 57. Kruttschnitt, C., & Gartner, R. (2005). Marking time in
Geller, A., Garnkel, I., Cooper, C., & Mincy, R. (2009). the golden state: Womens imprisonment in California.
Parental incarceration and childhood wellbeing: Cambridge: Cambridge University Press.
Implications for urban families. Social Science Lauritsen, J. L. (1998). The age-crime debate: Assessing
Quarterly, 90, 11861202. the limits of longitudinal self-report data. Social
Gelman, A., Fagan, J., & Kiss, A. (2007). An analysis of Forces, 77(1), 127154.
the New York City police departments stop-and- Levitt, S. (1996). The effect of prison population size on
frisk policy in the context of claims of racial bias. crime: Evidence from prison overcrowding and litiga-
Journal of the American Statistical Association, 102, tion. Quarterly Journal of Economics, 111, 319351.
813823. Liedka, R. V., Piehl, A. M., & Useem, B. (2006). The
Goffman, E. (1961). Asylums: Essays on the social situa- crime-control effect of incarceration: Does scale mat-
tion of mental patients and other inmates. Garden ter? Criminology and Public Policy, 5, 245276.
City: Anchor Books. Liu, S. (2015). Is the shape of the age-crime curve invari-
Goffman, A. (2014). On the run: Fugitive life in an ant by sex evidence from the national sample with
American city. Chicago: University of Chicago Press. exible non-parametric modeling. Journal of
Gonnerman, J. (2004, November 16). Million-dollar Quantitative Criminology, 31, 93123.
blacks: The neighborhood costs of Americas prison MacCoun, R., & Reuter, P. (1992). Are the wages of sin
boom. Village Voice. $30 an hour? An economic analysis of street-level
Gottfredson, M., & Hirschi, T. (1986). The true value of drug dealing. Crime and Delinquency, 38, 477491.
lambda would appear to be zero: An essay on career Manza, J., & Uggen, C. (2006). Locked out: Felon disen-
criminals, criminal careers, selective incapacitation, franchisement and American democracy. New York:
cohort studies, and related topics. Criminology, 24(2), Oxford University Press.
213234. Massoglia, M., & Uggen, C. (2010). Settling down and
Hagan, J. (1993). The social embeddedness of crime and aging out: Toward and interactionist theory of desis-
unemployment. Criminology, 31, 465491. tance and the transition to adulthood. American
Hirscheld, P. J. (2008). Preparing for prison? The crimi- Journal of Sociology, 116(2), 543582.
nalization of school discipline in the USA. Theoretical Massoglia, M., Remster, B., & King, R. D. (2011). Stigma
Criminology, 12(1), 79101. or separation? Understanding the incarceration-
Irwin, J. (1970). The felon. Berkeley: University of divorce relationship. Social Forces, 90(1), 133155.
California Press. Matsueda, R. L., Gartner, R., Piliavin, I., & Polakowski,
Irwin, J. (2005). The warehouse prison: Disposal of the M. (1992). The prestige of criminal and conventional
new dangerous class. Los Angeles: Roxbury occupations: A subcultural model of criminal activity.
Publishing. American Sociological Review, 57, 752770.
Jacobs, J. (1977). Stateville: The penitentiary in mass McAdam, D. (1988). Freedom summer. New York:
society. Chicago: University of Chicago Press. Oxford University Press.
Jacobs, D., & Helms, R. E. (1996). Toward a political McLanahan, S., & Percheski, C. (2008). Family structure
model of incarceration: A time series examination for and the reproduction of inequalities. Annual Review of
prison admission rates. American Journal of Sociology, Sociology, 34(1), 257276.
102, 323357.
Criminal Justice and the Life Course 317

Mitka, M. (2004). Aging prisoners stressing health care analysis tables for the pew public safety and mobility
system. Journal of the American Medical Association, project. Cambridge, MA: Harvard University.
292(4), 423424. Phelps, M. (2011). Rehabilitation in the punitive era: The
Moftt, T. E. (1993). Adolescence-limited and life- gap between rhetoric and reality in US prison pro-
course-persistent antisocial-behaviorA develop- grams. Law & Society Review, 45, 3368.
mental taxonomy. Psychological Review, 100(4), Ramakers, A., Apel, R., Nieubeerta, P., Dirkzwager, A., &
674701. Wilsem, J. V. (2014). Imprisonment length and post-
Mumola, C. J. (2000). Incarcerated parents and their chil- prison employment prospects. Criminology, 52(3),
dren. Washington, DC: Bureau of Justice Statistics, 499527.
US Government Printing Ofce. Reuter, P., MacCoun, R., & Murphy, P. (1990). Money
Myers, R. R., & Wakeeld, S. (2014). Sex, gender, and from crime: A study of the economics of drug dealing
imprisonment: Rates, reforms, and lived realities. In in Washington, D.C. Santa Monica: Rand Corporation.
R. Gartner & B. McCarthy (Eds.), The oxford hand- Roberts, J. V. (2004). Public opinion and youth justice.
book of gender, sex, and crime (pp. 572593). Crime and Justice: A Review of Research, 31, 495.
New York: Oxford University Press. Sampson, R. J. (2011). The incarceration ledger: Toward a
Nagin, D. S., Piquero, A. R., Scott, E. S., & Steinberg, L. new era in assessing societal consequences.
(2006). Public preferences for rehabilitation versus Criminology & Public Policy, 10, 819828.
incarceration of juvenile offenders: Evidence from a Sampson, R. J., & Laub, J. H. (1994). Crime in the mak-
contingent valuation survey. Criminology, 5(4), ing: Pathways and turning points through life.
627651. Cambridge, MA: Harvard University Press.
Nagin, D. S., Cullen, F. T., & Jonson, C. L. (2009). Sampson, R. J., & Loefer, C. (2010). Punishments
Imprisonment and reoffending. In M. Tonry (Ed.), place: The local concentration of mass incarceration.
Crime and justice (Vol. 38, pp. 115200). Chicago: Daedalus, 139, 2031.
University of Chicago Press. Schwartz, R., & Rieser, L. (2001). Zero tolerance as man-
National Commission on Correctional Health Care. datory sentencing. In W. Ayers, B. Dohrn, & R. Ayers
(2002). The health status of soon-to-be-released (Eds.), Zero tolerance: Resisting the drive for punish-
inmates. Chicago: National Commission on ment in our schools (Vol. 126136). New York: The
Correctional Health Care. New Press.
National Research Council. (2014). The growth of incar- Schwartz-Soicher, O., Geller, A., & Garnkel, I. (2011).
ceration in the United States: Exploring causes and The effect of paternal incarceration on material hard-
consequences. Committee on Law and Justice, ship. Social Service Review, 3, 447473.
Division of Behavioral and Social Sciences and Shanahan, M. J. (2000). Pathways to adulthood in chang-
Education. Washington, DC: National Academies ing societies: Variability and mechanisms in the life
Press. course perspective. Annual Review of Sociology, 26,
Neal, D., & Rick, A. (2014). The prison boom and the lack 667692.
of black progress after Smith & Welch (NBER Working Shannon, S. K. S., Uggen, C., Schnittker, J., Thompson, M.,
Paper No. 20283). Cambridge, MA. Wakeeld, S., & Massoglia, M. (2014). The growth,
Neugarten, B. L., Moore, J. W., & Lowe, J. C. (1965). Age scope, and spatial distribution of Americas criminal
norms, age constraints, and adult socialization. class, 19492010 (Working Paper). Minneapolis, MN:
American Journal of Sociology, 70(6), 710717. Department of Sociology, University of Minnesota.
Olivares, K., Burton, V. S., & Cullen, F. T. (1996). Sharkey, P. (2010). The acute effect of local homicides on
Collateral consequences of a felony conviction: A childrens cognitive performance. Proceedings of the
national study of legal codes 10 years later. Federal National Academy of Sciences, 107, 1173311738.
Probation, 60, 117. Sherman, L. W. (1993). Deance, deterrence, and irrele-
Page, J. (2004). Eliminating the enemy: The import of vance: A theory of criminal sanction. Journal of
denying prisoners access to higher education in Research in Crime and Delinquency, 30, 445473.
Clintons America. Punishment and Society, 6(4), Sherman, L. W., Gartin, P. R., & Buerger, M. E. (1989).
357378. Hot spots of predatory crime: Routine activities and
Pager, D. (2003). The mark of a criminal record. American the criminology of place. Criminology, 27, 2755.
Journal of Sociology, 108(5), 937975. Siegel, J. (2011). Disrupted childhoods: Children of
Petersilia, J. (2003). When prisoners come home: Parole women in prison. New Brunswick: Rutgers University
and prisoner reentry. New York: Oxford University Press.
Press. Simon, J. (2000). The society of captives in the era of
Pettit, B. (2012). Invisible men: Mass incarceration and hyper-incarceration. Theoretical Criminology, 4,
the myth of black progress. New York: Russell Sage. 285308.
Pettit, B., Sykes, B., & Western, B. (2009). Technical Simon, J. (2007). Governing through crime: How the war
report on revised population estimates and nlsy-79 on crime transformed American democracy and cre-
318 S. Wakeeld and R. Apel

ated a culture of fear. New York: Oxford University The challenge of becoming an adult. In D. W. Osgood,
Press. M. Foster, & C. Flanagan (Eds.), On your own without
Simon, J. (2014). Mass incarceration on trial: A remark- a net: The transition to adulthood for vulnerable pop-
able court decision and the future of prisons in ulations (pp. 114144). Chicago: University of
America. New York: New Press. Chicago Press.
Singer, M. I., Anglin, T. M., Song, L. Y., & Lunghofer, L. Vicusi, W. K. (1986). The risks and rewards of criminal
(1995). Adolescents exposure to violence and associ- activity: A comprehensive test of criminal deterrence.
ated symptoms of psychological trauma. Journal of Journal of Labor Economics, 4, 317340.
the American Medical Association, 273(6), 477482. Wacquant, L. (2000). The new peculiar institution: On
Skarbeck, D. (2014). The social order of the underworld: the prison as surrogate ghetto. Theoretical
How prison gangs govern the American penal system. Criminology, 4(3), 377389.
New York: Oxford University Press. Wacquant, L. (2001). Deadly symbiosis: When ghetto and
Skiba, R. J., Michael, R. S., Nardo, A. C., & Peterson, prison meet and mesh. Punishment and Society, 3(1),
R. L. (2002). The color of discipline: Sources of racial 95134.
and gender disproportionality in school punishment. Wakeeld, S. (2015). Accentuating the positive or elimi-
The Urban Review, 34(4), 317342. nating the negative? Paternal incarceration and care-
Smith, D. A. (1986). The neighborhood context of police giver-child parenting quality. Journal of Criminal Law
behavior. In J. Albert, J. Reiss, & M. Tonry (Eds.), and Criminology, 104(4), 905928.
Crime and justice: A review of research (Vol. 8, Wakeeld, S., & Uggen, C. (2010). Incarceration and
pp. 313341). Chicago: University of Chicago Press. stratication. Annual Review of Sociology, 36(1), 387
Sourcebook of Criminal Justice Statistics. (2014). Number 406. doi:10.1146/annurev.soc.012809.102551.
and rate (per 100,000 resident population in each Wakeeld, S., & Wildeman, C. (2011). Mass imprisonment
group) of sentenced prisoners under jurisdiction of and racial disparities in childhood behavioral problems.
state and federal corectional authorities on December Criminology & Public Policy, 10(3), 793817.
31. Washington, DC: U.S. Department of Justice, Wakeeld, S., & Wildeman, C. (2013). Children of the
Bureau of Justice Statistics. Available online at http:// prison boom: Mass incarceration and the future of
www.albany.edu/sourcebook/pdf/t6282009.pdf American inequality (Crime and public policy series).
Spelman, W. (2000). The limited importance of prison New York: Oxford University Press.
expansion. In A. Blumstein & J. Wallman (Eds.), The Welch, M., Price, E. A., & Yankey, N. (2002). Moral panic
crime drop in America (pp. 97129). New York: over youth violence: Wilding and the manufacture of
Cambridge University Press. menace in the media. Youth and Society, 34, 330.
Steiner, B., & Wooldredge, J. (2009). Rethinking the link Western, B. (2006). Punishment and inequality in
between institutional crowding and inmate miscon- America. New York: Russell Sage.
duct. The Prison Journal, 89(2), 205233. Western, B., & Beckett, K. (1999). How unregulated is the
Sykes, G. (1958/2007). The society of captives: A study of U.S. Labor market? The penal system as a labor mar-
maximum security prisons. Princeton: Princeton ket institution. American Journal of Sociology, 104(4),
University Press. 10301060.
Teplin, L. A. (1984). Criminalizing mental disorder: The Western, B., & Pettit, B. (2010). Incarceration & social
comparative arrest rate of the mentally ill. American inequality. Daedalus, 139(3), 819.
Psychologist, 39(7), 794803. Wildeman, C. (2009). Parental imprisonment, the prison
Terry v. Ohio (1968). 392: U.S. Supreme Court. boom, and the concentration of childhood advantage.
Turner, R. H. (1972). Deviance avowal as neutralization Demography, 46, 265280.
of commitment. Social Problems, 19, 308321. Wildeman, C. (2010). Parental incarceration and chil-
Turney, K. (2014). The consequences of paternal incar- drens physically aggressive behaviors: Evidence from
ceration for maternal neglect and harsh parenting. the fragile families and child wellbeing study. Social
Social Forces, 92, 16071636. Forces, 89, 285310.
Turney, K., & Haskins, A. R. (2014). Falling behind? Wildeman, C., & Muller, C. (2012). Mass imprisonment
Childrens early grade retention after paternal incar- and inequality in health and family life. Annual Review
ceration. Sociology of Education, 87(4), 241258. of Law and Social Science, 8, 1130.
Turney, K., & Wildeman, C. (2015). Detrimental for Wildeman, C., & Turney, K. (2014). Positive, negative, or
some? The heterogeneous effects of maternal incar- null? The effects of maternal incarceration on chil-
ceration on child wellbeing. Criminology & Public drens behavioral problems. Demography, 51,
Policy, 14, 125156. 10411068.
Tyler, T. R. (2006). Why people obey the law. Princeton: Wildeman, C., Schnittker, J., & Turney, K. (2012). Despair
Princeton University Press. by association? The mental health of mothers with
Uggen, C., & Wakeeld, S. (2005). Young adults reenter- children by recently incarcerated fathers. American
ing the community from the criminal justice system: Sociological Review, 77(2), 216243.
Criminal Justice and the Life Course 319

Wilson, J. Q., & Abrahamse, A. (1992). Does crime pay? tion, vocation, and work programs for adult offenders.
Justice Quarterly, 9, 359377. Journal of Research in Crime and Delinquency, 37,
Wilson, J. Q., & Kelling, G. L. (1982). Broken windows: 347368.
The police and neighborhood safety. Atlantic Monthly, Zimring, F. E. (2012). The city that became safe:
March, 249(3), 2938. New Yorks lessons for urban crime and its control.
Wilson, D. B., Gallagher, C. A., & MacKenzie, D. L. New York: Oxford University Press.
(2000). A meta-analysis of corrections-based educa-
Disaster and Life Course Processes

Jack DeWaard

1 Introduction the vulnerability, or potential for loss (Cutter


et al. 2003: 242), of persons and places to envi-
Disasters exist when the hazards, or forces of ronmental hazards.
harm (Schultz et al. 2007: 69), associated with Disasters are also routinely characterized as
an environmental event adversely affect the func- rapid-onset, rapid-desistance events (Kates et al.
tioning of persons and places.1 To the extent that 2006; Masten and Narayan 2012; Sampson and
persons and places are adversely affected by Laub 1996). However, if persons and places are
environmental hazards, it follows that they were adversely affected by environmental hazards for
vulnerable, or susceptible, to harm in the rst prolonged periods (Yule et al. 2000), then disas-
place (Adger 2006). It is well documented that ters must be understood as processes, not events.
persons and places are differentially vulnerable As processes that unfold in and over the life
to environmental hazards (Cutter 1996; Thomalla course, disasters encompass interconnected life
et al. 2006). The rst aim of this chapter is there- course moments and domains, and thus have the
fore to discuss and demonstrate the import of a potential to alter short- and long-run life course
life course perspective for studying differences in trajectories and outcomes. The second aim of this
chapter is to discuss and demonstrate this idea.
1
This chapter is organized as follows. First, I
The term environmental is used broadly throughout this
chapter to cover natural, technological, and man-made
provide a brief example to motivate the two ideas
events (Guha-Sapir et al. 2013). above. Second, I review and summarize past and
Readers may be struck by the broad scope of this deni- current research on disasters and the life course
tion; however, this is the denition of disaster used by a that attends to these ideas. While this review is not
number of research and relief organizations in the United exhaustive, it samples from multiple disciplines
States and worldwide, including the Center for Research
on the Epidemiology of Disasters (CRED), Doctors
and life course processes and domains. Third, in a
Without Borders, the Federal Emergency Management stylized demonstration, I illustrate these ideas in
Agency (FEMA), the Intergovernmental Panel on Climate the context of a simple dynamic population
Change (IPCC), the International Federation of Red Cross model. While not typically described as such, the
and Red Crescent Societies (IFRC), the Ofce of
U.S. Foreign Disaster Assistance (USAID), and the
model itself is a model of the life course, and
United Nations (U.N.), to name only a few. relates the timing of life course transitions to
J. DeWaard (*)
short- and long-run trajectories and outcomes. To
Department of Sociology, Minnesota Population connect my work in this demonstration to a real-
Center, University of Minnesota, world example, I consider several recent studies
Minneapolis, MN, USA on the potential impacts of climate change for
e-mail: jdewaard@umn.edu

Springer International Publishing Switzerland 2016 321


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_14
322 J. DeWaard

human migration, which suggest that disasters are which raises important questions about their pre-
not only rapid-onset, slow-desistance processes, disaster vulnerability. To illustrate, although inci-
but are in some cases slow-onset processes, as dence rates are difcult to come by, Rathore et al.
well (Leighton 2009; McLeman 2014; McLeman (2007) and Taquir et al. (2007) estimated that
and Hunter 2010). I conclude this chapter with a between 600 and 750 spinal cord injuries were
restatement of the two key ideas, followed by dis- sustained during this earthquake. Female home-
cussing potential avenues for future research on makers of middle age accounted for the majority
disaster and life course processes. of these victims. Not only does this gure repre-
sent an important life course transition from
uninjured to injured, Taquir et al. (2007) further
2 Motivating Example: showed that many of these cases developed into
The 2005 Kashmir paraplegia, which presented additional short-
Earthquake term complications, e.g., urinary tract infections,
pressure and stress ulcers, etc. In the long-run,
On October 8, 2005, an earthquake registering 7.6 Rathore et al. (2008) noted the scarcity of reha-
on the Richter scale and lasting 3045 s struck an bilitation resources in the region, which presents
area in the disputed Kashmir region northeast of a host of difculties for victims future physical
Muzaffarabad, Pakistan. As of November 2005, health, not to mention their functioning in other
an estimated 87,000 people had died as a result of life course domains, e.g., mental health, family
this earthquake. Another 138,000 and 3.5 million and work, etc.
people were injured and displaced, respectively The above example demonstrates that the
(Earthquake Engineering Research Institute 2006; 2005 Kashmir Earthquake was a disaster because
U.S. Geological Survey 2005). the hazards associated with this earthquake
There is likely universal agreement that the adversely affected the functioning of persons and
2005 Kashmir Earthquake was a disaster. places. Like other disasters, they did so unevenly
However, it was not a disaster because of the in accordance with the pre-disaster vulnerability
magnitude or duration or the earthquake. It was a of persons and places. Moreover, while the haz-
disaster because the hazards associated with this ards associated with this disaster occurred sud-
earthquake adversely affected the functioning of denly and lasted only seconds, the duration and
persons and places. This is evidenced by the desistance of this disaster were prolonged, espe-
number of people killed, injured, and displaced, cially for some (versus other) victims. A life
as well as by the extent of damage to housing and course perspective is uniquely suited for studying
infrastructure (Earthquake Engineering Research these ideas given a number of emphases, most
Institute 2006). Had the estimated 780,000 struc- especially the emphasis placed on the timing of
tures that were damaged or destroyed been able events and processes at specic stages of the life
to withstand this earthquake, for example, the course (Elder et al. 2003).
number of dead, injured, and displaced would
likely have been much lower. The implication,
then, is that had this earthquake not adversely 3 Disasters and the Life
affected the functioning of persons and places, it Course: A Selected Review
would not have been a disaster. This statement is and Summary of Previous
not intended to undermine the signicance of Research
what took place on October 8, 2005, most espe-
cially the loss of life, which is always unfortu- There are any number of ways to organize a
nate. Instead, it is meant to highlight the review and summary of previous research on
contingent and mostly social nature of disas- disasters and the life course. These include: by
ters (Alexander 2005: 27). disaster (Cherry 2009), disaster type (Guha-Sapir
Evidence suggests that persons and places et al. 2013), the characteristics of and differences
were differentially affected by this earthquake, between those adversely affected (Cherniack
Disaster and Life Course Processes 323

2008; Johnson and Galea 2009), and life course Cutter et al. 2003; Thomalla et al. 2006). This
domain (Johnson and Galea 2009; Shenk et al. directs attention to the characteristics and ori-
2009). Erring on the side of depth over breadth, I gins of differential vulnerability.
organize this review and summary into two parts. Past and current research has documented a
In the rst, I focus on studies that have applied range of salient characteristics and origins. To
(explicitly or implicitly) a life course perspective illustrate, consider the case of Hurricane Katrina.
to understand the origins of the differential vul- Studies have documented pronounced racial dif-
nerability of persons and places adversely ferences, for example, in the timing of evacua-
affected by environmental hazards. In the second, tions relative to when Hurricane Katrina made
I focus on studies that have applied a life course landfall (Elliott and Pais 2006), subsequent out-
perspective to understand disasters as processes migration from disaster-affected areas (Myers
that persist in time and desist slowly. As such, I et al. 2008), post-traumatic stress (Paxson et al.
do not claim that this review is exhaustive. 2012), return migration to disaster-affected areas
Instead, it is organized in such a way so that these (Fussell et al. 2010; Groen and Polivka 2010),
ideas are clearly illustrated within the context of and the repopulation of local areas and neighbor-
life course theory and empirical research. hoods (Chamlee-Wright and Storr 2009; Logan
2009; Pais and Elliott 2008). Ultimately, these
studies raise questions about the origins of these
3.1 Differential Vulnerability differences, e.g., past (and current) economic,
political, and social conditions that make people
In 2013, some 330 disasters worldwide were and places more or less susceptible to environ-
responsible for the deaths of 21,610 people; these mental hazards (Myers et al. 2008: 272).
disasters likewise adversely affected 96.5 million A life course perspective is uniquely suited for
people, and caused $118.6 billion in damages studying the differential vulnerability of persons
(Guha-Sapir et al. 2013). According to the Center and places to environmental hazards because it
for Research on the Epidemiology of Disasters locates differences in vulnerability in the life stage
(CRED), in order to qualify as a disaster, two of persons and places. Because life stages are char-
classes of environmental hazardsnatural and acterized by unique forms and accumulations of
technologicalmust be accompanied by one or human capital (broadly conceived) (KC et al.
more of the following: 10 or more fatalities, 100 2014), it follows that the origins of these differ-
or more people adversely affected, a declaration ences across life stages are diverse. Below, I illus-
of a state of emergency, or a call for assistance trate this in the context of three examples focusing
(Guha-Sapir et al. 2013).2 Excluding some types on selected sources of vulnerability in three life
of man-made hazards, e.g., wars or terrorist course stages. As mentioned above, this review is
attacks, environmental hazards are typically not intended to be exhaustive, but, instead, is
indiscriminate with respect to objects in their meant to take an in-depth look at selected sources
path(s). Disasters, however, reveal much about of vulnerability over the life course.
pre-existent (which is not to say primordial) char- To begin, consider past and current research
acteristics and differences that render some per- on the mental health of children and adolescents
sons and places more vulnerable, or susceptible, after a hazard event, compared to adults. Jones
to harm than others (Adger 2006; Cutter 1996; et al. (2009) conducted a review of 12 studies on
the effects of Hurricane Katrina on the mental
2
CRED further subdivides natural hazards into ve health of children and adolescents. Outcomes
groups: geophysical, meteorological, hydrological, clima- examined in these studies included intrusive
tological, and biological (Guha-Sapir et al. 2013). Others, thoughts, post-traumatic stress disorder (PTSD),
e.g., Logan and Xu (2011) and Schultz et al. (2007), focus
instead on the characteristics of hazards, both natural and
mood and anxiety disorders, depression, and
technological, with respect to such features as spatial somatic and behavioral problems. Likewise, sim-
scope, magnitude, duration, etc. ilar studies have been conducted on the effects of
324 J. DeWaard

Hurricane Katrina on the mental health of adults attachment to resources and resource conserva-
(Fussell and Lowe 2014; Paxson et al. 2012; tion (Cohan and Cole 2002; Fussell and Lowe
Rhodes et al. 2010). Outcomes considered in 2014), among others.
these studies also include PTSD, as well as other To illustrate, adopting a conservation of
forms of psychological distress. Generally, the resources framework (Hobfoll 1989), Fussell and
above studies show that Hurricane Katrina nega- Lowe (2014) examined the impact of housing
tively affected the mental health of children and displacement during and after Hurricane Katrina
adolescents, as well as of adults, with the severity on the subsequent mental health of low-income
of these effects appearing to diminish over time parents. The central idea underlying this study is
and conditioned by a number of aggravating and that [real or perceived] threats to already scarce
protective factors at the individual-, household-, resources, e.g., housing and material well-being,
and community-levels, including poverty, social are associated with secondary forms of stress,
support, local healthcare infrastructure, etc. e.g., PTSD, after a hazard event that go beyond
(Berggren and Curiel 2006; Weems and primary forms of stress that are associated with
Overstreet 2009). directly experiencing the hazard event itself. In
In many of the studies above, mental health analyzing three housing displacement proles,
outcomes are assessed in the context of a dose- Fussell and Lowe (2014: 138) showed that, rela-
response framework (Masten and Osofsky 2010), tive to low-income parents who returned home
where the mental health outcome of interest (the after Hurricane Katrina, low-income parents who
response) is situated in relation to the number, relocated (or were relocated) and those who were
intensity, or severity of hazards (the dose) unstably housed experienced signicantly ele-
(Masten and Narayan 2012: 234). While these vated levels of stress. Like the studies reviewed
studies provide convincing evidence that by Jones et al. (2009), this nding demonstrates
Hurricane Katrina affected the subsequent men- that Hurricane Katrina adversely affected the
tal health of children and adolescents, as well as functioning of low-income parents (through the
of adults, relatively few of these studies consider medium of housing displacement), however, it
the reasons why. This is an important consider- did so because low-income parents (and adults in
ation because, to the extent that persons were general) are attached to and seek to conserve
adversely affected by Hurricane Katrina, they resources, which are often scarce. In a different
were vulnerable, or susceptible, to harm in the setting (Hurricane Hugo) and at a different spa-
rst place (Adger 2006). What are the origins of tial scale (counties in South Carolina), Cohan and
this vulnerability? And how do these origins dif- Cole (2002: 16) effectively make the same argu-
fer for individuals at different life stages? ment, suggesting that the origins of vulnerability
Among children and adolescents, one impor- to environmental hazards in adulthood funda-
tant reason why the above effects were observed mentally involve processes of attachment to
is on account of the fact that children and adoles- scarce resources.
cents are undergoing a number of highly sensi- To briey step back at this point, the fact
tive transitions in the areas of cognition and that vulnerability to environmental hazards dif-
interpretation of experiences, emotion and emo- fers across life stages ultimately suggests that
tional understanding, self-regulation skills, the framework of dose-response is insufcient
knowledge, social connections and relationships, for analyzing outcomes after a hazard event.
physical size and strength, beliefs and faith, etc., This is so because observed outcomes reect
that are known to be consequential for both short- changes in response to the hazard event itself,
and long-run trajectories and outcomes (Masten as well as the extent to which persons and
and Narayan 2012: 241; see also Buchanan et al. places were vulnerable to begin with. This
2009). In contrast, with respect to adults, the ori- points to the need for an encompassing frame-
gins of vulnerability to environmental hazards work that situates dose-response in a more
are different, and seem to lie in processes of holistic life course sequence that considers
Disaster and Life Course Processes 325

pre-disaster vulnerability, e.g., Ingram and mental health outcomes among children and ado-
Luxtons (2005) vulnerability-stress model. lescents after Hurricane Katrina, Olteanu et al.
This observation is made especially clear in (2011) documented the importance of pre-
studies of elderly adults. Cherniack (2008) locates disaster psychological distress. Likewise, in the
the origins of vulnerability to environmental haz- study discussed earlier by Fussell and Lowe
ards at this life stage in the pre-disaster physical (2014), the most important predictor of post-
and mental health of elderly adults. About 80 % of disaster mental health among low-income par-
elderly adults (persons age 65 or older) in the ents was pre-disaster mental health. McFarlane
United States have a least one chronic condition; (1988) made a similar observation in his study of
some 50 % of elderly adults in the United States reghters exposed to a bushre disaster in South
have two or more chronic conditions (Aldrich and Australia in 1983, and demonstrated that three
Bensen 2008). These gures raise a number of pre-disaster variablesadverse life events,
important observations about the vulnerability of avoidance of thinking about problems, and past
elderly adults to environmental hazards. First, psychological treatmentwere strong predictors
elderly adults are more dependent on the avail- of post-disaster acute, chronic, and delayed-onset
ability, consistency, and quality of medical care. negative mental health outcomes. The impor-
Second, the nature of these dependencies depends tance of pre-disaster characteristics has also been
on a number of critical intersections, including shown in studies of psychological resilience after
the costs and administrative hurdles associated the terrorist attacks on September 11, 2001
with securing care, as well as the difculties that (Bonanno et al. 2007).
come with doing so with multiple chronic condi- By understanding the uniqueness and inter-
tions, especially those related to mental health connectedness of life stages, a life course per-
(Fultz et al. 2003). It is also well documented that spective is well-suited for studying differential
chronic conditions often lead to functional limita- vulnerability to environmental hazards and, thus,
tions and long-term disability (Vogeli et al. 2007). disasters for what they are, namely social phe-
This presents additional challenges, especially nomena (Perry 2007: 10).
when caring is provided in the context of family
relationships, i.e., linked lives, involving multi-
ple generations (Elder et al. 2003: 13). Given 3.2 Disasters as Slow-Desistance
these complex origins of vulnerability, it is there- Processes
fore not surprising that elderly adults are dispro-
portionately adversely affected by environmental Disasters are often characterized as rapid-onset,
hazards because they are relatively less equipped rapid-desistance events (Kates et al. 2006;
to maintain and/or seek out and secure new forms Masten and Narayan 2012; Sampson and Laub
of care during and after the hazard event itself 1996). While some types of disasters might t
(Cherniack 2008). this prole, as Gallopn (2006) notes, these char-
A life course perspective is uniquely suited for acteristics typically belong to the hazards, or the
locating differences in vulnerability to environ- forces of harm (Schultz et al. 2007: 69), that
mental hazards in the life stage of persons and are associated with environmental events, and are
places, and subsequently isolating the origins of, not essential properties of disasters themselves.
i.e., reasons for, these differences in operative life As social phenomena, disasters exist only when
course processes, e.g., developmental transitions persons and places are adversely affected by
and trajectories in childhood and adolescence. environmental hazards. To the extent that persons
An important implication for empirical research, and places are adversely affected for prolonged
then, is the need to thoroughly inventory and ana- periods, disasters must be understood as pro-
lyze the pre-disaster characteristics of persons cesses, not events.
and places (Jones et al. 2009; Masten and The suggestion that disasters are processes
Narayan 2012). For example, in an assessment of that persist in time and desist slowly intersects
326 J. DeWaard

with the concept of resilience (Cutter et al. 2008; hoods, communities, cities, etc., are frequently
Gallopn 2006; Paton and Johnston 2006), which decimated by environmental hazards. This
the IPCC (2012: 563) denes as the abilityto includes damage to structures and physical infra-
anticipate, absorb, accommodate or recover from structure (Burton and Hicks 2005; Nigg et al.
a hazardous event. A recent study by Kates et al. 2006), as well as to the social fabric of places
(2006) of the timing of reconstruction in New held together by networks of trust, mutual sup-
Orleans after Hurricane Katrina nicely illustrates port, reciprocity, and collaboration (Sampson and
the distinction between disaster and resilience Groves 1989: 780). Ethnographic research sug-
processes. Kates et al. (2006) identied three gests that the damage experienced at the level of
consecutive and overlapping periods to charac- place is an important factor in the persistence of
terize the time after Hurricane Katrina: emer- disasters in the lives of individuals. Chamlee-
gency, restoration, and reconstruction. Each Wright and Storr (2009: 621), for example, noted
period is illustrated with a density plot, wherein that decisions to return to the Ninth Ward in New
the x-axis corresponds to the number of weeks Orleans after Hurricane Katrina were strongly
after Hurricane Katrina and the y-axis corre- inuenced by bundle[s] of characteristics
sponds to the level of activity. While the latter attributed to home.
two periodsrestoration and reconstruction Other evidence also points to the fact that
are clearly rooted in the idea of resilience, and disasters can persist in the lives of individuals.
include such activities as the return of displaced Marmar et al. (1999), for example, surveyed 322
residents, restoration of levees, and issuing of rescue workers who were on duty during the col-
building permits, the rst periodemergency lapse of Interstate 880 in the 1989 Loma Prieta
is somewhat different. Earthquake. Peritraumatic dissociation and emo-
Kates et al. (2006: 14656) describe the emer- tional distress were assessed 1.9 and 3.5 years
gency period as one of substantial dewatering, out. They found that while most rescue workers
further accompanied by the completion of search showed improvement over time, those with
and rescue activities. Importantly, while the den- greater exposure to the collapse were at higher
sity plots for the restoration and reconstruction risk of chronic symptomatic distress. Likewise,
periods exhibit pronounced left-skews, the den- in a study of 217 survivors of the 1988 sinking of
sity plot for the emergency period is strongly the Jupiter, Yule et al. (2000) found that one-third
right-skewed, suggesting a transition from a pro- of survivors who had developed PTSD 1 year
cess that is ending (emergency) to one that is after this event were still suffering from PTSD
beginning (restoration). It is also illustrative that 58 years later. Similar ndings have also been
the respective right and left tails of these plots documented with respect to Hurricane Katrina
overlap. With respect to exactly what is ending in (Paxson et al. 2012).
the emergency period, this is largely left to inter- The above studies demonstrate that disasters
pretation. Of course, recalling from earlier in this persist in time depending on a number of factors.
chapter that disasters exist by virtue of the fact However, these studies do not pinpoint an exact
that persons and places are adversely affected by moment at which disasters desist. That said, it is
environmental hazards, it seems reasonable to reasonable to think that disasters do desist slowly,
interpret the emergency period as a process of and, in many cases, cease to exist. Evidence for
disaster persistence and, eventually, desistance this idea comes from a number of sources, includ-
that overlaps, but is nonetheless distinct from, ing memory research. Among elderly adults,
processes of resilience, e.g., restoration and Shenk et al. (2009) noted that isolated, i.e., occa-
reconstruction. sional, memories of disasters are associated with
Having made the above distinction, disasters positive ideation, e.g., the stimulation of other
persist in the time and desist slowly because the positive memories and memory retention.
adverse effects of environmental hazards are Likewise, in a meta-analysis of 25 studies of
often prolonged. Places, e.g., homes, neighbor- depression in elderly adults, Kraaij et al. (2002)
Disaster and Life Course Processes 327

examined depression as a function of the total system of region-to-region migration ows


number of negative life events, broken down into before, during, and after a hazard event. While I
several classes: death of signicant others, severe focus on migration for the reasons described
illness (on the part of self and others), negative below, it is important to point out that this model
socioeconomic circumstances, negative relation- can be applied to any process that involves a set
ships, and sudden unexpected events. Sudden of transitions among states, e.g., health and men-
unexpected events included disasters and war- tal health states (Palloni 2001; Yang and Waliji
time events, crime, car accidents, etc. Importantly, 2010), family and household contexts (Bumpass
they found that sudden life events were the only and Lu 2000), etc.
negative life eventsnot to be related to depres- To connect my work in this demonstration to a
sion scores (Kraaij et al. 2002: P91). substantive example, I consider several recent
Although past and current research has yet to studies on the potential impacts of climate change
determine exactly when and under what condi- for human migration (Leighton 2009; McLeman
tions disasters desist, the above studies indicate 2014; McLeman and Hunter 2010). These studies
that disasters do, in many cases, desist, with are rooted in a broader class of literature on envi-
existing research suggesting that the nature of ronmental migration, i.e., migration that is due to
desistance is likely slow (versus rapid) depending environmental factors and/or changes such as cli-
on whether and the extent to which adverse mate change (Bates 2002; El-Hinnawi 1985), that
effects associated with environmental hazards are seeks to incorporate features of the natural envi-
prolonged. ronment, e.g., rainfall decits (Nawrotzki et al.
2013) and natural capital, e.g., access to land and
landholdings (Hunter et al. 2014), into canonical
4 Disaster and Life Course migration accounts, e.g., push-pull theories of
Processes: A Demonstration neoclassical micro- and macro-economics (Black
et al. 2011, 2013; Findlay 2011). With respect to
In this section of the chapter, I provide a stylized the aims of this demonstration, these studies sug-
demonstration in which I illustrate the two ideas gest that persons and places are threatened by
discussed above in the context of a dynamic popu- both slow- and rapid-onset climate disasters
lation model, with roots in multiregional, or multi- (from droughts to oods to hurricanes to land-
state, demography (Rogers 1975, 1995; Palloni slides) (Leighton 2009: 337). Subsequently tak-
2001; Schoen 1988). Although not usually ing this observation into account, I conclude this
described as a model of the life course, the model demonstration by further showing that disaster
itself has a number of properties that are consistent processes that start and desist slowly (i.e., slow-
with the paradigmatic principles of life course onset, slow-desistance) are the most consequen-
theory (Elder et al. 2003: 10). Among these, the tial for long-run life course outcomes.
model relates the timing of life course transitions
to short- and long-run trajectories and outcomes.
In this demonstration, I show that (1) differen- 4.1 A Model of Migration over
tial vulnerability to environmental hazards across the Life Course
life course stages has different implications for
long-run life course outcomes in the presence of There are two primary outcomes of interest in
an environmental event that is experienced at a this demonstration. These include:
given age (or stage) in the life course, and (2) that
T0ii
disaster processes that persist in time and desist e0ii = (1)
slowly are more consequential for long-run life l 0i
course outcomes than rapid-onset, rapid desis-
T0ij
tance disaster events. These two ideas are illus- e0ij = (2)
trated in this demonstration in the context of a l 0i
328 J. DeWaard

Each outcome is a conditional life expectancy at l (0 + n ) = l (0) P (0) (5)


birth, and summarizes the average total number
of years that members of a birth cohort could The process in (5) then repeats consecutively
expect to live in, e0ii, and outside of, e0ij, the across all age intervals. Person-years lived in
region in which they were born over the course of each region over the life course are calculated
their lives. Each measure is constructed by track- from the population vectors at each age, e.g., l(0),
ing the migration transitions and deaths among l ( 0 + n ) , l ( 0 + 2n ) , etc., under the assumption
members of a cohort born in region i at each age that migration transitions and deaths occur at the
over their lives. The numerators in (1) and (2) midpoint of each age interval (Palloni 2001).
refer to the total number of person-years lived in These quantities are divided through by the size
regions i and j, respectively. The denominators of the birth cohort, l0i, to arrive at the summary
refer to the number of persons born in region i, measures in (1) and (2).
i.e., persons eligible to contribute person-years. Data for the P(x) matrices are taken from
Accordingly, the measures in (1) and (2) can be Wilson (2010) and the Australian Bureau of
viewed as long-run life course outcomes that Statistics, and are shown in Fig. 1. Migration data
reect transitions involving two life course pertain to male in-migration to Canberra,
domains (migration and to death, discussed in Australia, between 2005 and 2006. Mortality
detail below) at multiple life course moments data pertain to male deaths in southern Australia
(ages). Clearly, the model could be adapted to between 2004 and 2006.
accommodate any number and type of life course In this demonstration, I refer to the age sched-
domains. ules, or proles, of migration and mortality dis-
To estimate the quantities in (1) and (2), one played in Fig. 1 as typical because they exhibit
starts by considering a birth cohort in region i at several well-known regularities. Migration, for
exact age zero: example, typically exhibits peaks at early child-
hood, at entry into the labor force, and at retire-
l ( 0 ) = l 0i l 0j l 0d (3) ment (Rogers and Castro 1981). Mortality
likewise exhibits a bathtub shape, such that
where l 0i > 0 and l 0j = l 0d = 0 . This cohort is probabilities of dying are relatively high in the
then exposed to the prevailing age-specic rst year of life, decline thereafter, and increase
risks (probabilities) of migrating and dying, sum- rapidly at older ages.
marized in the transition matrix, P(0):
n p0ii p0ij p0id

n n 4.2 Disasters
jd
P ( 0 ) = n p0ji n p0
jj
n p0 (4)
0 0 1 In this section, I discuss how a rapid-onset, rapid-

desistance disaster event is incorporated into the
Each off-diagonal element in (4) denotes the model, and subsequently modied to approxi-
probability of migrating from one region (i or j) mate a disaster process that is rapid-onset,
to another or dying (d) between age 0 and age slow-desistance, followed by slow-onset, slow-
0 + n, where n is the width of the age interval. The desistance. In doing so, building on my discus-
elements in the main diagonal denote the proba- sion thus far in this chapter, the primary intuition
bilities of not migrating or dying. The nal row is that disasters exist by virtue of the fact that
expresses the fact that death is absorbing. they adversely affect persons (i.e., a birth cohort)
The process itself follows a simple rst-order in the model. Because I am considering two pro-
Markov process, such that the number of persons cesses, migration and mortality, this means that a
in each region, as well as the number of persons disaster exists in the model by disrupting the
who have died, at age 0 + n is contained in the typical age proles of migration and mortality,
updated population vector, l ( 0 + n ) : shown earlier in Fig. 1.
Disaster and Life Course Processes 329

.07
Probability
.035
0

0 10 20 30 40 50 60 70 80
Age
Migration Mortality

Fig. 1 Typical age schedules of migration and mortality

From a life course perspective, the extent of Accordingly, I proceed in exploratory fashion,
these disruptions depends on the stage in the life and consider what exposure, sensitivity, and
course when the disaster is experienced. Life adaptive capacity might look like at different
course stages are characterized by different risks stages in the life course. Envisioning each of
(probabilities) of migrating and dying, as well as these components on a scale ranging from zero to
by differential vulnerability to environmental one, rst, I assume that exposure increases as a
hazards. To operationalize vulnerability in the linear function of age. This is to say that older
model, I consider the denition used by the IPCC persons have greater previous exposure to envi-
(2012) and subsequently elaborated by McLeman ronmental hazards than younger persons. Next, I
(2014), which situates vulnerability as the prod- assume that persons near the beginning and end
uct of three factors: exposure, sensitivity, and of their lives are the most sensitive to environ-
adaptive capacity. Exposure refers to the pres- mental hazards. I therefore model sensitivity as
ence of persons and places that could potentially an exponentially decreasing function from age
be adversely affected by environmental hazards. zero to middle age, and as an exponentially
Sensitivity refers to the idea that some persons increasing function from middle age to the oldest
and places are more likely to be adversely age, which is 100 in this demonstration. I model
affected by environmental hazards than others. adaptive capacity in the opposite way, as an expo-
And adaptive capacity refers to the resources nentially increasing function from age zero to
availablethat can be used to prepare for and middle age, and as an exponentially decreasing
undertake actions to minimize the extent of from middle age to oldest age. I then calculate
harm potentially caused by environmental haz- vulnerability at each age as an average of expo-
ards (IPCC 2012: 556). sure, sensitivity, and the reciprocal of adaptive
Neither the IPCC (2012) nor McLeman (2014) capacity. Vulnerability scores are displayed in
provide a functional form of vulnerability. Fig. 2.
330 J. DeWaard

1
Vulnerability
.5
0

0 25 50 75 100
Age

Fig. 2 Vulnerability scores by age

Given the specications above, older persons enced at an early age increases the probability of
are more vulnerable to environmental hazards migration, which is already quite high to begin
than younger persons. However, those in the ear- in the typical schedule. This explains why the
liest stages of life are more vulnerable to environ- probabilities of migrating at older ages in the
mental hazards than those at, say, middle age. disaster age prole are hardly elevated at all,
The fact that vulnerability does not reach zero relative to corresponding probabilities in the
indicates that persons in all life course stages are, typical schedule. That is, despite the fact that
to some degree or another, vulnerable to environ- older persons are vulnerable to environmental
mental hazards. hazards (see Fig. 2), older persons are at a low
Using the vulnerability scores at each age, nvx, risk of migrating in the rst place per the typical
I then construct a set of alternative age schedules age prole. In contrast, the disaster age prole of
of migration and mortality. Hereafter, I refer to mortality indicates that older persons are at
these schedules as disaster age proles of migra- much greater risk of dying in the presence of a
tion and mortality. These are constructed by disaster because they are vulnerable to environ-
weighting the age-specic probabilities of mental hazards and at a greater risk of dying to
migrating and dying in the typical age proles, as begin with.
shown in Fig. 1, by a factor of: 1+ n v x . The typi- With this background in place, I consider the
cal and disaster age proles of migration and impact of a rapid-onset, rapid-desistance disaster
mortality are displayed in Fig. 3. event that is experienced at age x, and only age x,
Given the specications above, as is evident on the conditional expectations of life in (1) and
in Fig. 3, relative to the typical age prole of (2). This is achieved by substituting the probabil-
migration, the disaster age prole of migration is ities of migrating and dying from the disaster age
more pronounced at the earliest ages because proles, e.g., nDpxij and nDpxid, where D denotes
younger persons are vulnerable to environmental disaster, into the typical age proles for the
hazards, and an environmental event experi- corresponding probabilities, e.g., npxij and npxid,
Disaster and Life Course Processes 331

.07
Probability
.035
0

0 10 20 30 40 50 60 70 80
Age
Migration: Typical Schedule Mortality: Typical Schedule
Migration: Disaster Schedule Mortality: Disaster Schedule

Fig. 3 Typical and disaster age schedules of migration and mortality

respectively. Accordingly, a rapid-onset, rapid- taneously) return to the levels in the typical age
desistance disaster event approximates a shock, schedules, and so approximate a rapid-onset,
or disruption, to the typical age proles of migra- slow-desistance disaster process.
tion and mortality at age x. I consider the impact Finally, in recalling my earlier discussion of
of a rapid-onset, rapid-desistance disaster event the potential impacts of climate change for
at each and every age, one at a time, on the con- human migration, particularly the fact that per-
ditional expectations of life in (1) and (2), and sons and populations are threatened by both
subsequently graph these results. slow- and rapid-onset climate disasters
Next, I consider the impact of a rapid-onset, (Leighton 2009: 337; see also McLeman 2014;
slow-desistance disaster process. This is achieved McLeman and Hunter 2010), I conclude this
by carrying out the same substitution described demonstration by considering the impact of a
above at exact age x, followed by making similar slow-onset, slow-desistance disaster process that
substitutions at all subsequent ages. To illustrate, peaks at age x on the conditional expectations
given a disaster that begins at age 65, for those in of life in (1) and (2). This is achieved by applying
region i at exact age 66, the probability of migrat- the procedure described in the previous para-
ing from region i to region j, np66ij, is weighted by graph to all ages before and after the peak of the
D disaster.
n p66ij - n p66ij
a factor of . Likewise, the probabil-
66 - 65
ity of migrating from region i to region j at age 4.3 Results
D ij ij
n p73 - n p73
73, np73ij, is weighted by a factor of . Based on the typical age proles of migration and
73 - 65 mortality displayed in Fig. 1, the average number
In this way, the age-specic probabilities of of years that members of a birth cohort could
migration and mortality gradually (versus instan- expect to live in the region in which they were
332 J. DeWaard

born is e0ii = 57.95 . Persons could likewise erated from this demonstration that are of primary
expect to live e0ij = 21.32 years outside of their interest. In particular, the results displayed in Fig.
region of birth. These two estimates serve as the 4 evidence how three factorsthe risks of migra-
baselines against which all subsequent gures tion and mortality, vulnerability to environmental
will be compared. Additionally, as a quality hazards, and a disaster event that is slow-onset,
check, because e0ii and e0ij are additive (Rogers slow-desistancecombine to alter long-run life
1975, 1995; Palloni 2001; Schoen 1988), they course outcomes. Moreover, exactly how these
should sum to total life expectancy at birth, as three factors combine depends on when in the life
reported by the Australian Bureau of Statistics. course they actually come together.
The Australian Bureau of Statistics reports a g- In Fig. 5, I consider the impact of a rapid-
ure to one decimal point of 79.3 years. My esti- onset, slow-desistance disaster process that
mates sum to 57.95 + 21.32 = 79.27 years. occurs at age x and subsequently desists slowly
In Fig. 4, I consider the impact of a rapid- after age x on the conditional expectations of life
onset, rapid-desistance disaster event that occurs in (1) and (2). Like in Fig. 4, the estimates dis-
at age x on the conditional expectations of life in played in Fig. 5 are differences from the baseline
(1) and (2). The estimates presented in Fig. 4 are estimates of e0ii = 57.95 years and e0ij = 21.32
differences (in years) in the conditional expecta- years above.
tions life that include a rapid-onset, rapid- Relative to a rapid-onset, rapid-desistance
desistance disaster event experienced at age x and disaster event, a disaster process that is rapid-
the corresponding baseline estimates of onset, slow-desistance results in considerably
e0ii = 57.95 years and e0ij = 21.32 years above. less time lived in ones region of birth. This is
Given the stylized nature of this demonstra- especially the case for those at the earliest ages,
tion, I am not concerned with the magnitudes of and, to a lesser extent, those at peak working age
the differences shown in Fig. 4 because these will and the elderly. For those at the earliest ages and
vary depending on the set of inputs, i.e., typical at peak working age, much of these losses in the
and disaster age proles of migration and mortal- amount of time that could be expected to be lived
ity, used. That said, to walk through an example in ones region of birth is accounted for by gains
of how to read the differences in Fig. 4, consider lived outside of ones region of birth. In contrast,
a rapid-onset, rapid-desistance disaster event that among the elderly, elevated mortality reduces the
occurs at, say, age two. Over the life course, this number of years that could be expected to be
will reduce the number of years that persons lived both in and outside of ones region of birth.
could be expected to live in the region in which In substantive terms, relative to rapid-onset,
they were born by one-fth of 1 year, and increase rapid-desistance disaster events, disaster pro-
the number of years lived outside of the region in cesses that persist in time and desist slowly are
which they were born by slightly less than one- more consequential for long-run life course out-
fth of 1 year. Similarly, a rapid-onset, rapid- comes because the adverse effects of a hazard
desistance disaster event that occurs at age 25 event are effectively prolonged. While, in this
will result in changes of 0.04 and +0.03 years, demonstration, I am concerned with disruptions
respectively. The story is somewhat different for to the typical age proles of migration and mor-
persons who experience a rapid-onset, rapid- tality, the insights generated from Fig. 5 (and Fig.
desistance disaster event at older ages on account 4) apply to any life course process, e.g., marriage
of higher mortality, which reduces the number of and fertility (Cohan and Cole 2002), physical and
years that they could be expected to live both in mental health (Cherniack 2008; Yule et al. 2000),
and outside of their region of birth. etc. As is shown in Fig. 5, while it is common to
Beyond the magnitudes of the differences dis- envision disasters as instantaneous events (Kates
played in Fig. 4, it is the substantive insights gen- et al. 2006; Masten and Narayan 2012; Sampson
Disaster and Life Course Processes 333

1
Change in Years
0 -1

0 10 20 30 40 50 60 70 80 90 100
Age at Peak of Disaster
In Region of Birth Outside Region of Birth

Fig. 4 Differences (from baseline) in conditional life expectancies at birth by age at peak of disaster: rapid-onset,
rapid-desistance
1
Change in Years
0
-1

0 10 20 30 40 50 60 70 80 90 100
Age at Peak of Disaster
In Region of Birth Outside Region of Birth

Fig. 5 Differences (from baseline) in conditional life expectancies at birth by age at peak of disaster: rapid-onset,
slow-desistance
334 J. DeWaard

and Laub 1996), to the extent that the adverse tion, correspond to ages in the life course when
effects of a hazard event are prolonged, there are migration and mortality are pronounced.
consequential implications for long-run life
course outcomes, which clearly depend on the
life course stage in which these effects are 5 Discussion
experienced.
As such, in further considering the impacts of In adopting the view that disasters are inher-
slow-onset, slow-desistance disaster process, one ently social phenomena (Perry 2007: 10), in
should expect to see greater uctuations around this chapter, I developed and subsequently dem-
the baseline estimates of e0ii = 57.95 years and onstrated two ideas. First, because disasters
e0ij = 21.32 years because the adverse effects of a exist to the extent that environmental hazards
hazard event encompass an even greater portion of adversely affect the functioning of persons and
persons lives. These results are shown in Fig. 6. places, I considered the import of a life course
Clearly, slow-onset, slow-desistance disaster perspective for understanding how persons and
processes amount to extended perturbations to places are differentially vulnerable to such haz-
life course transitions, and thus to long-run life ards in the rst place. I focused especially on
course outcomes (Gallopn 2006: 294). As is evi- differential vulnerability by life course stage. I
dent, relative to at older ages, disasters are more further considered selected origins of differen-
consequential when experienced at earlier ages; tial vulnerability at various stages in the life
moreover, disasters are especially consequential course. In the process, I suggested that prevail-
when they intersect stages and moments in the ing conceptual and empirical models, e.g., dose-
life course when persons and places are particu- response, must be situated in a broader
larly vulnerable to environmental hazards and framework that includes consideration of pre-
undergoing sensitive transitions (Masten and disaster vulnerability, e.g., see Ingram and
Narayan 2012: 241), which, in this demonstra- Luxton (2005).
1
Change in Years
0
-1

0 10 20 30 40 50 60 70 80 90 100
Age at Peak of Disaster

In Region of Birth Outside Region of Birth

Fig. 6 Differences (from baseline) in conditional life expectancies at birth by age at peak of disaster: slow-onset,
slow-desistance
Disaster and Life Course Processes 335

Second, because disasters potentially encom- challenge from the vantage point of life course
pass interconnected life course moments and studies is to see these factors as bridges which
domains, I further considered the idea that disas- connect past life course transitions to future ones,
ters exist as processes, not events, such that the thereby shaping short- and long-run life course
adverse effects of environmental hazards persist trajectories and outcomes.
and desist slowly in the lives of those who have At a conceptual level, this suggests that the
experienced them. Although the point at which origins of differential vulnerability to environ-
the adverse effects of such hazards fully cease mental hazards are diverse. Moreover, it points to
and are replaced (instantaneously or gradually) the need for longitudinal and dynamical frame-
by processes of resilience and recovery is not works capable of linking the past to the future via
well understood to date, evidence suggests that the present. Ultimately, this intersects with exist-
disasters do, at some point, desist, and slowly at ing conceptualizations of disasters (and there are
that (Kates et al. 2006). Drawing on an emerging many (e.g., see Perry 2007)), which, rather than
body of research on the potential implications of events, are social processes that fundamentally
climate change for human migration (Leighton reect variability in the life course.
2009; McLeman 2014; McLeman and Hunter
2010), I also considered the idea that disaster
processes are, in some cases, slow-onset, as well. References
These ideas were illustrated in the context of a
stylized demonstration, which provided a plat- Adger, W. N. (2006). Vulnerability. Global Environmental
form grounded in the paradigmatic principles Change, 16, 268281.
Aldrich, N., & Bensen, W. F. (2008). Disaster prepared-
of life course theory to jointly consider the inter- ness and the chronic disease needs of vulnerable older
sections of life course processes and domains, adults. Preventing Chronic Disease: Public Health
vulnerability to environmental hazards, and the Research, Practice, and Policy, 5, 17.
nature of disaster onset and desistance (Elder Alexander, D. (2005). An interpretation of disaster in
terms of changes in culture, society, and international
et al. 2003: 10). Despite the stylized use of the relations. In R. W. Perry & E. L. Quarantelli (Eds.),
model in this chapter, the model itself is amena- What is a disaster? New answers to old questions
ble to empirical investigations of other life course (pp. 2538). Philadelphia, PA: XLibris Press.
processes and specic disasters provided that Bates, D. C. (2002). Environmental refugees? Classifying
human migration caused by environmental change.
suitable data are available. Population and Environment, 23, 465477.
With respect to other implications for future Berggren, R. E., & Curiel, T. J. (2006). After the storm:
research, I noted earlier in my discussion of dif- Health care infrastructure in post-Katrina New
ferential vulnerability to environmental hazards, Orleans. The New England Journal of Medicine, 354,
15491552.
that one important implication for future research Black, R., Adger, W. N., Arnell, N. W., Dercon, S.,
is the need to inventory and analyze the pre- Geddes, A., & Thomas, D. S. G. (2011). The effect of
disaster characteristics of persons and places environmental change on human migration. Global
(Jones et al. 2009; Masten and Narayan 2012). Environmental Change, 21S, S3S11.
Black, R., Arnell, N. W., Adger, W. N., Thomas, D. S. G.,
Previous research has documented a number of & Geddes, A. (2013). Migration, immobility and dis-
aggravating and protective factors at the individ- placement outcomes following extreme events.
ual-, household-, and community-levels which Environmental Science & Policy, 27S, S32S43.
condition the impact(s) of a hazard event on sub- Bonanno, G., Galea, S., Bucciarelli, A., & Vlahov, D.
(2007). What predicts psychological resilience after
sequent outcomes. While these factors have been disaster? The role of demographics, resources, and life
shown to be important during and after the haz- stress. Journal of Consulting and Clinical Psychology,
ard event itself, these factors are also important in 75, 671682.
investigations into the characteristics and origins Buchanan, T. K., Casbergue, R. M., & Baumgartner, J. J.
(2009). Young childrens demonstrated understanding
of differential vulnerability (Weems and of hurricanes. In K. E. Cherry (Ed.), Lifespan
Overstreet 2009). Rather than simply control- perspectives on natural disasters: Coping with Katrina,
ling for pre-disaster characteristics, the central Rita, and other storms (pp. 326). New York: Springer.
336 J. DeWaard

Bumpass, L., & Lu, H. (2000). Trends in cohabitation and Gallopn, G. C. (2006). Linkages between vulnerability,
implications for childrens family contexts in the resilience, and adaptive capacity. Global
United States. Population Studies, 54, 2941. Environmental Change, 16, 293303.
Burton, M. L., & Hicks, M. J. (2005). Hurricane Katrina: Groen, J. A., & Polivka, A. E. (2010). Going home after
Preliminary estimates of commercial and public sec- Hurricane Katrina: Determinants of return migration
tor damages. Huntington: Center for Business and and changes in affected areas. Demography, 47,
Economic Research, Marshall University. 821844.
Chamlee-Wright, E., & Storr, V. H. (2009). Theres no Guha-Sapir, D., Hoyois, P., & Below, R. (2013). Annual
place like New Orleans: Sense of place and commu- disaster statistical review 2013: The numbers and
nity recovery in the ninth ward after Hurricane Katrina. trends. Centre for Research on the Epidemiology of
Journal of Urban Affairs, 31, 615634. Disasters (CRED), Universit Catholique de Louvain.
Cherniack, E. P. (2008). The impact of natural disasters on Hobfoll, S. E. (1989). Conservation of resources: A new
the elderly. American Journal of Disaster Medicine, 3, attempt at conceptualizing stress. American
133139. Psychologist, 44, 513524.
Cherry, K. E. (2009). Lifespan perspectives on natural Hunter, L. M., Leyk, S., Nawrotzki, R., MacLaurin, G.,
disasters: Coping with Katrina, Rita, and other Twine, R., Collinson, M., & Erasmus, B. (2014). Rural
storms. New York: Springer. outmigration, natural capital, and livelihoods in rural
Cohan, C. L., & Cole, S. W. (2002). Life course transi- South Africa. Population, Space and Place, 20,
tions and natural disaster: Marriage, birth and divorce 402420.
following Hurricane Hugo. Journal of Family Ingram, R. E., & Luxton, D. D. (2005). Vulnerability-
Psychology, 16, 1425. stress models. In B. L. Hankin & J. R. Z. Abela (Eds.),
Cutter, S. L. (1996). Vulnerability to environmental haz- Development of psychopathology: A vulnerability-
ards. Progress in Human Geography, 20, 529539. stress perspective (pp. 3246). Beverly Hills: Sage
Cutter, S. L., Boruff, B. J., & Shirley, W. L. (2003). Social Publications.
vulnerability to environmental hazards. Social Science Intergovernmental Panel on Climate Change. (2012).
Quarterly, 84, 242261. Glossary of terms. In C. B. Field, V. Barros, T. F.
Cutter, S. L., Barnes, L., Berry, M., Burton, C., Evans, E., Stocker, D. Qin, D. J. Dokken, K. L. Ebi, M. D.
Tate, E., & Webb, J. (2008). A place-based model for Mastrandrea, K. J. Mach, G. K. Plattner, S. K. Allen,
understanding community resilience to natural disas- M. Tignor, & P. M. Midgley (Eds.), Managing the
ters. Global Environmental Change, 18, 598606. risks of extreme events and disasters to advance cli-
Earthquake Engineering Research Institute. (2006). The mate change adaptation (pp. 555564). Cambridge:
Kashmir earthquake of October 8, 2005: Impacts in Cambridge University Press.
Pakistan. EERI Special Earthquake Report, Oakland. Johnson, J., & Galea, S. (2009). Disasters and population
Elder, G. H., Jr., Johnson, M. K., & Crosnoe, R. (2003). health. In K. E. Cherry (Ed.), Lifespan perspectives on
The emergence and development of life course theory. natural disasters: Coping with Katrina, Rita, and
In J. T. Mortimer & M. J. Shanahan (Eds.), Handbook other storms (pp. 281326). New York/Frankfurt am
of the life course (pp. 322). New York: Kluwer Main: Springer/Campus Verl.
Academic/Plenum Publishers. Jones, R. T., Burns, K. D., Immel, C. S., Moore, R. M.,
El-Hinnawi, E. (1985). Environmental refugees. Nairobi: Schwartz-Goel, K., & Culpepper, B. (2009). The
United Nations Environmental Programme. impacts of Hurricane Katrina on children and adoles-
Elliott, J. R., & Pais, J. (2006). Race, class, and Hurricane cents: Conceptual and methodological implications
Katrina: Social differences in human responses to for assessment and intervention. In K. E. Cherry (Ed.),
disaster. Social Science Research, 35, 295321. Lifespan perspectives on natural disasters: Coping
Findlay, A. M. (2011). Migrant destinations in an era of with Katrina, Rita, and other storms (pp. 6596).
environmental change. Global Environmental Change, New York: Springer.
21S, S50S58. Kates, R. W., Colten, C. E., Laska, S., & Leatherman, S. P.
Fultz, N. H., Ofstedal, M. B., Herzog, A. R., & Wallace, (2006). Reconstruction of New Orleans after Hurricane
R. B. (2003). Additive and interactive effects of Katrina: A research perspective. Proceedings of the
comorbid physical and mental conditions on func- National Academy of Sciences, 103, 1465314660.
tional health. Journal of Aging and Health, 15, KC, S., Lutz, W., Loichinger, E., Muttarak, R., &
465481. Strienig, E. (2014). Reducing vulnerability in critical
Fussell, E., & Lowe, S. R. (2014). The impact of housing life course phases by enhancing human capital
displacement on the mental health of low-income par- (Occasional Paper, Human Development Reports).
ents after Hurricane Katrina. Social Science and New York: United Nations.
Medicine, 113, 137144. Kraaij, V., Arensman, E., & Spinhoven, P. (2002).
Fussell, E., Sastry, N., & VanLandingham, M. (2010). Negative life events and depression in elderly persons:
Race, socioeconomic status, and return migration to A meta-analysis. Journal of Gerontology, 57B, 8794.
New Orleans after Hurricane Katrina. Population and Leighton, M. (2009). Migration and slow-onset disasters:
Environment, 31, 2042. Desertication and drought. In F. Laczko &
Disaster and Life Course Processes 337

C. Aghazarm (Eds.), Migration, environmental and Palloni, A. (2001). Increment-decrement life tables. In
climate change: Assessing the evidence. Geneva: S. Preston, P. Heuveline, & M. Guillot (Eds.),
International Organization for Migration. Measuring and modeling population processes
Logan, J. R. (2009). Unnatural disaster: Social impacts (pp. 256273). Oxford: Blackwell Publishers.
and policy choices after Katrina. In R. D. Bullard & B. Paton, D., & Johnston, D. M. (2006). Disaster resilience:
Wright (Eds.), Race, place and environmental justice An integrated approach. Springeld: Charles
after Hurricane Katrina (pp. 249264). Boulder: C. Thomas Publisher.
Westview Press. Paxson, C., Fussell, E., Rhodes, J., & Waters, M. (2012).
Logan, J. R., & Xu, Z. (2011). Increasing population vul- Five years later: Recovery from post-traumatic stress
nerability to hurricane damage on the U.S. Gulf Coast, and psychological distress among low-income moth-
19502009 (Working Paper, Spatial Structures in the ers affected by Hurricane Katrina. Social Science and
Social Science). Providence: Brown University. Medicine, 74, 150157.
Marmar, C. R., Weiss, D. S., Metzler, T. J., Delucchi, K. Perry, R. W. (2007). What is a disaster? In H. Rodrguez,
L., Best, S. R., & Wentworth, K. A. (1999). E. L. Quarantelli, & R. R. Dynes (Eds.), Handbook of
Longitudinal course and predictors of continuing dis- disaster research (pp. 115). New York: Springer.
tress following critical incident exposure in emer- Rathore, M. F. A., Rashid, P., Butt, A. W., Malik, A. A.,
gency services personnel. Journal of Nervous and Gill, Z. A., & Haig, A. J. (2007). Epidemiology of spi-
Mental Disease, 187, 1522. nal cord injuries in the 2005 Pakistan earthquake.
Masten, A. S., & Narayan, A. J. (2012). Child develop- Spinal Cord, 45, 658663.
ment in the context of disaster, war, and terrorism: Rathore, M. F. A., Farooq, F., Muzammil, S., New, P. W.,
Pathways of risk and resilience. Annual Review of Epi, M. C., Ahmad, N., & Haig, A. J. (2008). Spinal
Psychology, 63, 227257. cord injury management and rehabilitation: Highlights
Masten, A. S., & Osofsky, J. (2010). Disasters and their and shortcomings from the 2005 earthquake in
impacts on child development. Child Development, Pakistan. Archives of Physical Medicine and
81, 10291039. Rehabilitation, 89, 579585.
McFarlane, A. C. (1988). The longitudinal course of post- Rhodes, J., Rouse, C. E., Waters, M., & Fussell, E. (2010).
traumatic morbidity: The range of outcomes and their The impact of Hurricane Katrina on the mental and
predictors. The Journal of Nervous and Mental physical health of low-income parents in New Orleans.
Disease, 176, 3039. American Journal of Orthopsychiatry, 80, 237247.
McLeman, R. A. (2014). Climate change and human Rogers, A. (1975). Introduction to multiregional mathe-
migration: Past experiences, future challenges. matical demography. New York: Wiley.
Cambridge: Cambridge University Press. Rogers, A. (1995). Multiregional demography: Principles,
McLeman, R. A., & Hunter, L. M. (2010). Migration in methods and extensions. London: Wiley.
the context of vulnerability and adaptation to climate Rogers, A., & Castro, L. J. (1981). Model migration
change: Insights from analogues. Wiley schedules (RR-81-30). Laxenburg: International
Interdisciplinary Reviews: Climate Change, 1, Institute for Applied Systems Analysis.
450461. Sampson, R. J., & Groves, W. B. (1989). Community
Myers, C. A., Slack, T., & Singelmann, J. (2008). Social structure and crime: Testing social disorganization
vulnerability and migration in the wake of disasters: theory. American Journal of Sociology, 94, 774802.
The case of Hurricanes Katrina and Rita. Population Sampson, R. J., & Laub, J. H. (1996). Socioeconomic
and Environment, 29, 271291. achievement in the life course of disadvantaged men:
Nawrotzki, R. J., Riosmena, F., & Hunter, L. M. (2013). Military service as a turning point, circa 19401965.
Do rainfall decits predict U.S. bound migration American Sociological Review, 61, 347367.
from rural Mexico? Evidence from the Mexican cen- Schoen, R. (1988). Modeling multigroup populations.
sus. Population Research and Policy Review, 31, New York: Plenum.
129158. Schultz, J. M., Espinel, Z., Galea, S., & Reissman, D. B.
Nigg, J. M., Barnshaw, J., & Torres, M. R. (2006). (2007). Disaster ecology: Implications for disaster
Hurricane Katrina and the ooding of New Orleans: psychiatry. In R. J. Ursano, C. S. Fullerton,
Emergent issues in sheltering and temporary housing. L. Weisaeth, & B. Raphael (Eds.), Textbook of disaster
The Annals of the American Academy of Political and psychiatry (pp. 6996). Cambridge: Cambridge
Social Science, 604, 113128. University Press.
Olteanu, A., Amberger, R., Grant, R., Davis, C., Abramson, Shenk, D., Ramos, B., Kalaw, K. J., & Tufan, I. (2009).
D., & Asola, J. (2011). Persistence of mental health History, memory, and disasters among older adults: A
needs among children affected by Hurricane Katrina life course perspective. Traumatology, 15, 3543.
in New Orleans. Prehospital Disaster Medicine, 36, Taquir, S. F., Mirza, S., Gul, S., Ghaffar, H., & Zafar, A.
36. (2007). Complications in patients with spinal cord
Pais, J. F., & Elliott, J. R. (2008). Places as recovery injuries sustained in an earthquake in northern
machines: Vulnerability and neighborhood changes Pakistan. Journal of Spinal Cord Medicine, 30,
after major hurricanes. Social Forces, 86, 14151453. 373377.
338 J. DeWaard

Thomalla, F., Downing, T., Spanger-Siegfried, E., Han, from Hurricane Katrina. In K. E. Cherry (Ed.),
G., & Rockstrm, J. (2006). Reducing hazard vulner- Lifespan perspectives on natural disasters: Coping
ability: Towards a common approach between disaster with Katrina, Rita, and other storms (pp. 2744).
and risk reduction and climate adaptation. Disasters, New York: Springer.
30, 3948. Wilson, T. (2010). Model migration schedules incorporat-
United States Geological Survey. (2005). M7.6 northern ing student migration peaks. Demographic Research,
Pakistan earthquake of 8 October 2005. Reston: 23, 191222.
National Earthquake Information Center. Yang, Y., & Waliji, M. (2010). Increment-decrement life
Vogeli, C., Shields, A. E., Lee, T. A., Gibson, T. B., table estimates of happy life expectancy for the U.S.
Marder, W. D., Weiss, K. B., & Blumenthal, D. population. Population Research and Policy Review,
(2007). Multiple chronic conditions and implica- 29, 775795.
tions for quality, care management, and costs. Yule, W., Bolton, D., Udwin, O., Boyle, S., ORyan, D., &
Journal of General Internal Medicine, 22, 391395. Nurrish, J. (2000). The long-term psychological
Weems, C. F., & Overstreet, S. (2009). An ecological- effects of a disaster experience in adolescence: The
needs-based perspective of adolescent and youth emo- incidence and course of PTSD. Journal of Child
tional development in the context of disaster: Lessons Psychology and Psychiatry, 41, 503511.
Part III
Health and Development
Through the Life Course
Early Childhood Poverty: Short
and Long-Run Consequences Over
the Lifespan

Ariel Kalil, Greg J. Duncan,


and Kathleen M. Ziol-Guest

1 Introduction est quintile score at the 69th percentile on math


and literacy (Waldfogel and Washbrook 2011).
Using a poverty line of about $22,000 for a fam- Gaps in conduct problems and attention/hyperac-
ily of four, the U.S. Census Bureau counted more tivity are also apparent albeit less pronounced.
than 15 million U.S. children living in poor fami- Duncan and Magnuson (2005) examined
lies in 2012. Poor children begin school well teacher-reported gaps in attention and engage-
behind their more afuent age mates and, if any- ment in rst and fth grade. They showed that
thing, lose ground during the school years. children from the top income quintile are reported
Although poorer and more afuent children make by teachers to be far more engaged in school (the
similar gains on standardized achievement tests gap is approximately two-thirds of a standard
during each school year, poor children fall behind deviation) compared with their counterparts in
during the summers when school is not in ses- the bottom quintile. Children from poor families
sion, particularly in the early grades, while their also go on to complete less schooling, work less
more afuent counterparts continue to forge and earn less. Understanding the origins and per-
ahead. As a result, the gap between them widens sistence of these differences in fortunes is a vital
over time (Entwisle et al. 2003). At age 4, chil- step for understanding the intergenerational
dren from families in the poorest income quintile reproduction of poverty and ensuring the pros-
score on average at the 32nd percentile of the perity of future generations.
national distribution on math and the 34th percen- Social scientists have been investigating links
tile in a test of literacy while children in the rich- between family poverty and subsequent child
outcomes for decades (see Mayer 1997 for a
review). As in many research areas, early empiri-
cal studies were typically based on point-in-time
A. Kalil (*) cross-sectional data. The creation of nationally
Harris School of Public Policy, University
of Chicago, Chicago, IL, USA representative longitudinal data sets in the late
e-mail: akalil@uchicago.edu 1960s and 1970s enabled researchers to test
G.J. Duncan more rened and dynamic models of links
School of Education, University of California-Irvine, between childrens poverty experiences and later
Irvine, CA, USA outcomes, which predicted, for example, that
K.M. Ziol-Guest persistent poverty is more harmful than transient
Steinhardt School of Culture, Education, poverty. Although large-scale random-
and Human Development, New York University, assignment social experiments conducted in the
New York, NY, USA

Springer International Publishing Switzerland 2016 341


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_15
342 A. Kalil et al.

United States in the 1970s manipulated family influence of economic loss during the Great
income, most of them focused on the question of Depression (Elder 1974). According to this per-
how additional income affected adult work effort spective, poor families face signicant economic
rather than on child well-being (Maynard and pressure as they struggle to pay bills and pur-
Murnane 1979). chase important goods and services, and these
Almost universally neglected in the poverty economic pressures, coupled with other stressful
indicators and impacts literatures are careful life events that are more prevalent in the lives of
thought about the timing of economic hardship poor families, creates high levels of psychologi-
across childhood and adolescence. Emerging cal distress, including depressive and hostile feel-
research in neuroscience and developmental psy- ings, in poor parents (Kessler and Cleary 1980;
chology suggests that poverty early in a childs McLeod and Kessler 1990).
life may be particularly harmful. Not only does
the astonishingly rapid development of young This perspective has been broadened by recent
childrens brains leave them sensitive (and vul- behavioral economics work showing that condi-
nerable) to environmental conditions, but the tions of poverty and scarcity not only create psy-
family context (as opposed to schools or peers) chological distress, but also deplete important
dominates childrens everyday lives. cognitive resources (Spears 2011). Studies, most
After a brief review of theoretical perspectives of which have been conducted in developing
and both experimental and nonexperimental evi- countries, nd that making economic decisions
dence linking poverty with childhood outcomes, under conditions of scarcity reduces adults sub-
we highlight emerging research based on newly sequent behavioral self-control and renders them
available data providing both poverty measures less able to regulate their own behavior in order
as early as the prenatal year and adult outcomes to pursue less immediate goals.
measured in the fourth decade of life. Our review Psychological distress spills over into marital
and discussion of the existing literature pays par- and co-parenting relationships. As couples strug-
ticular attention to the issue of causal evidence. gle to make ends meet, their interactions tend to
become more hostile and conicted, and this
leads them to withdraw from each other (Brody
2 Theoretical Perspectives and Stoneman 1994; Conger and Elder 1994).
Parents psychological distress and conict, in
What are the short and long-run consequences of turn, are linked with parenting practices that are
growing up in a poor household? Economists, on average more punitive, harsh, inconsistent,
sociologists, developmental psychologists, and and detached, as well as less nurturing, stimulat-
neuroscientists emphasize different pathways by ing, and responsive to childrens needs. Such
which poverty may inuence childrens develop- lower-quality parenting is likely to elevate chil-
ment. The two main theoretical frameworks drens physiological stress responses, and ulti-
describing these processes are: family and envi- mately harms childrens development (Conger
ronmental stress, on the one hand and resources et al. 2002; Conger and Dogan 2007; McLoyd
and investment on the other. A third perspective, 1990). At the same time, methodologically
which we also describe, is social selection. strong studies that have manipulated family
income have shown few impacts on parenting
Family and Environmental Stress behavior or parent-child relationships (Morris
Perspective Economically disadvantaged fami- et al. 2001).
lies experience higher levels of stress in their Furthermore, although the biological links
everyday environments than more afuent fami- between low income and stress are compelling,
lies, and these disparities may affect childrens no methodologically strong studies have linked
development. The family stress model was rst poverty and elevated and prolonged stress reac-
developed by Glen Elder to document the tions in children. Some rigorous studies have
Early Childhood Poverty: Short and Long-Run Consequences Over the Lifespan 343

examined these connections in mothers. One of deprived or traumatic environments often develop
these linked expansions of the Earned Income differently. Traumatic stress that arises from
Tax Credit (EITC) to data from the National child maltreatment, for example, produces mea-
Health Examination and Nutrition Survey and surable changes in brain structures and is likely
found that when compared with mothers with to impart long-lasting disadvantages for adult
just one child, low-income mothers with two or mental and physical health and labor market
more children (who received a greater increase in functioning (Danese et al. 2007).
income under the program expansion) experi- Based on insights from this emerging neuro-
enced larger reductions in risky biomarkers and science literature, Cunha et al. (2006) propose an
self-reported better mental health (Evans and economic model of development in which pre-
Garthwaite 2010). A study of the impacts of school cognitive and socio-emotional capacities
increases in the Canadian Child Benet also are key ingredients for human capital acquisition
found improvements in maternal mental health. during the school years. In their model, skill
However, studies of the impacts of other welfare begets skill and early capacities can affect the
and anti-poverty programs that increased both likelihood that later school- age human capital
income and maternal employment did not show investments will be successful and productive.
similar improvements in mental health (Duncan This model predicts that economic deprivation in
et al. 2009). early childhood creates disparities in school read-
The Family Stress perspective has undergone iness and early academic success that widen over
major conceptual and empirical advances in the course of childhood.
recent years. On the conceptual side, a narrow Complementary studies in psychology and
focus on environmental toxins and parental men- social epidemiology illustrate that both in utero
tal health and parenting has been broadened by environments and early childhood experiences
neurobiological evidence on the importance of can have long-run impacts on adult physical and
maintaining tolerable levels of stress for both par- mental health (Sapolsky 2004; Strauss 1997).
ents and children, and a cognitive psychological The fetal origins hypothesis posits a program-
perspective on links between stress, information ming process whereby stimulants and insults
processing and decision-making. Increasingly during the prenatal period have long-lasting
sophisticated studies suggest linkages between implications for physiology and disease risk
income support and maternal stress. (Barker et al. 2002). Chronic stress from grow-
In particular, emerging evidence from neuro- ing up poor could also play a role in dysregula-
science and social epidemiology suggests that the tion across multiple physiological systems
timing of child poverty matters, and that for some whose effects persist (or possibly compound)
outcomes later in life, particularly those related into adulthood. For example, Evans and
to attainment and health, poverty early in a Schamberg (2009) showed that childhood pov-
childs life may be particularly harmful. Both erty increases allostatic load, a biological index
human and animal studies highlight the critical of the cumulative wear and tear on the body, dur-
importance of early childhood for brain develop- ing the teenage years. Moreover, the longer the
ment and for establishing the neural functions children had lived in poverty, the higher their
and structures that will shape future cognitive, allostatic load.
social, emotional and health outcomes (Knudsen Allostatic load is caused by the mobilization
et al. 2006). Essential properties of most of the of multiple physiological systems in response to
brains architecture are established very early in chronic stresses in the environment. Thus child-
life by genes and, importantly, early experience. hood poverty may actually reset the immune
Childrens brains are programmed to be wired system so that inammation processes become
efciently based on everyday interactions with dysregulated, resulting in higher levels and pro-
sights, sounds and supportive caregivers longed production of proinammatory cytokines
(Sapolsky 2004). The brains of children in (i.e., chemical signals that can cause blood
344 A. Kalil et al.

vessels to leak, leading to the swelling and red- such as books, computers, high-quality child
ness that is associated with inammation; Miller care, summer camps, and private school tuition.
et al. 2009). They investigate total expenditures per child,
enrichment expenditures per child and per-child
Resource and Investment Perspective Household expenditures in selected categories for different
production theory has played a central role in income groups and for families with either only
how economists conceive of family inuences on preschoolers or children age 6 or more.
child development. Gary Beckers A Treatise on Adjusted for family size and composition,
the Family (1991) posits that child development they nd that child enrichment expenditures are
is produced from a combination of endow- highly income elastic, constituting only 3 % of
ments and parental investments. Endowments total expenditures for families in the bottom
include genetic predispositions and the values expenditure quintile, and 9 % of total expendi-
and preferences that parents instill in their chil- tures for families in the top expenditure quintile.
dren. Parents preferences, such as the impor- Trips and child care expenditures are most
tance they place on education and their orientation important for the high-income families, while
toward the future, combined with their resources, expenditures are spread more evenly across cat-
shape parental investments. egories for low-income families. Moreover, 40
years ago, low-income families spent about $850
Economists argue that time and money are the (in 2011 dollars) on child enrichment expendi-
two basic resources that parents invest in their tures such as books, computers, high-quality
children. For example, investments in high- child care, summer camps and private school
quality child care and education, housing in good tuition, while higher-income families spent more
neighborhoods, and rich learning experiences than $3,500, already a substantial difference
enhance childrens development, as do invest- (Duncan and Murnane 2011). By 20052006,
ments of parents time. Links among endow- low-income families had increased their expen-
ments, investments, and development appear to ditures to over $1,300, but high-income families
differ by the domain of development under con- had increased theirs much more, to more than
sideration (e.g. achievement, behavior, health). $9,000 per child. The differences in spending
Household production theory suggests that between the two groups had almost tripled in the
children from poor families lag behind their eco- intervening years. The largest spending differ-
nomically advantaged counterparts in part ences were for activities such as music lessons,
because their parents have fewer resources to travel, and summer camps (Duncan and Murnane
invest in them. Compared with more afuent par- 2011).
ents, poor parents are less able to purchase inputs One of the most important parenting differences
for their children, including books and educa- between advantaged and disadvantaged parents is
tional materials at home, high-quality child care in how much time the parent spends with the child.
settings and schools, and safe neighborhoods. Annette Lareaus (2003) qualitative study of family
Economically disadvantaged parents may also life reported that middle-class parents target their
have less time to invest in their children, owing to time with children toward developmentally enhanc-
higher rates of single-parenthood, nonstandard ing activities. In her study, middle-class families
work hours, and less exible work schedules (whose jobs, by her denition, require college-level
(Smolensky and Gootman 2003). This too may skills) engage in a pattern of concerted cultiva-
have negative consequences for children. tion to actively develop childrens talents and
Evidence suggests that the amount of cognitive skills. By contrast, in lower-class families, Lareau
stimulation in the home environment varies with identied a pattern that she calls the accomplish-
changes in family income (Votruba-Drzal 2006). ment of natural growth, wherein parents attend to
Kaushal et al. (2011) provide new evidence, childrens material and emotional needs but pre-
based on recent Consumer Expenditure Surveys, sume that their talents and skills will develop with-
of the nature of child enrichment expenditures out concerted parental intervention.
Early Childhood Poverty: Short and Long-Run Consequences Over the Lifespan 345

Numerous quantitative studies not only show emotional support) than did race/ethnicity, with
large differences in the time investments of children in poor families having less stimulating
advantaged and disadvantaged parents but also and responsive parent-child interactions than
that these gaps persist even when other differ- children in higher SES families. Differences were
ences across families, such as employment hours relatively consistent across ages. These ndings
and schedules, are accounted for (Guryan et al. reect parenting patterns across income groups
2008). Using education as a marker of economic found in smaller studies (e.g., Conger and Conger
advantage, Guryan et al. (2008) use data from the 2000; Dodge et al. 1994; McLoyd 1990).
American Time Use Survey to show that condi-
tional on hours worked in the labor market moth- Selection Perspective Researchers seeking to
ers with a college education spend 6 more hours identify the causal inuences of income face for-
per week in child care activities than women who midable challenges. They must take seriously
did not nish high school and twice the amount threats to internal validity from a variety of
of time as mothers who graduated from high sources, including bias from simultaneous causa-
school. These researchers posited that economi- tion and potential omitted variables (Sobel 1998).
cally advantaged parents view time with children A key challenge involves isolating the effects of
as an investment behavior with which to income from other disadvantages that poor fami-
increase childrens human capital (for either lies face. A rst issue is that socioeconomic char-
altruistic or selsh reasons) and do not view mar- acteristics tend to cluster together, so that children
ket alternatives as highly effective substitutes for who live in poverty are likely to have parents
their own time investments. Kalil et al. (2012) with lower levels of education, which makes it
further show that highly-educated mothers are difcult to isolate the unique inuence of any one
more efcient in their parental time invest- indicator of SES. A second problem is that family
ments by tailoring their specic activities to chil- income or low parental education are linked with
drens developmental stage. Their work shows several other family circumstances such as
that highly educated mothers shift the composi- greater prevalence of single-parent families, low
tion of their time in ways that promote childrens levels of parents cognitive skills, and poor
development at different developmental stages, parental mental health. Each of these factors may
for example by emphasizing play and teaching have an independent negative effect on children.
when children are in preschool and management Thus, to ascribe a causal effect to poverty requires
of life outside the home for adolescents. ruling out other explanations for the associations
The most commonly used measure of parent- between socioeconomic factors and childrens
ing behavior in large national studies is the Home development.
Observation for Measurement of the Environment
(HOME), an observational and parent-report Some researchers have argued that income
measure of the quantity and quality of cognitive effects are largely the spurious result of unmea-
stimulation and emotional support children sured differences that are correlated both with
receive in the home. The HOME captures all of income and child outcomes (Mayer 1997). In
the parenting behaviors found to differ between other words, unmeasured characteristics such as
more and less advantaged children including par- parental mental health or motivation that contrib-
ents promotion of language and literacy, engage- ute to greater income and earnings may also
ment in enriching activities, and warmth and enhance child development, leading to a spurious
affection. Bradley et al. (2001) analyzed HOME correlation between SES and child development.
scores in a large nationally representative data set This threat of omitted variable bias is an impor-
(the National Longitudinal Survey of Youth- tant concern in most nonexperimental research.
Child Survey) to show that poverty status had a Studies that use rigorous statistical techniques to
larger association with both Teaching and address bias issues tend to uncover smaller effect
Parental Responsiveness (subscales that emerged sizes than studies that do not (Duncan 2006;
across ages that tap cognitive stimulation and Holmlund et al. in press).
346 A. Kalil et al.

Duncan (2006) describes a continuum for from three of the sites (Gary, Indiana, and rural
evaluating the methodological rigor of studies areas in North Carolina and Iowa) investigate
aimed at estimating poverty and incomes inu- impacts on achievement gains for children in
ence on child development. On one end are cor- elementary school; two of the three found signi-
relational studies that analyze associations cant impacts (Maynard and Murnane 1979;
between concurrent measures of family income Maynard 1977). In contrast, no achievement dif-
and child outcomes, with few adjustments for ferences were found for adolescents. Impacts on
confounding factors. These studies are common, school enrollment and attainment for youth were
but likely plagued by biases. On the other end are more uniformly positive, with both the Gary and
experiments in which families are randomly the New Jersey sites reporting increases in school
assigned to receive additional income, without enrollment, high school graduation rates, or years
any strings attached. If implemented correctly, of completed schooling. Second- through eighth-
experiments provide unbiased estimates of grade teachers rated student comportment in
income effects, but such studies are exceptionally the two rural sites; results showed income-
rare. Between these two extremes, ranging from induced improvements in one of the sites but not
less to more rigorous, are natural experiments, the other.
studies that employ econometric techniques to Taken together, these studies appear to sug-
reduce omitted variable bias (e.g., xed effects, gest that income is more important for the school
instrumental variables regression), and longitudi- achievement of pre-adolescents than adolescents
nal studies (see also Duncan et al. 2004). but may also matter for the school attainment of
Regardless of the timing of low income, iso- adolescents. None of the results from the Negative
lating its causal impact on childrens well- being Income Tax experiments bear on the early is
is very difcult. The best way to identify how best hypothesis, because none tracked the pos-
much money itself really matters is to conduct an sible achievement impacts for children who had
experiment that compares families that receive not yet entered school when the income treat-
some additional money with families that are oth- ment was being administered.
erwise similar, but do not receive such money. Experimental welfare reform evaluation stud-
We are not aware of any experimental or quasi- ies undertaken during the 1990s incentivized
experimental studies that compare the effect of parental employment by providing income sup-
family income contemporaneously, in early ports to working-poor parents through wage sup-
childhood, with the effect of family income later plements. Moreover, some measured the test
in childhood. But methodologically-strong stud- scores of children who were just entering school
ies have estimated the causal effects of overall during the time that the programs were in place.
childhood family income on later outcomes. We Morris et al. (2005) analyzed data from seven
turn to these studies next. random-assignment welfare and antipoverty poli-
cies, all of which increased parental employment,
while only some of them increased family income
3 Empirical Evidence (Morris et al. 2005).
The combined impacts of higher income and
The strongest causal evidence in the literature more maternal work on childrens school achieve-
relates income increases to childrens school ment varied markedly by the childrens age.
achievement and attainment. The only large-scale Treatment-group children between the ages of 4
randomized interventions to alter family income and 7 when the programs took effect, many of
directly were the U.S. Negative Income Tax whom made the transition into elementary school
Experiments, which were conducted between during the programs, scored signicantly higher
1968 and 1982 with the primary goal of identify- on achievement tests than their control group
ing the inuence of guaranteed income on par- counterparts. A sophisticated statistical analysis
ents labor force participation. Published reports of the data on these younger children suggests
Early Childhood Poverty: Short and Long-Run Consequences Over the Lifespan 347

that a $3,000 annual income boost is associated tributed about $6,000 each year to all adult tribal
with a gain in achievement scores of about one- members. Akee and colleagues (2010) compared
fth of a standard deviation (Duncan et al. 2011). Native American children with non-Native
In contrast, there were no impacts on either American children, before and after the casino
teacher- or parent-reported behavior problems opened, and found that receipt of casino pay-
(Duncan et al. 2009). ments increased the educational attainment of
The achievement of children age 811 did not poor Native American youth by nearly a year and
appear to be affected by the programs, and the reduced criminal behavior and drug use.
achievement of children who were 12 and 13 dur- Finally, one study used the oil boom in
ing the programs seemed to be hurt by the pro- Norway in the beginning of the 1970s as a source
grams efforts to increase family income and of exogenous variation in family income.
parental employment (Morris et al. 2005). Instrumenting childhood family income with
Another recent study took advantage of the being born in the region and cohorts affected by
increasing generosity of the U.S. EITC between the oil boom, Lken (2010) nds little or no lin-
1993 and 1997 to compare childrens test scores ear effect of overall childhood family income
before and after it was expanded (Dahl and (mean for age 113) on educational attainment.
Lochner 2012). Most of the children in this study Lken et al. (2012) builds on Lken (2010), and
were between the ages of 8 and 14 and none was develops a non-linear instrumental variable
younger than 5. The authors found improvements method. They nd that the linear estimator misses
in low-income childrens achievement in middle the signicant effect of family income because it
childhood that coincided with the EITC assigns little weight to the large and positive mar-
expansion. ginal effects in the lower part of the income dis-
A second study, conducted in Canada, took tribution. They look at effects on education (and
advantage of variations in the generosity of the IQ for men) and conclude that childhood family
National Child Benet program across income (mean over age 212) has large positive
Canadian provinces to estimate income impacts effects for children from families in the lower
on child achievement. Among children age part of the income distribution.
610 residing in low-income families, policy- Several lessons emerge from these experimen-
related income increases had a positive and sig- tal and quasi-experiment studies. First, achieve-
nicant association with math scores and a ment gains are selective and depend at least in
negative link with the likelihood of a child part on the childrens age when income gains
receiving a diagnosis of a learning disability. were received. Children in their preschool years
For 4- to 6-year-olds, the income increases or making the transition to school and elementary
were associated with higher scores on a test of school students generally enjoyed the most con-
receptive vocabulary for boys, but not for girls. sistent achievement increases. For adolescents,
Turning to behavior, higher benets led to less the achievement changes were mixed, with vari-
aggression among 4- to 10-year-olds, but did ous studies nding positive, null, and even nega-
not appear to affect other behavioral dimen- tive impacts. Second, in the case of adolescents,
sions assessed in the study. income appears to affect educational attainments
Researchers have rarely used rigorous experi- such as high school graduation and completed
mental or strong quasi-experimental designs to years of schooling rather than test scores. Given
study childrens psychological and behavioral the high costs of post-secondary education, the
health. An exception is work by Costello et al. effect of family income on completed schooling
(2003), who were able to take advantage of the is not surprising. Third, we know far more about
Great Smoky Mountain Study of Youth, which how poverty reduction affects achievement and
gathered longitudinal data on child outcomes schooling outcomes than we do about its effects
during the introduction of a casino by a tribal on behavior problems including childbearing and
government in North Carolina. The casino dis- criminal activity.
348 A. Kalil et al.

Virtually none of the experimental literature sample of U.S. families and their children since
on income effects has been able to estimate the 1968 (Duncan et al. 2010). The study is based on
impacts of changes in family income during the children born between 1968 and 1975 and col-
very earliest years of a childs life the time lected information on their economic fortunes
when children are developing rapidly and may be between ages 25 and 37. Health conditions were
very sensitive to family and home conditions. assessed in 2006, when these individuals were
Nor have these studies been able to examine the between the ages of 30 and 37.
consequences of income changes during child- The PSID measured income in every year of a
hood for outcomes measured in adulthood. This childs life from the prenatal period through age
is particularly unfortunate, since the goals of 15. This enabled Duncan et al. (2010) to measure
policies directed at children are often couched in poverty across several distinct periods of child-
terms of lifetime impacts a middle-class stan- hood, distinguishing income early in life (prena-
dard of living or higher labor market earnings. tal through fth year) from income in middle
childhood and adolescence. The simple associa-
tions between income early in life and adult out-
4 Linking Early Poverty comes are striking (Table 1). Compared with
to Adult Outcomes children whose families had incomes of at least
twice the poverty line during their early child-
None of the past income literature has been able hood, poor children completed 2 fewer years of
to relate family income early in a childs life to schooling, earned less than half as much, worked
adult attainments, largely because no single study 451 fewer hours per year, and received $826 per
had collected data on both early childhood year more in food stamps as adults. Poor males
income and later adult outcomes. However, were twice as likely to be arrested. For females,
recent research has made this link using data poverty was associated with a more than vefold
from the Panel Study of Income Dynamics, increase in the likelihood of bearing a child out of
which has followed a nationally representative wedlock prior to age 21. As for health, poor chil-

Table 1 Adult outcomes by poverty status between the prenatal year and age 5 (Duncan et al. 2010)
Early childhood income Early childhood income Early childhood
below the ofcial U.S. between one and two income more than
poverty line times the poverty line twice the poverty line
Mean or % Mean or % Mean or %
Completed schooling (years) 11.8 12.7 14.0
Adult earnings between ages 25 $17.9 $26.8 $39.7
and 37 (in $10,000)
Annual work hours between ages 25 1,512 1,839 1,963
and 37
Food stamps between ages 25 and 37 $896 $337 $70
Ever arrested (men only) 26 % 21 % 13 %
Nonmarital birth (women only) 50 % 28 % 9%
Poor health in 2005 13 % 13 % 5%
Obese in 2005 (BMI>30) 45 % 32 % 26 %
Hypertension in 2005 25 % 10 % 9%
Arthritis in 2005 7% 7% 3%
Diabetes in 2005 4% 6% 2%
Work-limiting hypertension in 2005 4% 2% 2%
Note: The sample consists of individuals born between 1968 and 1975 in the PSID. Earnings and food stamp values are
in 2005 dollars
Early Childhood Poverty: Short and Long-Run Consequences Over the Lifespan 349

dren were nearly three times as likely to report other controls, are included in all regressions).
poor overall health as adults and more than twice The $3,000 amount was chosen for the interpre-
as likely to report various activity-limiting health tation of coefcients because it is well within the
conditions, and members of this group were 19 range of an actual U.S. policy the Earned
percentage points more likely to be overweight. Income Tax Credit. Given that a linear function
Looking beyond these simple correlations, was t to the entire income range up to $25,000,
Duncan and colleagues regressed the adult out- estimated impacts of income increments smaller
comes listed in Table 1 on three childhood stage- or larger than $3,000 can be obtained with pro-
specic measures of family income average portionate reductions or increases in the impacts
income between the prenatal year and age 5, shows in the gures.
average income between ages 6 and 10, and aver- Turning rst to their central measure of labor
age income between ages 11 and 15 plus an market productivity average annual earnings
extensive list of background controls. The back- between ages 25 and 37 Duncan and colleagues
ground controls consist of birth year, race, sex, found that for children growing up in families with
whether the childs parents were married and liv- average early childhood incomes below $25,000, a
ing together at the time of the birth, mothers age $3,000 annual boost to family income between the
at birth, region, number of siblings, parent prenatal year and age 5 was associated with a 17 %
schooling, parent test score, cleanliness of the increase in adult earnings (Fig. 1). For children
house, parents expectations for child, parent growing up in higher-income households (more
achievement motivation, parent locus of control than $25,000 per year), a $3,000 boost to family
and parent risk avoidance. To account for the income was statistically signicant but was esti-
possibility that income effects are nonlinear, two mated to increase adult earnings by only about 2
coefcients were estimated for each childhood %. None of the income increments later in child-
stage, the rst reecting the estimated effect of an hood was estimated to have statistically signicant
additional $3,000 of annual income in the given effects on later earnings.
stage for children whose income during that stage Results for work hours are broadly similar to
averaged less than $25,000 and the second those for earnings, showing a highly signicant
reecting comparable effects for higher-income estimated impact of early, but not later, childhood
children (all three sets of income variables, plus income. In this case, a $3,000 annual increase in

20
17%
< $25,000 $25,000 +
15

10

5 ns
2%
ns ns ns
0
Prenatal to age 5 Age 6-10 Age 11-15

Fig. 1 Percentage increase in adult earnings associated with a $3,000 annual increase in childhood income (ns means
not statistically signicant at p < 0.10. Reproduced from Duncan et al. 2010)
350 A. Kalil et al.

200
$152

100

-$93 ns ns ns
0
Annual work Food stamp Schooling Arrests Nonmarital
hours receipt birth
-100

-200

Fig. 2 Impacts on various adult outcomes associated with a $3,000 annual increase in prenatal to age 5 income, for
incomes <$25,000 (ns means not statistically signicant at p < 0.10. Reproduced from Duncan et al. 2010)

the prenatal to age-5 average income of low- nonmarital births) outcomes account for links
income families is associated with 152 additional between early income and adult labor market
work hours per year after age 25. This is shown as productivity, what does? Consistent with the
the rst bar in Fig. 2. Other results presented in Fig. early origins work in social epidemiology and
2 show that the boost in adult productivity associ- neuroscience, it appears that early income has
ated with additional income in early childhood also long-term effects on work-limiting health condi-
led to signicantly less food stamp receipt. tions (Ziol-Guest et al. 2012).
Earnings are the product of work hours and the Regression results are shown in Fig. 3. As
hourly wage rate. There is clearly a strong rela- with earnings and work hours, each of the health
tionship between early income and work hours, conditions was regressed on stage-specic
but it is also important to determine how impor- childhood income, demographic control vari-
tant early income is for the hourly wage rate. In ables. As before, the income associations are
results not shown, Duncan et al. (2010) found no allowed to be nonlinear, with one linear segment
connection between early income and hourly t across average annual incomes within a given
earnings virtually all of the earnings effect was childhood stage up to $25,000 and another t to
carried by increases in labor supply rather than incomes above $25,000. Only the coefcients
the wage rate. Accordingly, it is perhaps not sur- on the low-income segment for early childhood
prising that early income was not signicantly are shown in Fig. 3 (in only one case for
related to completed schooling, the most potent incomes above $25,000 for ages 1115 in the
determinant of hourly wage rates. Nor were there diabetes regression was the coefcient more
signicant impacts of early poverty on problem than twice its standard error.) Given the dichoto-
behavior being arrested or incarcerated (for mous nature of the health outcomes, the models
males) or having a nonmarital birth (for females); are estimated with logistic regression. The bars
family income during adolescence seemed to in Fig. 3 represent the percentage reductions in
matter more for these outcomes. the odds of a given condition associated with a
So if neither the human capital (schooling and $3,000 increase in annual income between the
wage rates) nor the behavioral (lack of arrests or prenatal year and age 5.
Early Childhood Poverty: Short and Long-Run Consequences Over the Lifespan 351

10

ns ns
0

Work limiting health conditions


Obesity

Hypertension

Arthritis

Diabetes
Poor health
-10

-20
-20%
-30
-29%
-40
-33%
-50 -46%
Fig. 3 Percentage reductions in odds of various health <$25,000 (ns means not statistically signicant at
conditions in 2006 associated with a $3,000 annual p < 0.10. Reproduced from Duncan et al. 2010)
increase in prenatal to age 5 income, for family income

Figure 3 shows a remarkable pattern of effects adolescent years. Although some studies have
on the emerging (mid- to late-30s) adult health produced suggestive associations between
problems. Although increments to early income early-life income and adult health, the more gen-
do not appear to affect self-rated overall adult eral preponderance of conicting results raises
health or diabetes, $3,000 increments to low more questions than answers.
income early in life are associated with a 20 % Given the conicting empirical results, the cur-
reduction in the odds of obesity, a 29 % reduction rent literature raises more questions than it
in the odds of reporting hypertension, a 46 % answers, pointing to the need for more research.
reduction in the odds of reporting arthritis, and a Specically, future work should consider differ-
33 % reduction in the odds of reporting a health- ences in incomes effect by developmental stage
related work limitation. Although more research and the domain of health outcomes. Another fruit-
is obviously needed, these health pathways ful direction would be to construct trajectories of
involving stress and inammation appear to be income over the early lifecourse and link these
very promising linkages between poverty early in trajectories to later life outcomes (Shanahan and
life and adult labor market productivity. Hofer 2010). Future studies could also investigate
the relevance of cumulative processes by which
early disadvantages beget yet more disadvantages
5 Directions for Future with a compound-interest-like effect (DiPrete and
Research Eirich 2006). A greater research emphasis on
mechanisms would also provide important
Despite recent growth in research addressing insights. That is, one might wish to know whether
links between income and child and adult health, increases to income correlates with specic types
it is difcult to draw causal connections because of changes in low-income families quality of
health has received much less attention by parenting, nutritional adequacy, various invest-
researchers than achievement and behavioral ments that might help us to further understand
development. Moreover, most studies seeking to the gradients in adult outcomes for their children.
link pre-adult family income with later health has Finally, methodological considerations, such as
drawn their income measures from childrens differences in measurement, may also be impor-
352 A. Kalil et al.

tant. Indeed, prior studies suggest that income This research suggests that greater policy atten-
gradients tend to be more pronounced for more tion should be given to remediating situations
subjective measures of child health, such as par- involving deep and persistent poverty in utero
ent-report measures, and are less evident in more and occurring early in childhood. In terms of
objective measures, such as biomarkers or physi- indicators, it is crucial to track rates of poverty
cian diagnosis (Currie and Lin 2007). among children especially deep poverty occur-
An alternative, if somewhat expensive, strat- ring early in childhood to inform policy discus-
egy would be to launch a neuroscience study sions regarding childrens well-being.
devoted to assessing the impact of experimental In the case of welfare policies, sanctions and
manipulation of SES. Suppose low income fami- other regulations denying benets to families
lies with newborns were recruited into a 5-year with very young children would appear particu-
study of brain development and randomly larly harmful. Not only do young children appear
assigned to treatment or controls groups. Further to be most vulnerable to the consequences of
suppose that experimental families received an deep poverty, but mothers with very young chil-
extra $12,000 income in the form of monthly dren are also least able to support themselves
payments of $333, and control-group families through employment in the labor market.
receiving a nominal $20 monthly payment. The More effective would be income transfer poli-
$333 monthly income supplements constitute a cies that provided more income to families with
substantial income increase for a family with an young children. In the case of work support pro-
income near the poverty line. And a simple power grams like the Earned Income Tax Credit, this
calculation shows that 1,000 total cases, evenly might mean extending more generous credits to
divided between experimental and control groups families with young children. In the case of child
is sufcient to provide 80 % power to detect a .20 tax credits, this could mean ensuring that the full
sd impact on child cognitive outcomes. Non- credit was refundable and also providing rela-
experimental studies reviewed above suggest that tively larger credits to families with young
this income increase would be sufcient to boost children.
test scores by around .25 sd; careful thought Interestingly, several European countries gear
would need to be given to judge whether more the time-limited benets provided by their assis-
sophisticated measures of brain functioning tance programs to the age of children. In
might be expected to change by this much, and to Germany, a modest parental allowance is avail-
take into account concerns about sample attrition. able to a mother working fewer than 20 h per
Nevertheless, the reorientation of the eld to the week until her child is 18 months old. France
goal of studying experimentally or quasi- guarantees a minimum income to most of its citi-
experimentally induced variation in changes in zens, including families with children of all ages.
SES components will vastly increase both the Supplementing this basic support is the Allocation
specicity and the certainty of our knowledge de Parent Isol (API) program for lone parents
about how SES affects the neurocognitive devel- with children under age 3. In effect, the API pro-
opment that provides the foundation for the cog- gram acknowledges a special need for income
nitive and non-cognitive skill development that in support during this period, especially if a parent
turn shape later life outcomes. wishes to care for very young children and forgo
income from employment. The state-funded
child care system in France beginning at age 3
6 Conclusions and Policy alleviates the problems associated with a parents
Implications transition into the labor force.
In emphasizing the potential importance of
Early childhood is a particularly sensitive period policies to boost income in early childhood, we
in which economic deprivation may compromise do not mean to suggest that this is the only policy
childrens health and employment opportunities. path worth pursuing. Obviously investments later
Early Childhood Poverty: Short and Long-Run Consequences Over the Lifespan 353

in life and those that provide direct services to formation. In E. Hanushek & F. Welch (Eds.),
Handbook of the economics of education (pp. 697
children and families may also be well advised.
812). Amsterdam: North-Holland.
Economic logic requires a comparison of the Currie, J., & Lin, W. (2007). Chipping away at health:
costs and benets of the various programs that More on the relationship between income and child
seek to promote the development of disadvan- health. Health Affairs, 26, 331344.
Dahl, G. B., & Lochner, L. (2012). The impact of family
taged children throughout the life course. In this
income on child achievement: Evidence from the
context, expenditures on income-transfer and earned income tax credit. American Economic Review,
service-delivery programs should be placed side 102, 19271956.
by side and judged by their benets and societys Danese, A., Pariante, C., Caspi, A., Taylor, A., & Poulton,
R. (2007). Childhood maltreatment predicts adult
willingness to pay for the outcomes they pro-
inammation in a life-course study. Proceedings of the
duce, relative to their costs. National Academy of Sciences, 23(104), 3191324.
DiPrete, T., & Eirich, G. (2006). Cumulative advantage as
a mechanism for inequality: A review of theoretical
and empirical developments. Annual Review of
References Sociology, 32, 271297.
Dodge, K. H., Pettit, G. S., & Bates, J. E. (1994).
Akee, R. K. Q., Copeland, W. E., Keeler, G., Angold, A., Socialization mediators of the relation between socio-
& Costello, E. J. (2010). Parents incomes and chil- economic status and child conduct problems. Child
drens outcomes: A quasi-experiment. American Development, 65, 649665.
Economic Journal: Applied Economics, 2, 86115. Duncan, G. J. (2006). Income and child well-being:
Barker, D., Forsn, T., Eriksson, J., & Osmond, C. (2002). Thirty-fourth Geary Lecture, 2005 (Vol. 34). Dublin:
Growth and living conditions in childhood and hyper- Economic and Social Research Institute.
tension in adult life: A longitudinal study. Journal of Duncan, G., & Magnuson, K. (2005). Test score gaps: The
Hypertension, 20, 19511956. contributions of family economic and social condi-
Becker, G. S. (1991). A treatise on the family. Cambridge, tions. The Future of Children, 15, 3552.
MA: Harvard University Press. Duncan, G. J., & Murnane, R. (Eds.). (2011). Whither
Bradley, R. H., Corwyn, R. F., McAdoo, H. P., & Garcia opportunity: Rising inequality, schools, and childrens
Coll, C. (2001). The home environments of children in life chances. New York: Russell Sage.
the United States. Part 1: Variations by age, ethnicity, Duncan, G. J., Magnuson, K., & Ludwig, J. (2004). The
and poverty status. Child Development, 72, endogeneity problem in developmental studies.
18441867. Research in Human Development, 1, 5980.
Brody, G. H., & Stoneman, Z. (1994). Financial resources, Duncan, G., Gennetian, L., & Morris, P. (2009). Parental
parent psychological functioning, parent cocaregiving, pathways to self-sufciency and the well-being of
and early adolescent competence in rural two-parent younger children. In C. Heinrich & K. Scholz (Eds.),
African-American families. Child Development, 65, Making the work-based safety net work better:
590605. Forward-looking policies to help low-income families
Conger, R. D., & Conger, K. J. (2000). Resilience in mid- (pp. 117148). New York: Russell Sage.
western families: Selected ndings from the rst Duncan, G. J., Ziol-Guest, K. M., & Kalil, A. (2010).
decade of a prospective longitudinal study. Journal of Early childhood poverty and adult attainment, behav-
Marriage & Family, 64, 361373. ior, and health. Child Development, 81, 306325.
Conger, R., & Dogan, S. (2007). Social class and social- Duncan, G. J., Morris, P. A., & Rodrigues, C. (2011).
ization in families. In J. Grusec & P. Hastings (Eds.), Does money really matter? Estimating impacts of fam-
Handbook of socialization theory and research ily income on young childrens achievement with data
(pp. 433460). New York: Guilford Press. from random-assignment experiments. Developmental
Conger, R., & Elder, G. H., Jr. (1994). Families in trou- Psychology, 47, 12631279.
bled times: Adapting to change in rural America. Elder, G. H. (1974). Children of the great depression.
New York: Aldine de Gruyter. Chicago: University of Chicago Press.
Conger, R., et al. (2002). Economic pressure in African Entwisle, D., Alexander, K., & Olson, L. (2003). The rst
American families: A replication and extension of the grade transition in life course perspective. In
family stress model. Developmental Psychology, 38, J. Mortimer & M. Shanahan (Eds.), Handbook of the
179193. life course (Vol. 1, pp. 229250). New York: Kluwer.
Costello, E. J., Compton, S. N., Keeler, G., & Angold, A. Evans, W., & Garthwaite, C. (2010). Giving mom a break:
(2003). Relationships between poverty and psychopa- The impact of higher EITC payments on maternal
thology: A natural experiment. JAMA, the Journal of health (Working Paper 16296). Cambridge, MA:
the American Medical Association, 290(15), National Bureau of Economic Research.
20231029. Evans, G., & Schamberg, M. (2009). Childhood poverty,
Cunha, F., Heckman, J. J., Lochner, L., & Masterov, D. chronic stress, and adult working memory. Proceedings
(2006). Interpreting the evidence on life cycle skill
354 A. Kalil et al.

of the National Academy of Sciences, 106, parenting, and socioemotional development. Child
65456549. Development, 61, 311346.
Guryan, J., Hurst, E., & Kearney, M. (2008). Parental edu- Miller, G. E., Chen, E., Fok, A. K., Walker, H., Lim, A.,
cation and parental time with children. Journal of Nicholls, E. F., et al. (2009). Low early-life social
Economic Perspectives, 22, 2346. class leaves a biological residue manifested by
Holmlund, H., Lindahl, M., Plug, E. (in press). The causal decreased glucocorticoid and increased proinamma-
effect of parents schooling on childrens schooling: A tory signaling. Proceedings of the National Academy
comparison of estimation methods. Journal of of Sciences, 106, 1471614721.
Economic Literature. Milligan, K., & Stabile, M. (2011). Do child tax benets
Kalil, A., Ryan, R., & Corey, M. (2012). Diverging destinies: affect the well-being children? Evidence from
Maternal education and the developmental gradient in Canadian child benet expansions. American
time with children. Demography, 49, 13611383. Economic Journal-Economic Policy, 3, 175205.
Kaushal, N., Magnuson, K., & Waldfogel, J. (2011). How Morris, P., Huston, A., Duncan, G., Crosby, D., & Bos, H.
is family income related to investments in childrens (2001). How welfare and work policies affect children:
learning? In G. J. Duncan & R. J. Murnane (Eds.), A synthesis of research. New York: MDRC.
Whither opportunity? Rising inequality, schools, and Morris, P., Duncan, G. J., & Clark-Kauffman, E. (2005).
childrens life chances (pp. 187206). New York: Child well-being in an era of welfare reform: The sen-
Russell Sage. sitivity of transitions in development to policy change.
Kessler, R., & Cleary, P. (1980). Social class and psycho- Developmental Psychology, 41, 919932.
logical distress. American Sociological Review, 45, Sapolsky, R. (2004). Mothering style and methylation.
463478. Nature Neuroscience, 7, 791792.
Knudsen, E., Heckman, J., Cameron, J., & Shonkoff, J. Shanahan, M., & Hofer, S. (2010). Molecular genetics,
(2006). Economic, neurobiological, and behavioral aging and well-being: Sensitive period, accumulation
perspectives on building Americas future workforce. and pathway models. In R. Binstock & L. George
Proceedings of the National Academy of Sciences, (Eds.), Handbook of aging and the social sciences (7th
103, 1015510162. ed., pp. 135147). Amsterdam: Elsevier.
Lareau, A. (2003). Unequal childhoods: Class, race, and Smolensky, E., & Gootman, J. (Eds.). (2003). Working
family life. Berkeley: University of California Press. families and growing kids: Caring for children and
Lken, K. (2010). Family income and childrens educa- adolescents. Washington, DC: National Academies
tion: Using the Norwegian oil boom as a natural Press.
experiment. Labour Economics, 17, 118129. Sobel, M. E. (1998). Causal inference in statistical models
Lken, K., Mogstad, M., & Wiswall, M. (2012). What lin- of the process of socioeconomic achievement A case
ear estimators miss: The effects of family income on study. Sociological Methods & Research, 27,
child outcomes. American Economic Journal: Applied 318348.
Economics, 4, 135. Spears, D. (2011). Economic decision-making in poverty
Mayer, S. (1997). What money cant buy: The effect of depletes behavioral control. The B.E. Journal of
parental income on childrens outcomes. Cambridge, Economic Analysis & Policy 11(1) (Contributions).
MA: Harvard University Press. Article 72.
Maynard, R. A. (1977). Some aspects of income distribu- Strauss, R. (1997). Effects of the intrauterine environment
tion: The effects of the rural income maintenance on childhood growth. British Medical Bulletin, 53,
experiment on the school performance of children. 8195.
The American Economic Review, 67, 370375. Votruba-Drzal, E. (2006). Economic disparities in middle
Maynard, R., & Murnane, R. (1979). The effects of a neg- childhood: Does income matter? Developmental
ative income tax on school performance: Results of an Psychology, 42, 11541167.
experiment. Journal of Human Resources, 14, Waldfogel, J., & Washbrook, E. (2011). Early years pol-
463476. icy. Child Development Research, 2011, 12 pages.
McLeod, J., & Kessler, R. (1990). Socioeconomic status Ziol-Guest, K. M., Duncan, G. J., Kalil, A., & Boyce, T.
differences in vulnerability to undesirable life events. (2012). Early childhood poverty, immune-mediated
Journal of Health and Social Behavior, 31, 162172. disease processes and adult productivity. Proceedings
McLoyd, V. C. (1990). The impact of economic hardship of the National Academy of Sciences, 106(Supplement
on black families and children: Psychological distress, 2), 1728917293.
Does the Body Forget? Adult
Health, Life Course Dynamics,
and Social Change

Mark D. Hayward and Connor M. Sheehan

1 Introduction cohorts born at different points in history. In turn,


changes in the nature and timing of life course
The life course approach is a powerful lens for experiences can lead to a host of changes in pop-
understanding the origins of adult health prob- ulation health. Major conditions that were once
lems. Most highly prevalent adult health prob- important in dening the health of the population
lems are a long time in the making and reect the may wane (e.g., the long-run decline in heart dis-
combinations of exposures stretching from the ease or the near eradication of polio) while other
prenatal environment and childhood, to adoles- conditions may become more prevalent (e.g.,
cence, and adulthood. Because the social condi- some cancers). The concept of the birth cohort is
tions, institutions, and stratication systems of critical in understanding these trends, due to
societies are powerful forces that shape the nature cohort differences in the nature and timing of
of individuals life courses, adult health is funda- exposures. For example, given the role of child-
mentally a reection of these lifetime social hood vaccinations in improving childhood health
forces. and survival in the twentieth century in the United
Social conditions, institutions, and stratica- States (Andre et al. 2008; Centers for Disease
tion systems, however, are far from static. Over Control and Prevention 1999), combined with the
the course of the twentieth century, for example, dramatic decline in adult smoking (Fenelon and
the social capacity for health in many countries, Preston 2012), more recent birth cohorts have
particularly more developed countries, has dra- experienced fewer and less exposure to lifetime
matically improved through technological inno- health risks than earlier cohorts, contributing in
vation, dramatic growth in biomedical knowledge, important ways to historical declines in U.S.
and improvements in social institutional resources adult mortality (Yang 2008). Thus, the health of
(Easterlin 1997), and these societal forces have birth cohorts, often portrayed as age-specic
had rippling consequences for changes in the trends in health, may change due to fundamental
nature and timing of life course experiences for shifts in the nature and timing of life course expe-
riences. A key issue that is developed in this
chapter is that life course inuences on adult
health are largely endogenous to the historical
M.D. Hayward (*) C.M. Sheehan context and that we are only at the initial stages
Population Research Center and Department
of understanding how life course inuences on
of Sociology, University of Texas at Austin,
Austin, TX, USA health are being transformed by changes in the
e-mail: mhayward@prc.utexas.edu social capacity for health.

Springer International Publishing Switzerland 2016 355


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_16
356 M.D. Hayward and C.M. Sheehan

The idea that life course inuences are 2 From Biological Risk
embedded in a historical context is hardly new, to Mortality: A Population
and it has long been recognized as a central tenet Health Perspective
in the life course conceptual framework (Elder
et al. 2003; Elder 1994, 1998; Riley 1987). Most research on the life course origins of adult
Despite general agreement over this conceptual health focuses on single diseases or domains of
framework, however, researchers applying this health. Examples include studies of cardiovascu-
framework to adult health outcomes nonetheless lar disease, diabetes, physical and cognitive func-
face important practical and conceptual chal- tioning, disability and all-cause mortality. Only
lenges that warrant serious consideration. Three recently have life course researchers begun to
challenges are considered in this essay. First, we take a more integrative view of adult health and
take up the thorny question of how life course examined how life course experiences inuence
researchers ought to dene health. We offer a pop- multiple facets of adult health (e.g., how life
ulation health perspective as one possible frame- course factors inuence the interplay of function-
work to understand how life course exposures ing and mortality to determine healthy life expec-
give rise perhaps differentially to a portfolio tancy (Montez and Hayward 2014)). Life course
of health outcomes encompassing biological risk, researchers have largely ignored the idea of how
morbidity, functioning and disability, and mortal- exposures may give rise to a cascade of adult
ity. Second, we argue that life course studies of health conditions starting, say, from morbidity, to
adult health would benet signicantly by having disability, and then to mortality or not. Research
a clearer biologically informed framework of how often assumes that such a cascade exists, it under-
life course exposures from childhood into adult- lies more endogenous health outcomes such as
hood potentially contribute to the development of disability and mortality, and is unidirectional
adult health conditions. In the discussion below, despite warnings that these assumptions need not
we discuss how life course exposures shape health necessarily hold (Verbrugge and Jette 1994).
and health disparities through both developmental Health at the population level is a multidimen-
and aging processes. Finally, this chapter makes a sional concept, and it is difcult to describe the
case that life course research on adult health must life course origins of population health in a sim-
attend more explicitly to the historical context to ple way. Increasingly, denitions of health refer-
better understand trends and differences in the life ence the core domains of physiological
course pathways leading to adult health problems. dysregulation (e.g., metabolic functioning), con-
Dramatic changes have occurred across current ditions (e.g., disease conditions such as diabetes),
birth cohorts represented in the adult population functioning (e.g., physical and cognitive decits),
in their prenatal, childhood and adult exposures, and important facets of well-being and health
yet these changes are rarely central in life course potential (e.g., ability to live independently,
studies of health. The development of conceptual developmental potential). The measurement of
models requires sensitivity to the fact that strati- these domains is complex and so too are the rela-
cation systems, institutions, social conditions, tionships among these domains. Physiological
technology, and even the epidemiological envi- dysregulation, disease onset, functioning loss,
ronment are changing, which has implications for and frailty, for example, are all parts of the pro-
appropriate measures and analytical strategies. cess of health change that can but need not
These issues are raised to better guide the precede death (Crimmins et al. 2010; Crimmins
development of future life course studies of and Beltrn-Snchez 2011; Martin et al. 2010).
health. Researchers are increasingly awash in a The result is that population trends in these
wealth of data from around the world. Whether domains need not move in the same direction,
they have the conceptual and analytical tools to and social group differences may vary, depending
make the most of these data is not clear. on the domain of interest.
Does the Body Forget? Adult Health, Life Course Dynamics, and Social Change 357

Life course exposures and behaviors need not health disparities as well as trends in population
affect all parts of the process of health change in health and their life course origins. For example,
the same way. Some life course exposures may with regard to racial/ethnic health disparities,
be specic to certain parts of the process of health native-born U.S. older adult whites and blacks
change, and this is important in considering poli- appear to experience a health process in which
cies and interventions to improve health. For increased physiological dysregulation heightens
example, it is plausible that some childhood disease onset, which in turn heightens functional
health problems (e.g., infectious conditions that loss, disability and death (Crimmins et al. 2004;
heighten inammation) give rise to adult morbid- Hayward et al. 2000; Hayward and Heron 1999).
ity conditions (e.g., coronary heart disease) that In this case, blacks experience higher risk com-
results in a cascade of increased risks of func- pared to whites throughout the health process.
tional problems and mortality from cardiovascu- Adult Hispanics in the U.S., particularly older
lar disease (CVD). In this hypothetical example, foreign-born Hispanics, however, appear to expe-
childhood health problems association with rience relatively low rates of certain chronic con-
CVD-related mortality stems primarily from ditions as well as mortality compared to blacks
increasing disease incidence. and whites (Cantu et al. 2013; Lariscy et al.
In contrast, other life course factors may come 2014), while their disability rates substantially
into play through the process of health changes. exceed those for the other racial/ethnic groups
Educational attainment, for example, not only is (Hayward et al. 2014a; Melvin et al. 2014). The
associated with a lower risk of heart disease inci- etiologies of these health domains appear to vary
dence but potentially educational attainment may in important ways across the racial/ethnic groups,
differentiate the risks of functional problems and pointing to the need to carefully identify how life
mortality among persons with heart disease. course factors and pathways may come into play
Educational attainment positively inuences differently across the groups. How we think
human agency and nancial resources, factors about health disparities in this example necessar-
that allow persons to better manage disease pro- ily depends on the domain of health of interest.
cesses, alter the environment to minimize physi- A population health portfolio of measures,
cal challenges, and actively seek the best care thus, allows for a more comprehensive assess-
(Mirowsky and Ross 2003). This point was illus- ment of health disparities. The same is also the
trated by Manton and colleagues (1997), who case for assessing trends in population health.
observed that higher educational attainment was
associated with a lower risk of functional prob-
lems, and a lower risk of mortality after condi- 3 A Biologically Informed
tioning on functional status using data from the Conceptual Framework
National Long-Term Care Study. In this case, it for the Life Course Origins
appears that a major social resource acquired of Adult Health
relatively early in life, educational attainment,
exacerbates disparities throughout the disable- Life course research on adult health often lacks a
ment process (Manton et al. 1997; Merkin et al. biologically informed conceptual framework of
2009). how exposures and events inuence health. Life
Although life course studies of health are course researchers appropriately attend to issues
beginning to investigate how life course factors such as the timing of exposures, turning points in
inuence health processes involving multiple a lifetime, human agency, and linked lives (Elder
health domains and their connections (e.g., et al. 2003; Elder 1994, 1998), but the conceptual
Freedman et al. 2008; Manton et al. 1997; Montez connections of these more social types of
and Hayward 2014), the body of research as a concepts with adult health are less evident in the
whole is relatively small. The lack of progress literature. This problem potentially results in
ultimately has implications for understanding under- or mis-conceptualized frameworks,
358 M.D. Hayward and C.M. Sheehan

conicting results, and poorly understood For heuristic purposes, the developmental
mechanisms and pathways. As will be discussed trajectories shown in Fig. 1 are expressed in
later, these problems become compounded when terms of physical and cognitive capacity, which
turning to questions about historical change or are dependent on the major biological systems. In
cross-national studies of life course inuences on this example, trajectories are shown for four peo-
health. ple, but myriad trajectories are possible in the
Although the bulk of research often fails to population. For simplicity, persons A and B are
consider a strong biologically informed concep- assumed to share the same pattern of childhood
tual framework, such frameworks exist. One such growth in capacity as do persons C and D. Persons
framework is the developmental framework of A and B begin to diverge in early adulthood, with
life course health. The core idea is that processes person B exhibiting a steeper rate of decline. The
of the major biological systems (e.g., endocrine, pattern of decline in capacity for persons C and D
immune, neurological, respiratory) display a sim- similarly begin to diverge in early adulthood.
ilar pattern of development a steady curvilinear The gure conveys a number of important
growth in functional capacity during childhood ideas to keep in mind when considering the fac-
followed by maintenance and eventually some tors that come into play in understanding differ-
decline in adulthood (Halfon and Hochstein ences/disparities in adult health.
2002). Figure 1 summarizes the basic life course
trajectory of functional capacity that has been Differences in capacity begin as early as in
used to summarize a number of biological sys- utero and are evident at birth. This speaks to
tems (Halfon and Hochstein 2002; Kuh 2007; the seminal work by Barker and his colleagues
Miller et al. 2011). Halfon and Hochstein (2002) (Barker 1997, 1998, 1999, 2004).
and Miller and colleagues (2011) provide more Childhood is a period of growth in capacity,
detailed discussions of the micropathways and differences in capacity widen because of
underlying these types of functional trajectories. differences in inputs throughout childhood
Here, we focus on the utility of this framework (e.g., SES, diet, family relationships). This is
for understanding the life course origins of adult the period in which cohort morbidity pheno-
health. types (e.g., lifelong health risks that accrue

Determinants of loss (e.g., genes, disease, SES, exercise, weight, smoking,


disease, marital conflict, life events
45
Determinants of
gain (e.g., SES, 40
genes, birth weight,
LEVEL OF CAPACITY

diet, disease, 35
exercise, muscle 30
and cognitive
development, 25
family relations)
20
15
10
5
0

AGE

A B C D

Fig. 1 Trajectories of physical and cognitive capacity over the life course (20 = hypothetical level below which adult
limitations occur)
Does the Body Forget? Adult Health, Life Course Dynamics, and Social Change 359

through early life exposures such as infection mechanisms, there is evidence that factors such
and inammation) are established (Crimmins as maternal smoking, infections, and malnutri-
and Finch 2006; Finch and Crimmins 2004). tion may initiate the atherosclerosis process even
The maximum level and period of growth are before birth (Leduc et al. 2010; Napoli et al.
highly dependent on the level of inputs. In the 1999; Palinski and Napoli 2002). During child-
gure, for example, persons A & B reach a hood, factors such as fatty diets, infections, and
maximum value of 39 at age 30 compared to a obesity may accelerate atherogenesis (Charakida
maximum value of 33 at age 25 for persons C et al. 2007; Crimmins and Finch 2006; Gurven
& D. In this sense, both the length and degree et al. 2008; Juonala et al. 2005). Adult exposures
of development are highly malleable to inputs that elevate inammation and other pathways
(Finch and Crimmins 2004). such as smoking, obesity, socioeconomic status,
Declines in capacity can start at different ages and psychosocial stress further accelerate the
and exhibit varying rates of decline. Resources process (Libby et al. 2002; Roberts et al. 2010;
such as quality jobs, marriages, and the avoid- Wang et al. 2007), while statins may control the
ance of disease result in less step declines, further development of atherosclerosis (Blum
e.g., persons A and C. The lack of resources and Shamburek 2009). The key point is that the
and risk factors (e.g., smoking) contribute to combination of early and later-life factors fre-
relatively accelerated rates of decline for per- quently comes into play in inuencing the devel-
sons B and D. Indeed, it is possible that despite opment of adult diseases (see Dannefer Opening
early life developmental advantages, adult cir- the Social: Sociological Imagination in Life
cumstances can result in rates of decline that Course Studies, this volume). Different combi-
negate their earlier advantage; note the cross- nations of exposures of the life course can lead to
ing of the trajectories for persons B and C. a signicantly degree of heterogeneity in disease
The combination of lifetime gains and losses development such as illustrated in Fig. 1.
results in growing heterogeneity of capacity Although this idea is not novel, it nonetheless
across most of the adult life course the has broader implications. For example, it points
period of life where losses occur in capacity. to the need for life course researchers to attend to
Assuming a hypothetical score of 20 as an the combinations of lifetime exposures that may
indicator of limitations, the ages at which put adults at greater or lessor risk of various
people reach this level differ greatly. Moreover, health outcomes. In addition, this idea suggests
the pathways can differ greatly for persons that, depending on the inputs over a lifetime, the
who become limited at the same age. Note the balance of childhood and adult inuences on
trajectories for persons B and C. health disparities can change. This is a very
important idea in that the developmental ori-
In a very real sense, growth and decline in gins of adult health are highly dynamic and
capacity both contribute to health disparities, mutable to a variety of inputs across the lifetime.
pointing directly to the importance of both child- No phase is inherently more important that the
hood and adulthood for adult health disparities. other in terms of inuencing adult health dispari-
In this sense, the body does not appear to forget. ties. That said, life course stratication processes
An example might help to convey this idea. often reinforce childhood advantages and disad-
Atherosclerosis is the cause of myocardial infarc- vantages in adulthood, compounding the effects
tion, stroke and ischemic heart disease. of childhood (see also chapter Life Course Lens
Atherosclerosis involves on-going inammatory on Aging and Health by Ferraro, this volume).
response at all stages of the disease, and research There is no lottery in early adulthood that ran-
has documented an array of risk factors across domly assigns adolescents to adult trajectories of
the entire life course inuencing inammation. resources and risks. Moreover, some combinations
Although debate continues about the specic role of life course exposures are likely to be relatively
of the prenatal environment and the biological common while others quite rare due to life course
360 M.D. Hayward and C.M. Sheehan

stratication processes and social change. As disease exposure and rising obesity. This period
discussed below, the malleability of life course was characterized rst by massive increases in
trajectories in physical and cognitive capacity has post-secondary educational institutions and the
signicant implications for understanding social prevalence of post-secondary education post
group differences and trends in the life course WWII. This trend was followed in the later part
origins of adult health. of the century by the stalling of post-secondary
attainment and the rise in student load debt.
These are only a few examples but they lead to
4 The Importance of Social important questions about how researchers mea-
Change for Life Course sure key concepts relating to the types of inputs
Inuences on Adult Health over cohorts lifetimes (e.g., childhood health,
socioeconomic resources) shown in Fig. 1.
Much of the research on the life course inuences As Ryder noted 50 years ago, [t]he principal
on adult health has occurred in the context of motor of contemporary social change is techno-
single birth cohorts or a relatively narrow band of logical innovation (Ryder 1965, p. 851).
birth cohorts, e.g., the British Cohort Studies and Technological change is fundamentally impor-
the U.S. Health and Retirement Study. In a very tant for the level of social capacity for population
real sense, much of what we know or think we health (Easterlin 1997). It is embedded in social
know about the life course origins of adult institutions and denes the stock of knowledge
health is necessarily framed by the experiences of and institutional resources that individuals in the
individuals in these studies. With the explosion population have access to and can act on to gar-
of international studies built on the life course ner health advantages. The idea is similar to
framework of the Health and Retirement Study Fogels concept of technophysio evolution (Fogel
(including the Study on Global Ageing and 2004; Fogel and Costa 1997) which reects the
Health (SAGE)), we also have an expanding synergistic association between technological
array of studies to conduct comparative research and physiological improvements in the modern
on a global scale. era. Although technological change is sometimes
Yet, there are a number of challenges both in thought to be felt most strongly by persons about
interpreting what we think we know and what we to make lifelong choices (Ryder 1965), the
might learn from the proliferation of new data. developmental trajectories shown earlier in Fig. 1
Perhaps the key issue confronting researchers is illustrate that technological change may have
that research is often insensitive to the specic implications for inputs of changes in physical
historical conditions that characterized the and cognitive capacity at all stages of the life
cohorts experiences at particular ages in the life course.
course. Stratication systems and their metrics The pace of technological change is an issue
may change across cohorts (and differ across that is rarely considered in life course studies of
countries). For example, occupation was more heath, yet this phenomenon has enormous impli-
strongly tied to resources that garnered health cations for understanding how the inputs to the
advantages in the early part of the twentieth cen- growth and decline in capacity ultimately inu-
tury in the U.S. compared to educational attain- ence adult health (Goldin and Katz 2009;
ment (Hayward and Gorman 2004; Preston and McEniry 2014; Palloni and Souza 2013). To
Haines 1991). However, educational attainment, illustrate how rapid technological change can
particularly advanced education, has grown in be and how the pace of change (and variation in
importance for reducing mortality in the U.S. in the types of inputs) can vary across countries
the last half of the twentieth century (Hayward Table 1 shows infant mortality rates (IMR) for a
et al. 2014b; Masters et al. 2012; Montez et al. select group of countries for a 50-year period,
2011). The twentieth century in the United States 19502010 (National Center for Health Statistics
also was a period of large reductions in infectious 2014, p. Table 16). Infant mortality rates provide
Does the Body Forget? Adult Health, Life Course Dynamics, and Social Change 361

Table 1 Infant mortality rates for selected countries, 19602010 (infant deaths per 1,000 live births) (National Center
for Health Statistics 2014)
International rankings
Country 1960 1970 1980 1990 2000 2010 1960 2010
Australia 20.2 17.9 10.7 8.2 5.2 4.1 5 20
Austria 37.5 25.9 14.3 7.8 4.8 3.9 19 19
Belgium 31.4 21.1 12.1 8.0 4.8 3.6 17 12
Canada 27.3 18.8 10.4 6.8 5.3 12
Chile 120.3 79.3 33.0 16.0 8.9 7.4 27 27
Czech Republic 20.0 20.0 16.9 10.8 4.1 2.7 4 5
Denmark 21.5 14.2 8.4 7.5 5.3 3.4 8 9
Finland 21.0 13.2 7.6 5.6 3.8 2.3 6 1
France 27.7 18.2 10.0 7.3 4.5 3.6 13 12
Germany 35.0 22.5 12.4 7.0 4.4 3.4 18 9
Greece 40.1 29.6 17.9 9.7 5.9 3.8 20 15
Hungary 47.6 35.9 23.2 14.8 9.2 5.3 23 23
Ireland 29.3 19.5 11.1 8.2 6.2 3.8 15 15
Israel 24.2 15.6 9.9 5.5 3.7 14
Italy 43.9 29.6 14.6 8.1 4.3 3.4 22 9
Japan 30.7 13.1 7.5 4.6 3.2 2.3 16 1
Korea 45.0 3.2 7
Mexico 92.3 52.6 19.4 14.1 26 29
Netherlands 16.5 12.7 8.6 7.1 5.1 3.8 2 15
New Zealand 22.6 16.7 13.0 8.4 6.3 5.5 10 24
Norway 16.0 11.3 8.1 6.9 3.8 2.8 1 6
Poland 56.1 36.4 25.4 19.4 8.1 5.0 24 22
Portugal 77.5 55.5 24.3 10.9 5.5 2.5 25 3
Slovak Republic 28.6 25.7 20.9 12.0 8.6 5.7 14 25
Spain 43.7 28.1 12.3 7.6 4.3 3.2 21 7
Sweden 16.6 11.0 6.9 6.0 3.4 2.5 3 3
Switzerland 21.1 15.1 9.1 6.8 4.9 3.8 7 15
Turkey 189.5 145.0 117.5 51.5 31.6 7.8 28 28
United Kingdom 22.5 18.5 12.1 7.9 5.6 4.2 9 21
United States 26.0 20.0 12.6 9.2 6.9 6.1 11 26

an excellent indicator of the nature of epidemio- and a knowledge base and socioeconomic
logical environment and the social capacity for resources that allow parents to minimize risk.
population health (Easterlin 1997). High infant As is evident, infant mortality dropped dra-
mortality rates reect an epidemiological envi- matically for all selected countries since 1960,
ronment characterized by high rates of infectious and there has been a marked convergence in the
exposure, malnutrition, poor sanitation, few rates (and early life epidemiological conditions)
parental socioeconomic resources to avoid risks, by 2010. Yet, some countries experienced greater
an absence of modern health care, and a lack of change than others, due largely to adverse epide-
knowledge about how to minimize risk (Guyer miological environments in 1960. This is clearly
et al. 2000). In comparison, low infant mortality the case for Chile, Turkey and Mexico. Even
rates reect an environment with low levels of within Europe and countries that are part of the
exposure to infections, well developed sanitation Survey of Health, Ageing, and Retirement in
systems and clean water, advanced health care, Europe (SHARE) there was substantial
362 M.D. Hayward and C.M. Sheehan

variation in the rate of decline. The implications exposure over the 1960 birth cohorts entire life-
of these types of patterns for life course studies of time. Although atherosclerosis is affected by fac-
adult health are substantial. Continuing with the tors other than infection, the patterns shown in
example of early life infection and the develop- Table 1 illustrate the types of heterogeneity in the
ment of atherosclerosis, for birth cohorts born in range and pace of technological change reected
1960, exposure to infections was much lower in in adult populations and its potential impact on
some countries (e.g., Netherlands, Norway, and an important disease process.
Sweden) with already low levels of IMR com- Palloni and Souza (2013) also make a similar
pared to exposure in other countries (e.g., Turkey, argument in their study of adult mortality in Latin
Mexico, Chile, Portugal) with high IMRs. This American and Caribbean countries. Using a sce-
suggests that the development of atherosclerosis nario approach, their study considers how
started much earlier in the life course and was cohorts exposure to adverse early conditions
faster in countries such as Chile, Mexico, inuences old-age mortality. Their results illus-
Portugal, and Turkey compared to countries such trate that to the extent that exposure to adverse
as Norway, the Netherlands, and Sweden. In early life conditions is high, it is difcult to offset
addition, if one looks at the rates across years these effects even with advances in medical and
within a country, the pattern points to the lifetime other technology. More generally, the pace of
development of atherosclerosis stemming from technological change in combination with early
infection. Turkey, for example, exhibits a pattern life exposures shown in Table 1 relates strongly
of a relatively high level of infectious exposure to the heterogeneity of lifetime inputs to the gain
across most of the lifetime of the 1960 birth and loss of physical and cognitive capacity shown
cohort, pointing to the likely important role of in Fig. 1.
infection for atherosclerosis over most of the As noted earlier, technological change and its
lifetime. Chile exhibits high rates of exposure in pace are an integral part of the larger process of
childhood and early adulthood for this cohort, yet change in the social capacity for health in the
the exposure drops dramatically in those years population (Easterlin 1997). To get a better sense
indicative of prime adulthood. At the other of how the social capacity for health has changed
extreme, countries such as Sweden, Norway and in the United States, Table 2 provides a timeline
the Netherlands exhibited low levels of infectious of major social, biomedical, and technological

Table 2 Timeline of twentieth century major social, biomedical, and technological changes inuencing population
health
19011910
Public Health Service Established Beta blockers invented to lower BP and risk of heart
attacks and strokes
US Meat Inspection Act Highway Safety Act
Existence of different blood types discovered National Trafc and Motor Vehicle Act
19111920
Refrigerators for home use Mumps vaccine licensed
Rabies and typhoid vaccines licensed in US Mammography developed
Spanish u epidemic First humans walked on the moon
Vitamins discovered EPA & OSHA established
19211930 19711980
Tetanus vaccine developed DDT banned
Diphtheria & tetanus vaccines recommended Divorce rates began to escalate
for children
(continued)
Does the Body Forget? Adult Health, Life Course Dynamics, and Social Change 363

Table 2 (continued)
TB vaccine developed Car airbags invested
Insulin (best treatment of diabetes) was discovered WIC program made permanent
Penicillin discovered Safe Drinking Water Act
Frozen food processes discovered First drive-thru McDonalds
Microsoft founded
19311940 Apple 1 was built
TVA founded & Rural Electrication Act Pneumonia vaccine licensed
Social Security Act Antiviral drugs developed
Minimum Wage 19811990
A shift in home to hospital deliveries FDA approved rst commercial statin
DDT invented Women earned more bachelors degree than men for the
1st time
19411950 Commercial internet providers emerged
Mass production of penicillin Start of US obesity epidemic
Inuenza vaccine licensed Food labeling legislation enacted
Risk factors identied for heart disease Human Genome Project begins
CDC established 19912000
Water uoridation started Introduction of pneumococcal conjugate vaccine and
routine rotavirus vaccination
19511960 Rapid rise in Americans internet use
Polio epidemic with 57,628 cases reported in Google was incorporated
US; vaccine licensed
Fluoride toothpaste BlackBerry smart phone was launched
Commercial nuclear energy begins Genetically engineered crops developed for commercial
use
Interstate Highway System launched Large increase in immigration from Latin America
Sputnik and the start of the space race 20012010
2.2 million vets had used the GI Bill to attend college The genome is sequenced
& 5.6 million used benets for other training programs
19611970 Strides made in stem cell research
National uoride recommendations Major strides made in controlling HIV
Measles vaccine licensed Targeted cancer therapies
Beginning of 12-years escalation in post-secondary First hybrid car
educational enrollment
U.S. Surgeons report on smoking and health Smartphone use rapidly grew
Head Start
Higher Education Act provided need based nancial
assistance
Medicare and Medicaid Established
Cigarette warning labels
FDA approved the rst combined oral contraceptive

changes (both positive and negative) that reect a though not all led to gains in social capacity.
centurys change in the social capacity for health Prior to 1930, many of the changes were related
in the U.S. to biomedical discoveries blood types, vaccines
Scanning across the timeline, it is evident that for a number of conditions, and penicillin. Some
each decade brought important social, techno- technological changes, however, improved
logical, and biomedical changes, most of which individuals control over their health-related
364 M.D. Hayward and C.M. Sheehan

environment refrigerators for home use and How, then, have these changes in social
frozen food processes. And, the early part of the capacity for population health inuenced the
twentieth century saw some of the very rst insti- changes in the way that the life course inuences
tutional developments that had population-level adult health? A recent study by Hayward and col-
health consequences the establishment of the leagues (2014b) provide some insights into this
Public Health Service and the U.S. Meat question. Their study examined long-term trends
Inspection Act. in the ways in which educational attainment was
The 1930s marked an era of widespread insti- associated with adult mortality. Congruent with
tutional developments the expansion of electri- past research (Montez et al. 2012), their results
cal power in American homes, Social Security provided evidence that the shape of the associa-
and the setting of a minimum wage. In the 1940s, tion reects increasingly lower mortality for very
major changes occurred in public health initia- highly educated Americans compared to their
tives the mass production of penicillin, the low-educated counterparts over the latter part of
inuenza vaccine, the discovery of the risk fac- the twentieth century and continuing into the
tors for heart disease, and water uoridation. The twenty-rst century. Importantly, these results
Centers for Disease Control and Prevention point to an array of mechanisms associated with
(CDC) was also established in the 1946. advanced education that account for this pat-
In the latter half of the twentieth century, there tern increases in psychosocial resources, valu-
was an acceleration of institutional changes that able information and support for healthy
improved the social capacity for population lifestyles, quality medical care, wealth, sense of
health, especially in the 1960s the War on control and cognitive skills (Cutler and Miller
Poverty, the Civil Rights Act, the Voting Rights 2005; Cutler and Lleras-Muney 2010). These are
Act, the Food Stamp Act, Medicare and Medicaid, the types of resources made possible by the his-
and the Older Americans Act all came into exis- torical change in the social capacity for health in
tence in the 1960s. Profound changes in public the United States shown in Table 2. In this sense,
health policies occurred as well e.g., the the long-term association between education and
Surgeon Generals report on smoking and health. mortality is endogenous to the larger macro
Educational levels also skyrocketed in the changes in the social capacity for health.
American population, fueled by the growth in An important tenet in life course conceptual
post-secondary educational institutions, the GI frameworks is the importance of historical
Bill, and the Higher Education Act. Biomedical context. Here, we have seen that the historical
technologies grew rapidly in the latter part of the context is likely to matter a great deal and that
twentieth century as well e.g., the widespread rapid changes have occurred in the social capac-
use of beta blockers and statins. A number of ity for health. In addition, current surviving
technological advances were made that funda- American birth cohorts are likely to differ in
mentally changed how individuals accessed important ways in their lifetime exposures to
health-related information, including personal important features of the social capacity for pop-
computing and the internet. At the same time, ulation health. The types of exposures and the
other changes point to losses in social capacity pace of change in social capacity necessarily will
e.g., the growth of fast food as a part of the shape how birth cohorts physical and cognitive
American diet. The key point of Table 2 is that capacities are changed over the lifetimes, ulti-
the twentieth century was characterized by a pro- mately inuencing trends and disparities in the
found and enormously fast-paced improvement major domains of population health. The impor-
in the social capacity in population health that tant of broader societal forces for health can thus
was fueled by multiple and reinforcing types of only be appreciated by studies that include a
changes changes in technology, social institu- range of cohorts and that analyze the data with
tions, and biomedical knowledge. sensitivity to between-cohort differences.
Does the Body Forget? Adult Health, Life Course Dynamics, and Social Change 365

5 Conclusions think we know is necessarily constrained by a


set of macro-level boundary conditions reecting
This essay has laid out three conceptual issues the levels of technology, social institutions, and
that we view as challenges in making signicant the stock of knowledge in the population, the
advances in understanding the life course origins degree of change in these factors, and the rate at
of health. Of critical importance, we believe, is which change occurs. Second, the boundary con-
the need for researchers to carefully consider ditions set by social capacity point to the mallea-
adult health outcomes as part of a larger process bility of associations and pathways inuencing
of health change that encompasses biological adult health across time and place. Some associa-
risk, morbidity, functioning and disability and tions are likely to become more causal over time
mortality. Viewing health outcomes as part of this while other associations may become less causal
larger process is fundamentally important in or become trivial in magnitude (Hayward et al.
gaining a clearer understanding how life course 2014b). Third, adult health is clearly being
conditions inuence multiple facets of health shaped by human control at many levels by
perhaps differentially. individuals, families, communities, and larger
A second conceptual challenge is that life societal institutions. This issue speaks directly to
course conceptual frameworks are often not bio- Fogel and Costas (1997) concept of techno-
logically informed when examining health out- physio evolution. Humans have gained unprece-
comes. Little attention is thus devoted to the ways dented control over their bodies and thus shape
in which childhood and adult conditions get the trends and disparities that are evident in the
under the skin to inuence important health population. Although it is tempting to think that
processes. Here, we have revisited a developmen- this phenomenon points to ever improving popu-
tal framework in which the processes of major lation health, human control also means that
biological systems are characterized by a growth social capacity may decline in important ways.
in capacity during childhood, followed by main- The rise in obesity (see also chapter Life Course
tenance and eventual decline. We have described Lens on Aging and Health by Ferraro, this vol-
how the health of adults is forged by exposures ume), growing economic inequality, environmen-
over an entire lifetime through the gains and tal degradation, and economic segregation are
losses in capacity. only a few examples of macro forces that threaten
Finally, we have argued that although the life long-term historical improvements in health.
course framework conceptually acknowledges Moreover, there are important social group dif-
the importance of historical context, much of life ferences, especially racial/ethnic disparities, that
course research on health is a historical. The clas- point to inequality in the ways that groups are
sic debates that characterize much of the current able to marshal resources to garner health advan-
life course research on health the role of child- tages (Hayward et al. 2014b; Montez et al. 2012).
hood SES and health, the balance of childhood As is hopefully evident, future improvements
and adulthood in inuencing health, the path- in population health will come from myriad
ways linking early life with adult health are all interacting and reinforcing macro inuences
endogenous in fundamentally important ways to making up the social capacity for population
the historical context. Here, historical context has health. These inuences shape birth cohorts life-
been framed as the social capacity for health in a time gains and losses of physical and cognitive
society to better understand the macro-level capacity, and, in turn, these gains and losses play
forces in play. out in terms of trends and disparities in the key
This third challenge, the endogeneity of life domains of population health. Life course
course studies of health to macro social condi- researchers have laid important groundwork to
tions, has especially important implications for examine these processes in detail, but much more
how we think about the empirical results of cur- can be achieved by explicitly attending to
rent research. First, much of what we know or understanding how domains of health interact,
366 M.D. Hayward and C.M. Sheehan

how gains and losses of physical and cognitive Crimmins, E., & Finch, C. (2006). Infection, inamma-
tion, height, and longevity. Proceedings of the
capacity shape trends and disparities in health
National Academy of Sciences, 103(2), 498503.
outcomes, and how this basic process is embed- doi:10.1073/pnas.0501470103.
ded in a societys ever-changing social capacity Crimmins, E. M., Hayward, M. D., & Seeman, T. (2004).
for population health. Race/ethnicity, socioeconomic status and health. In
N. B. Anderson, R. A. Bulatao, & B. Cohen (Eds.),
Critical perspectives on racial and ethnic differences
Acknowledgment This work was partially supported by in health in later life (pp. 310352). Washington, DC:
grants, R01 HD05396 (PI, Robert A. Hummer), 5 R24 National Academy Press.
HD042849 (PI, Mark D. Hayward) and 5 T32 HD007081 Crimmins, E., Kim, J. K., & Vasunilashorn, S. (2010).
(PI, Jennifer Glass) awarded to the Population Research Biodemography: New approaches to understanding
Center at the University of Texas at Austin by the Eunice trends and differences in population health and mortal-
Kennedy Shriver National Institute on Child Health and ity. Demography, 47(Supplement), S41S64.
Human Development. Cutler, D. M., & Lleras-Muney, A. (2010). Understanding
differences in health behaviors by education. Journal
of Health Economics, 29(1), 128. doi:10.1016/j.
jhealeco.2009.10.003.
References Cutler, D., & Miller, G. (2005). The role of public health
improvements in health advances: The twentieth-
Andre, F., Booy, R., Bock, H., Clemens, J., Datta, S., century United States. Demography, 42(1), 122.
John, T., & Schmitt, H. (2008). Vaccination greatly doi:10.1353/dem.2005.0002.
reduces disease, disability, death and inequity world- Easterlin, R. A. (1997). Growth triumphant, the twenty-
wide. Bulletin of the World Health Organization, 86, rst century in historical perspective. Ann Arbor: The
140146. University of Michigan Press.
Barker, D. J. P. (1997). Maternal nutrition, fetal nutrition, Elder, G. H. J. (1994). Time, human agency, and social
and disease in later life. Nutrition, 13(9), 807813. change: Perspectives on the life course. Social
Barker, D. J. P. (1998). Mothers, babies, and health in Psychology Quarterly, 57, 415.
later life (2nd ed.). London: Churchill Livingstone. Elder, G. H., Jr. (1998). The life course as developmental
Barker, D. (1999). Early growth and cardiovascular dis- theory. Child Development, 69(1), 112.
ease. Archives of Disease in Childhood, 80(4), 305. Elder, G., Jr., Johnson, M., & Crosnoe, R. (2003). The
Barker, D. J. P. (2004). The developmental origins of well- emergence and development of life course theory. In
being. Philosophical Transactions: Biological J. Mortimer & M. Shanahan (Eds.), Handbook of the
Sciences, 359(1449), 13591366. life course (pp. 319). New York: Springer.
Blum, A., & Shamburek, R. (2009). The pleiotropic Fenelon, A., & Preston, S. (2012). Estimating smoking-
effects of statins on endothelial function, vascular attributable mortality in the United States.
inammation, immunomodulation and thrombogene- Demography, 49(3), 797818. doi:10.1007/
sis. Atherosclerosis, 203(2), 325330. doi:10.1016/j. s13524-012-0108-x.
atherosclerosis.2008.08.022. Finch, C. E., & Crimmins, E. M. (2004). Inammatory
Cantu, P. A., Hayward, M. D., Hummer, R. A., & Chiu, exposure and historical changes in human life-spans.
C. T. (2013). New estimates of racial/ethnic differ- Science, 305(5691), 17361739. doi:10.1126/
ences in life expectancy with chronic morbidity and science.1092556.
functional loss: Evidence from the National Health Fogel, R. W. (2004). The escape from hunger and prema-
Interview Survey. Journal of Cross-Cultural ture death, 17002100. New York: Cambridge
Gerontology, 28(3), 283297. doi:10.1007/ University Press.
s10823-013-9206-5. Fogel, R. W., & Costa, D. L. (1997). A theory of techno-
Centers for Disease Control and Prevention. (1999). CDC physio evolution, with some implications for forecast-
on infectious diseases in the United States: 190099. ing population, health care costs, and pension costs.
Population and Development Review, 25(3), 635640. Demography, 34(1), 49.
doi:10.2307/172366. Freedman, V. A., Martin, L. G., Schoeni, R. F., &
Charakida, M., Deaneld, J. E., & Halcox, J. P. (2007). Cornman, J. C. (2008). Declines in late-life disability:
Childhood origins of arterial disease. Current Opinion The role of early- and mid-life factors. Social Science
in Pediatrics, 19(5), 538545. 510.1097/ & Medicine, 66(7), 15881602. doi:10.1016/j.
MOP.1090b1013e3282eff1585. socscimed.2007.11.037.
Crimmins, E. M., & Beltrn-Snchez, H. (2011). Mortality Goldin, C. D., & Katz, L. F. (2009). The race between
and morbidity trends: Is there compression of morbid- education and technology. Boston: Harvard University
ity? The Journals of Gerontology Series B: Press.
Psychological Sciences and Social Sciences, 66B(1), Gurven, M., Kaplan, H., Winking, J., Finch, C., &
7586. doi:10.1093/geronb/gbq088. Crimmins, E. M. (2008). Aging and inammation in
Does the Body Forget? Adult Health, Life Course Dynamics, and Social Change 367

two epidemiological worlds. The Journals of 1982 to 1991. Journal of Aging and Health, 9(4),
Gerontology Series A: Biological Sciences and 419450. doi:10.1177/089826439700900401.
Medical Sciences, 63(2), 196199. Martin, L. G., Schoeni, R. F., & Andreski, P. M. (2010).
Guyer, B., Freedman, M. A., Strobino, D. M., & Sondik, Trends in health of older adults in the United States:
E. J. (2000). Annual summary of vital statistics: Past, present, future. Demography, 47(Supplement),
Trends in the health of Americans during the 20th S17S40.
century. Pediatrics, 106(6), 13071317. doi:10.1542/ Masters, R. K., Hummer, R. A., & Powers, D. A. (2012).
peds.106.6.1307. Educational differences in U.S. adult mortality: A
Halfon, N., & Hochstein, M. (2002). Life course health cohort perspective. American Sociological Review,
development: An integrated framework for developing 77(4), 548572. doi:10.1177/0003122412451019.
health, policy, and research. Milbank Quarterly, 80(3), McEniry, M. (2014). Early life conditions and rapid
433479. doi:10.1111/1468-0009.00019. demographic changes in the developing world.
Hayward, M. D., & Gorman, B. (2004). The long arm of Dordrecht: Springer.
childhood: The inuence of early life conditions on Melvin, J., Hummer, R., Elo, I., & Mehta, N. (2014). Age
mens mortality. Demography, 41, 87107. patterns of racial/ethnic/nativity differences in disabil-
Hayward, M. D., & Heron, M. (1999). Racial inequality in ity and physical functioning in the United States.
active life among adult Americans. Demography, 36, Demographic Research, 31(17), 497510.
7791. Merkin, S. S., Karlamangla, A., Crimmins, E., Hayward,
Hayward, M., Crimmins, E., Miles, T., & Yu, Y. (2000). M., & Seeman, T. (2009). Education differentials by
The signicance of socioeconomic status in explain- race in the diagnosis and management of hypercholes-
ing the racial gap in chronic health conditions. terolemia. International Journal of Public Health,
American Sociological Review, 65, 910930. 54(3), S1S1.
Hayward, M., Hummer, R., Chiu, C.-T., Gonzlez- Miller, G. E., Chen, E., & Parker, K. J. (2011).
Gonzlez, C., & Wong, R. (2014a). Does the hispanic Psychological stress in childhood and susceptibility to
paradox in U.S. adult mortality extend to disability? the chronic diseases of aging: Moving toward a model
Population Research and Policy Review, 33(1), 8196. of behavioral and biological mechanisms.
doi:10.1007/s11113-013-9312-7. Psychological Bulletin, 137(6), 959997. doi:10.1037/
Hayward, M. D., Hummer, R. A., & Sasson, I. (2014b). a0024768.
Trends and group differences in the association Mirowsky, J., & Ross, C. E. (2003). Education, social sta-
between educational attainment and U.S. adult mortal- tus, and health. New York: Aldine De Gruyter.
ity: Implications for understanding educations causal Montez, J. K., & Hayward, M. D. (2014). Cumulative
inuence. Social Science & Medicine. doi:10.1016/j. childhood adversity, education, and active life expec-
socscimed.2014.11.024. tancy among U.S. adults. Demography, 51,
Juonala, M., Jrvisalo, M. J., Mki-Torkko, N., Khnen, 415435.
M., Viikari, J. S. A., & Raitakari, O. T. (2005). Risk Montez, J. K., Hummer, R. A., Hayward, M. D., Woo, H.,
factors identied in childhood and decreased carotid & Rogers, R. G. (2011). Trends in the educational gra-
artery elasticity in adulthood: The cardiovascular risk dient of U.S. adult mortality from 1986 to 2006 by
in Young Finns Study. Circulation, 112(10), 1486 race, gender, and age group. Research on Aging, 33(2),
1493. doi:10.1161/circulationaha.104.502161. 145171. doi:10.1177/0164027510392388.
Kuh, D. (2007). A life course approach to healthy aging, Montez, J. K., Hummer, R. A., & Hayward, M. D. (2012).
frailty, and capability. Journal of Gerontology: Educational attainment and adult mortality in the
Medical Sciences, 62(7), 717721. United States: A systematic analysis of functional
Lariscy, J. T., Hummer, R. A., & Hayward, M. D. (2014). form. Demography, 49, 315336. doi:10.1007/
Hispanic adult mortality in the United States: A s13524-011-0082-8.
review, new estimates, and comparisons to other popu- Napoli, C., Glass, C. K., Witztum, J. L., Deutsch, R.,
lation groups. Demography. doi:10.1007/ DArmiento, F. P., & Palinski, W. (1999). Inuence of
s13524-014-0357-y. maternal hypercholesterolaemia during pregnancy on
Leduc, L., Levy, E., Bouity-Voubou, M., & Delvin, E. progression of early atherosclerotic lesions in child-
(2010). Fetal programming of atherosclerosis: hood: Fate of Early Lesions in Children (FELIC)
Possible role of the mitochondria. European Journal study. The Lancet, 354(9186), 12341241.
of Obstetrics & Gynecology and Reproductive Biology, doi:10.1016/S0140-6736(99)02131-5.
149(2), 127130. doi:10.1016/j.ejogrb.2009.12.005. National Center for Health Statistics. (2014). Health,
Libby, P., Ridker, P. M., & Maseri, A. (2002). Inammation United States, 2013: With special feature on prescrip-
and atherosclerosis. Circulation, 105(9), 11351143. tion drugs. Hyattsville.
doi:10.1161/hc0902.104353. Palinski, W., & Napoli, C. (2002). The fetal origins of
Manton, K. G., Stallard, E., & Corder, L. (1997). atherosclerosis: Maternal hypercholesterolemia, and
Education-specic estimates of life expectancy and cholesterol-lowering or antioxidant treatment during
age-specic disability in the U.S. elderly population pregnancy inuence in utero programming and
368 M.D. Hayward and C.M. Sheehan

postnatal susceptibility to atherogenesis. The FASEB American Journal of Epidemiology. doi:10.1093/


Journal, 16(11), 13481360. doi:10.1096/ aje/kwq193.
fj.02-0226rev. Ryder, N. B. (1965). The cohort as a concept in the study
Palloni, A., & Souza, L. (2013). The fragility of the future of social change. American Sociological Review,
and the tug of the past: Longevity in Latin America 30(6), 843861. doi:10.2307/2090964.
and the Caribbean. Demographic Research, 29(21), Verbrugge, L. M., & Jette, A. M. (1994). The disablement
543578. process. Social Science & Medicine, 38(1), 114.
Preston, S. H., & Haines, M. R. (1991). Fatal years: Child Wang, H. X., Leineweber, C., Kirkeeide, R., Svane, B.,
mortality in late nineteenth century America. Schenck-Gustafsson, K., Theorell, T., & Orth-
Princeton: Princeton University Press. Gomr, K. (2007). Psychosocial stress and athero-
Riley, M. W. (1987). On the signicance of age in sociol- sclerosis: Family and work stress accelerate
ogy. American Sociological Review, 52(1), 114. progression of coronary disease in women. The
doi:10.2307/2095388. Stockholm Female Coronary Angiography Study.
Roberts, C. B., Couper, D. J., Chang, P. P., James, Journal of Internal Medicine, 261(3), 245254.
S. A., Rosamond, W. D., & Heiss, G. (2010). doi:10.1111/j.1365-2796.2006.01759.x.
Inuence of life-course socioeconomic position on Yang, Y. (2008). Trends in U.S. adult chronic disease mor-
incident heart failure in blacks and whites: The tality, 19601999: Age, period, and cohort variations.
Atherosclerosis Risk in Communities Study. Demography, 45, 387416.
Living Healthier and Longer: A Life
Course Perspective on Education
and Health

Monica Kirkpatrick Johnson, Jeremy Staff,


John E. Schulenberg, and Megan E. Patrick

1 Introduction how education relates to health-risk behaviors,


mental and physical health, and mortality, selec-
Current research using diverse datasets clearly tively summarizing past research as a foundation
shows that adults in the United States with higher for our later observations. From there we high-
levels of education enjoy better health, both light ways in which taking a life course perspec-
physically and psychologically, and they live sig- tive directs scholars and summarize issues that
nicantly longer (see reviews by Baker et al. arise as a result.
2011; Eide and Showalter 2011; Institute of To illustrate how a life course perspective can
Medicine 2012, 2014; Pampel et al. 2010; Ross be used to examine educational disparities in
and Mirowsky 2010; Woolf and Aron 2013). This health, we then briey present analyses based
association is strong and enduring, and spans a upon longitudinal data from the Monitoring the
diverse set of health indicators. Indeed, it is one Future (MTF) study. Beginning in 1976, the MTF
of the most powerful relationships in social sci- study has been conducting annual surveys with
ence research (Lynch 2003:309). nationally representative samples of high school
A large and vibrant body of scholarship exists seniors, and following a randomly selected sub-
on this basic relationship and our purpose here is sample of each cohort into adulthood with mail
not to thoroughly review it. Instead, our aim is to surveys biennially through age 30 and then every
help advance the scholarship in this area in new 5 years through middle adulthood (Johnston et al.
and needed directions by further incorporating a 2014a, b). Finally, we summarize the key issues
life course perspective. We begin with a brief covered over the course of this chapter.
summary of some of the major perspectives on

2 Education as a Cause
of Health
M. Kirkpatrick Johnson (*)
Department of Sociology, Washington State A large body of research promotes the idea that
University, Pullman, WA, USA
e-mail: monicakj@wsu.edu education shapes health in benecial ways; that
is, higher educational attainment contributes to
J. Staff
The Pennsylvania State University, better health and longevity. A key issue is the
State College, PA, USA extent to which selection plays a strong role in
J.E. Schulenberg M.E. Patrick this connection between education and health;
University of Michigan, Ann Arbor, MI, USA the literature has given attention to this issue and

Springer International Publishing Switzerland 2016 369


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_17
370 M. Kirkpatrick Johnson et al.

the search for a common cause of education and and Wu 1995). Because earnings are positively
health continues on. We will say more about this associated with these other job conditions, stud-
later, but for now we note that studies that attend ies that measure only earnings or income may
to selection effects in various ways nd that edu- also be capturing the effects of these conditions.
cational attainment independently contributes to Alongside educations effects through place-
health and longevity. Furthermore, numerous ment in the labor market, education is also
mechanisms have been identied linking greater thought to build knowledge and cognitive skills
education to a wide range of physical and mental literacy, numeracy, reasoning, and learning how
health outcomes, supporting a causal to learnall of which enhance risk assessment
perspective. and decision-making skills, enabling people to
One key mechanism involves the strong link live healthier lives (Baker et al. 2011, 2012;
between educational attainment and the condi- Mirowsky and Ross 2003). Indeed, human capi-
tions of peoples employment. Educational tal theorys explanation for the better labor mar-
attainment has a long-standing and strong, posi- ket returns of more educated people is that
tive association with workers earnings, as well education imparts skills that are valuable to
as other valued conditions of jobs (Day and employers. These same skills are resources that
Newberger 2002; Goldin and Katz 2008; Hout people who value health marshal to live lives in
2012; Kalleberg 2011). With respect to earnings, ways that promote their health (Mirowsky and
nancial resources are linked with better health Ross 2003). These skills also underlie a sense of
(Cutler and Lleras-Muney 2010; Mirowsky and having control over ones life. A strong sense of
Ross 2003; Pampel et al. 2010; Ross and Wu control provides motivation and condence to
1995). Controlling for earnings, or household pursue goals, tackle challenges and solve prob-
income more broadly, substantially reduces but lems (Mirowsky and Ross 2003). More educated
does not usually eliminate the positive relation- people tend to have more of this efcacy, or
ship of education with health (Cutler and Lleras- learned effectiveness, which enables people to
Muney 2010; Mirowsky and Ross 2003; Ross better gather and act on information, overcome
and Wu 1995). It does not appear to be the case obstacles to live a healthy life, and confront
that money simply buys better health, however. health problems when they arise (Pampel et al.
Economic resources can facilitate healthy behav- 2010). Consistent with this argument, personal
iors (e.g., gym memberships, purchasing fresh control is positively associated with health and
and healthier foods), but other healthy behaviors explains a signicant portion of the relationship
are not costly (e.g., wearing a seatbelt) and some between education and health (Mirowsky and
unhealthy behaviors are quite costly (e.g., smok- Ross 2003; Ross and Mirowsky 1999; Ross and
ing, heavy daily drinking). Observing that the Wu 1995).
primary differences in health by income appear Another way in which education is thought to
between low to middle incomes, Mirowsky and facilitate better health is through greater access to
Ross (2003) argue that it is primarily economic and utilization of health care, arising in part
hardship that undermines health through its acti- through nancial resources and the cognitive and
vation of the bodys stress response, often social psychological resources just outlined. The
intensely and over long periods of time (see also picture is not entirely straightforward, however.
Pampel et al. 2010). Utilizing check-ups and screenings is generally
With respect to other job conditions, auton- not associated with better health (Ross and Wu
omy, exibility and intrinsically engaging fea- 1995), nor is medical insurance coverage a factor
tures of jobs such as task variety, learning (Mirowsky and Ross 2003). Looking internation-
opportunities and problem solving are also linked ally, health disparities have not been reduced in
to better health and likewise explain part of the countries instituting universal health care sys-
association between education and health tems either (Glied and Lleras-Muney 2008;
(Mirowsky and Ross 2003; Qiu et al. 2012; Ross Mirowsky and Ross 2003). In Great Britain, for
Living Healthier and Longer: A Life Course Perspective on Education and Health 371

example, Mirowsky and Ross (2003) point out allows them to merge otherwise unrelated habits
that instituting national care equalized medical and ways into a healthy lifestyle. Thus between
services but did not reduce health disparities. Yet, nancial resources, social psychological
at the population level, mortality declines across resources, or social support and health outcomes
countries are signicantly related to medical also lies behavioral patterns that complete the
advances (Beckett 2000). In addition, educational links back to education levels.
gradients in mortality rise with improvements in Lifestyle differences may be driven by
health technologies (Glied and Lleras-Muney attempts of higher status individuals to distin-
2008). Glied and Lleras-Muney (2008) nd that guish themselves from the lower classes and vice
mortalitys strongest inverse relationship to edu- versa (Pampel et al. 2010). Tastes vary by class,
cation is for deaths attributable to diseases for according to this argument, with implications for
which there have been greater advances in medi- exercise levels, diet, smoking, and so on. Social
cal technology. They argue that this pattern arises networks, which are becoming increasingly
because better educated individuals adopt newer homogamous by education level in society over
technologies more quickly than less educated time, are also involved in differential sanctioning
individuals. with respect to health-related behaviors in ways
The higher levels of social support more edu- that may perpetuate behavioral differences
cated people enjoy is another mechanism through (Pampel et al. 2010).
which education is thought to operate. The range of mechanisms through which
Educational attainment is positively associated greater education is thought to improve health,
with general indicators of social support as well the range of mental and physical health condi-
as the likelihood of getting and staying married tions for which the relationship is observed,
(Goldstein and Kenny 2001; Martin n.d.; along with the durability of the relationship over
McLanahan 2004), both of which are good for time, have led to calls to consider education a
ones health (Mirowsky and Ross 2003; Pampel fundamental cause of health. As Link and
et al. 2010; Ross and Mirowsky 1999; Ross and Phelan (1995) argue, A fundamental cause
Wu 1995). Social support, through marriage and involves access to resources, resources that help
other social relationships, improves mental and individuals avoid diseases and their negative con-
physical health and buffers the negative effects of sequences through a variety of mechanisms.
stress (Thoits 1995, 2011). Thus, even if one effectively modies intervening
Finally, lifestyle factors, including what are mechanisms or eradicates some diseases, an
considered health-risk and health-maintenance association between a fundamental cause and
behaviors, loom large as mediators of the effects disease will reemerge. The resources of educa-
of education on health (Cutler and Lleras-Muney tion are exible, according to this argument, such
2010; Mirowsky and Ross 2003; Pampel et al. that no matter what strategies of risk avoidance
2010). For instance, education is inversely asso- and protective actions are conducive to good
ciated with smoking and positively associated health in a given place and time, education-
with wearing seatbelts, physical activity, and eat- related resources allow people to adopt those
ing a healthy diet (Cutler and Lleras-Muney strategies (Link et al. 2008).
2010; Ross and Wu 1995), and these factors Despite the consistency with which studies
account for around one-quarter of the association link education and health, it remains unclear in
between education and health (Pampel et al. the literature whether the health benets of edu-
2010). Indeed, most of the mechanisms discussed cation are linear or depend on degree attainment.
above also implicate participation in health-risk Though some research has found a positive rela-
and health-maintenance behaviors to some tionship between years of schooling and health
extent, as well as stress processes. Mirowsky and (Cutler and Lleras-Muney 2010; Ross and
Ross (2003:52), for example, suggest that the Mirowsky 1999), others have found that health
learned effectiveness of more educated people returns depend on degree attainment (Liu et al.
372 M. Kirkpatrick Johnson et al.

2011; Rosenbaum 2012). Psychologically, each well as to be better assessors of risky behaviors.
additional year of education may improve health Education, particularly the receipt of 4-year col-
incrementally as students gain a better under- lege degrees, also improves occupational out-
standing of health risks and become more agentic comes and facilitates stable and supportive
and skillful in addressing health problems. The unions with partners of similar educational cre-
health benets that come from greater material dentials. Each of these enable more educated
resources, however, may depend more on degree people to lead a healthier lifestyle, and to avoid
completion than total years of schooling. Though and buffer stress across the life course.
research has shown some health benets of
2-year degree completion (Rosenbaum 2012),
college dropouts might show limited health ben- 3 A Life Course Perspective
ets of postsecondary education due to the stress- on Education and Health
ors associated with high student loan debt and
their higher likelihood of oundering in the labor As is apparent from even our brief discussion
market (Vuolo et al. 2012). At least one study above, the relationship between education and
indicates college dropouts have equivalent or health has been well-studied, and recent studies
worse health outcomes than high school gradu- continue to show that educational attainment is
ates, as do those earning non-academic, positively related to self-rated health and psycho-
vocational-technical degrees after high school logical wellbeing and inversely related to mortal-
(Zajacova et al. 2012). Moreover, little research ity (Glied and Lleras-Muney 2008; Herd 2010;
has addressed the health benets of educational Herd et al. 2007; Masters et al. 2012; Montez
attainment past the 4-year bachelor degree et al. 2011, 2012; Ross and Mirowsky 2010,
despite the growing number who have attended 2011; Scharoun-Lee et al. 2011; Wheaton and
and completed graduate and professional pro- Clarke 2003; Yang 2008; Zheng et al. 2011). And
grams (the number of masters degrees conferred as we will discuss below, the life course perspec-
has increased by 50 % in the past 10 years, doc- tive is readily visible in many studies on this
toral degrees by 34 %; Snyder and Dillow 2012). issue. Given this, can the life course perspective
Research that uses more standard treatments of offer anything further to stimulate new lines of
educationby counting years of attendance or inquiry and advance our understanding of the
comparing outcomes for high school graduates social processes involved? We argue that it can,
and 4-year degree recipientsmay miss differ- and discuss several possibilities in the remainder
ences in health outcomes between those who of this chapter. First, we highlight aspects of the
start college but do not complete a degree (drop- life course perspective that are particularly perti-
outs) and those who have an associates degree nent to our observations.
(a degree that has become increasingly common) For purposes here, we focus on two important
or differences in health between students who and interrelated themes from the life course per-
attend graduate or professional school versus spective: historical time and life-long develop-
those who take more than 4 years to complete a ment. Recognizing that the life course perspective
bachelor degree (something that is also becoming offers many orienting concepts and principles to
more common). scholars (Elder et al. 2003), we suggest that one
In summary, though the functional form of of its core features is its emphasis on time. Time
educational effects is understudied, education is is marked by several clocks, one is historical,
widely theorized as a fundamental cause of health another developmental-aging, and a third is social
and mortality through the creation of hierarchies (Elder 1998). Multiple clocks are implicated
in social, psychological, informational, and mate- whenever we think about when experiences hap-
rial resources (Link and Phelan 1995). Young pen, in what order, or for how long, or when we
people are taught in school to be more agentic, think about stability and change. These clocks
future oriented, masterful, and conscientious as are also tied to one another, as captured in the life
Living Healthier and Longer: A Life Course Perspective on Education and Health 373

course perspectives emphasis on the intersection of the poor may have been a key issue, later, other
of biography and history. Thinking about time is factors arose as central mediating mechanisms.
fundamental to taking a life course perspective, For example, Link and Phelan (1995) note how
and within that, taking the long view of lives as smoking was not linked to socioeconomic status
embedded in historical times illuminates new until the 1960s, when information about the
avenues of investigation helpful in building a health risks of smoking became available. The
deeper understanding of social processes. negative association emerged as smokers from
The life course principle of time and place higher socioeconomic backgrounds quit and
(Elder et al. 2003) emphasizes historical inu- higher socioeconomic status non-smokers
ences on lives. This time-oriented lens is evident became less likely to start. Differences in other
in the extant research on education and health in health-risk behaviors by education and occupa-
the focus on historical trends, as well as cohort x tional status also emerged over time. Mirowsky
age patterns, in the education-health gradient. and Ross (2005) suggest that education became
Indeed, attention to historical time in empirical more closely tied to health because variation in
patterns was a major factor prompting the argu- health over history has become more a product of
ment that education is a fundamental cause of choices in how to live. Education, they argue,
health. Health disparities by socioeconomic sta- equips people to make and implement choices
tus in the late 1800s and early 1900s were thought with better health results.
to result from differences in basic living condi- While the reasons may not yet be fully under-
tions tied to housing, sanitation, nutrition and stood, a consideration of history has substantially
related factors (Link and Phelan 1995; Lynch advanced our understanding of health disparities.
2003). Major improvements in public health Indeed, the basic idea that the landscape of health
(e.g., improved sanitation, widespread immuni- and health care changes with time is a dening
zation) reduced mortality rates early in the twen- feature of the fundamental cause perspective. As
tieth century. Although many believed health and Link and Phelan (1995:87) explain, If no new
mortality disparities would be reduced as well, diseases emerged (such as AIDS), no new risks
they actually increased across the twentieth cen- developed (such as pollutants), no new knowl-
tury, particularly over the latter half of the twen- edge about risks emerged (as about cigarette
tieth century, as the leading health problems and smoking in the 1950s and 1960s), and no new
causes of death became chronic and degenerative treatments were developed (such as heart trans-
diseases (Mirowsky and Ross 2003, 2008; Lynch plants), the concept of fundamental social causes
2003; Link and Phelan 1995).1 Whereas at the would not apply.
turn of the century the terrible living conditions No clearer example exists illustrating the use
of life course themes in research on this subject
1
Some scholars suggest that health disparities widened than that in the debate about whether educational
even in the face of major public health improvements; oth- disparities in health, which generally widen with
ers describe initial declines in disparities before widening age, continue to widen into old age or whether
disparities over the latter half of the Twentieth Century.
they eventually narrow (Herd 2006; House et al.
For example, Mirowsky and Ross (2003) argue that dis-
parities widened even with public health improvements 1994; Lynch 2003; Miroswky and Ross 2003).
early in the century. Lynch (2003) claims public health On the one hand, the cumulative advantage
improvements disproportionately beneted those of perspective posits that disparities continue to
higher socioeconomic status, consistent with this view.
grow, as education-based inequalities continue to
Schnittker (2004), focusing on income disparities, argues
disparities were reduced as people shared broadly in the widen into old age (Mirowsky and Ross 2003).
benets of the public health infrastructure improvements, On the other hand, the age-as-leveler hypothe-
but widened as later advances in knowledge and technol- sis posits that disparities eventually narrow in old
ogy required more effort and choice on the part of indi-
age. This is thought to occur as the limits of
viduals to improve health. Masters et al. (2012) also
describes a narrowing at midcentury followed by widen- bodily functioning and life expectancy eventually
ing gap thereafter. trump the advantages of being highly educated,
374 M. Kirkpatrick Johnson et al.

and as the social institutions of old age (e.g., appear minimal. These studies highlight the need
social security, Medicare) provide a baseline of to examine the cumulative advantage and age-as-
support to the least advantaged, narrowing the leveler hypotheses with multi-cohort longitudinal
gap in health-protective resources (Beckett 2000; data.
Herd 2006; House et al. 1994). There has also The notion that human development is a life-
been concern that apparent narrowing results long process is a second key principle in life
from selective mortality; that is, because those in course research, and scholars in this tradition
the poorest of health die at earlier ages, the pool have called for studies that look at individuals
of individuals alive at older ages may represent a over substantial periods of time. As Elder and
particularly healthy group. However, several colleagues (2003) note, development is a life-
studies conclude that selective mortality is not a long process, as individuals show biological as
major factor (Beckett 2000; Herd 2006). Dupre well as social and psychological changes through-
(2007) argues that we need to distinguish patterns out the life course. Yet, the literature on how edu-
at the individual and aggregate level, offering cation affects health has been almost exclusively
support for the idea that selective mortality leads focused on adulthood. Though there is much to
to leveling at the aggregate level, but that indi- be gained by tracking how education relates to
vidual changes in health with age by education adult health from midlife to older ages, there are
level follow a cumulative advantage process. additional ways of thinking about time that can
Sorting out age, period, and cohort patterns shape the research questions we pursue in devel-
has been a fundamental aspect of comparing the oping our understanding of how and why educa-
cumulative advantage and age-as-leveler hypoth- tion promotes health. For instance, by taking the
eses (Goesling 2007; Lauderdale 2001; Lynch long view of human development we can ask
2003; Mirowsky and Ross 2003, 2008). Goesling how much of the education-health association
(2007), for instance, found that educational dis- observed in middle to late adulthood results from
parities in self-reported health have increased experiences much earlier in life, such as motiva-
among adults over age 70 in recent years but have tions, abilities, and orientations that are expressed
narrowed among younger adults (i.e., those ages in childhood and adolescence. Similarly, we can
3049). One issue this literature reveals quite assess the long-term inuence of child and ado-
clearly is that data capable of adjudicating these lescent health and health-risk behaviors on sub-
arguments is rare. The research to date typically sequent educational attainment. By following
makes use of cross-sectional data (e.g., Goesling children and adolescents through the transition to
2007; Schnittker 2004) or longitudinal data with adulthood and beyond, and by considering pro-
limited cohorts or limited waves of data (e.g., cesses across multiple generations, we can more
Lynch 2003; Lauderdale 2001) instead of multi- accurately place individuals in developmental
ple cohorts of longitudinal data over many years, and historical time as well as better understand
which are needed to fully assess the cumulative the social pathways and processes linking educa-
advantage hypothesis. Yet existing research tion and health.
points to the value of such data. Lynch (2003), for Studying lives over an extended period of time
instance, shows that the effect of education on can also help researchers identify short-term
self-rated health strengthens across age (in a qua- impacts of education on health and health-risk
dratic pattern in cross-sectional data, but a posi- behaviors. College attendance, in particular, may
tive linear pattern in panel data). This fanning ultimately be benecial to health in the long run
with age pattern is also becoming stronger but leads to short-term increases in substance use
across cohorts. Importantly, Lynch shows that (Schulenberg and Maggs 2002) and other health-
without taking cohort differences into account, risk behaviors (poor diet, inadequate sleep, lack
the growing disparities with age appear much of exercise). In some instances, these health-risk
more minimal; likewise, without taking the age behaviors during the college years may have
differences into account, cohort differences long-term negative health consequences,
Living Healthier and Longer: A Life Course Perspective on Education and Health 375

especially if they lead to dropout, accidents, or their educations, and the short-term processes
substance use problems. involved, with longer-term processes and mid- to
The ages and life stages at which college is late-life health outcomes; and (3) assessing his-
pursued may moderate these patterns. According torical trends in the mediating mechanisms and
to the life course principle of timing, the conse- their implications for health disparities.
quences of transitions and events depend on
when they occur in a persons life (Elder et al.
2003). For instance, older rst-year college stu- 4 Merging Selection
dents may have a lower risk of experiencing a and Causal Processes into
dramatic increase in substance use upon college a Long-Term,
entrance compared to more traditional rst- Multigenerational View
year students. On the other hand, students who
pursue college degrees after making other impor- Childrens and adolescents experiences, orienta-
tant role transitions do not have the benets in- tions, and social backgrounds have played a
hand of their degree as they navigate work and somewhat limited role in research on education
family experiences. Assessing whether the tim- and health. The primary attention given to this
ing of education matters for health is especially early period of life is directed toward assessing
important as the age of school completion the causal direction between education and adult
increasingly is stretching well into adulthood. health, or whether a cause-effect relationship is
Currently, over a third of postsecondary students justied at all.
are age 25 or older (Snyder and Dillow 2012). First, scholars have considered whether the
Finally, building on these themes, we must relationship between education and health may
address whether the potential mechanisms link- be spurious due to family socioeconomic back-
ing education and health have changed histori- ground and parental support, prenatal and child-
cally. This may be especially important to hood risk factors, and genetic endowments that
consider as the pool of college goers has diversi- give rise to both educational and health outcomes
ed and expanded over the past three decades in (Pampel et al. 2010). Part of the positive associa-
the United States,2 the links between education tion between educational attainment and adult
and other social institutions have changed, and health does appear to be generated through such
educational disparities have shifted as new threats processes. Parents education and other child-
to population health have emerged. As we dis- hood background advantages have long-term
cuss later in the chapter, these broad changes in effects on both education and health (Ross and
mechanisms may lead to increased educational Mirowsky 2011), and controlling for these fac-
disparities in health. At the same time, the land- tors substantially reduces estimates of education
scape and nature of education has changed, with effects for adults in some studies (Behrman et al.
a diversication of institution types at both the 2011; Fujiwara and Kawachi 2009). Adolescent
secondary and tertiary levels, and the possibility academic performance and interest also predict
that educational quality is eroding. health declines later in life and partially account
In the sections that follow, we discuss further for educational attainments impact (Herd 2010).
these three general possibilities for further incor- Second, scholars have considered whether the
poration of a life course perspective: (1) merging causal process works primarily in the other direc-
selection and causal effect processes into a long- tion, with health inuencing educational attain-
term, multigenerational view; (2) linking health ment. Research shows how poor childhood and
across the years in which people largely achieve adolescent health reduce educational attainment,
especially for non-Hispanic whites (Haas and
2 Fosse 2008; Jackson 2009; Palloni 2006).
College enrollment in the United States is currently at an
all-time high (20.6 million students) and is expected to Moreover, other research designs creating proxy
increase 14 % by 2019 (Snyder and Dillow 2012). measures for early health differences through
376 M. Kirkpatrick Johnson et al.

birth weight, childhood height, or differences in environment (e.g., Surgeon Generals warning
birth weight among twins, or research that capi- against smoking in 1964).
talizes on exogenous shocks in quasi-experimental Yet, as George (2003) articulated beautifully
designs (e.g., exposure to a major inuenza out- in the previous Handbook, life course perspec-
break, the Chernobyl disaster, the Swedish phase tives push scholars to move beyond viewing
out of lead in gasoline), document signicant social selection as a methodological nuisance,
effects of early health on later education levels and social selection and social causation as com-
and occupational attainment (Eide and Showalter peting hypotheses or even fully distinct. George
2011). (2003:675) writes:
One of the greatest challenges in addressing From a life course perspective, the distinctions
the impact of education on health and health-risk between social selection and social causation are
behaviors is the lack of sufciently compelling much less important. An investigator can have an
long-term data and analytic strategies to offer outcome of interest (e.g., health, employment sta-
tus) and wish to identify the social factors associ-
insight into causal relations and mechanisms ated with that outcome. But the distinction between
(Palloni 2006). Given that random assignment to selection and causation is relatively moot. The
different levels of education is not possible, focus is on the long-term processes and pathways
scholars have used a range of methodological that result in the outcome of interest, not categoriz-
ing those processes and pathways as selection or
strategies to assess whether these arguments causation. Many investigators who address social
about spurious factors and causal order erase the selection and social causation from the traditional
possibility of education having a causal impact perspective apparently fail to understand that they
on health. Some studies have taken the approach are attempting to eliminate the life course from
their inquiries. Using statistical procedures to esti-
of identifying and measuring childhood health mate a coefcient that represents selection essen-
and other early conditions and including them as tially means that the investigator is bundling the
controls in regression analyses. Studies control- study participants pasts into a neat little package
ling for many of these factors, including child- that is ignored substantively and interpretively. This
is one of the ways in which social and behavioral
hood socioeconomic status, intelligence, and research tends to be ahistoricalit ignores not only
health, do continue to document education effects the social history, but also personal history.
on adult health and mortality (Hayward and
Gorman 2004; Herd 2010; Link et al. 2008; Liu By putting concerns with selection and causal
et al. 2011). In a recent study, Schafer and col- order at the apex in our studies, we are not giving
leagues (2013) used these and other childhood appropriate attention to developmental processes
factors in propensity score models by which indi- and mechanisms that come over time and are
viduals were matched on all prior constructs and responsible for the myriad longitudinal connec-
varied only in whether they completed college; tions (Schulenberg 2006). That is, it is the
propensity score models are viewed as among the ongoing interaction between individuals and con-
strongest approaches to attending to selection texts that contributes to the direction of develop-
effects and in these models they found that com- ment, so it is inaccurate to suggest that picking up
pleting college did indeed contribute to health. midstream in such relations is simply charting
Researchers have also used quasi-experimental processes that are playing out their inevitable
and instrumental variable approaches to more course set by unmeasured third variables
rigorously tackle selection issues and these stud- (Schulenberg 2006, p 111). Obviously, selection
ies generally provide support for a causal inter- and causal order deserve careful attention, but
pretation (Eide and Showalter 2011). For our argument here is that, from a life course per-
example, studies have capitalized on siblings spective, the intricacies of long-term effects that
and twins shared genetic and environmental his- shape lives in personal and historical time also
tories, and other studies have compared cohorts deserve careful attention.
across changes in compulsory education laws or Fully solving the problems of selection and
signicant changes in the informational causal order and demonstrating the unequivocal
Living Healthier and Longer: A Life Course Perspective on Education and Health 377

effect of education level on health might help jus- graduation saw no health benet to earning the
tify investments in education, but would still not degree. These ndings echo earlier ones by Brand
give us a complete picture. It is not simply the and Xie (2010) that the earnings pay-off of a col-
case that educational attainment affects health via lege education is also greatest for those least likely
a range of mechanisms that scholars can map, but to earn a college degree.
that (a) parents education levels and health affect The resource substitution model also predicts
childhood and adolescent health which affects more education improvements in health for
educational attainment, (b) parents education women than men. Social scientists have long
levels and health affect childrens eventual edu- shown that women have fewer resources (e.g.,
cational attainment via childhood health but also power, authority, earnings) than men, making
parenting behavior and other mechanisms, and education potentially more important to their
that (c) educational attainment has further impor- health. For example, this hypothesis argues that
tant implications for health, and both attainment men more often have redundant resources, but in
and health in turn presumably shape these out- the absence of higher education, earnings and
comes for future children. Rarely is this process, power can also facilitate better health. Educational
which unfolds over generations and very long attainment is more important for women because
periods of individuals lives, considered as a without it they are otherwise likely to have fewer
whole (Bauldry et al. 2012; Hayward and Gorman resources than men. The ndings of several
2004). Yet extrapolating across studies that look recent studies provide some support for these
at different periods of the life course or incorpo- hypothesized conditional effects by gender (Ross
rate information on socioeconomic origins, it and Mirowsky 2006, 2010b). More research is
would appear this is the larger story. Viewing it needed on how these background characteristics
that way (and better yet, investigating it) enables moderate education effects on health, especially
connection to important questions about the as the pool of students pursuing education beyond
intergenerational transmission of advantage. high school has undergone substantial transfor-
What are the patterns in disparity not just during mation in recent years.
the adult years, but beginning much earlier in In summary, continued conceptual and empir-
life? What is the causal reach of these early life ical attention to how education and health affect
health disparities? Do they extend into mid life one another within and across generations is
and old age? Perhaps most importantly, for whom needed. Several studies show that poor health and
does education have the greatest long-term health health-risk behaviors as assessed in infancy,
benets? childhood, and adolescence are associated with
Recent evidence that education has heteroge- reduced educational attainment in adulthood. In
neous effects on health highlights the importance addition, parental socioeconomic status, among
of these interconnected questions. Arguing a other conditions of the early life course, is linked
resource substitution model, that education ben- to health, as well as childrens eventual educa-
ets most those who otherwise have fewer tional attainment. As such, health prior to the
resources, Ross and Mirowsky (2011) nd condi- completion of education and parental socioeco-
tional effects of personal educational attainment nomic status are considered key controls in
by parental educational attainment. The benets of assessments of the effects of education on health,
ones own education matter more for those whose although there has been some recent interest in
parents had completed less schooling. Likewise, the long-term processes linking conditions in
Schafer and colleagues (2013) nd that those peo- childhood to health in young and later adulthood
ple least likely to complete college (based on (e.g., Bauldry et al. 2012; Hayward and Gorman
parental socioeconomic status but also a range of 2004). These studies point to a need for contin-
factors) gain the most with respect to later health ued research that uses longitudinal data within an
outcomes when they do complete college. Indeed historical and intergenerational framework and
the group with the highest propensity for college analytic methodologies that more fully map
378 M. Kirkpatrick Johnson et al.

relationships between adolescent achievements, range of health-risk and health-protective behav-


orientations, socioeconomic advantages, and iors. The transition from adolescence to young
other background factors (both observed and adulthood is a time of change in many of these
unobserved), educational pathways, and health. behaviors, with aggregate levels of physical
Concern with, and controls for selection and activity and healthy eating declining, and alcohol
causal order are obviously important, but repre- and other substance use, and participation in
sent only part of the overall picture of how educa- risky sexual behavior, rising (Bauldry et al. 2012;
tion and health are interrelated across the life Eaton et al. 2007; Institute of Medicine 2014;
course. Johnston et al. 2014b; Schulenberg and Maggs
2002). Research on education and health-related
behaviors during this period focuses on short-
5 The Transition to Adulthood, term effects and uses a mix of between-person
Short-Term Effects and Links and within-person (that is, individual change over
with Adulthood time) modeling strategies. This research is still
quite limited, but suggests greater attention be
If childhood and adolescence have played a nar- paid to the potential short-term and sometimes
row or limited role in research on education and detrimental effects of schooling, especially the
health, young adulthood is practically absent uptick in health-risk behaviors that occurs during
altogether (Bauldry et al. 2012). The heavy focus college (i.e., heavy drinking, drug use, poor eat-
on adulthood, particularly mid- to late-life adult- ing habits, weight gain, etc.).
hood, is understandable from the perspective that Educational status, for example currently
most serious physical health problems are not attending or not attending college, is related to
very prevalent until these life stages. A focus on substance use, though there is some conicting
adulthood proper also enables a simple consider- evidence across studies. Some studies have docu-
ation of nal education levels in analytic mod- mented that college students engage in heavier
els. Indeed, a common (though by no means and more frequent alcohol and marijuana use
universal) strategy across studies is to restrict than young adults who are not attending college
analysis to those who have reached a specic age (e.g., Schulenberg and Patrick 2012; Staff et al.
in adulthood authors deem to be one in which 2010; White et al. 2006), while others have found
further education can be assumed to be trivial, fewer differences between college and noncol-
typically age 25 or 30 (Cutler and Lleras-Muney lege drinking levels (e.g., Chen et al. 2004; Lanza
2010; Goesling 2007; Lynch 2003; Ross et al. and Collins 2006). Research shows that atten-
2012; Zajacova et al. 2012). That choice effec- dance at 4-year universities, but not 2-year uni-
tively requires analysis strategies to model versities, increases engagement in heavy drinking
between-person variationthat is, looking at as well as illicit drug use (Patrick et al. 2012;
health outcomes in adulthood as they differ Staff et al. 2010). Evidence also suggests that
across persons with different education levels. It greater increases in alcohol use across the transi-
also means the effects of education on health are tion to college are associated with a lower likeli-
necessarily long-term effects. For the most part, hood of dropping out, and students who drop out
the study of education and health is separated of college decrease their drinking after leaving
from the time in which education occurs. The college (Schulenberg and Patrick 2012), likely
period in which educational differences emerge because alcohol use can be a proxy for engage-
is left out. ment in college social life (Carter et al. 2010;
There is little attention to the effects of educa- Schulenberg and Maggs 2002). One recent study
tion on health as it is differentially gained, largely, comparing patterns of heavy drinking from ado-
though not exclusively, in the late teens and twen- lescence through the transition to adulthood in
ties. There has been research emerging, however, the Add Health data indicates that non-Hispanic
on post-secondary educational patterns and a whites who attend and graduate from a 4-year
Living Healthier and Longer: A Life Course Perspective on Education and Health 379

college or university experience a steeper rise health-related behaviors that may be relevant or
with age in heavy drinking and ultimately reach that have been implicated in adult health dispari-
higher levels of heavy drinking compared to ties and, as such, additional work is needed.
those who do not attend college at all (Chen and Furthermore, a continued focus on health-risk
Jacobson 2013). Those obtaining 2-year degrees and health-protective behaviors may be quite
or attending some college without earning a appropriate for this part of the life course, again
degree fall in between. No group differences given the age patterns of disease and other health
remain signicant by age 3031, however. For problems. Yet there is a need to link these short-
Blacks, in contrast, pathways through school term effects with longer-term patterns of behav-
were much less associated with trajectories of ior and to later health outcomes to complement
heavy drinking. By the mid to late 20s, however, the study of adult health and the often contempo-
Blacks who had attended some post-secondary raneous measures of mediating mechanisms
education but did not earn either a 2- or 4-year (e.g., job quality, social support, and lifestyle fac-
degree (college-withdrawers) had signicantly tors). Do effects of education visible in the short-
higher levels of heavy drinking than other groups term have direct implications for later health?
that persisted until at least age 3031. The study Might the health-risk behaviors associated with
was not designed to assess whether problematic college attendance partially offset long-term
drinking led to or resulted from college health gains from education, perhaps especially if
drop-out. the young person attends but does not nish col-
As discussed earlier with respect to the rela- lege? Do short-term changes in health behavior
tion between education and health more broadly, affect later health through persistence in behav-
studies of this age period indicate that inuence ioral patterns? Upticks in health-risk behaviors,
between education and health-related behaviors like heavy drinking during the years of college
can go both ways. Prospective associations attendance, may have little long-term conse-
between health behaviors and college attendance quence so long as they do not lead to alcohol use
and graduation demonstrate the effect of health disorders, contribute to accidental injury, or cause
behaviors on educational attainment. For exam- school difculties; indeed, although certainly not
ple, substance use in high school is associated a recommended practice, heavy drinking and
with lower likelihood of college attendance and other risky behaviors during this time may have
graduation, even after controlling for academic more social benets than negative consequences
performance and expectations (Bachman et al. (Institute of Medicine 2014; Schulenberg and
2008, 2011; Patrick et al. 2013; Siennick and Maggs 2002).
Staff 2008). Patterns for high school alcohol use How health-related behaviors tied to school-
differ, with drinking positively related to educa- ing are linked to longer-term trajectories of
tional attainment (Patrick et al. 2013). This behavior is an important piece of this puzzle.
echoes again that alcohol use can be a marker for Piecing empirical patterns across the largely
social engagement (Crosnoe 2011). Obesity in cross-sectional studies, the picture associated
adolescence is also associated with college atten- with college-going is that college students engage
dance, though only among girls (Crosnoe 2007). in heavy drinking at higher rates while in school,
The lower likelihood of college entrance among but that sometime within the post-college decade
obese girls, which was especially prominent or so rates drop off enough to reverse the com-
when girls attended high schools where obesity parison entirely, such that less educated adults
was less common, was partially explained by are more likely to drink heavily than those having
increased internalizing symptoms, self- pursued higher education. Trajectories of obesity,
medication, and disengagement from academics healthy diet, and physical activity need to be
in high school (Crosnoe 2007). examined from the ages when educational
Assessments of the short-term effects of edu- attendance patterns begin to diverge. Importantly,
cation have not yet encompassed the full range of education may be tied not only to each, but to the
380 M. Kirkpatrick Johnson et al.

way in which they come together into more or permanent one. In more recent cohorts it has
less healthy lifestyles over time. become more common to pursue postsecondary
Despite lower rates of disease among young education at older ages. Life course scholars high-
adults compared to older adults, it is also impor- light how the developmental consequences of key
tant to examine health itself among young adults transitions and events depend on when they occur
as it relates to education. Are disparities in health in life as well as the historical context. The
immediately observable as further pursuit of edu- increased average age of college students raises
cation differentially establishes the adult educa- important questions about whether education
tion levels often measured for assessments of gained at later stages in life has the same benecial
health later in the life course? Or are differences at impact on health in general and for different
this stage largely due to earlier experiences also cohorts. Students over the traditional age may be
tied to educational pursuits (e.g., childhood socio- less susceptible than younger students to the risky
economic status), with health effects due to educa- behaviors of college life such as binge drinking,
tion emerging over time in the years after education thus minimizing the potential for short-term nega-
levels are achieved and as adult lives are organized. tive effects; and more recently, they may have
Low rates of specic diseases may make them more like-aged peers thus increasing social inte-
impractical measures of (ill) health at young ages, gration effects. However, education might also
but self-rated health, mental health, and general make less of a difference the later in life it is
measures of activity limitation show sufcient earned, as long-term behavioral patterns may be
variation (Amato and Kane 2011; Bauldry et al. more difcult to disrupt later in life and the fewer
2012; Eaton et al. 2007; Schulenberg et al. 2005). number of years lived after education is achieved
In pursuing these questions focused on educa- for which its associated resources can operate to
tion and health during the young adult years, we promote health. Miech and colleagues (2015) also
need to consider a number of interrelated issues. show that even among young adults, the sequenc-
First, are the mechanisms theorized to link educa- ing of postsecondary education with respect to
tion and health and health-related behaviors the entering other adult social roles moderates the
same for long- and short-term processes? Second, impact of education on health. They argue that
how long does it take for them to materialize and people who earn their degrees before entering
matter? Though postsecondary students do not marriage and parenthood have greater resources in
usually have access yet to the higher paying jobs place to avoid the weight gain that these familial
they will eventually hold, nor are they usually mar- transitions can often bring. Consistent with their
ried (school attendance is inversely related to the argument, they nd the probability of obesity
likelihood of marriage), they do have access to much higher among those who married before
health services and mental health counseling, completing college compared to those completing
school-organized social activities, and perhaps the college rst. For black men only, the probability of
nancial resources of their parents. These school- obesity is also higher among those who become
related resources may protect students from poten- parents before completing college. Heterogeneity
tial long-term risk. The cognitive and social in the health benets of education by when educa-
psychological benets of education would also tion is earned in the life course is beginning to be
seem to be candidates for mediators that are evi- considered, but to date most other studies have
dent in the short-term and work similarly to relied on inadequate data to fully address the issue.
improve health-related behaviors and health in the In summary, examining health disparities tied
short- and long-term. More careful attention to the to education during the late adolescence and young
school-going years would potentially yield addi- adult years has two primary advantages. First,
tional insights into how the mechanisms believed more and more youth are extending their educa-
to link education and health actually operate. tions past high school, and research shows that the
In addition, it is important to note that the rst likelihood of engaging in unhealthy behaviors
transition out of education is increasingly not a increases during periods of college attendance.
Living Healthier and Longer: A Life Course Perspective on Education and Health 381

Whether this unprecedented growth in college more supportive relationships, and imparts cogni-
going, and the short-term health risks associated tive skills and other resources that are conducive to
with it, will lead to long-term population health good health. Social scientists have documented
problems is unknown. Perhaps these short-term important ways in which education (particularly a
health risks are offset by on campus health and 4-year college degree) is becoming increasingly
social services (when available), or maybe college tied to several of these potential mediators, each
going has long-term negative health consequences suggesting changing linkages between education
only among those who do not graduate from col- and health for more recent cohorts.
lege, or among those who develop more serious Prior research has clearly documented the
health problems (e.g., addiction). Second, postsec- health benets of completing high school versus
ondary students today are on average older com- dropping out or temporarily stopping out of high
pared to previous generations. This changing school, but college is now the central focus of
demographic suggests young people are staying in students futures, both from their own point of
school longer and/or increasingly returning to col- viewas reected in their educational plans
lege after periods of work. Health studies that view (Goyette 2008)as well as at a broad cultural
educational attainment as static, both developmen- levelas reected in the college for all era in
tally and historically, potentially miss key mecha- the United States (Pascarella and Terenzini 2005;
nisms of education duration and timing and of Rosenbaum 2011). President Obama has set the
interlinked social and employment experiences on goal that the United States has the highest pro-
long-term health outcomes. portion of college graduates (including 2-year
degrees) in the world by 2020. In addition, the
long-term health benets of a postsecondary
6 Assessing Historical Trends degree may be as strong or stronger than the
in the Mediating effect of high school graduation (compared to not
Mechanisms and Their graduating from high school) with respect to
Implications for Health mortality, chronic disease, self-rated physical
Disparities health, psychological wellbeing, and health-risk
behaviors (Baum et al. 2010; Palloni 2006).
The fundamental cause perspective posits that the Given this, we focus our illustrative examples of
relationship between education and health is endur- historical change in mediating mechanisms as
ingeven though the mechanisms involved may they relate to postsecondary levels of education.
emerge and fade and diseases may change. That is, First, employment status and conditions are
the links between education and health are often among the major theorized mediators linking edu-
overlapping (e.g., those with higher education lev- cation and health, yet the relationship between
els benet in myriad and often redundant ways), education and these work-related experiences is
and across historical time, some mechanisms cap- changing in fundamental ways. Postsecondary
ture more of the mediation and others capture less. educational credentials are increasingly needed to
This assertion awaits testing with multi-cohort secure a good job in the contemporary economy.
long-term data. But meanwhile, the life course per- In particular, the employment and earnings gap
spective complements this tenet of fundamental between those who earn 4-year degrees and those
cause by directing needed attention to historical who obtain less education is substantial and grow-
changes in the intervening mechanisms that can ing (Day and Newburger 2002; Goldin and Katz
serve to potentiate or counter the long-term effects 2008; Grusky et al. 2011; Institute of Medicine
of earlier education-based experiences. As men- 2014). If these education differences in economic
tioned previously, educational attainment is thought outcomes continue to widen we may soon witness
to be positively linked to health and inversely a health gradient that is more clearly based upon
linked to mortality because education facilitates the achievement of certain educational milestones,
securing better jobs, increases the likelihood of such as the receipt of a Bachelors degree.
382 M. Kirkpatrick Johnson et al.

Second, the links between education and points out, is over-time improvement in writing
family formation are changing. Demographers a single, knowledge-free essay a good measure of
have identied the diverging destinies in fam- what is learned in college? (1001). Furthermore,
ily life for those with and without college degrees time spent studying alone was not signicantly
(McLanahan 2004), nding that the latter are less (p < 0.05) associated with gains in cognitive
likely to marry and more likely to experience scores in Arum and Roskas study, raising ques-
union dissolution and family instability. Stable tions as to whether time spent studying is a good
sources of social support may be increasingly measure of educational quality or more generally
more difcult to achieve for less educated people. whether the decline in studying is a cause for
If multiple pathways through which education concern. Thus, whether quality of college educa-
matters (employment, family formation) are tion is declining is an open question, along with
increasingly concentrated among the college- the effects of any decline on health benets.
educated in recent cohorts, that has important A fourth historical change of potential impor-
implications for health trends. tance is the expansion and diversication of the
Third, it is important to address possible pool of postsecondary students (Hout 2012). As
changes in educational quality when predicting mentioned before, this expansion has included stu-
whether education will have commensurate ben- dents age 25 or older. The number of degrees con-
ets for health in contemporary cohorts of youth ferred has also risen dramatically, with the greatest
in comparison to their predecessors in decades gains among female and minority students. Since
past. Though concerns about the current quality of research has shown that the long-term health ben-
higher education are not new, there appears to be ets to education are greatest among those who
growing worries that the experience of going to have been least likely to attend college (e.g.,
college, for both traditional advantaged students Schafer et al. 2013), the increased heterogeneity in
and non-traditional older students from more college going is likely to lead to greater educa-
disadvantaged backgrounds, has changed for the tional disparities in health. On the other hand, if
worse (Arum and Roksa 2011; Ct and Allahar the quality of higher education has been eroding
2011; Professor X 2011). Issues such as grade substantially over time, educational disparities
ination, the retreat from public nancing of could diminish. Nontraditional and minority stu-
higher education (leading to more use of adjunct dents are also more likely to attend 2-year vs.
professors), increasing class sizes, and limited 4-year programs, and less selective schools among
time students spend studying (while getting good the 4-year programs, and the relative benets of
grades nonetheless) all raise the question as to education across these institution types is not well
whether college degrees will have the same pay- understood. Research is needed that accounts for
off in terms of work benets, cognitive growth, these broad changes in postsecondary education,
and health for todays young people and for future health outcomes, and mechanisms among recent
generations (if these trends continue). U.S. cohorts. Specically, given that historical
Documenting changes in the quality of higher change in college going and its links to family for-
education is certainly not easy to do. Arum and mation and employment outcomes are evident
Roksa (2011) examine both historical changes in across the period from the 1970s to 2010s, research
time spent studying among college students as is needed comparing cohorts moving from adoles-
well as whether postsecondary students show cence to midlife from these decades (Hout 2012).
cognitive gains while attending school. The Though the effect of a college degree on health
authors nd that time spent studying has indeed may result from access to better jobs, higher earn-
historically declined, and cognitive gains among ings, and greater social support, or from enhanced
recent college students are modest, leading them cognitive resources and self-direction that lead to a
to conclude that learning on college campuses is healthier lifestyle (Baker et al. 2011; Link and
now limited. Yet, the cognitive test was based Phelan 1995; Mirowsky and Ross 2003), these
upon one essay question from the Collegiate varied potential mechanisms (i.e., economic,
Learning Assessment, and as Farkas (2011) social psychological and interpersonal, and health
Living Healthier and Longer: A Life Course Perspective on Education and Health 383

lifestyle) for education-health relationships are Patrick et al. (2013) showed that substance use
often not considered in tandem. Thus, in addition during high school predicted a greater likelihood
to research attending to historical change, research of never attending college (for cigarettes and ille-
is needed that approaches these explanatory mech- gal drugs), of graduating from a 2-year rather
anisms synergistically (Leigh and Dhir 1997; than a 4-year school (for cigarettes), and of drop-
Schnittker 2004), assessing the relative predictive ping out versus graduating from a 4-year school
power of each in explaining education effects on (for cigarettes, marijuana, and other illegal
health, and illuminating their potentially changing drugs). Interestingly, high school alcohol use was
roles and explanatory power across cohorts. positively related to longer-term educational
In summary, as recommended by a recent attainment. Consistent with other recent research
Institute of Medicine (2014) report on health and (e.g., Jackson 2009), these ndings provide rea-
well-being during the transition to adulthood, son for giving attention to reciprocal effects over
continued research is needed addressing whether time when educational experiences are happen-
educational disparities in health are changing ing, as well as controlling for adolescent health-
over time as new health threats emerge, economic risk behaviors when looking at education-health
conditions and general patterns of family forma- relationships in adulthood.
tion change, and college education expands and In a series of papers, we have examined how
diversies. According to the rising importance family transitions account for changes in wellbe-
hypothesis, educational disparities in health will ing and health-risk behaviors during the transition
emerge at younger ages and last longer for more to adulthood (i.e., from ages 19 to 28). Using
recent cohorts (Goesling 2007; Mirowsky and analyses of within-person change to control for all
Ross 2008). As Montez and colleagues (2011) time-stable sources of spuriousness, Staff and col-
noted, this may be due to several mechanisms we leagues (2010) showed how marriage and cohabi-
note above, such as the increasing importance of tation lead to reductions in health-risk behaviors.
education for wage attainment or education being These analyses of within person change also
more positively associated with marriage and revealed potential negative health effects in the
staying married, which in turn creates a network short term during periods of school attendance.
of more educated people (Breen and Salazar For instance, heavy drinking and marijuana use
2011). However, the recent economic downturn increase when young people are attending college
in the United States may have exacerbated health but then decline sharply after degree completion.
declines or may have led to health declines at In new analyses reported here for the rst time,
younger ages (due in part to the inability of many we used MTF data for the 19762010 senior year
recent college graduates to nd jobs or jobs that cohorts to examine how college attendance and
match their educational credentials). completion predict changes in overall self-
reported health. To illustrate these ndings, Fig. 1
shows predicted probabilities of self-reported
7 Some Illustrative Examples physical health by age, student status, and degree
from the Monitoring attainment using analyses of within-person
the Future Study change (again to control for unobserved time-sta-
ble sources of spuriousness).3 As shown in Fig. 1,
Previous analyses using MTF have stressed the
important role adolescent health-risk behaviors 3
At each wave, respondents were asked, Overall, relative
play in inuencing educational attainment. For
to people your age, do you think your physical health over
instance, we have found that adolescent sub- the past year has been (responses range on a ve-point
stance use reduces the likelihood of both college scale from much poorer than average to much better
matriculation and completion above and beyond than average). For the purposes of these analyses, this
outcome was dichotomized (1 = somewhat or much better
the standard controls, such as academic perfor-
than average). These ndings are based upon a subsample
mance, expectations, and socioeconomic back- of 11,441 MTF respondents who completed surveys from
ground (Bachman et al. 2008, 2011, 2013). ages 19 to 28 (encompassing 38,231 person waves).
384 M. Kirkpatrick Johnson et al.

Fig. 1 Predicted probability of .50


better-than-average health in the
past year by college matriculation .48

Better-than-average health in past year


and degree attainment using Non-student
analyses of within-person change .46
(n = 11,441)
.44

.42
Student then
.40 two-year
graduate by
.38 age 21
.36
Student then
.34 four-year
graduate by
.32 age 23

.30
19 20 21 22 23 24 25 26 27 28
Age

the predicted probability of better than average an excellent illustration of how life course
health signicantly declines by 3 % each year thinking has fundamentally changed traditional
from ages 19 to 28. When students attend college, areas of study and stimulated new debates (e.g.,
health increases by 4 %, but this change is not sta- cumulative advantage vs. age as leveler
tistically signicant.4 However, completing a hypotheses of health disparities). But we believe
degree is linked to improved self-reported health, more is possible. By fully embracing the princi-
but it depends on the degree. Health increases by ple that development is life-long, and attending
23 % when students graduate from 4-year pro- to issues of timing and historical context, we can
grams.5 Interestingly, 2-year degree completion build an understanding of how education and
does not signicantly increase health in the same health are related across generations as well as
way as 4-year degree completion,6 highlighting a reciprocally in the early life course, with impli-
need for more research on whether different types cations for later life. The life stages of child-
of degrees matter more for health beyond years of hood, adolescence, and younger and older
school attained. adulthood are linked such that theoretical
advancement requires attention to earlier life
stages. Our understanding of the mechanisms
8 Conclusion linking educational pursuits and attainments to
health in the short- and long-term will grow, and
A life course perspective is already evident in the we will identify the processes linking short- and
sociological and gerontological study of educa- long-term change. Paying more attention to his-
tion and health. In many ways the topic provides torical change as it intersects lives at every age
particularly via changes in educational
4
Estimate = 0.039; Robust standard error = 0.029; Odds institutions and the interrelationships of educa-
ratio = 1.039; p = 0.191. tional, familial, and labor market institutions
5
Estimate = 0.203; Robust standard error = 0.041; Odds during young adulthood to midlifewill assist
ratio = 1.225; p = 0.001. in assessing the implications of social change for
6
Estimate = 0.069; Robust standard error = 0.048; Odds
future disparities in health.
ration = 1.071; p = 0.152.
Living Healthier and Longer: A Life Course Perspective on Education and Health 385

References adulthood. Journal of Studies on Alcohol and Drugs,


71, 742750.
Chen, P., & Jacobson, K. C. (2013). Longitudinal relation-
Amato, P. R., & Kane, J. B. (2011). Life-course pathways
ships between college education and patterns of heavy
and the psychosocial adjustment of young adult
drinking: A comparison between Caucasians and
women. Journal of Marriage and Family, 73, 279
African-Americans. Journal of Adolescent Health, 53,
295. doi:10.1111/j.1741-3737.2010.00804.x.
356362.
Arum, R., & Roksa, J. (2011). Academically adrift:
Chen, C. M., Dufour, M. C., & Yi, H. (2004). Alcohol con-
Limited learning on college campuses. Chicago:
sumption among young adults ages 1824 in the
University of Chicago Press.
United States: Results from the 20012002 NESARC
Bachman, J. G., OMalley, P. M., Schulenberg, J. E.,
survey. Alcohol Research & Health, 28, 269280.
Johnston, L. D., Freedman-Doan, P., & Messersmith,
Ct, J. E., & Allahar, A. (2011). Lowering higher educa-
E. E. (2008). The education-drug use connection:
tion: The rise of corporate universities and the fall of
How successes and failures in school relate to adoles-
liberal education. Buffalo: University of Toronto
cent smoking, drinking, drug use, and delinquency.
Press.
New York: Lawrence Erlbaum Associates/Taylor &
Crosnoe, R. (2007). Gender, obesity, and education.
Francis.
Sociology of Education, 80, 241260.
Bachman, J. G., Staff, J., OMalley, P., Schulenberg, J. E.,
Crosnoe, R. (2011). Fitting in, standing out: Navigating
& Freedman-Doan, P. (2011). Student work intensity:
the social challenges of high school to get an educa-
New evidence on links to educational attainment and
tion. New York: Cambridge University Press.
problem behaviors. Developmental Psychology, 47,
Cutler, D. M., & Lleras-Muney, A. (2010). Understanding
344363.
differences in health behaviors by education. Journal
Bachman, J. G., Staff, J., OMalley, P., & Freedman-
of Health Economics, 29(1), 128. doi:10.1016/j.
Doan, P. (2013). Race-ethnicity and socioeconomic
jhealeco.2009.10.003.
status moderate how student paid work intensity is
Day, J. C., & Newburger, E. C. (2002). The big payoff:
related to scholastic performance and substance use.
educational attainment and synthetic estimates of
Developmental Psychology, 49, 21252134.
work-life earnings (Current population reports
Baker, D. P., Leon, J., Smith Greenaway, E. G., Collins, J.,
pp. 23210). Washington, DC: U.S. Census
& Movit, M. (2011). The education effect on popula-
Bureau U.S. Department of Commerce, Economics
tion health: A reassessment. Population and
and Statistics Administration.
Development Review, 37, 307332.
Dupre, M. E. (2007). Educational differences in age-
Baker, D. P., Salinas, D., & Eslinger, P. J. (2012). An envi-
related patterns of disease: Reconsidering the cumula-
sioned bridge: Schooling as a neurocognitive develop-
tive disadvantage and age-as-leveler hypotheses.
mental institution. Developmental Cognitive
Journal of Health and Social Behavior, 48, 115.
Neuroscience, 2S, S6S17.
Eaton, D. K., Kann, L., Okoro, C. A., & Collins, J. (2007).
Bauldry, S., Shanahan, M. J., Boardman, J. D., Miech,
Selected health status indicators and behaviors of
R. A., & Macmillan, R. (2012). A life course model of
young adults, United States2003. American Journal
self-rated health through adolescence and young adult-
of Health Education, 38, 6675.
hood. Social Science & Medicine, 75, 13111320.
Eide, E. R., & Showalter, M. H. (2011). Estimating the
Baum, S., Ma, J., & Payea, K. (2010). Education pays,
relation between health and education: What do we
2010: The benets of higher education for individuals
know and what do we need to know? Economics of
and society. Washington, DC: The College Board.
Education Review, 30, 778791.
Beckett, M. (2000). Converging health inequalities in later
Elder, G. H. (1998). The life course and human develop-
lifeAn artifact of mortality selection? Journal of
ment. In R. M. Lerner (Ed.), Handbook of child psy-
Health and Social Behavior, 41, 106109.
chology, volume 1: Theoretical models of human
Behrman, J. R., Kohler, H., Jensen, V. M., Pedersen, D.,
development (pp. 939991). New York: Wiley.
Petersen, I., Bingley, P., & Christensen, K. (2011).
Elder, G. H., Johnson, M. K., & Crosnoe, R. (2003). The
Does more schooling reduce hospitalization and delay
emergence and development of life course theory. In
mortality? New evidence based on Danish twins.
J. T. Mortimer & M. J. Shanahan (Eds.), Handbook of
Demography, 48, 13471375.
the life course (pp. 319). New York: Kluwer
Brand, J. E., & Xie, Y. (2010). Who benets most from
Academic/Plenum Publishers.
college? Evidence for negative selection in heteroge-
Farkas, G. (2011). Book review of academically adrift:
neous economic returns to higher education. American
Limited learning on college campuses, by R. Arum
Sociological Review, 75(2), 273302.
and J. Roksa. American Journal of Sociology, 117,
Breen, R., & Salazar, L. (2011). Educational assortative
10001002.
mating and earnings inequality in the United States.
Fujiwara, T., & Kawachi, I. (2009). Is education causally
American Journal of Sociology, 117, 808843.
related to better health? A twin xed-effect study in
Carter, A. C., Brandon, K. O., & Goldman, M. S. (2010).
the USA. International Journal of Epidemiology, 38,
The college and noncollege experience: A review of
13101322.
the factors that inuence drinking behavior in young
386 M. Kirkpatrick Johnson et al.

George, L. K. (2003). Life course research: Achievements Johnston, L. D., OMalley, P. M., Bachman, J. G.,
and potential. In J. T. Mortimer & M. J. Shanahan Schulenberg, J. E., & Miech, R. A. (2014a). Monitoring
(Eds.), Handbook of the life course (pp. 671680). the future national survey results on drug use, 1975
New York: Kluwer Academic/Plenum Publishers. 2013: Volume I, secondary school students. Ann
Glied, S., & Lleras-Muney, A. (2008). Technological Arbor: Institute for Social Research, The University of
innovation and inequality in health. Demography, Michigan.
45(3), 741761. Johnston, L. D., OMalley, P. M., Bachman, J. G.,
Goesling, B. (2007). The rising signicance of education Schulenberg, J. E., & Miech, R. A. (2014b).
for health? Social Forces, 85, 16211644. Monitoring the future national survey results on drug
Goldin, C. D., & Katz, L. F. (2008). The race between use, 19752013: Volume 2, college students and adults
technology and education. Cambridge: Harvard. ages 1955. Ann Arbor: Institute for Social Research,
Goldstein, J. R., & Kenney, C. T. (2001). Marriage delayed The University of Michigan.
or marriage forgone? New cohort forecasts of rst Kalleberg, A. L. (2011). Good jobs, bad jobs: The rise of
marriage for U.S. women. American Sociological polarized and precarious employment systems in the
Review, 66(4), 506519. doi:10.2307/3088920. United States, 1970s to 2000s. New York: Russell
Goyette, K. A. (2008). College for some to college for all: Sage.
Social background, occupational expectations, and Lanza, S. T., & Collins, L. M. (2006). A mixture model of
educational expectations over time. Social Science discontinuous development in heavy drinking from
Research, 37(2), 461484. ages 18 to 30: The role of college enrollment. Journal
Grusky, D. B., Western, B., & Wimer, C. (2011). The great of Studies on Alcohol, 67(4), 552561.
recession. New York: Russell Sage. Lauderdale, D. S. (2001). Education and survival: Birth
Haas, S. A., & Fosse, N. (2008). Health and the educa- cohort, period, and age effects. Demography, 36(4),
tional attainment of adolescents: Evidence from the 551561.
NLSY97. Journal of Health and Social Behavior, Leigh, J. P., & Dhir, R. (1997). Schooling and frailty
49(2), 178192. among seniors. Economics of Education Review, 16,
Hayward, M. D., & Gorman, B. K. (2004). The long arm 4557.
of childhood: The inuence of early-life social condi- Link, B. G., & Phelan, J. C. (1995). Social conditions as
tions on mens mortality. Demography, 41, 87107. fundamental causes of disease. Journal of Health and
Herd, P. (2006). Do functional health inequalities decrease Social Behavior, 35 (Extra Issue), 8094.
in old age? Educational status and functional decline Link, B. G., Phelan, J. C., Miech, R., & Westin, E. L.
among the 19311941 birth cohort. Research on (2008). The resources that matter: Fundamental social
Aging, 28, 375392. causes of health disparities and the challenge of intel-
Herd, P. (2010). Education and health in late-life among ligence. Journal of Health and Social Behavior, 49,
high school graduates: Cognitive versus psychological 7291.
aspects of human capital. Journal of Health and Social Liu, S. Y., Buka, S. L., Linkletter, C. D., Kawachi, I.,
Behavior, 51(4), 478496. Kubzansky, L., & Loucks, E. B. (2011). The associa-
Herd, P., Goesling, B., & House, J. S. (2007). tion between blood pressure and years of schooling
Socioeconomic position and health: The differential versus educational credentials: Test of the sheepskin
effects of education versus income on the onset versus effect. Annals of Epidemiology, 21, 128138.
progression of health problems. Journal of Health and Lynch, S. M. (2003). Cohort and life course patterns in the
Social Behavior, 48, 223238. relationship between education and health: A hierar-
House, J. S., Lepkowski, J. M., Kinney, A. M., Mero, chical approach. Demography, 40(2), 309331.
R. P., Kessler, R. C., & Herzog, A. R. (1994). The Martin, S. P. (n.d.). Growing evidence for a divorce
social stratication of aging and health. Journal of divide? Education and marital rates in the U.S. since
Health and Social Behavior, 35, 213234. the 1970s (Working Paper). New York: Russell Sage
Hout, M. (2012). Social and economic returns to college Foundation.
education in the United States. Annual Review of Masters, R. K., Hummer, R. A., & Powers, D. A. (2012).
Sociology, 28, 327400. Educational differences in U.S. adult mortality: A
Institute of Medicine. (2012). How far have we come in cohort perspective. American Sociological Review,
reducing health disparities?: Progress since 2000: 77(4), 548572.
Workshop summary. Washington, DC: The National McLanahan, S. (2004). Diverging destinies: How children
Academies Press. fare under the second demographic transition.
Institute of Medicine. (2014). Investing in the health and Demography, 41, 607627.
well-being of young adults. Washington, DC: The Miech, R., Shanahan, M. J., Boardman, J., & Bauldry, S.
National Academies Press. (2015). The sequencing of a college degree during the
Jackson, M. I. (2009). Understanding links between ado- transition to adulthood: Implications for obesity.
lescent health and educational attainment. Journal of Health and Social Behavior, 56(2),
Demography, 46, 671694. 28195.
Living Healthier and Longer: A Life Course Perspective on Education and Health 387

Mirowsky, J., & Ross, C. E. (2003). Education, social sta- Ross, C. E., & Mirowsky, J. (2010a). Why education is the
tus, and health. Hawthorne: Aldine de Gruyter. key to socioeconomic differentials in health. In C. E.
Mirowsky, J., & Ross, C. E. (2005). Education, cumulative Bird, P. Conrad, A. M. Fremont, & S. Timmermans
advantage, and health. Ageing International, 30, 2762. (Eds.), The handbook of medical sociology. Nashville:
Mirowsky, J., & Ross, C. E. (2008). Education and self- Vanderbilt University Press.
rated health: Cumulative advantage and its rising Ross, C. E., & Mirowsky, J. (2010b). Gender and the
importance. Research on Aging, 30, 93122. health benets of education. The Sociological
Montez, J. K., Hummer, R. A., Hayward, M. D., Woo, H., Quarterly, 51, 119.
& Rogers, R. G. (2011). Trends in the educational gra- Ross, C. E., & Mirowsky, J. (2011). The interaction of
dient of U.S. adult mortality from 1986 through 2006 personal and parental education on health. Social
by race, gender and age group. Research on Aging, 33, Science & Medicine, 72, 521599.
145171. Ross, C. E., & Wu, C. (1995). The links between educa-
Montez, J. K., Hummer, R. A., & Hayward, M. D. (2012). tion and health. American Sociological Review, 60,
Educational attainment and adult mortality in the 719745.
United States: A systematic analysis of functional Ross, C. E., Masters, R., & Hummer, R. A. (2012).
form. Demography, 49, 315336. Education and the gender gaps in health and mortality.
Palloni, A. (2006). Reproducing inequalities: Luck, wal- Demography, 49, 11571183.
lets, and the enduring effects of childhood health. Schafer, M. H., Wilkinson, L., & Ferraro, K. (2013).
Demography, 43, 587615. Childhood (mis)fortune, educational attainment, and
Pampel, F. C., Krueger, P. M., & Denney, J. T. (2010). adult health: Contingent benets of a college degree.
Socioeconomic disparities in health behaviors. Annual Social Forces, 93, 10071034.
Review of Sociology, 36, 349370. Scharoun-Lee, M., Gordon-Larsen, P., Adair, L. S.,
Pascarella, E. T., & Terenzini, P. T. (2005). How college Popkin, B. M., Kaufmann, J. S., & Suchindran, C. M.
affects students: Volume 2. A third decade of research. (2011). Intergenerational proles of socioeconomic
San Francisco: Jossey Bass. disadvantage and obesity during the transition to
Patrick, M. E., Wightman, P., Schoeni, R. F., & adulthood. Demography, 48, 625651.
Schulenberg, J. E. (2012). Socioeconomic status and Schnittker, J. (2004). Education and the changing shape of
substance use among young adults: A comparison the income gradient in health. Journal of Health and
across constructs and drugs. Journal of Studies on Social Behavior, 45, 286305.
Alcohol and Drugs, 73, 772782. Schulenberg, J. E. (2006). Understanding the multiple
Patrick, M. E., Schulenberg, J. E., & OMalley, P. M. contexts of adolescent risky behavior and positive
(2013). High school substance use as a predictor of col- development: Advances and future directions. Applied
lege attendance, completion, and dropout: a national Developmental Science, 10, 107113.
multi-cohort longitudinal study. Youth & Society. Schulenberg, J. E., & Maggs, J. L. (2002). A develop-
NIHMSID528548. doi:10.1177/0044118x13508961 mental perspective on alcohol use and heavy drinking
Professor X. (2011). In the basement of the Ivory Tower: during adolescence and the transition to young adult-
Confessions of an accidental academic. New York: hood. Journal of Studies on Alcohol. Supplement, 14,
Penguin. 5470.
Qui, H., Bures, R., & Shehan, C. L. (2012). The inconsis- Schulenberg, J. E., & Patrick, M. E. (2012). Historical and
tent mediating effects of psychosocial work character- developmental patterns of alcohol and drug use among
istics on the education-health relationship. Social college students: Framing the problem. In H. R. White
Science & Medicine, 75, 15391546. & D. Rabiner (Eds.), College drinking and drug use
Rosenbaum, J. E. (2011). The complexities of college for (pp. 1335). New York: Guilford.
all: Beyond fairy-tale dreams. Sociology of Education, Schulenberg, J., OMalley, P. M., Bachman, J. G., &
84(2), 113117. Johnston, L. D. (2005). Early adult transitions and
Rosenbaum, J. (2012). Degrees of health disparities: their relation to well-being and substance use. In R. A.
Health status disparities between young adults with Settersten, F. F. Furstenberg, & R. G. Rumbaut (Eds.),
high school diplomas, sub-Baccalaureate degrees, and On the frontier of adulthood: Theory, research, and
Baccalaureate degrees. Health Services and Outcomes public policy (pp. 417453). Chicago: University of
Research Methodology, 12(2), 156168. Chicago Press.
Ross, C. E., & Mirowsky, J. (1999). Rening the associa- Siennick, S. E., & Staff, J. (2008). Explaining the educa-
tion between education and health: The effects of tional decits of delinquent youths. Criminology, 46,
quantity, credential, and selectivity. Demography, 36, 609636.
445460. Snyder, T. D., & Dillow, S. A. (2012). Digest of educa-
Ross, C. E., & Mirowsky, J. (2006). Sex differences in the tion statistics 2011 (NCES 2012001). Washington,
effect of education on depression: Resource multipli- DC: National Center for Education Statistics,
cation or resource substitution. Social Science & Institute of Education Sciences, U.S. Department of
Medicine, 63, 14001413. Education.
388 M. Kirkpatrick Johnson et al.

Staff, J., Schulenberg, J. E., Maslowsky, J., Bachman, White, H., McMorris, B. J., Catalano, R. F., Fleming, C. B.,
J. G., OMalley, P. M., Maggs, J. L., & Johnston, L. D. Haggerty, K. P., & Abbott, R. D. (2006). Increases in
(2010). Substance use changes and social role transi- alcohol and marijuana use during the transition out of
tions: Proximal developmental effects on ongoing tra- high school into emerging adulthood: The effects of leav-
jectories from late adolescence through early ing home, going to college, and high school protective
adulthood. Development and Psychopathology, 22, factors. Journal of Studies on Alcohol, 67, 810822.
917932. Woolf, S. H., & Aron, L. (2013). U.S. health in interna-
Thoits, P. A. (1995). Stress, coping, and social support tional perspective: Shorter lives, poorer health.
processes: Where are we? What next? Journal of Washington, DC: National Academies Press.
Health and Social Behavior, 35, 5379. Yang, Y. (2008). Social inequalities in happiness in the
Thoits, P. A. (2011). Mechanisms linking social ties and United States, 19722004: An age-period-cohort anal-
support to physical and mental health. Journal of ysis. American Sociological Review, 73, 204226.
Health and Social Behavior, 52, 145161. Zajacova, A., Rogers, R. G., & Johnson-Lawrence, V.
Vuolo, M., Staff, J., & Mortimer, J. T. (2012). Weathering (2012). Glitch in the gradient: Additional education
the great recession: Psychological and behavioral tra- does not uniformly equal better health. Social Science
jectories in the transition from school to work. & Medicine, 75(11), 20072012. doi:10.1016/j.
Developmental Psychology, 48, 17591773. socscimed.2012.07.036.
Wheaton, B., & Clarke, P. (2003). Space meets time: Zheng, H., Yang, Y., & Land, K. C. (2011). Variance func-
Integrating temporal and contextual inuences on tion regression in hierarchical age-period-cohort mod-
mental health in early adulthood. American els: Applications to the study of self-reported health.
Sociological Review, 68, 680706. American Sociological Review, 76(6), 955983.
Life Course Lens on Aging
and Health

Kenneth F. Ferraro

1 Introduction have emerged in recent decades. Chief among


them is a life course lensa perspective or way
A lens is a transparent object used to help form an of thinking that has spawned the development of
image of something. Humans with 20-20 vision new theories and methods to analyze the aging
have optimal visual function for everyday life, process and the health changes that accompany it
but many humans rely on glasses or contact (Cain 1964; Elder 1974). Although many schol-
lenses to enhance their visual function. Even with ars nd the medical lens with its focus on geriat-
corrective lenses, however, there are untold phe- ric syndromes to be very helpful, scholars from a
nomena that remain undetectable to the human variety of elds are turning to the life course lens.
eye. Thus, we turn to telescopes, microscopes, or Indeed, it could be argued that within some elds,
an imaging device to help us to capture images the life course lens is emerging as one of the most
that were previously undetected. Use the appro- useful ways to study health during adulthood and
priate lens and a new world appears, whether in later life. Instead of a focus on the health prob-
outer space or within the cell. Indeed, imaging lems of later life, the life course lens emphasizes
devices are transforming science in remarkable the long-term antecedents of both survival and
ways. The same can be said of the scientic study the health status of the survivors. By adopting
of aging and health. How we form our images of this lens to study aging and health, scientists
what it means to grow older is essential to our from multiple disciplines including sociologists,
understanding of concepts such as normal aging, epidemiologists, economists, and psychologists,
public health, and health inequality. have deposited breakthrough discoveries in the
There are many lenses that can be used to vault of knowledge about aging and health.
observe the myriad changes in health as one Indeed, the application of the life course lens has
grows older. A medical lens emphasizes geriatric helped to transform our knowledge of the inter-
syndromes and has long held sway in the study of section of aging and health.
aging and health. Other valuable lenses, however, The purposes of this chapter are to critically
evaluate why the life course lens has been widely
adopted for studying aging and health and to
review how it has aided our understanding of
K.F. Ferraro (*) health dynamics. Doing so enables one to (a)
Department of Sociology and Center on Aging and evaluate the degree to which the life course lens
the Life Course, Purdue University, 700 West State is challenging long-held views of aging and
Street, West Lafayette, IN 47907-2059, USA
e-mail: ferraro@purdue.edu

Springer International Publishing Switzerland 2016 389


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_18
390 K.F. Ferraro

health and (b) suggest theoretical and method- especially by evolutionary biologists, there are
ological avenues for future application. three basic life stages: pre-reproductive, repro-
ductive, and post-reproductive. What we think of
as aging is qualitatively different in these three
2 What Do We Mean by Aging? stages. It could be argued that the eld of geriat-
rics focuses on post-reproductive aging.
Other chapters in this volume apply a life course The three stages convey fundamental differ-
lens to the study of health during adolescence and ences in how the organism is changing over time
mid-life, which could also be characterized as the but do not imply they are tightly tethered to
study of aging and health. Thus, it may be use- chronological age. By surviving, a child accumu-
ful to briey clarify what is meant in this chapter lates time from the date of birth (we celebrate
by the term aging. Some denitions are tightly anniversaries of the birth date), but biological,
linked to chronological aging: to age is to change psychological, and social changes during child-
with the passage of time, as reected in birth- hood are quite distinct from those occurring dur-
days. Infants age, adolescents age, and older peo- ing the reproductive or post-reproductive periods.
ple age. The value of this denition lies in its Although the focus of gerontology has been
universality, but that can also be a limitation. adaptation during the last stage, there is a rapidly
Aging reduces to chronological aging, an essen- developing body of research revealing vital links
tial concept for our endeavor, but one that fails to between the pre- and the post-reproductive
convey what interests most scholars when study- periods.
ing what it means to grow older. Using a life course lens, therefore, is inconsis-
Gerontology and geriatrics systematically tent with an exclusive focus on post-reproductive
examine senescence or the process of gradual aging. Essential to the life course lens is recog-
declines in function. Given the attention to dete- nizing the interrelatedness of various life stages
rioration in function, and its accompanying and the resulting continuities and discontinuities
greater risk for mortality, one can conceptualize that are apparent from the long view of growing
biological aging as the rate of such decline. older (see also chapter by Does the Body Forget?
Biological aging affects all members of a species, Adult Health, Life Course Dynamics, and Social
not just those aficted by disease. Juxtaposing Change Hayward and Sheehan, this volume). At
chronological aging with biological aging pro- its core, the life course lens provides a view of the
vides a vantage point for identifying the factors entire period of time from conception to death
that accelerate or decelerate the pace of aging. (Ferraro 2011; Kuh and Ben-Shlomo 1997). In
Scholars refer to accelerated aging or weath- this sense, the study of aging using a life course
ering as premature declines in physiological sta- lens is more closely linked to family medicine
tus and function, typically assessed by earlier than to geriatrics (Daaleman and Elder 2007). As
declines than those experienced in the population a result, the term aging is used herein to focus on
(Geronimus et al. 2010). Identifying the factors adulthood, especially the post-reproductive
that lead to deviations from average decline pro- period, while acknowledging its interrelatedness
vides essential information for detecting a weath- to both the pre-reproductive stage and the repro-
ering effect and developing effective interventions ductive stage.
to slow the processes of senescence. Moreover, the use of life stages in this chapter
The reference in this chapter to aging and is distinct from many stage depictions. A strict-
health is intended to focus attention on adult stage approach to aging reies the periods of the
health outcomes while recognizing their early life course and looks for solutions to many later-
antecedents. Drawing from the biological sci- life health problems within the post-reproductive
ences, it can be argued that reproduction is a period, typically uncovering proximal causes. A
fulcrum for dening life course trajectories and strict-stage approach gives priority to periods of
population aging (Ferraro and Shippee 2009, the life course, presuming that most people pro-
p. 337). In many biological studies of organisms, ceed through the stages in a similar sequence and
Life Course Lens on Aging and Health 391

often at predictable ages. What may be considered assume that historical context is not consequential
a soft-stage approach to aging, by contrast, to the outcome. Second, the age differences
emphasizes the interrelatedness of the stages described above are actually contrasts across
albeit with detailed examination of outcomes noncontiguous age categories. The gaps between
measured in later life. In short, a soft-stage the three age groups are presumed to be inconse-
approach acknowledges qualitative differences in quential to the outcome or simply reected in the
the three basic stages, many of which are biologi- midpoint of estimates between the two closest
cally driven, but the interrelatedness of the three categories. This assumption is also a threat to
stages and the environmental processes that external validity because it presumes that age and
shape the biological processes. In this chapter, I cohort differences are linear. Disordered cohort
use a soft-stage approach to focus on the early ow, resulting from variations in fertility (e.g.,
origins of adult health outcomes. Although there Baby Boom), means that a linear assumption is
remains profound interest in the study of later probably unwarranted. Cross-sectional studies of
(post-reproductive) life, it is counterbalanced by a survey sample also confound age and cohort,
interest in the long-term antecedents of health but the typical age range enables one to at least
status in the later years. compare contiguous age groups.
A related research approach, with even stron-
ger assumptions, is the extreme-contrast design
3 Sources of Evidence for assessing age differences (Getzmann et al.
for Research on Aging 2015). Why even bother with the middle-age
and Health group if one is truly interested in contrasts
between youngpresumably healthyadults
3.1 Age Contrasts: Types and older adults? Why not just compare persons
and Meaning of Age 1825 to those who are age 65 and older? The
Differences two problems associated with the three-category
age contrast are exacerbated in the extreme-
Many scholars of aging and health make use of contrast design. The engines of health change
age contrasts to study the processes involved with over time (see also chapter by Does the Body
growing older. Commonly used experimental Forget? Adult Health, Life Course Dynamics,
designs entail studying age contrasts such as and Social Change Hayward and Sheehan, this
young people (1825 year olds, frequently col- volume). Whether considering water uorida-
lege students), middle-age persons (4049), and tion, consumption of unsaturated fats, or noise
older people (65 and older). Whether used in psy- exposure, there are major social trends that inu-
chology, hearing science, or toxicology, the goal ence health and health behavior. Extreme-contrast
is to identify age differences in levels of function- designs make ferreting out the inuence of such
ing or some phenomenon under investigation social trends exceedingly difcult. As such, it is
(Abdala et al. 2014). This may be a wholly convenient to assume that they are trivial. In the
acceptable design for some research questions, process, one misses social and historical forces
but its limitations are substantial. that are essential to the life course lens and may
First, the purported age differences are con- actually be the keys to health improvement.
founded with cohort differences. Failure to
account for cohort differences, reecting histori-
cal context, entails an assumption that differences 3.2 Longitudinal Study of Aging
in functioning (or the outcome under study) and Health
between persons 18 and 25 years old and those in
middle or older categories are untainted by social By contrast, the life course lens values historical
change. Historical context is very important to context, seeks to explicate cohort variability, and
many outcomes of interest, but external validity questions assumptions of linear age or cohort
may be threatened by age-contrast studies that differences. As a result, emphasis is given to
392 K.F. Ferraro

longitudinal formulations of data. Again, potential bias derived from retrospective


cross-sectional studies may be wholly appropri- questions (Alwin 2007). For persons of advanced
ate for some research questions, but most research age, there are salient concerns about recall accu-
questions focused on aging and health require racy, but this is minimized by effective question
more than associations between point estimates wording (Henkel 2014; Schryer and Ross 2014)
of phenomena. Integrating some type of longitu- and adjusting estimates for respondent attributes
dinal or temporally-sequenced data is critical for that have been shown to inuence reporting:
opening vistas of understanding regarding how adult self-rated health, psychological disorder,
health status changes as one grows older. and SES (Vuolo et al. 2014). Although long-term
The most commonly-used designs involve prospective studies are the gold standard for
repeated measures, and a growing number of studying aging and health, information from ret-
these studies involve many follow-ups over fairly rospective questions is having a demonstrable
long time periods. Indeed, several large longitu- impact on the study of aging and inequalities in
dinal surveys of representative samples now pro- health, especially when coupled with a prospec-
vide exceptional data for studying aging and tive study.
health. The net effect is that the study of aging
and health has catapulted during the past four
decades from a eld reliant on age differences in 3.3 Contributions from the Long-
health to observed changes in health. Some stud- Term Study of Aging
ies have had fairly short windows of assessing and Health
change in repeated measures, but the trend toward
the analysis of rich longitudinal data has been in There are at least three main vantage points for
effect for years and will probably accelerate the long-term study of aging and health, and each
(Ferraro and Kelley-Moore 2003a). Both the approach has generated a vibrant research litera-
length of time participants are studied and the ture: centenarians, early origins, and family lin-
number of measurement occasions are rising in eage. The life course lens can be used to clarify
studies such as the Health and Retirement Study, many processes related to aging and health, but
National Longitudinal Study of Adolescent its application is distinct based on the vantage
Health, and Panel Study of Income Dynamics. point.
These large probability surveys are enabling
investigators to make unparalleled observations 3.3.1 Centenarians: The Remarkable
about how health changes during adulthood Rearview Mirror
and the health inequalities that emerge over the There are no studies of which I am aware that
life course (e.g., McDade et al. 2014; Shuey and were designed to prospectively study centenari-
Willson 2014). ans from the day of birth to 100 years. The practi-
There has also been growing interest in asking cal challenges of such a study are obvious; thus,
retrospective questions to incorporate experi- investigators do the opposite: they study people
ences from childhood. As discussed below, there who have lived to 100 (or perhaps 95), chroni-
are studies tracking people from birth to later life, cling their lives, summarizing their exposures
but representative samples are largely limited to and, if possible, integrating vital or health
European studies (e.g., Jrvelin et al. 2004; records. Centenarian studies frequently capitalize
Wadsworth et al. 2006). Catching the wave of life on a mixture of retrospective and prospective
course epidemiology, however, many U.S. stud- data, but sampling persons 100 years old means
ies have included questions probing childhood that the prospective tracking is relatively limited.
health, socioeconomic status (SES), and misfor- Instead, the emphasis in centenarian studies is
tune. There is measurement error in all proce- documenting the validity of reported age, captur-
dures asking respondents for information, but ing notable life events and transitions, and mea-
research during the past decade has probed suring current characteristics of the phenotype at
Life Course Lens on Aging and Health 393

age 100 or more. The overall approach, therefore, together to extend longevity and enhance health
is a rearview mirror. in adulthood. Genes are indispensable to the cal-
The Okinawan Centenarian Study begun in culus of longevity, especially because there is
1975 is the longest-running investigation of its growing evidence that centenarians are more
type, and many other centenarian studies have likely to be nested within long-lived families, but
appeared since then, including U.S. studies in the life course circumscribes the exposures and
New England (Perls and Silver 1999) and Georgia choices known to inuence health.
(Cho et al. 2012), as well as a host of studies
across the globe (e.g., Heidelberg, Southern Italy, 3.3.2 Early Origins: Tracking the Way
Sweden, and Tokyo). We Age
Beyond their remarkable similaritythat of Although the rearview mirror approach has
reaching 100centenarians are a diverse lot. In yielded far-reaching ndings about aging and
the New England Centenarian Study, the investi- health, a different genre of researchthe early
gators identied three basic categories: escaper origins approachhas delivered unprecedented
(people who reach 100 without any major disease insights into life course health. Instead of relying
or illness: 13 % of the sample); delayer (major on a reconstruction of the life course from retro-
disease onset at age 80 or later: 45 %); and survi- spective data, long-term tracking of human lives
vor (major disease onset before age 80: 42 %) has opened a window for studying health over
(Perls and Silver 1999; Sebastiani and Perls large periods of time. To begin, it altered a
2012). widely-held denition of the life course.
Although one might hypothesize that leading For decades, gerontologists exhorted others to
an advantaged life should increase the odds of examine the entire lifefrom birth to deathbut
living to 100, Stathakos et al. (2005, p. 514) research in recent decades has led to a revision:
reported that Greek centenarians have experi- from conception to death.
enced special hardships at some point of their There is a rapidly growing body of research
lives, extending from poverty and starvation examining the early origins of adult health, and
(mostly during war) to participation in battles, this line of inquiry has been stimulated by the
captivity, or exile. Buettner (2012, p. 193) simi- work of Barker (1997, 2001) on the fetal origins
larly reported that many long-lived persons expe- of adult health (see also chapter by Mental
rienced hard times and remarked that in Costa Health Avison, this volume). Thus, the range of
Rica an early life of hardship had tempered life course analyses has been extended in con-
men to embrace physically demanding work and temporary epidemiology because it also covers
deal with life challenges. The evidence from gestation. Indeed, Barkers research revealed that
these studies is not derived from a prospective or multiple health problems, which emerged during
control-group design, but perhaps there is some- adulthood, actually originated in the womb. Low
thing about overcoming misfortune that is associ- birth weight, which often reects some type of
ated with exceptional longevity. Regardless of negative prenatal exposure such as fetal
their life circumstances, scholars agree that malnutrition, has been shown to predict a host of
adaptation to the challenges of aging is also a maladies in middle and later life such as insulin
key protective factor for healthy aging and lon- resistance, cardiovascular disease, and lung can-
gevity (Willcox et al. 2010, p. 2). cer (e.g., Barker 1997; Barker et al. 2002;
Tracing the life course of centenarians has yet Eriksson et al. 2010). This research program has
to reveal any magic bullets for exceptional lon- grown, especially in England and Finland, via
gevity. No single factor such as diet, physical clever record linkage of middle- and older-age
activity, water consumption, mental outlook, or persons to their birth records and body measure-
genes satisfactorily explains who lives to 100 nor ments during infancy (some studies also capital-
the health variability among centenarians. Rather, ized on records of mothers height and weight
the literature points to multiple factors operating and/or fathers occupation). And some of these
394 K.F. Ferraro

studies augmented the life course data by splic- related to a host of adult risks, including adult
ing in hospital records (e.g., Finland maintains cardiovascular disease, depression, chronic pain,
national registers of hospital discharges by cause; and mortality (Cohen et al. 2010; Goosby 2013;
Barker et al. 2002). Findings from this line of ORand and Hamil-Luker 2005; Pudrovska and
inquiry have been transformative on the study of Anikputa 2014; Schafer et al. 2013). Childhood
aging and health. health has also received attention, generally
Even more impressive in terms of research revealing life course continuity: poor health dur-
design are long-term studies of the life course ing childhood heightens the risk of poor health in
that were structured to track one or more birth adulthood and later life (Blackwell et al. 2001;
cohorts over decades. One exemplar is the British Case and Paxson 2011; Latham 2014). Although
National Survey of Health and Development acquired immunity occurs for some infectious
(NSHD), which capitalized on a maternity study diseases, there are relatively few studies that nd
to track all births during 1 week in March 1946. evidence for it affecting chronic disease risk.
This means that the NSHD has data on individu- Researchers have systematically studied
als from gestation to about 70 yearsa treasure whether various types of negative exposures,
for applying the life course lens to health referred to as misfortune or adversity, inuence
(Wadsworth et al. 2006). Another birth cohort health status in later lifeand the picture is
study, tracking persons born in northern Finland daunting. Dozens of studies reveal the long-term
during 1966 also reveals that birth weight and consequences of child abuse on multiple diseases
infant growth are associated with the risk of mul- during adulthood including, but not limited to,
tiple diseases during middle age (Jrvelin et al. cancer (Felitti et al. 1998; Fuller-Thomson and
2004). Brennenstuhl 2009; Morton et al. 2012), hyper-
Most studies from the United States and tension (Stein et al. 2010), myocardial infarction
Canada rely on surveys of adults with retrospec- (heart attack) (Morton et al. 2014; ORand and
tive questions to tap experiences during child- Hamil-Luker 2005), ulcers (Springer 2009), and
hood or adolescence (Felitti et al. 1998) or mental distress (Edwards et al. 2003). Beyond
retrospective questions posed to the mothers of child abuse per se, there is growing evidence that
research subjects for information on infancy being raised in a risky family is associated with a
(birthweight, breastfeeding; McDade et al. 2014). host of physical and mental health problems
A notable exception is the Berkeley Guidance (Brown et al. 2009a, 2010; Nock and Kessler
study that sampled 248 infants in 19281929 and 2006; Repetti et al. 2002; Schafer and Ferraro
followed them for 70 years, but the sample was 2012; Williamson et al. 2002).
small and restricted to mostly White middle-class Although there are studies reporting no rela-
families in California (Wink et al. 2007). tionship between noxious exposures during
Although current U.S. studies cannot match childhood and adult health (e.g., Korpimki et al.
the duration of prospective design available in the 2010), most of the published studies show nota-
British NSHD, multiple prospective studies of ble links. The emergent research questions for
adolescents or adults are providing scholars with the next decade, therefore, include: (1) What are
data that answer important questions about expo- the most consequential types of misfortune inu-
sures that are related to life course health. These encing adult health? (2) Are the presumed health
studies include, but are not limited to, Americans consequences of negative life exposures inexora-
Changing Lives, Health and Retirement Study, ble? (3) What psychosocial resources enable
National Longitudinal Studies, National Survey people to blunt or compensate for the negative
of Midlife Development in the United States, exposures? (4) What biological mechanisms are
Project Talent, and Wisconsin Longitudinal involved in how these negative exposures inu-
Study. ence adult health?
Tremendous interest has been shown in paren- Vincent Felitti (2002, p. 44), a physician
tal SES, generally revealing that it is inversely researcher and lead investigator of the Adverse
Life Course Lens on Aging and Health 395

Childhood Experiences Study of enrollees in the to inuence health, sometimes described as


Kaiser Health Plan in San Diego summarized the stress-diathesis processes (Johnson and Krueger
extant ndings by concluding that: 2005a, b; Shanahan et al. 2008).
Our ndings are of direct importance to the every- Second, although lacking information on
day practice of medicine and psychiatry because genetics per se, there is growing interest in study-
they indicate that much of what is recognized as ing linked lives within families. Most of these
common in adult medicine is the result of what is studies either purposively study multiple family
not recognized in childhood.
members over one or more generations
(Fingerman et al. 2012) or tap linked lives via ret-
3.3.3 Family Lineage: rospective questions on parental characteristics
Intergenerational Health (Quesnel-Valle and Taylor 2012). Other studies
The rst two approaches for studying aging and of family lineage merge related data sets or com-
health provide glimpses of the third approach. ponents of a large study such as the Panel Study
Felittis claim above, for instance, emphasizes of Income Dynamics (e.g., Davis et al. 2008).
the role of families and households in setting dis- These types of studies are not only adding to the
tinct paths for adolescent and adult develop- stock of knowledge on parental inuence but also
mentand the health consequences of early to the reciprocal inuence between grandparents
exposures, whether positive or negative. And and grandchildren. Examples include how grand-
with centenarians, there is the recognition that parents SES (including investment in college
longevity is nested within families. Indeed, a savings plans) may inuence the status attain-
proposition of cumulative inequality theory iden- ment of grandchildren and how raising a grand-
ties the central role of family lineage on health child inuences health and nancial well-being
and functioning: Inuenced by genes and envi- of grandparents (see respectively, Jger 2012;
ronment, family lineage is critical to status dif- Hayslip et al. 2015).
ferentiation early in the life course (Ferraro and Finally, any discussion of how family lineage
Shippee 2009, p. 337). is central to studying aging and health also needs
There are different literatures that bring into to consider the well-established fact that parental
focus the inuence of family lineage on aging SES is an important inuence on ones
and health. First, whereas parents endow their SES. Although there is clearly intergenerational
children with both traits and experiences that mobility, there is considerable continuity in SES
shape their life chances, many studies search for across generations. And since family-of-origin
a common genetic source of health risks. SES is related to health, the intergenerational
Although scholars may debate the relative inu- transmission of SES is a contributing factor to the
ences of nature, nurture, and the combination of intergenerational transmission of health. As such,
the two, there is no denying the genetic origins of studies of the inuence of family lineage on
many physical and mental diseases, cognitive health need to adjust adequately for SES. Failure
ability, and behaviors. In this sense, not only is to do so may lead to overestimating the inuence
there a genetic component to the diseases that of other factors relative to SES. This literature
compromise health, but also the health-risk reveals that there are independent effects of SES
behaviors that lead to or reinforce poor health on health, but also that childhood health, nutri-
(Corcoran 1995; Shanahan and Hofer 2005; tion, and schooling partially mediate the SES
Wickrama et al. 1999, 2005). Both behavioral relationship across generations (Carvalho 2012).
genetics studies of twins (in search of shared
genetic and environmental overlap) and studies 3.3.4 Three Vantage Points
of how specic alleles are related to behavior and for the Study of Aging
health risks are helping to explicate the role of and Health
environmental contingencies shaping exposures After a brief review of these three vantage points,
and how such exposures combine with genetics it may be helpful to integrate and graphically
396 K.F. Ferraro

summarize these three positions for viewing prospectively studying the grandchildren or
aging and health. A lens enables one to see some- great-grandchildren of centenarians.
thing more clearly, but photography or magnetic
resonance imaging is very dependent upon the
vantage point. I depict the three vantage points in 4 Compression of Morbidity?
Fig. 1. In the top section of the diagram, labeled
Generation 1, the rst isosceles triangle is labeled One of the major issues confronting scholars of
A, which represents the eld of vision for study- aging during the past three decades has focused
ing centenarians. Viewing from the right side of on the consequences of the public health success
the diagram (vertex with the smallest-degree referred to as population aging. It is source of
angle), one assumes the vantage point of age 100 great satisfaction that many societies are seeing a
or older to gaze back on the life course. By con- rise in life expectancy, largely spurred by
trast, B moves the vantage point to the left side of improvements in nutrition and public hygiene as
the diagram and represents the early origins study well as reduced fertility. Although population
of life course; one begins as early as gestation, aging is a remarkable success, it also raises ques-
birth, or childhood and looks forward to prospec- tions about what this means for the quality of
tively track the life course. The idea is that the later life. How might we describe the years that
same person can be studied retrospectively or are being added to life expectancy? Is it an addi-
prospectively, and one would no doubt ask dis- tion of years of healthy life, perhaps disability-
tinct questions when trying to uncover the ante- free aging? Or might the aging of the population
cedents of health and functioning from two usher in more years of life constrained by mor-
vantage points. Of course, only a small fraction bidity and disability?
of humans become centenarians, but one could These questions prompted Fries (1980) to prof-
theoretically modify the vantage points on the fer the idea of a compression of morbidity, in
x-axis to suit the age range under consideration. which he envisioned a delay in the onset of chronic
The lower half of Fig. 1 replicates these two disease, enabling many people to live a larger pro-
vantage points (A and B) but the triangles and portion of their lives free from disease and disabil-
vantage points are shifted to the right on the his- ity. Sociological study of the compression of
torical time axis to depict a second generation. C, morbidity has given fairly limited attention to dis-
which crosses over Generations 1 and 2, repre- ease onset but warmly embraced efforts to exam-
sents the family lineage approach where at a ine whether disability has been compressed.
given point in time, one can examine health Two studies of disability were especially inu-
across generations. We know that family medical ential in shaping the eld and discussions of the
history is important to predict the health of subse- compression of morbidity. First, Verbrugge and
quent generations, and scholars have begun to Jette (1994) described a disablement process, by
integrate multiple generations to examine disease which diseasesas diverse as arthritis, cancer, dia-
risk, functioning, and longevity. Studies imple- betes, emphysema, and hypertensionled to func-
menting a family-lineage approach open up new tional disability. For older people, the expectation
ways to study not only genetic inuences but also was that the disablement process had a strong pull.
shared environments and the inuence of SES Although not universal, the disablement process,
across generations. once begun, offers only a modicum of hope to
Figure 1 is intended to summarize the three escape it. Many presumed that the task is to mini-
vantage points, each of which is extremely valu- mize or stall the decline in functional ability. The
able, and foster conversation about integrating authors identied factors that might speed or slow
them in future research. Additional generations disablement, with little attention given to revers-
may be added to enrich the study of family health ing the process (Verbrugge and Jette 1994, p. 1).
while attending to social change on the historical A nding about reversing the disablement
axis. The gure may also suggest the utility of process, however, emerged about the same
Life Course Lens on Aging and Health 397

0 10 20 30 40 50 60 70 80 90 100

0 10 20 30 40 50 60 70 80 90 100

Fig. 1 Three vantage points for the study of aging and health: A Centenarians, B Early origins, and C Family lineage

time. Manton et al. (1993, p. S194) reported 4.1 Interpreting the Evidence
that disability actually declined among chroni-
cally disabled community-dwelling and institu- Although efforts to replicate these ndings con-
tionalized elderly populations. Many scholars tinue, three conclusions are apparent from the
challenged this nding, which was based on extant literature. First, there is now considerable
analyses of the National Long-Term Care evidencefrom studies of both individual- and
Survey, and attempted replication. To date, the cohort-changethat disability among older
ndings are mixed, with some studies reporting adults, measured via activities of daily living
partial support, especially among the oldest (ADL), declined during the mid-1980s to 2000
members of the population (Crimmins et al. (Freedman et al. 2004; Manton 2008). Second,
2009; Freedman et al. 2004; Manton et al. 2008; more recent evidence reveals that the trend of a
Martin et al. 2009). disability decline has stalled since 2000 and is
398 K.F. Ferraro

unlikely to resume because adults poised to and what will likely happen when one considers
enter late life over the next decade have slightly the health consequences of the rise in life expec-
higher rates of activity limitations (Freedman tancy. They convincingly outlined how there are
et al. 2013:669). Third, whereas ADL disability inklings of a compression of morbidity on some
taps whether or not a person is functionally diseases or conditions. Examples include a recent
capable of independent living, there are many decline in rates of myocardial infarction as well
implications of this nding for projecting the as an increase in survivorship after a heart attack.
future of aging and health. The way in which dis- A decline in some types of cancer incidence has
ability was measured focused on more advanced also been observed recently, and cancer survival
forms of disability, and some surveys actually continues to rise. On the other hand, the preva-
limited their samples to persons who were lence of diabetes, arthritis, and some types of
screened to have some disability (e.g., compo- cancer (e.g., liver) has risen in recent yearsevi-
nents of the National Long Term Care Survey). dence that is counter to the claims of a compres-
The reported decline in disability, although sion of morbidity. Perhaps the correct answer is
important, needs to be interpreted as a short-term that a compression of morbidity has been
trend among persons who previously reported observed in the U.S. on some diseases but not on
some ADL disability. In other words, the decline others. Regardless, it seems that survival after
was a real; but it was temporary and a proverbial disease onset has improved, consistent with a
tip of the iceberg in the study of disability among compression of disability interpretation.
adults (Verbrugge 1986). This pattern of ndings also led Crimmins and
So has a compression of morbidity occurred? Beltrn-Snchez (2011) to offer some further
Can we expect more years of active life expec- interpretation. First, diseases are both less lethal
tancy? Will the life course two decades from now and less disabling; they have become more
have a similar period of morbidity and disability chronic but perhaps less progressive (p. 82). It
or will it be further compressed? Answers to could be argued that we are better at medicating
these questions are challenging scholars of aging people with various diseases to keep them func-
and health to interpret the extant data in light of tioning with their conditions. Second, they con-
recent trends. tend that what we are witnessing is not that
In a follow-up essay about three decades after people are less likely to get a disease but that
the provocative idea, Fries understandably argued people are better at living with each disease,
that a compression of morbidity had occurred thereby aiding longevity. As a result, they argue
(Fries et al. 2011). Although his 1980 article that there is little evidence that a compression of
focused on the compression of morbidity (i.e., morbidity has occurred.
disease), Fries et al. (2011) subsequently enlarged The fact that both the prevalence and inci-
the denition of morbidity to include disability. dence of diabetes continues to rise is also indica-
Classical epidemiology holds that morbidity tive of another trend of epic proportions: the
refers to disease and that disability refers to an increased prevalence of obesity. There is ample
inability in physical function. Thus, simply evidence to expect that an increasing number of
changing the denition of morbidity to include older people in developed nations will be able to
disability seems to be a strained approach to maintain their physical function. At the same
interpreting the evidence. time, this buoyant view of the later years may be
tempered by the rising prevalence of obesity.
Many scholars contend that the growing rate of
4.2 Informing Future Projections obesity may well halt any further compression of
of Compression of Morbidity morbidity (Manton 2008). The rising prevalence
of obesity, and severe obesity, is yielding an epi-
Crimmins and Beltrn-Snchez (2011) argued demic of diabetes mellitus (Ogden et al. 2006).
for a more nuanced view of what has happened Given that obesity is a potent risk factor for
Life Course Lens on Aging and Health 399

disability (Ferraro and Kelley-Moore 2003b), the Are health trajectories inuenced by parental sta-
increase in the prevalence of obesity over the past tus and health? Do changing status hierarchies
30 years may compromise further improvements modulate health trajectories? Interest has grown
in population health, even those related to main- in theoretical articulation of these relationships
taining physical function. Unfortunately, the as manifest in cumulative inequality theory
most recent evidence shows that prevalence of (Ferraro and Shippee 2009; Schafer et al. 2011)
both childhood and adult obesity has not abated, and efforts to link stress process theory to the life
suggesting little relief regarding this ominous course lens (Pearlin 1989, 2010). Persons of low
trend (Ogden et al. 2014). SES generally have greater lifetime exposure to a
variety of risks that may compromise health.
Given that SES shapes the life course from con-
5 Advancing Life Course ception to death, many investigators seek to
Research on Aging splice together long-term exposures to socioeco-
and Health nomic stressors and evaluate the potential health
consequences (e.g., Lynch et al. 1997; Shippee
Although we have learned much in recent decades et al. 2012). Most of this research is informed by
about life course health, there are so many impor- a soft-stage approach to aging, revealing continu-
tant research questions that merit sober scientic ity and change in both status hierarchies and
consideration, both theoretically and empirically. health, not just how SES is related to health at
In an effort to advance this genre of research one period of life.
and apply the life course lensI articulate sev- At the same time, interest in multi-level mod-
eral areas of inquiry that hold great promise for els has grown exponentially (Bollen and Curran
understanding aging and health as well as laying 2006; Singer and Willett 2003). These types of
the foundation for effective interventions to statistical procedures are useful for elegantly
improve public health in adulthood and later life. analyzing large data arrays of the life course.
Rich data from detailed and long-term studies of
the life course provide exceptional opportunities
5.1 Cumulative Inequality for studying how health inequality develops over
in Health time. Yet, investigators seek parsimony in repre-
senting the life course. The response of many
It is noteworthy in many respects that the term scholars has been to embrace multi-level models
inequality has entered more widely into the pub- for trajectory analyses, including parallel trajec-
lic lexicon in recent years. No doubt research on tories. Using such methods means that one may
the social gradient in health has played a role in simultaneously track changes in social status and
this popularity. There has been growing interest changes in health and functioning. The life course
in how inequality inuences health, sparked by lens leads one to consider related trajectories, and
studies examining the relationship between SES these and other statistical procedures have
and health (House et al. 1990; Link and Phelan empowered investigators to see important conti-
1995; Link et al. 2008; Marmot 2005; Marmot nuities and discontinuities in the life course
et al. 1991). (Lynch 2003).
What is distinctive in recent scholarship on the Cumulative inequality theory also calls atten-
topic, however, is the integration of the life course tion to selection processes, most notably mortal-
lens to transform the study of the social gradient ity and non-response attrition, when studying
in health from one focused on social status at a trajectories. According to Ferraro et al. (2009,
given time to more dynamic models of social p. 428), the premature mortality associated with
hierarchies and health. Research questions now accumulated risksselective survivalwill
include: How do status hierarchies early in life result in compositional change to a population.
inuence social status and health in later life? Cohorts shrink in a nonrandom manner,
400 K.F. Ferraro

sometimes reducing population heterogeneity. If the aging process by integrating this type of
selection is not accounted for when modeling the information into behavioral and social science
relationships between status hierarchies and studies. Studying biomarkers also may reveal
health, it is possible that conclusions may be mis- psychosomatic pathways in the response to
directed by what appears to be decreasing stressors (Seplaki et al. 2004).
inequality over time. Thus, trajectory analyses, The use of biomarkers is especially useful to
which capitalize on impressive data arrays of the life course scholars because it can reveal some
life course, also need to ask how selection may aspects of the rate of biological aging and dys-
inuence the ndings and conclusions. Persons regulated patterns of stress reactivity. Specic
with high levels of stress exposure (e.g., forced to substances such as cortisol or C-reactive protein
relocate due to nancial strain) and/or many (CRP)or a complex of indicators such as allo-
health problems may be more difcult to recruit static loadmay help us reliably characterize
into surveys but also more likely to retain in long- persons as having higher or lower levels of health
term longitudinal follow-ups. They are important risk than persons of comparable age (Friedman
for accurately characterizing cumulative inequal- et al. 2012; Seeman et al. 2002). With that infor-
ity in a sample, and our analytic methods need to mation, we are able to capture some measure of
test for these forms of selection. biological aging or weathering, and most of these
biomarkers should not be biased by self-report.
Biomarkers also enable investigators to be
5.2 Biomarkers for Early more specic about how social factors inuence
Detection and Pathway health. Although the phrase get under the skin
Specicity is widely used, it could be argued that it remains
a black box (see also chapter by Mental
A second approach to studying aging and health Health Avison, this volume). Biomarkers pro-
that has received considerable attention in the vide the opportunity to be more specic in expli-
past decade involves the integration of biomark- cating biopsychosocial processes.
ers (or biomeasures) into social science research. The inuence of social context on a variety of
Family lineage approaches to studying aging and indicators during adulthood is striking. Consider,
health, including behavioral genetics and for example, the role of social context on CRP, an
genome-wide association studies (GWAS), have acute-phase protein, which is widely used as a
integrated genetic information into the study of marker of chronic inammation (Herd et al.
aging and health, but life course epidemiology is 2012). Browning and colleagues (2012) exam-
also tapping a wide array of other biological ined whether a sharp rise in the crime rate might
markers that reect current physiological condi- be associated with elevated CRP. In this study of
tion. Many biomarkers can be collected in a rising crime rates in Dallas, TX, the authors
minimally-invasive way from blood (lancet observed that a spike in the burglary rate was
prick), saliva, or toenail clippings. Of course, associated with elevated CRP among men but not
these sources can also be used to identify DNA for women. A different type of stressor was
and epigenetics (methylation of DNA), but the examined in a study of Brazilian immigrants
study of aging and health has gravitated to those dwelling in Boston. Holmes and Marcelli (2012)
biomarkers that are sensitive to the bodys observed that CRP was higher in unauthorized
response to stimuli (e.g., cortisol, alpha amylase, residents than in legal residents, even after adjust-
C-reactive protein). ing for variability in neighborhood socioeco-
In the process, we learn more about how the nomic status.
person responds and adapts to exposures, whether Other studies have examined religion as a
positive or negative. The underlying premise is source of integration and support as potentially
that biological systems are responsive to the protective against rising CRP in later life. Using
effects of context, and we will learn more about data from the National Social Life, Health, and
Life Course Lens on Aging and Health 401

Aging Project, Hill et al. (2014) found that well-being (Carp 1967; Hochschild 1978;
attending religious services was associated with Marshall 1975). Although many of these early
lower levels of CRP as well as EpsteinBarr studies capitalized on relocation to a single senior
virus. Given notable differences in the social community, the recent work is observing contex-
organization of predominantly Black and White tual variability to see how variation in place is
congregations, Ferraro and Kim (2014) observed associated with variation in aging and health. In
that religious attendance protected against ele- short, the study of clustered observations has
vated CRP for Black adults only. Moreover, this advanced considerably in recent years.
relationship was observed in cross-sectional and Social scientists, including sociologists, epide-
longitudinal analyses: religious attendance was miologists, and anthropologists, are well aware of
negatively associated with initial levels of CRP the inuence of social context on health across the
but also led to a decrease or slower rise in CRP life course, but scholarship in the past 15 years
among Black adults who frequently attended reli- has accentuated this tenet in new ways. Research
gious services. has identied important inuences on physical
These ndings reveal not only how risks such and mental health due to social capital (Snelgrove
as neighborhood crime or living as an unauthor- et al. 2009), neighborhood characteristics (Clarke
ized immigrant are associated with a marker of et al. 2008; Lee and Ferraro 2007), and even
inammation but also how social resources can architectural features of the neighborhood that
protect against rising CRP. The contribution of facilitate social interaction (Brown et al. 2009b).
biomarkers for research on aging and health is These are especially important ndings when
immense because they can signal health risks considering the health of older people: limited
long before diseases are diagnosedand this is mobility means that the importance of the built
particularly important for assessing risk in popu- environment and local social ties are magnied.
lations with limited access to medical care. Older adults may be particularly vulnerable when
Whereas CRP is highly predictive of cardiac social capital is low. This is illustrated well by
events such as myocardial infarction and stroke Klinenbergs (2002) nding that mortality was
for men and women, it is a harbinger of major highest for older adults with limited social inte-
cardiovascular risks (Ridker et al. 1998; Sesso gration when facing a Chicago heat wave.
et al. 2003). Integrating biomarkers across the Relocation is another important consideration
life course, moreover, can avoid the problems because it calls attention to person-environment
associated with self-reported health and identify t. Older adults are less likely than younger
upstream indicators of health risks. adults to relocate to independent residences.
When they do, however, there are major differ-
ences between moves that are due to pull fac-
5.3 Social and Physical Context tors (e.g., amenities, milder climate) and those
as Inuences on Aging that are involuntary or entail highly constrained
and Health choices (push factors such as poor health and
cognitive impairment). Environments are also
The terse review of how biomarkers are increas- important because they are markers for physical
ingly being integrated to study aging and health environmental exposures, creating opportunities
also revealed a cross-cutting theme that has come for social scientists to collaborate with toxicolo-
into sharper focus in recent decades: how social gists to examine the life course conuence of
and physical context inuences aging and health. social and environmental exposures. Residential
This is not a new interest; it has long been history conveys exposure to water, air, noise,
observed in medical sociology, sociology of crime, and social capital; thus, integrating envi-
aging, and developmental science. Classic stud- ronmental change into life course epidemiology
ies of life in age-segregated communities illumi- may yield considerable dividends for the eld
nated how social forces shape health and (Wheaton and Clarke 2003).
402 K.F. Ferraro

At the same time, investigators should not focus on the early origins of adult health is
limit environmental change to studying reloca- becoming paradigmatic for many scholars of the
tion only. The irony of aging in place is that life courseand even those who identify them-
residential stability is often associated with selves as gerontologists. The study of life course
neighborhood change. Because of their generally health is inuencing multiple elds and leading
lower rates of relocation, older adults often see to both methodological and theoretical advances
appreciable neighborhood change. Familiarity in subelds that are not blatantly related to health.
with ones residence may be salubrious, espe- At the same time, it is difcult to pinpoint the
cially if there are positive connections to the timing of causal processes because in utero fac-
community, but residential stability increases the tors are probably related to subsequent risks and
odds of observing neighborhood decline, which resources (Shanahan and Hofer 2011).
may affect morale and physical health (Brown It is understandable that much of the research
et al. 2003). This is another important arena for during recent decades has focused on explicating
future research. the exposures and the health consequences of
As noted before, some of these emergent or those exposures. Now that we have accumulated a
reinvigorated areas of study are evolving on both strong foundation of empirical generalizations on
theoretical and methodological fronts. Many the- the topic, perhaps it is time to give more attention
ories in sociology and economics have some to (a) interrupting chains of risk or (b) redirecting
component that addresses multiple levels of anal- etiological pathways by providing resources that
ysis such as macro, meso, and micro structures can neutralize or reverse the anticipated effects of
and processes. Multi-level models discussed ear- negative early exposures (DiPrete and Eirich
lier in the context of trajectory analysis also 2006; Ferraro and Shippee 2009). Studying cumu-
occupy a prominent role in studies of contextual lative inequality is important in its own right, but
inuences on aging and health. The challenge it also demands more attention to what types of
will be incorporating so many levels in multi- resources are best suited for interrupting chains of
level models (involving both ecological and tem- riskand for which populations subgroups.
poral clustering) while rendering ndings and Some interventions are effective for men but not
conclusions that are useful and succinct. for women or vice versa. The same can be said for
ethnicity and cultural groups. Thus, the call for
greater attention to ameliorating the effects of
6 Concluding Comments: negative exposures is also a call for sensitivity to
Interrupting Chains of Risk heterogeneity in the aging process.

The past two decades have ushered in a spectacu-


lar array of life course research on health. Using References
different vantage points and longitudinal designs,
the scientic community has opened new vistas Abdala, C., Dhar, S., Ahmadi, M., & Luo, P. (2014).
Aging of the medial olivocochlear reex and associa-
for understanding how social factors inuence
tions with speech perception. Journal of the Acoustical
aging and health. The ndings have transformed Society of America, 135(2), 754765.
the eld by giving more theoretical attention and Alwin, D. F. (2007). Margins of error: A study of reliabil-
more rigorous analyses of data on individuals ity in survey measurement. Hoboken: Wiley
Interscience.
over the life course. The role of context is unmis-
Barker, D. J. (1997). Maternal nutrition, fetal nutrition,
takable. Whether context is seen as place or time, and disease in later life. Nutrition, 13(9), 807813.
the aging process inuences health through mul- Barker, D. J. P. (Ed.). (2001). Fetal origins of cardiovas-
tiple mechanisms. Those mechanisms begin at cular and lung disease. New York: Marcel Dekker.
Barker, D. J., Eriksson, J. G., Forsn, T., & Osmond, C.
conception, and we now know that the in utero
(2002). Fetal origins of adult disease: Strength of
environment is consequential to more than just effects and biological basis. International Journal of
early survival and child development. Indeed, a Epidemiology, 31(6), 12351239.
Life Course Lens on Aging and Health 403

Blackwell, D. L., Hayward, M. D., & Crimmins, E. M. adult health. Annals of the New York Academy of
(2001). Does childhood health affect chronic morbid- Sciences, 1186, 3755.
ity in later life? Social Science & Medicine, 52(8), Corcoran, M. (1995). Rags to rags: Poverty and mobility
12691284. in the United States. Annual Review of Sociology, 21,
Bollen, K. A., & Curran, P. J. (2006). Latent curve mod- 237267.
els: A structural equation perspective. Hoboken: Crimmins, E. M., & Beltran-Sanchez, H. (2011). Mortality
Wiley. and morbidity trends: Is there a compression of mor-
Brown, B., Perkins, D. D., & Brown, G. (2003). Place bidity? Journal of Gerontology: Social Sciences,
attachment in a revitalizing neighborhood: Individual 66B(1), 7586.
and block levels of analysis. Journal of Environmental Crimmins, E. M., Hayward, M. D., Hagedorn, A., Saito,
Psychology, 23(3), 259271. Y., & Brouard, N. (2009). Change in disability-free
Brown, D. W., Anda, R. F., Tiemeier, H., Felitti, V. J., life expectancy for Americans 70 years old and older.
Edwards, V. J., Croft, J. B., & Giles, W. H. (2009a). Demography, 46(3), 627646.
Adverse childhood experiences and the risk of prema- Daaleman, T. P., & Elder, G. H., Jr. (2007). Family medi-
ture mortality. American Journal of Preventive cine and the life course paradigm. Journal of the
Medicine, 37(5), 389396. American Board of Family Medicine, 20(1), 8592.
Brown, S. C., Mason, C. A., Lombard, J. L., Martinez, F., Davis, M. M., McGonagle, K. A., Schoeni, R. F., &
Plater-Zyberk, E., Spokane, A. R., & Szapocznik, Stafford, F. P. (2008). Grandparental and parental obe-
J. (2009b). The relationship of built environment to sity inuences on childhood overweight: Implications
perceived social support and psychological distress in for primary care practice. Journal of American Board
Hispanic elders: The role of eyes on the street. of Family Medicine, 21(6), 549554.
Journal of Gerontology: Social Sciences, 64B(2), DiPrete, T. A., & Eirich, G. M. (2006). Cumulative advan-
234246. tage as a mechanism for inequality: A review of theo-
Brown, D. W., Anda, R. F., Felitti, V. J., Edwards, V. J., retical and empirical developments. Annual Review of
Malarcher, A. M., Croft, J. B., & Giles, W. H. (2010). Sociology, 32, 271297.
Adverse childhood experiences are associated with the Edwards, V. J., Holden, G. W., Felitti, V. J., & Anda, R. F.
risk of lung cancer: A prospective cohort study. BMC (2003). Relationship between multiple forms of child-
Public Health, 10, 20. hood maltreatment and adult mental health in commu-
Browning, C. R., Cagney, K. A., & Iveniuk, J. (2012). nity respondents: Results from the adverse childhood
Neighborhood stressors and cardiovascular health: experiences study. American Journal of Psychiatry,
Crime and C-reactive protein in Dallas, USA. Social 160(8), 14531460.
Science & Medicine, 75(7), 12711279. Elder, G. H., Jr. (1974). Children of the great depression:
Buettner, D. (2012). The blue zones: 9 lessons for living Social change in life experience. Chicago: University
longer from the people whove lived the longest. of Chicago Press.
Washington, DC: National Geographic Books. Eriksson, J. G., Thornburg, K. L., Osmond, C., Kajantie,
Cain, L. D., Jr. (1964). Life course and social structure. In E., & Barker, D. J. P. (2010). The prenatal origins of
R. E. L. Faris (Ed.), Handbook of modern sociology lung cancer. I. The fetus. American Journal of Human
(pp. 272309). Chicago: Rand McNally. Biology, 22(5), 508511.
Carp, F. M. (1967). The impact of environment on old Felitti, V. J. (2002). The relationship between adverse
people. The Gerontologist, 7(2 Part 1), 106108. childhood experiences and adult health: Turning gold
Carvalho, L. (2012). Childhood circumstances and the into lead. The Permanente Journal, 6, 4447.
intergenerational transmission of socioeconomic sta- Felitti, V. J., Anda, R. F., Nordenberg, D., Williamson,
tus. Demography, 49(3), 913938. D. F., Spitz, A. M., Edwards, V., Koss, M. P., & Marks,
Case, A., & Paxson, C. (2011). The long reach of child- J. S. (1998). Relationship of childhood abuse and
hood health and circumstance: Evidence from the household dysfunction to many of the leading causes
Whitehall II Study. The Economic Journal, 121(554), of death in adults: The adverse childhood experiences
F183F204. (ACE) study. American Journal of Preventive
Cho, J., Martin, P., Margrett, J., MacDonald, M., Poon, Medicine, 14(4), 245258.
L. W., & Johnson, M. A. (2012). Cohort comparisons Ferraro, K. F. (2011). Health and aging: Early origins,
in resources and functioning among centenarians: persistent inequalities? In R. A. Settersten & J. L.
Findings from the Georgia Centenarian Study. Angel (Eds.), Handbook of sociology of aging
International Journal of Behavioral Development, (pp. 465475). Dordrecht: Springer.
36(4), 271278. Ferraro, K. F., & Kelley-Moore, J. A. (2003a). A half cen-
Clarke, P., Ailshire, J. A., Bader, M., Morenoff, J. D., & tury of longitudinal methods in social gerontology:
House, J. S. (2008). Mobility disability and the urban Evidence of change in the Journal. Journal of
built environment. American Journal of Epidemiology, Gerontology: Social Sciences, 58B(5), S264S270.
168(5), 506513. Ferraro, K. F., & Kelley-Moore, J. A. (2003b). Cumulative
Cohen, S., Janicki-Deverts, D., Chen, E., & Matthews, disadvantage and health: Long-term consequences of
K. A. (2010). Childhood socioeconomic status and
404 K.F. Ferraro

obesity? American Sociological Review, 68(5), Hayslip, B., Blumenthal, H., & Garner, A. (2015). Social
707729. support and grandparent caregiver health: One-year
Ferraro, K. F., & Kim, S. (2014). Health benets of reli- longitudinal ndings for grandparents raising their
gion among Black and White older adults? Race, reli- grandchildren. Journal of Gerontology: Social
giosity, and C-reactive protein. Social Science & Sciences, 70B, in press. doi:10.1093/geronb/gbu165
Medicine, 104, 9299. Henkel, L. A. (2014). Do older adults change their eyewit-
Ferraro, K. F., & Shippee, T. P. (2009). Aging and cumula- ness reports when re-questioned? The Journal of
tive inequality: How does inequality get under the Gerontology: Psychological Sciences, 69(3),
skin? The Gerontologist, 49(3), 333343. 356365.
Ferraro, K. F., Shippee, T. P., & Schafer, M. H. (2009). Herd, P., Karraker, A., & Friedman, E. (2012). The social
Cumulative inequality theory for research on aging patterns of a biological risk factor for disease: Race,
and the life course. In V. L. Bengtson, D. Gans, N. M. gender, socioeconomic position, and C-reactive pro-
Putney, & M. Silverstein (Eds.), Handbook of theories tein. Journal of Gerontology: Social Sciences, 67B(4),
of aging (2nd ed., pp. 413433). New York: Springer. 503513.
Fingerman, K. L., Pillemer, K. A., Silverstein, M., & Hill, T. D., Rote, S. M., Ellison, C. G., & Burdette, A. M.
Suitor, J. J. (2012). The baby boomers intergenera- (2014). Religious attendance and biological function-
tional relationships. The Gerontologist, 52(2), ing: A multiple specication approach. Journal of
199209. Aging and Health, 26(5), 766785.
Freedman, V. A., Crimmins, E., Schoeni, R. F., Spillman, Hochschild, A. R. (1978). The unexpected community:
B. C., Aykan, H., Kramarow, E., Land, K., Lubitz, J., Portrait of an old age subculture. Berkeley: University
Manton, K., Martin, L. G., Shinberg, D., & Waidmann, of California.
T. (2004). Resolving inconsistencies in trends in old- Holmes, L. M., & Marcelli, E. A. (2012). Neighborhoods
age disability: Report from a technical working group. and systemic inammation: High CRP among legal
Demography, 41(3), 417441. and unauthorized Brazilian migrants. Health & Place,
Freedman, V. A., Spillman, B. C., Andreski, P. M., 18(3), 683693.
Cornman, J. C., Crimmins, E. M., Kramarow, E., House, J. S., Kessler, R. C., & Herzog, A. R. (1990). Age,
Lubitz, J., Martin, L. G., Merkin, S. S., Schoeni, R. F., socioeconomic status, and health. The Milbank
Seeman, T. E., & Waidmann, T. A. (2013). Trends in Quarterly, 68(3), 383411.
late-life activity limitations in the United States: An Jger, M. M. (2012). The extended family and childrens
update from ve national surveys. Demography, 50, educational success. American Sociological Review,
661671. 77(6), 903922.
Friedman, E. M., Karlamangla, A. S., Almeida, D. M., & Jrvelin, M. R., Sovio, U., King, V., Lauren, L., Xu, B.,
Seeman, T. E. (2012). Social strain and cortisol regula- McCarthy, M. I., Hartikainen, A. L., Laitinen, J.,
tion in midlife in the US. Social Science & Medicine, Zitting, P., Rantakallio, P., & Elliott, P. (2004). Early
74(4), 607615. life factors and blood pressure at age 31 years in the
Fries, J. F. (1980). Aging, natural death, and the compres- 1966 northern Finland birth cohort. Hypertension,
sion of morbidity. New England Journal of Medicine, 44(6), 838846.
303(3), 130135. Johnson, W., & Krueger, R. F. (2005a). Higher perceived
Fries, J. F., Bruce, B., & Chakravarty, E. (2011). life control decreases genetic variance in physical
Compression of morbidity 19802011: A focused health: Evidence from a national twin study. Journal
review of paradigms and progress. Journal of Aging of Personality and Social Psychology, 88(1),
Research, 261702, 110. 165173.
Fuller-Thomson, E., & Brennenstuhl, S. (2009). Making a Johnson, W., & Krueger, R. F. (2005b). Predictors of
link between childhood physical abuse and cancer. physical health: Toward an integrated model of genetic
Cancer, 15(14), 33413350. and environmental antecedents. Journal of
Geronimus, A. T., Hicken, M. T., Pearson, J. A., Seashols, Gerontology: Social Sciences, 60B(Special Issue 1),
S. J., Brown, K. L., & Cruz, T. D. (2010). Do US 4252.
Black women experience stress-related accelerated Klinenberg, E. (2002). Heat wave: A social autopsy of
biological aging? A novel theory and rst population- disaster in Chicago. Chicago: University of Chicago
based test of Black-White differences in telomere Press.
length. Human Nature, 21(1), 1938. Korpimki, S. K., Sumanen, M. P. T., Sillanmaki, L. H., &
Getzmann, S., Falkenstein, M., & Wascher, E. (2015). Mattila, K. J. (2010). Cancer in working-age is not
ERP correlates of auditory goal-directed behavior of associated with childhood adversities. Acta
younger and older adults in a dynamic speech per- Oncologica, 49(4), 436440.
ception task. Behavioural Brain Research, 278, Kuh, D., & Ben-Shlomo, Y. (1997). Introduction: A life
435445. course approach to the aetiology of adult chronic dis-
Goosby, B. J. (2013). Early life course pathways of adult ease. In D. Kuh & Y. Ben-Shlomo (Eds.), A life course
depression and chronic pain. Journal of Health and approach to chronic disease epidemiology (pp. 314).
Social Behavior, 54(1), 7591. New York: Oxford University Press.
Life Course Lens on Aging and Health 405

Latham, K. (2014). The long arm of childhood health: adulthood? Journal of Aging and Health, 24(6),
Linking childhood disability to late midlife mental 948984.
health. Research on Aging, 37(1), 82102. Morton, P. M., Mustillo, S. A., & Ferraro, K. F. (2014).
Lee, M. A., & Ferraro, K. F. (2007). Neighborhood resi- Does childhood misfortune raise the risk of acute
dential segregation and physical health among myocardial infarction in adulthood? Social Science &
Hispanic Americans: Good, bad, or benign? Journal of Medicine, 104, 133141.
Health and Social Behavior, 48(2), 131148. Nock, M. K., & Kessler, R. C. (2006). Prevalence of and
Link, B. G., & Phelan, J. (1995). Social conditions as fun- risk factors for suicide attempts versus suicide ges-
damental causes of disease. Journal of Health and tures: Analysis of the National Comorbidity Survey.
Social Behavior, 35(special issue), 8094. Journal of Abnormal Psychology, 115(3), 616623.
Link, B. G., Phelan, J. C., Miech, R., & Westin, E. L. Ogden, C. L., Carroll, M. D., Curtin, L. R., McDowell,
(2008). The resources that matter: Fundamental social M. A., Tabak, C. J., & Flegal, K. M. (2006). Prevalence
causes of health disparities and the challenge of intel- of overweight and obesity in the United States, 1999
ligence. Journal of Health and Social Behavior, 49(1), 2004. JAMA, 295(13), 15491555.
7291. Ogden, C. L., Carroll, M. D., Kit, B. K., & Flegal, K. M.
Lynch, S. M. (2003). Cohort and life course patterns in the (2014). Prevalence of childhood and adult obesity in
relationship between education and health. the United States, 20112012. Journal of the American
Demography, 40(2), 309331. Medical Association, 311(8), 806814.
Lynch, J. W., Kaplan, G. A., & Shema, S. J. (1997). ORand, A. M., & Hamil-Luker, J. (2005). Processes of
Cumulative impact of sustained economic hardship on cumulative adversity: Childhood disadvantage and
physical, cognitive, psychological, and social func- increased risk of heart attack across the life course.
tioning. New England Journal of Medicine, 337(26), Journal of Gerontology: Social Sciences, 60B(Special
18891895. issue 2), S117S124.
Manton, K. G. (2008). Recent declines in chronic disabil- Pearlin, L. I. (1989). The sociological study of stress.
ity in the elderly U.S. population: Risk factors and Journal of Health and Social Behavior, 30(3),
future dynamics. Annual Review of Public Health, 29, 241256.
91113. Pearlin, L. I. (2010). The life course and the stress pro-
Manton, K. G., Corder, L. S., & Stallard, E. (1993). cess: Some conceptual comparisons. Journal of
Estimates of change in chronic disability and institu- Gerontology: Social Sciences, 65B(2), S207S215.
tional incidence and prevalence rates in the U.S. Perls, T. T., & Silver, M. H. (1999). Living to 100.
elderly population from 1982, 1984, and 1989 New York: Basic Books.
National Long Term Care Survey. Journal of Pudrovska, T., & Anikputa, B. (2014). Early-life socio-
Gerontology: Social Sciences, 48(4), S153S166. economic status and mortality in later life: An integra-
Manton, K. G., Gu, X., & Lowrimore, G. R. (2008). tion of four life-course mechanisms. The Journals of
Cohort changes in active life expectancy in the U.S. Gerontology Series B: Psychological Sciences and
elderly population: Experience from the 19822004 Social Sciences, 69(3), 451460.
National Long-Term Care Survey. Journal of Quesnel-Valle, A., & Taylor, M. (2012). Socioeconomic
Gerontology: Social Sciences, 63(5), S269S281. pathways to depressive symptoms in adulthood:
Marmot, M. (2005). Social determinants of health Evidence from the National Longitudinal Survey of
inequalities. The Lancet, 365(9464), 10991104. youth 1979. Social Science & Medicine, 74(5),
Marmot, M. G., Smith, G. D., Stansfeld, S., Patel, C., 734743.
North, F., Head, J., White, I., Brunner, E., & Feeney, A. Repetti, R. L., Taylor, S. E., & Seeman, T. E. (2002).
(1991). Health inequalities among British civil ser- Risky families: Family social environments and the
vants: The Whitehall II study. The Lancet, 337(8754), mental and physical health of offspring. Psychological
13871393. Bulletin, 128(2), 330366.
Marshall, V. W. (1975). Socialization for impending death Ridker, P. M., Buring, J. E., Shih, J., Matias, M., &
in a retirement village. American Journal of Sociology, Hennekens, C. H. (1998). Prospective study of
80(5), 11241144. C-reactive protein and the risk of future cardiovascular
Martin, L. G., Freedman, V. A., Schoeni, R. F., & Andreski, events among apparently healthy women. Circulation,
P. M. (2009). Health and functioning among baby 98(8), 731733.
boomers approaching 60. Journal of Gerontology: Schafer, M. H., & Ferraro, K. F. (2012). Childhood mis-
Social Sciences, 64B(3), 369377. fortune as a threat to successful aging: Avoiding dis-
McDade, T. W., Metzger, M. W., Chyu, L., Duncan, G. J., ease. The Gerontologist, 52(1), 111120.
Gareld, C., & Adam, E. K. (2014). Long-term effects Schafer, M. H., Ferraro, K. F., & Mustillo, S. A. (2011).
of birth weight and breastfeeding duration on inam- Children of misfortune: Early adversity and cumula-
mation in early adulthood. Proceedings of the Royal tive inequality in perceived life trajectories. American
Society B: Biological Sciences, 281(1784), 19. Journal of Sociology, 116(4), 10531091.
Morton, P. M., Schafer, M. H., & Ferraro, K. F. (2012). Schafer, M. H., Wilkinson, L. R., & Ferraro, K. F. (2013).
Does childhood misfortune increase cancer risk in Childhood (mis) fortune, educational attainment, and
406 K.F. Ferraro

adult health: Contingent benets of a college degree? course model. Social Science & Medicine, 69(1),
Social Forces, 91(3), 10071034. 138146.
Schryer, E., & Ross, M. (2014). Does the age-related posi- Stathakos, D., Pratsinis, H., Zachos, I., Vlahaki, I.,
tivity effect in autobiographical recall reect differ- Gianakopoulou, A., Zianni, D., & Kletsas, D. (2005).
ences in appraisal or memory? Journal of Gerontology: Greek centenarians: Assessment of functional health
Psychological Sciences, 69(4), 548556. status and life-style characteristics. Experimental
Sebastiani, P., & Perls, T. T. (2012). The genetics of Gerontology, 40(6), 512518.
extreme longevity: Lessons from the New England Stein, D. J., Scott, K., Abad, J. M. H., Aguilar-Gaxiola, S.,
centenarian study. Frontiers in Genetics, 3(277), 17. Alonso, J., Angermeyer, M., Von Korff, M. (2010).
Seeman, T. E., Singer, B. H., Ryff, C. D., Love, G. D., & Early childhood adversity and later hypertension: Data
Levy-Storms, L. (2002). Social relationships, gender, from the World Mental Health Survey. Annals of
and allostatic load across two age cohorts. Clinical Psychiatry, 22(1), 1928.
Psychosomatic Medicine, 64(3), 395406. Verbrugge, L. M. (1986). From sneezes to adieux: Stages
Seplaki, C. L., Goldman, N., Weinstein, M., & Lin, Y.-H. of health for American men and women. Social
(2004). How are biomarkers related to physical and Science and Medicine, 22(11), 11951212.
mental well-being? Journal of Gerontology: Biological Verbrugge, L. M., & Jette, A. M. (1994). The disablement
Sciences, 59A(3), 201217. process. Social Science and Medicine, 38(1), 114.
Sesso, H. D., Buring, J. E., Rifai, N., Blake, G. J., Gaziano, Vuolo, M., Ferraro, K. F., Morton, P. M., & Yang, T.-Y.
J. M., & Ridker, P. M. (2003). C-reactive protein and (2014). Why do older people change their ratings of
the risk of developing hypertension. Journal of the childhood health? Demography, 51(6), 19992023.
American Medical Association, 290(22), 29452951. Wadsworth, M., Kuh, D., Richards, M., & Hardy, R.
Shanahan, M. J., & Hofer, S. M. (2005). Social context in (2006). Cohort prole: The 1946 National Birth
geneenvironment interactions: Retrospect and pros- Cohort (MRC National Survey of Health and
pect. Journal of Gerontology: Social Sciences, Development). International Journal of Epidemiology,
60(Special Issue 1), 6576. 35(1), 4954.
Shanahan, M. J., & Hofer, S. M. (2011). Molecular genet- Wheaton, B., & Clarke, P. (2003). Space meets time:
ics, aging, and well-being: Sensitive period, accumu- Integrating temporal and contextual inuences on
lation, and pathway models. In R. H. Binstock & L. K. mental health in early adulthood. American
George (Eds.), Handbook of aging and the social sci- Sociological Review, 68(5), 680706.
ences (7th ed., pp. 135147). San Diego: Elsevier. Wickrama, K. A., Conger, R. D., Wallace, L. E., & Elder,
Shanahan, M. J., Vaisey, S., Erickson, L. D., & Smolen, A. G. H., Jr. (1999). The intergenerational transmission
(2008). Environmental contingencies and genetic pro- of health-risk behaviors: Adolescent lifestyles and
pensities: Social capital, educational continuation, and gender moderating effects. Journal of Health and
dopamine receptor gene DRD21. American Journal of Social Behavior, 40(3), 58272.
Sociology, 114(S1), S260S286. Wickrama, K. A. S., Conger, R. D., & Abraham, W. T.
Shippee, T. P., Wilkinson, L. R., & Ferraro, K. F. (2012). (2005). Early adversity and later health: The intergen-
Accumulated nancial strain and womens health over erational transmission of adversity through mental dis-
three decades. Journal of Gerontology: Social order and physical illness. Journal of Gerontology:
Sciences, 67(5), 585594. Social Sciences, 60(Special Issue 2), S125S129.
Shuey, K. M., & Willson, A. E. (2014). Economic hard- Willcox, D. C., Willcox, B. J., & Poon, L. W. (2010).
ship in childhood and adult health trajectories: An Centenarian studies: Important contributors to our
alternative approach to investigating life-course pro- understanding of the aging process and longevity.
cesses. Advances in Life Course Research, 22, 4961. Current Gerontology and Geriatrics Research,
Singer, J. D., & Willett, J. B. (2003). Applied longitudinal 484529, 16.
analysis: Modeling change and event occurrence. Williamson, D. F., Thompson, T. J., Anda, R. F., Dietz,
New York: Oxford University Press. W. H., & Felitti, V. (2002). Body weight and obesity in
Snelgrove, J. W., Pikhart, H., & Stafford, M. (2009). A adults and self-reported abuse in childhood.
multilevel analysis of social capital and self-rated International Journal of Obesity and Related
health: Evidence from the British Household Panel Metabolic Disorders, 26(8), 10751082.
survey. Social Science & Medicine, 68(11), Wink, P., Ciciolla, L., Dillon, M., & Tracy, A. (2007).
19932001. Religiousness, spiritual seeking, and personality:
Springer, K. W. (2009). Childhood physical abuse and Findings from a longitudinal study. Journal of
midlife physical health: Testing multi-pathway life Personality, 75(5), 10511070.
Mental Health

William R. Avison

A fundamental premise of the sociology of men- relatively recent with the vast majority of work in
tal health is that the social distribution of psycho- this area having been produced over the past 15
logical distress and mental disorder is not years.
uniform that there is a social gradient of mental In this chapter, I rst present a selective review
health and disorder. Over the past six decades, a of the signicant scientic advances that have
substantial body of research has documented dif- emerged from the synthesis of the life course per-
ferences in the prevalence of mental health prob- spective with the sociology of mental health. In
lems by social characteristics such as gender, particular, I focus on Leonard Pearlins well-
education, race/ethnicity, income, socio- known synthesis of the stress process with the
economic status, employment status, marital sta- life course perspective. I also examine a number
tus, family structure, and other social statuses and of important currents in life course research in
roles. This body of scientic knowledge has been mental health.
exceptionally well-characterized in a recent com- For the past decade, life course research in
pendium (Ansehensel et al. 2013). mental health has embraced a number of concep-
One of the most important developments in tual and methodological approaches that have
the sociology of mental health has been its syn- emphasized trajectories of mental health over the
thesis with the life course perspective. Linda life course. In the second section of this chapter,
George (2013), one of the foremost proponents I will raise some challenges that confront this line
of life course research in mental health and ill- of investigation and suggest some new directions
ness, has pointed out that this synthesis has been that might stimulate fresh ways of conducting
such research.
The third broad topic to be considered in this
chapter concerns the intersection of physiologi-
W.R. Avison, Ph.D., FCAHS (*) cal processes with socio-environmental inu-
Departments of Sociology, Paediatrics, and ences on mental illness. In recent years, the
Epidemiology & Biostatistics, The University of
developmental origins of health and disease
Western Ontario, London, ON, Canada
(DOHaD) has become a focus for many research-
Division of Childrens Health and Therapeutics,
ers interested in the ways in which fetal and early
Childrens Health Research Institute, London, ON,
Canada postnatal health may be associated with adult
health outcomes, including mental health. Other
Lawson Health Research Institute, London, ON,
Canada researchers have focused on the role of social
e-mail: wavison@uwo.ca experiences in childhood in stimulating

Springer International Publishing Switzerland 2016 407


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_19
408 W.R. Avison

pro-inammatory responses in individuals that Pearlin notes, this idea is consistent with both
ultimately manifest themselves in elevated risk Durkheims view of suicide as a nature conse-
for symptoms of mental health problems. Still quence of social attachments and Mertons idea
others researchers have considered the roles of that anomie is a normative result of the disjunc-
gene-environment interactions and correlations tion between aspirations and access to opportuni-
and the ways in which gene-environment pro- ties. It is also consistent with Selyes (1956)
cesses may play out across the life course. All of classic statement about stress as a normal expe-
these models are essentially life course models, rience. Third, the origins of stress are essentially
but the sociology of the life course and the sociol- social in nature: stressors emerge out of social
ogy of mental health have largely been absent experience. This directs sociological stress
from these considerations. researchers toward more proximal than distal
sources of social stress and to a greater emphasis
on social context than on history or biology. As
1 The Stress Process in Life we shall see, recent developments that have
Course Context incorporated life course principles as well as the
consideration of biological and physiological
Over three decades ago, Leonard Pearlin and his mechanisms have drawn some stress researchers
colleagues (1981) described a conceptual model attention to new directions in stress research that
that has continued to stimulate research in the consider more distal processes.
sociology of mental health. The introductory Pearlins model of the stress process contains
paragraph of this article is a succinct summary of four major components. Sources of stress include
the stress process: stressful life events as well as other dimensions
The process of social stress can be seen as combin- such as role-related strains, daily hassles, and life
ing three major conceptual domains: the sources of traumas. These sources of stress are potentially
stress, the mediators of stress, and the manifesta- interactive in their effects on health outcomes;
tions of stress. Each of these extended domains that is, life events can intensify or exacerbate
subsumes a variety of subparts that have been
intensively studied in recent years. Thus, in the existing strains and vice versa. In addition, stress-
search for sources of stress, considerable interest ful life experiences can create new role strains
has been directed to life events and to chronic life and, conversely, role strains can generate new
strains, especially the former; in work concerned stressful life events.
with conditions capable of mediating the impact of
stressful circumstances, coping and social supports Manifestations of stress include an array of
have had a rather dramatic rise to prominence; and possible health outcomes. Sociologists have
as for stress and its symptomatic manifestations, tended to focus on symptoms of mental illness or
the expanding volume of research ranges from the measures of psychological distress. These mea-
microbiological substrates of stress to its overt
emotional and behavioural expressions (p. 337). sures typically include symptoms of depressive
illness; however, some sociologists have exam-
In his reconsiderations of the stress process, ined the effects of stressors on diagnosable disor-
Pearlin (1989, 1999) further elaborated the para- ders (Avison 2001; Brown and Harris 1978;
digm to take into account the social context in Turner et al. 1995) and others have extended the
which the stress process occurs and to add other study of stress to include alcohol consumption
features into the model. According to Pearlin, the and drug dependence (Aneshensel et al. 1991;
foundation of the stress process rests on three key Turner et al. 2006a), and physical illness (Brown
assumptions. First, the process is a dense causal and Harris 1989).
web that involves dynamic interconnections The third component of the stress process
among the components of the model. Changes in model, mediators of stress, refers to a broad range
one set of factors produce changes in others. of factors including social support, psychosocial
Second, social stress is a typical experience of resources such as mastery, self-esteem, matter-
ordinary life; it is not unusual or abnormal. As ing, interpersonal dependency, and coping
Mental Health 409

strategies. These mediators are hypothesized to ideas from the stress process paradigm was also
function as pathways that connect exposure to recognized by Elder et al. (1996). They system-
stress to its manifestations. So, for example, indi- atically explore these possibilities and lay out a
viduals exposed to stress experience the erosion number of possibilities for life course research in
of their sense of control over their lives. In turn, the stress process that complement the position
this decline in mastery manifests itself in symp- advocated by Pearlin.
toms of distress or depression. These ideas have been elaborated further.
These mediating factors may also operate as Pearlin et al. (2005) specify elements of the stress
moderators of the stress-distress relationship. process that may affect stress and health across
The classic examples are those in which the the life course. These include the effects of eco-
impact of stress on mental health outcomes is nomic strains and discriminatory experiences,
reduced in the presence of higher levels of psy- stress proliferation, and the intersection of status
chosocial resources, social support, or coping attainment and stress exposure. This synthesis of
resources. These moderating effects, often the stress process with the life course has been
referred to as stress-buffering effects, typically stimulating to research in the sociology of mental
reect processes in which the consequences of health (Wheaton 2010).
stress are mitigated by individuals abilities to There is now a growing body of theoretical
cope or otherwise deal with stress (Avison and ideas on stress and the life course (George 2007,
Cairney 2003; Thoits 2010; Turner and Turner 2013; McLeod and Pavalko 2008) and a tremen-
2013; Wheaton 1985). dous increase in the number of empirical investi-
The interplay among sources of stress, media- gations of these ideas. Recently, Turner and
tors and moderators of stress, and outcomes all Schieman (2008) assembled a wide-ranging set
occurs in a social context that is dened by the of papers that explores the interface of the stress
social and economic statuses and roles that indi- process with the life course. The theoretical rich-
viduals occupy. ness and empirical rigor of this work make it
Another important elaboration emerged when clear that sociologists of mental health have
Pearlin and Skaff (1996) suggested a number of embarked on some very ambitious programs of
ways in which principles central to the life course research to examine how the stress process
perspective could be integrated with key ele- unfolds across the life course.
ments of the stress process to examine how indi- To date, much of the research in the sociology
viduals exposure to stressors changes as people of mental health that incorporates a life course
move through the life course, and their lives are perspective has focused on turning points or tran-
restructured. As their statuses and roles change, sitions in individuals lives in adulthood. Life
so too do the stressors they encounter and the course research on stress and mental health has
mediating resources to which they have access. provided innumerable insights into issues such as
So, for example, the stressors that young people transitions from adolescence to adulthood
experience in school give way to stressors associ- (Conger et al. 1992; Gore and Aseltine 2003;
ated with marriage, parenting, and their work Wickrama et al. 2008a), family transitions and
roles as they move through the life course. These, family structure (Avison 2010; Avison and
in turn, may be replaced by the stress of retire- Davies 2005; Avison et al. 2008), the intersection
ment or of caregiving for a spouse with a chronic of work and family (Menaghan 1997; Menaghan
illness. As Pearlin (1983) points out, the different and Parcel 1995), and aging and mental health
roles that we occupy expose us to different role (Schieman et al. 2001).
tasks, to the possibility of interpersonal conicts, There have also been signicant sociological
role conicts, role captivity, and role studies of stress and mental health in childhood
restructuring. (see Menaghan 2009, for a review). These include
The possibility that principles of the life studies of trajectories of poverty and childrens
course perspective could be integrated with core mental health (McLeod and Edwards 1995;
410 W.R. Avison

McLeod and Shanahan 1993, 1996; Reiss 2013; as child abuse or neglect (Horwitz et al. 2001;
Strohschein 2005); Menaghan and Parcels Kessler and Magee 1994), childhood sexual
(1994, 1995) research on parental occupational assault (Wineld et al. 1990), and witnessing vio-
circumstances and childrens mental health; lence as a child (Kessler and Magee 1993) have
Avison and McAlpines (1992) exploration of also been documented.
gender, stress, and mental health in adolescence; Turner and Lloyd (1995) and Turner and
Umberson et al. (2005) work on stress in child- Turner (2005) have described how traumatic
hood and adulthood; and the body of research on experiences and adversities in childhood and
the effects of neighborhoods on childrens and adolescence, together with stressful experiences
adolescents mental health (Aneshensel and in adulthood, constitute a cumulative burden of
Sucoff 1996; Wheaton and Clarke 2003). These adversity that exerts powerful inuences on indi-
investigations conrm that childrens stressful viduals mental health. It seems clear that the
experiences arise out of the social context of experiences of childhood, adolescence, and early
everyday life. adulthood are formative for adult mental health.
There is substantial evidence that these pat-
terns are observable across different societies. In
1.1 Research Dividends their analyses of the 1946 British birth cohort in
from the Synthesis the National Survey of Health and Development,
of the Stress Process with Life Colman et al. (2014) document how adversities
Course Principles as early as those originating in the rst years of
life are associated with symptoms of depression
There have been several clear advances in in adolescence and over 50 years later in mid-
research on mental health that are direct conse- adulthood. In the same cohort, Kuh et al. (2002)
quences of Pearlins synthesis of the stress pro- nd that women in midlife with trajectories of
cess with ideas from the life course perspective. increasing symptoms of mental health problems
Although space does not permit a detailed dis- were more likely to have a history of adversities
cussion of these dividends, they deserve a brief and mental health problems across their earlier
comment. life courses. Slopan et al. (2010) analyzed data
Foremost among these advances has been the from the South Africa Stress and Health Study
realization that the long arm of childhood and report that childhood adversities were sig-
(Hayward and Gorman 2004; see also Ferraro, nicantly associated with the onset of anxiety
chapter Life Course Lens on Aging and Health disorders but not mood disorders. Lee et al.
this volume; Hayward and Sheehan, chapter (2011) also nd that family childhood adversities
Does the Body Forget? Adult Health, Life are signicantly related to the onset of an array of
Course Dynamics, and Social Change this vol- psychiatric disorders among a sample of over
ume) extends its reach to mental health. It is now 5,000 adults living in Beijing or Shanghai.
clear that adversities and traumas in childhood Hatch and Dohrenwend (2007) report that the
have consequences for mental health in adult- prevalence of many of these adversities declines
hood. Adversities such as parental divorce (Harris with age. This suggests that a comprehensive
et al. 1986, 1990; McLeod 1991; Ross and understanding of mental health in adulthood will
Mirowsky 1999), growing up in a single-parent benet signicantly from stress process analyses
family (Barrett and Turner 2005), and experienc- that take a life course perspective by incorporat-
ing poverty in childhood (Gibb et al. 2012; Luo ing experiences from childhood and
and Waite 2005; Wheaton and Clarke 2003) are adolescence.
all signicantly associated with the risk of a vari- From a life course perspective, these various
ety of mental disorders as well as elevated symp- patterns may be studied as pathways or trajecto-
toms and psychological distress. The effects on ries that link early adversities to subsequent men-
adult mental health of traumatic experiences such tal health problems in adulthood. For example, in
Mental Health 411

a study of single and married mothers, Davies process framework to examine how the experi-
et al. (1997) have argued that childhood adversi- ence of physical disabilities (as manifest in activ-
ties contribute to early onset of depression that is ity limitations and pain) exerted a substantial
associated with a higher risk of separation and impact of individuals levels of psychological
divorce among these women. In turn, single distress and that constructs central to the stress
mothers greater exposure to stressors in their process mastery, social support, and stressful
lives increased their risk of recurrence of major life events played important roles in predicting
depressive disorder. levels of depression among individuals with
For life course researchers, the challenge is to physical disabilities. The importance of both of
identify the mediating processes that link early these studies for life course research relates to the
experiences with later mental health. Some well-known association of age with physical ill-
researchers conclude that children who have lost ness and disability. Subsequent research (Bierman
a parent experience subsequent difculties in and Pearlin 2011; Meeks et al. 2000; Miech and
developing supportive, close attachments with Shanahan 2000; Turner et al. 2006b; Yang 2004;
their own spouses (McLeod 1991). Others argue Yang and George 2005) has explored the rela-
that parental divorce impedes educational attain- tionship between physical disability and symp-
ment which in turn contributes to economic hard- toms of depression across different age groups.
ship, unhappy personal relationships, and feelings A central idea in the life course perspective is
of mistrust. All of these contribute to elevated that individuals are socially imbedded in a web of
levels of psychological distress. relationships. This notion of linked lives presup-
A related theme concerns the impact of early poses that any individuals life experiences are
mental illness on subsequent mental health. We importantly inuenced by the social networks in
have known for some time that most mental dis- which they participate. For the stress process para-
orders have a relatively early age of onset (Kessler digm, this has been most clearly articulated by
et al. 2005) and that early onset is associated with Aneshensel and Pearlins program of research on
elevated risk of disorder in adulthood (McGue the caregiving role. In a series of important papers
and Iacono 2005; Rutter et al. 2006). Moreover, (Aneshensel et al. 1993, 2005; Pearlin et al. 1997)
early mental illness also has negative conse- and a research monograph (Aneshensel et al.
quences for educational attainment (Chen and 1995), they described how the unexpected role of
Kaplan 2003; McLeod and Fettes 2007) and early caregiver exposed individuals to elevated levels of
age of marriage and parenthood (Forthofer et al. stressors and how this exposure was patterned by
1996; Wade and Pevalin 2004). Thus, from a life the life course. This research provided a model for
course perspective, experiences of adversity or incorporating the concept of linked lives in stress
trauma in childhood or the early onset of mental process research on mental health.
health problems may set in motion a series of
socio-economic, intrapsychic, and social psycho-
logical processes that inuence subsequent men- 2 Trajectories of Mental Health
tal health in adulthood. and Illness
A third dividend that has emerged from the
synthesis of the stress process with life course Leonard Pearlins synthesis of the life course per-
principles concerns the relationship between spective with the stress process paradigm stands
physical health and mental health. Two indepen- as a signicant turning point in the sociology of
dent studies set the groundwork for research in mental health. His thinking fundamentally altered
this area. In a four-wave study over the course of the way that stress process researchers think
one year, Aneshensel et al. (1984) documented about mental health over the life course. One
the reciprocal inuences of physical illness and consequence of this has been the burgeoning
depressive symptoms on one another. In a related interest in both role trajectories as well as trajec-
vein, Turner and Noh (1988) employed a stress tories of mental health over the life course.
412 W.R. Avison

Macmillan and Copher (2005: 859) argue that More recent life course studies of mental
role trajectories take place over an extended health have constructed trajectories of mental
period of time and index temporal involvement in health typically, measures of depressive symp-
major institutions through schooling, paid toms or psychological distress across multiple
employment, marriage, and parenthood. time points and have identied social condi-
Trajectories are marked at the beginning and end tions and processes that are associated with dif-
by transitions. Sociologists of mental health ferent trajectories. For example, Wickrama et al.
have been interested in role trajectories, but they (2008b) document how childhood adversities
have also examined trajectories of mental health. associated with family socio-economic disadvan-
These trajectories refer either to temporal pat- tage are associated with higher initial levels of
terns of diagnosed disorder or, more commonly, depression in adolescence. Early transitions into
to levels of psychological distress or symptoms adult roles further increase levels of depressive
of mental health problems over time. symptoms in adulthood among these individuals.
In the area of stress and mental health, one of Strohschein (2005) reports that household
the rst comprehensive considerations of trajec- income is associated with trajectories of chil-
tories can be found in Gotlib and Wheatons drens mental health problems.
(1997) edited collection of studies on the experi- Spence et al. (2011) examine racial disparities
ence of stress over the life course and their effects in trajectories of depression among women aged
on mental health outcomes. They focus on the 5281. Their analyses reveal that African
ways in which trajectories of stressors or other Americans experience higher levels of symptoms
social experiences and turning points in individu- than their White counterparts at all ages; that is,
als lives inuence their mental health outcomes. there is no apparent convergence in African
Wheaton et al. (1997) document how various American and white womens levels of depres-
childhood adversities may accumulate over time sion in older age. This sustained difference over
and how such trajectories contribute to psycho- the life course is attributable to racial differences
logical distress later in the life course. Menaghan in socio-economic status and to increasing rates
(1997) shows how the relatively stable, unchang- of physical health problems among African
ing trajectories of paid work in many families American women in older age groups.
have important consequences for childrens emo- Walsemann et al. (2009) nd that trajectories
tional well-being. Where parents have stressful of depressive symptoms decline from early adult-
work lives for extended periods of time, chil- hood through middle age among Whites, African
drens educational performance and emotional Americans, and Hispanics. The slopes of these
well-being may suffer. These studies built the trajectories differ, however, as a function of dif-
foundation for subsequent life course research in ferential socio-economic disadvantage. By their
the sociology of mental health. mid-30s, Hispanics and Whites have similar lev-
Other researchers have expanded the consid- els of depressive symptoms that are lower on
eration of trajectories to include intrapsychic fac- average than for African Americans.
tors. So, for example, Barrett and White (2002) Johnson et al. (2014) examine how intimate
describe how increasing perceptions or feelings partner violence (IPV) inuences trajectories of
of masculinity over time appear to decrease depressive symptoms. They report that both vic-
depressive symptoms in early adulthood among timization and perpetration are associated with a
both males and females. They argue that these trajectory of increasing symptomatology for both
increasing trajectories of masculinity protect men and women. They also nd that cumulative
both young men and young women from feelings exposure to IPV has little discernible inuence
of depression. In this sense, social psychological on trajectories of depressive symptoms, suggest-
processes are seen to be core elements of a life ing that there is a recency effect in which feelings
course approach to the study of mental health of depression are more likely to reect the recall
(McLeod 2012). of the most recent victimization.
Mental Health 413

Other investigators have concluded from these Education, Health, and Historical Change this
studies that it is reasonable to expect that trajec- volume). Although there are several studies that
tories of social experiences, whether socioeco- follow children for several waves over 1015
nomic factors or various stressors, may inuence years, there are few large scale studies with cred-
trajectories of mental health outcomes. So, for ible measures of mental health that follow chil-
example, in a multi-wave study over 7 years, dren well into adulthood. The most notable
George and Lynch (2003) have documented how exception is the well-known Dunedin
trajectories of increasing exposure to stressors Multidisciplinary Health and Development Study
are associated with trajectories of growth in (commonly known as the Dunedin Longitudinal
depressive symptoms among African American Study) which has followed a cohort of 1,037
respondents but not whites. babies born in New Zealand between April 1,
Avison et al. (2008) report that specic trajec- 1972, and March 31, 1973, over 12 assessments
tories of family structure are associated with tra- spanning 40 years. Most impressively, the study
jectories of increasing depressive symptoms has maintained a retention rate of 96 % over this
among a sample of mothers who were inter- time, thus avoiding many of the typical problems
viewed over a 14-year period. They nd that the associated with differential attrition bias. The
prolonged experience of single motherhood is Dunedin Longitudinal Study includes a rich array
associated with growth in depressive symptoms. of demographic and socio-economic data as well
Re-partnering is related to slightly less pro- as a broad range of measures of physical health
nounced growth in symptoms, but these women and functioning. It has also routinely included a
still have more pronounced trajectories of battery of psychometrically sound measures of
increasing symptoms that do mothers who were mental health.
partnered over the course of the study. Longitudinal studies of mental health like the
These examples of the analyses of trajectories Dunedin Study are rare. The feasibility of such
associated with mental health are illustrative but projects in countries with much larger geographic
by no means exhaustive of the work that has been areas and more heterogeneous populations is an
produced in this area of research. They reveal the open question. Although so-called conception
complexities involved in taking into account to death studies have been discussed in North
change over time, both in terms of the social con- America, the nancial and logistical challenges
ditions of peoples lives and the stressors and to mounting such a massive investigation are
adversities that they encounter, and in terms of substantial.
their expressions of psychological distress and Second, many studies of trajectories of mental
mental health problems. health report that large proportions of the samples
studied display relatively unchanging levels of
mental health problems, psychological distress or
2.1 Cautions About Research depressive symptoms over time (for example,
on Trajectories Avison et al. 2008; Broidy et al. 2003; McLeod
and Fettes 2007; Moftt et al. 1996) and that the
Despite the important advances that have been proportions of samples with stable, high levels of
made in studying mental health across the life problems typically exceed those with increasing
course, there remain several limitations to this or decreasing levels over time. This constitutes a
work that require some fresh thinking to stimu- challenge to researchers because these smaller
late new ways of studying mental health across subsamples threaten the statistical power of these
the life course. Some of these limitations are research designs. Accordingly, it is often difcult
methodological; others are more conceptual. to nd statistically signicant risk and protective
First, the span of most data sets available to factors that differentiate among these trajectories.
study mental health across the life course is rela- A third challenge to life course research is
tively limited (see also Johnson et al., chapter more conceptual in nature. Some researchers
414 W.R. Avison

have argued that social conditions are predictive trace the patterns of adversity and mental health
of certain kinds of subsequent mental health tra- over long periods of time. Nevertheless, she also
jectories. For example, Meadows (2009) has doc- points out that there several quantitative analytic
umented how differences in family structure techniques that are useful in estimating recipro-
among men in the Fragile Families and Child cal pathways among socio-economic factors,
Wellbeing Study appear to have consequences for stressful experiences, and metal health.
their trajectories of mental health. Others have Other researchers have employed natural
argued that trajectories of social experience are experiments to generate causal evidence that
associated with mental health outcomes mea- social experiences affect mental health outcomes
sured subsequently. Thus, McLeod and Shanahan (see also Johnson et al., chapter Education,
(1993) demonstrate that the persistence of pov- Health, and Historical Change this volume). A
erty predicts childrens symptoms of depression notable example of this is Costello and col-
and anxiety over and above the effect of current leagues (Costello et al. 2003) Great Smokey
poverty. Still other investigators have suggested Mountains Study of a sample of American Indian
that trajectories of social experience may inu- and white children. During the course of this
ence trajectories of mental health outcomes. study, a casino was opened on an Indian reserva-
McLeod and Shanahan (1996) extended their tion and every American Indian received an
earlier work to show how childrens experience income supplement with an annual increase.
of persistent poverty is associated with increased Costello et al. report that the income intervention
antisocial behavior over the same period of time. that moved families out of poverty was responsi-
Avison et al. (2008) have documented the asso- ble for a signicant decline in conduct and oppo-
ciation of patterns of stability and change in fam- sitional disorder among children, but no
ily structure over a 14-year period are correlated signicant improvement in anxiety or
with mothers trajectories of psychological depression.
distress. Huang et al. (2013) have described a natural
These examples point to a very complex rela- experiment in China in which exposure to the
tionship between social experience and mental 19591961 Chinese Famine in utero or in the
health over the life course. The challenge for life early post-natal period was examined for its
course researchers is to unravel this complexity. effects on mental health some 40 years later.
Presumably, some of the association between tra- They conclude that such early adversities
jectories that characterize individuals position in increased the risk of mental illness among women
the social structure and trajectories of mental but decreased the risk among men. They attribute
health represents classic social causation pro- this gender difference to selection effects in utero
cesses and some will reect social selection that favor survival only among the most robust
effects. To date, life course investigators have not males.
developed specic conceptual frameworks that These two examples highlight both the poten-
allow us to draw rm conclusions about the rela- tial of natural experiments to test causal hypoth-
tive impact of selection and causation in inuenc- eses as well as the particular challenges to them.
ing these trajectories. Even when it appears that the assignment of
The dense causal webs that connect various subjects to the exposed and non-exposed condi-
social experiences with mental health at various tions is random, certain systematic selection
points in the life course highlight the challenges effects may compromise these natural experi-
of specifying causal pathways in life course ments. Nevertheless, these two studies provide
research on mental health. There have been diver- excellent examples of the utility of such designs
gent responses to this complexity. George (2013, in life course research in mental health.
2014) has argued that the concepts of social Still other researchers have tried to unravel the
selection and social causation are not especially complexity that is inherent in the study of the
meaningful to life course researchers when they pathways among social experiences, stressors,
Mental Health 415

mediators and moderators, and mental health out- psychiatric disorder has a cumulative effect on
comes by employing new analytic techniques. individuals lives. Georges (2013, 2014) think-
An interesting example of this is Longest and ing has also been very inuential in developing a
Thoits (2012) use of a congurational approach life course approach to diagnosed disorder. She
to examine the multiple contingent pathways argues that episodes of mental illness connect
within the stress process. They demonstrate how with individuals trajectories of work and family
Fuzzy Set Qualitative Comparative Analysis experiences. These experiences may be impor-
(fsQCA) is able to identify complex models that tant inuences on patterns of remission and
capture the complicated patterns that are charac- recurrence.
teristic of the interplay of key elements of the
stress process.
There is another issue concerning trajectories 2.2 New Directions in the Study
of mental health outcomes that deserves atten- of Trajectories of Mental
tion. The vast majority of studies on trajectories Health
that has emanated from the sociology of mental
health has employed continuous measures of If we think about the many developments in
psychological distress or indices of symptom- research on the stress process and the progress
atology. Primarily, this has been a methodologi- that has been made in examining trajectories of
cal rather than a substantive decision. In survey mental health, a number of new directions for
research, collection of data sufcient to generate future investigation arise. Most of these reect
diagnoses of specic psychiatric disorder is conceptual developments that can be easily syn-
costly and time consuming for the researcher and thesized with current theory in life course
may be burdensome to respondents. Accordingly, research in the sociology of mental health.
we have few life course studies that consider tra- It seems clear that adolescence and early
jectories of psychiatric illness. adulthood are segments of the life course where
In a 23-year prospective study, Eaton et al. levels of psychological distress or psychiatric
(2008) report that approximately half of individ- symptomatology are especially elevated. There is
uals who experience their rst depressive episode general agreement that childrens experiences
recover and suffer no further depression; another may be important, not only for creating set points
35 % experience at least one recurrent episode, for trajectories of mental health, but also for
and 15 % report unremitting, continuous depres- inuencing the arc of these trajectories. If we
sion. It is noteworthy that these results were accept this, then a life course analysis of mental
based on interviews with 92 of 1071 participants health ought to consider the social and social psy-
in the 1993 Epidemiologic Catchment Area sur- chological conditions of children, the stressful
vey in Baltimore. Thus, less than 10 % of the sur- experiences that arise out of these conditions, and
vey sample had a rst episode. This highlights the processes that mediate and moderate the
the methodological challenge of constructing tra- stress process in childhood (Avison 2010). The
jectories of mental disorder: the incidence of rst relevant social and social psychological condi-
onset of disorder is low and the patterns of recur- tions of childhood include family structure, race
rence seem either to be episodic or constant. and ethnicity, and parental socio-economic sta-
Thus, the sparseness of data on the course of tus, among others. Of course, these conditions
major depression does not lend it itself particu- intersect and have complex effects on childrens
larly well to analyses of trajectories. lives. Increasing heterogeneity of family structure
There are, however, important ways in which in terms of the different kinds of family types
a life course approach to diagnosed psychiatric (married parents with children, cohabiting par-
disorder can be undertaken. Aneshensel (2013) ents with children, blended families, single-
provides a very interesting analysis of mental ill- parent families, etc.) and increasing cultural and
ness as a career to highlight how the course of ethnic diversity in modern societies suggest that
416 W.R. Avison

the effects of structural characteristics on chil- to be done in developing a stress universe for
drens exposure to stress and their subsequent children (see Avison 2010).
mental health will be contingent on one another. At the same time, stress researchers have also
At the same time, it seems clear that childrens acknowledged that the meaning of stressful expe-
neighborhood environments affect their exposure rience is an important factor that inuences any
to stress and their emotional well-being. particular stressors impact on an individual.
Aneshensel and Sucoff (1996) have documented McLeod (2012) has argued that the cultural and
how neighborhoods are critical environments for structural conditions in which people live inu-
children and adolescents. Wheaton and Clarke ence their experiences of stressors. She also sug-
(2003) extend these ideas to incorporate life gests that the ways in which individuals make
course elements by demonstrating that the neigh- sense of their stressful experiences occurs in the
borhoods in which children grow up have conse- context of meaning negotiations with other peo-
quences for child and adolescent mental health ple in their lives. Thus, the importance of linked
which manifests itself further along the life lives for the negotiation of meaning may have
course in terms of psychological distress in implications for life course studies of stress and
adulthood. mental health.
One of most important developments in stress The meaning of stressors has also been
has been the specication of the vast array of dif- addressed from a somewhat different perspective.
ferent kinds of stressors to which individuals are Turner and Avison (1992) and Reynolds and
exposed. Pearlins observation that social stress Turner (2008) have employed crisis theory to
involves more than stressful life events and ought demonstrate that the impact of stressors on indi-
to include other dimensions such as role related viduals mental health is inuenced by whether
and ambient strains set stress research on this such events are perceived to be crisis experiences.
path. Wheatons (1994) concept of a stress uni- In addition, they nd some support for the idea
verse the idea that individuals may be exposed that events that are successfully resolved in the
to a large number of different kinds of stressful minds of individuals have less effect on subse-
experience has been a hallmark of this perspec- quent psychological distress.
tive. Turner et al. (1995) demonstrated the value As work progresses in this area, there is also a
of including multiple dimensions of stress to conceptual challenge to life course researchers.
account for social differences in mental health To date, most models of trajectories of stress
outcomes. Turner and Avison (2003) extended accumulation over the life course tend to be more
this line of work to show how a more comprehen- strongly driven by methodological considerations
sive estimate of stress, including measures of than by any conceptual model of stress accumu-
traumatic experiences and personal adversities as lation. The idea of cumulative adversity has much
well as indicators of discrimination stress, was a to recommend itself as a way of thinking about
much more powerful explanation of differences the experience of stress over the life course, but
in psychological distress associated with racial/ relatively little work has focused on what form
ethnic, social class, and to a lesser extent, gender such accumulation might take. For example, we
than was a simple checklist measure of stressful do not know whether stressors build up as a lin-
life events. ear, additive function or whether they do so expo-
Stress researchers have continued to expand nentially. Moreover, although we have evidence
this universe by developing inventories to index that the impact of stressful life events on psycho-
adversities and signicant stressful experiences logical distress deteriorates with the passage of
in childhood and adolescence. Although we now time (Avison and Turner 1988), we do not have
have evidence that stressors experienced early in any systematic, corresponding knowledge about
the life course have consequences for psycholog- traumatic stressors or childhood adversities. This
ical distress in adulthood, there is still much work seems to have limited the ways in which we con-
Mental Health 417

sider trajectories of stress accumulation across manner. These turning points or milestones sig-
the life course. nicantly change the course and sequelae of
DiPrete and Eirich (2006) have provided s mental illness for individuals. One of the most
comprehensive review of the ways in which the inuential turning points or milestones is early
concept of cumulative advantage have been onset of psychiatric disorder.
employed in sociological research. They also Research in mental health has clearly docu-
describe in some detail the statistical models that mented how the early onset of a psychiatric dis-
can be used to estimate cumulative advantage in order is associated with poorer prognosis,
a number of different contexts. These models including more frequent recurrence, longer dura-
appear to have considerable potential for better tions of episodes (Harrington et al. 1990;
specifying processes of stress accumulation over Lewinsohn et al. 1994). Turnbull et al. (1990)
the life course. document lower levels of socioeconomic attain-
A third opportunity for life course research on ment and higher rates of separation and divorce.
trajectories of mental health has emerged with These patterns have also been reported by Wade
the expanded knowledge that we have about the and Pevalin (2004) for subsequent divorce,
link between physical health and mental health. Davies et al. (1997) for single parenthood, and
For decades, sociologists of mental health have McLeod and Fettes (2007) for educational attain-
known of the interplay between physical and ment. Thus, the experience of a mental health
mental health (Aneshensel et al. 1984; Turner problem early in ones life may constitute a turn-
and Noh 1988). There is now a substantial body ing point or milestone that signicantly alters
of research that has examined the important ones life chances.
effects of physical illness on depressive symp- Another central principle of the life course
toms in older age (Gayman et al. 2008; Schieman perspective is the concept of linked lives, the idea
and Plickert 1997; Yang 2006). Bierman and that individuals lives are interconnected and that
Pearlin (2011) examine the ways in which trajec- one persons life experiences have consequences
tories of physical limitations inuence levels of for others. Using this concept as an organizing
psychological distress in later life. Their results principle, Avison and Comeau (2013) have
point to the important roles that mobility and reviewed two distinct bodies of literature related
functional limitations play in mediating the rela- to the consequences of mental illness for fami-
tionship between SES and psychological distress lies. The rst concerns the intergenerational
in later life. Although they do not examine how transmission of mental illness; the second focuses
trajectories of limitations are related to trajecto- on the mental health of individuals who provide
ries of mental health because of the complexity care to family members who have chronic
of such analyses, they clearly suggest that this is illness.
a next step in this line of inquiry. There is substantial evidence of a signicant
Other opportunities for integrating important correlation between parental mental illness and
life course concepts into the sociology of mental childrens emotional and behavioral problems.
health have presented themselves in recent years. Although much of the research in this area has
If we think about trajectories of mental health been conducted by developmentalists and psy-
over the life course, it seems reasonable to expect chiatrists, it seems clear that there are numerous
that there may be certain turning points or mile- opportunities for life course researchers in men-
stones along these trajectories that are important tal health to examine this relationship. Avison
to identify and understand. George (2014) has and Comeaus review is organized around the
provided some interesting insights into these mediating and moderating mechanisms that have
events or experiences that alter life trajectories. been described by Goodman and Gotlib (2002) in
She argues that certain experiences alter the their Integrative Model for the Transmission of
direction of an existing trajectory in a permanent Risk. This model identies four sets of mediators
418 W.R. Avison

that link parental mental disorder with childrens et al. 1996; Pearlin et al. 1997; Turner and Catania
mental health problems: (1) heritability; (2) 1997). All of these studies reinforce the idea that
innate dysfunctional neuroregulation; (3) expo- there is much that a life course perspective can
sure to maladaptive maternal behaviors and cog- bring to the study of the linked lives of individu-
nitions; and (4) elevated exposure to stressors. als with chronic illnesses.
Although there is ample evidence of genetic
sources of familial aggregation of mental disor-
der, the identication of candidate genes has been 3 Mental Health, Biology,
a relatively slow process. Goodman and Gotlib Social Experience,
suggest that innate dysfunctional neuroregula- and the Life Course
tory processes may also link maternal and child
depression. Goodman (2007) has reviewed a sub- In the sociology of mental health, there has
stantial body of research documenting how ele- always been a certain tension concerning the
vated neuroendocrine secretions during roles that biological or physiological processes
pregnancy among anxious or depressed mothers might play in explaining social disparities in
may expose their children prenatally to elevated mental health outcomes. It is one thing to assert
risks for neurodevelopmental challenges. The that there are biological determinants of mental
third class of mediators consists of qualities of illness; it is quite another thing to argue that bio-
parenting that may elevate childrens risk of logical processes account for social differences in
depression (see Dix and Meunier 2009, for a the prevalence of mental illness.
review) as a result of lower levels of competence Recently, there have been at least three devel-
in parenting including parental withdrawal, emo- opments which suggest mechanisms that at least
tional negativity toward children, and ineffective partially explain how biological processes may
discipline. The fourth group of mediators con- mediate the association between social status and
sists of stressors. These mediators include mari- mental health outcomes: (1) the developmental
tal or family discord (Cummings et al. 2005; origins of health and disease; (2) childhood
Papp 2010; Whisman and Kaiser 2008) as well as adversity and proinammatory response; and (3)
acute and chronic stressors reported by children genetics and epigenetics. Because these pro-
(Grant et al. 2003; Hammen et al. 2003). cesses unfold over time, the importance of under-
The second area that Avison and Comeau standing them as life course phenomena is
(2013) survey concerns the mental health of indi- critical.
viduals providing care to family members who
suffer from chronic illness. There is a long his-
tory in the sociology of mental health of studying 3.1 The Barker Hypothesis
the families of individuals with psychiatric disor- and the Developmental
ders (Clausen and Yarrow 1955; Greenberg et al. Origins of Health and Disease
1997; Noh and Avison 1988; Noh and Turner
1987). More recent studies have explicitly Over the past two decades, there has been an
adopted a stress process framework that is con- exponential growth in the body of research that
sistent with the life course perspective (Awad and focuses on the developmental origins of health
Voruganti 2008; Townsend et al. 2006). and disease (DOHaD). Much of this is research
Perhaps the most impressive advances in life that links low birth weight (LBW) or births small
course approaches to caregiving have emerged for gestational age (SGA) to subsequent elevated
from programs of research on individuals provid- risk of various diseases in adulthood. The devel-
ing care to family members with Alzheimers dis- opmental hypothesis or the Barker hypothesis, as
ease and other late-life dementias (Aneshensel it has become known, proposes that fetal malnu-
et al. 1995; Pearlin et al. 2001) and studies of trition (and perhaps fetal oxygen deprivation)
caregiving to individuals with AIDS (Moskowitz results in changes in fetal growth, metabolism,
Mental Health 419

and vasculature that adversely affect renal, pan- and short maternal stature. For life course
creatic, and cardiac development in utero and in researchers, the Barker hypothesis has substan-
infancy as well as subsequent changes to the tial signicance because it suggests that there are
hypothalamic- pituitary- adrenal (HPA) axis. All important social structural factors that may con-
of these early consequences of fetal growth tribute to restricted fetal growth that in turn mani-
restriction are known precursors of hypertension, fests itself in disease in adulthood.
renal failure, and insulin resistance (elements of It may also have implications for life course
the metabolic syndrome), and coronary heart dis- research on mental health. Thompson et al.
ease (Barker 1995, 1998, 2006; Hales and Barker (2001) analyzed data from a sample of individu-
2001). als born in Hertfordshire county in England in the
Barker (2006) gives three reasons why indi- 1920s who were reinterviewed in their 60s or
viduals born small are vulnerable to disease in 70s. Current social class, class at birth, recent
later life. Fetal growth restriction results in lim- loss of a loved one, and current experience of an
ited functional capacity of critical organs such as illness causing pain or limiting activities were all
the kidneys and heart. Second, abnormal fetal/ signicant predictors of depression among this
childhood hormonal and metabolic levels maxi- sample. After controlling for these factors, low
mize blood glucose concentrations for the brain birth weight was a signicant predictor of depres-
at the cost of limiting such transport into muscles sion among men but not women. This association
and muscle growth. Third, individuals born SGA is consistent with the Barker hypothesis, at least
may be vulnerable to environmental adversities for men, although it is unclear why there was a
across the life course. gender difference. In an earlier study of the same
Although much of Barkers formulation cohort, Barker et al. (1995) report higher rates of
clearly implicates processes that play out across suicide among individuals born small, a nding
the life course, this last point is especially impor- that again seems to support the hypothesis about
tant to sociologists because it highlights potential the developmental origins of mental health
interactive effects of LBW and socio-economic problems.
inuences on subsequent adult disease. Barker Other studies have reported similar ndings.
et al. (2005) have demonstrated with data from a Brown and colleagues (1995, 2002) document
birth cohort from Helsinki that the inverse asso- elevated rates of affective disorder in adulthood
ciation between household income in adulthood among children whose mothers were exposed to
and coronary heart disease was limited to men the Dutch Hunger Winter of 19441945 during
who were small at birth. The Whitehall studies their second trimester. Among a cohort born in
(Marmot and Wilkinson 2001) have suggested Helsinki between 1934 and 1944, Raikkonen
that lower socio-economic status may expose et al. (2007) report signicant associations
individuals to ongoing stress that elevates serum between gestational age at birth and scores on
cortisol and creates changes in neuroendocrine the Beck Depression Inventory at mean age 61
and HPA pathways that are linked to coronary and Center for Epidemiologic Studies-
heart disease and the metabolic syndrome. Depression scores at mean age 63. These asso-
There is a large literature that documents ciations remained signicant after controlling
higher rates of intrauterine growth restriction for social class at birth, educational attainment,
(IUGR) among infants born to mothers with gender, birth weight, and body mass index as an
lower income, lower SES, or lower education adult.
(see Kramer et al. 2000, for a review). In an Other studies have shown that trajectories of
attempt to identify the important mediators of the symptoms of depression and anxiety over the life
SES-IUGR relationship, Kramer et al. (2000) course are associated with birth weight. With
conclude that the factors that mediate this rela- data from the British National Survey of Health
tionship between SES and IUGR are smoking and Development, Colman et al. (2007) exam-
during pregnancy, low gestational weight gain, ined trajectories of symptoms of depression and
420 W.R. Avison

anxiety among a cohort of babies born in 1946 Thus far, tests of the Barker hypothesis have
and followed over 53 years. They nd that lower not fully considered the possibility that the focal
birth weight and delays in early development are relationship between SGA or IUGR and adult
associated with higher levels of symptoms in disease may reect a linked series of social cau-
middle age. sation and social selection processes beginning in
Some caution is necessary in interpreting utero and unfolding across the life course. Given
these ndings. First, it is unclear whether the that babies of small gestation age or low birth-
modest controls for mothers socio-economic sta- weight are more likely to be born to lower SES
tus at birth and in adulthood are sufcient to mothers, it seems likely that the childs experi-
explain away the life course effects of SES on ence of social disadvantage throughout child-
mental health. It seems unlikely that these statis- hood and adolescence may expose him/her to a
tical controls can capture all the social structural variety of stressors that contribute to poor physi-
inuences in an individuals life. As Link and cal and mental health. This is certainly consistent
Phelan (1995) and Phelan and Link (2005) have with recent work on cumulative disadvantage
so clearly demonstrated in their explication of (Hayward and Gorman 2004; ORand and Hamil-
fundamental causes, the effects of social class on Luker 2005) that describes processes that might
health may be pervasive and operate along a well specify the mediating pathways that connect
dense series of causal pathways. Accordingly, intrauterine growth restriction to adult mental
controlling for only one or two dimensions of health. McLeod and Pavalko (2008) have
social structure at selected points in the life described how childrens mental health problems
course is unlikely to be a convincing demonstra- may have consequences for school failure and
tion that the effect of birth weight on adult mental George (2007) has suggested how early mental
health is not a function of the persistent inuence illness has consequences for achievement in
of SES on pregnancy, childhood, adolescence, adulthood and for adult mental illness. It is a rea-
and adulthood. sonable hypothesis that the connections between
There is some evidence from studies of infant IUGR and adult mental health might better be
mortality that low birth weight is a correlate of characterized as the interplay of social experi-
adverse birth outcomes, but it is not on the causal ences and physiological processes across the life
pathway to infant mortality. This argument is course. This is consistent with Avisons (2010)
referred to as the Wilcox-Russell hypothesis description of the social and mental health expe-
(Wilcox 2001; Wilcox and Russell 1983a, b). riences of single mothers as chains of adversity.
Essentially, this challenge to the Barker hypoth- This is not to deny the Barker hypothesis, but
esis suggests that certain risk factors (such as rather to suggest that there are other pathways
maternal age, maternal smoking) may inuence connecting SGA and IUGR to adult mental health
both birthweight and infant mortality, but there is status.
scant evidence that population shifts in birth- It may be, then, that children born small may
weight produce concomitant changes in infant be at risk of emotional and behavioral problems
mortality (Gage et al. 2009). This suggests the as a function of psychophysiological processes
need to test thoroughly for the possibility that associated with the HPA system and family envi-
social disadvantage may produce disparities in ronments that are not conducive to healthy emo-
birthweight and, later, disparities in health across tional development. These problems manifest
the life course. The same argument might well themselves in academic challenges and failures
apply to other outcomes such as mental health. that create trajectories of low achievement in
Thapar and Rutter (2009) have urged caution in adulthood that, in turn, contribute to elevated
interpreting the correlation of prenatal risk fac- symptoms of psychiatric disorder or psychologi-
tors with adult psychiatric disorders as a causal cal distress. Life course researchers of mental
pathway. Clearly, this represents an interesting health have important roles to play in unraveling
opportunity for future research. these complex pathways.
Mental Health 421

3.2 Understanding How Social these processes for mental health has been estab-
Experience Gets Under lished by a substantial body of research docu-
the Skin menting how inammation and depression
cluster in individuals who have experienced
The stress process paradigm has made an enor- childhood adversities or traumas (Danese et al.
mous contribution to our understanding of how 2008; Miller and Cole 2012).
individuals locations in the social structure of In this context, one other development may
society exposes them to more or less stress and also have important implications for life course
how that stress is manifest as psychological dis- approaches to research on mental health. The
tress or mental disorder. The synthesis of the accumulated research on the effects of the social
stress process model with the life course perspec- environment on inammation has given rise to
tive has contributed to increased interest in the the eld of human social genomics (Cole 2009,
ways that adversities and stressors in earlier life 2014). The central feature of this eld is the
have consequences over the life course. For many hypothesis that everyday life experiences affect
scientists, what has been missing from these mod- gene expression.
els is the specication of a biologically plausible Shanahan (2014) has provided a provocative
set of pathways that connect socioeconomic status discussion of the opportunities for researchers
and stress exposure to subsequent mental health who are interested in health across the life course.
outcomes (see also Ferraro, chapter Life Course Essentially, he argues that progress in under-
Lens on Aging and Health this volume). standing how environmental stressors affect gene
Miller et al. (2011) have been at the forefront transcription can provide unique insights into the
of research that links stress in childhood to ways in which socioeconomic disadvantage
chronic diseases in aging. In their comprehensive translates into ill health through a series of
review of the literature on this issue, they develop genetic processes. He describes how complex,
a biological embedding model in which child- multilevel research designs incorporating periph-
hood stress results in proinammatory responses eral blood sampling and longitudinal data collec-
as a result of various molecular biological pro- tion about socio-environmental could be
cesses. Over the life course, these inammatory employed to examine various life course models
responses are exacerbated by behavioral tenden- of health outcomes.
cies including hypersensitivity to threat, mistrust Although his proposal focuses largely on phys-
of others, and conictual social relationships, all ical health, there seems to be no reason that men-
of which contribute to emotional and behavioral tal health outcomes would be excluded. After all,
problems over the life course. both the DOHaD hypothesis and the idea of pro-
Chen and Miller (2013) and Miller and Chen inammatory responses to the social environment
(2013) have written extensively on how socio- have been linked to mental health outcomes.
economic status patterns the association between All of this research has important implications
stressors and inammatory processes. They argue for life course studies of mental health. First, the
that social disadvantage exposes individuals to experience of childhood adversity appears to be a
the dynamic interplay of an array of stressors at turning point that sets in motion these physiolog-
the neighborhood level (exposure to violence, ical processes that culminate in the coupling of
decits in social capital), the family level (parent- inammatory responses and symptoms of depres-
ing problems, family conict, the absence or dis- sion. Second, this process takes time: it plays out
ruption of routine), and the individual level (risky over the life course. Third, the process seems to
health behaviors, mistrust). These stressors con- be inuenced importantly by individuals rela-
tribute to elevated inammatory response through tionships with others (their parents, their spouses,
processes described earlier. the quality of their social relationships). Thus,
Although much of Miller and Chens work has the importance of linked lives seems to be high-
been focused on physical health, the relevance of lighted in these processes.
422 W.R. Avison

For life course researchers in mental health, the interplay between the genome and social
this current of research offers a number of inter- experience has been discussed widely in recent
esting opportunities. For example, the associa- years (Kendler et al. 2011; Rutter 2007; Seabrook
tion of inammatory response with depression and Avison 2010; Shanahan and Hofer 2005).
(Pace et al. 2006) may provide insights into the There appear to be some important implica-
important relationship between some kinds of tions of this work for thinking about mental
physical disability and depression in older age. health in life course perspective. Shanahan and
Since many chronic diseases have a strong Hofer (2011) provide a very informative discus-
inammatory component, there may be reason to sion of how one can think about models that
explore how early adversity may exacerbate the incorporate sensitive periods, the accumulation
link between physical disability and mental of stress or the erosion of social capital, or path-
health problems. ways. Studies that report a G E interaction
As Cohen et al. (2010) have argued, the asso- where the environmental factor is a measure of
ciation between childhood SES and physical ill- life stress in childhood have been interpreted by
ness in adulthood has a number of alternative some as examples of models invoking a sensitive
explanations. The correlation could be a function period. A question that ows logically from these
of the aggregation of risky genes in select fami- investigations of sensitive periods is whether the
lies that elevates their risk of both poverty and same G E interaction can be observed among
poor health. It could be a function of the accumu- individuals for stressors that occur later in the life
lation of adversity over the life course. It might course. Uher (2014) has concluded that the
also be the result of adversity in childhood that weight of evidence suggests that the G E inter-
sets off recurrent chains of physiological pro- action involving 5-HTTLPR is specic to the
cesses as described by Miller and Chen. These timing and type of adversity (childhood maltreat-
same alternative explanations also seem plausible ment vs. adulthood life events) and outcome (per-
for understanding childhood disadvantage and sistent depression rather than single time-limited
adult mental health. For life course researchers, depressive episodes) and depends on high-quality
these possibilities generate a windfall of research epidemiological methodology, including a
questions that need to be addressed. detailed objective assessment of environmental
exposures (p. 6).
This clearly seems to specify an important
3.3 Genes, Epigenetic Processes, role for life course researchers as we continue to
and Social Genomics unravel the complex interplay of the roles of
social experiences and molecular genetics in
Research on the inuence of genes on mental shaping mental health outcomes. It suggests that
health outcomes has a long scientic history, but there is a specicity to gene-environment interac-
the modern era of investigation of this topic tions that requires testing over the life course.
seems to have begun with the report by Caspi To these considerations, one must consider the
et al. (2003) in Science of a gene-environment potential impact of epigenetic processes. The rela-
interaction (G E) in which a variant in the tively recent recognition that epigenetic processes
serontonin transporter gene (5-HTTLPR) exacer- play important roles with respect to gene expres-
bated the effects of child maltreatment and life sion has revolutionized the way in which we think
events on depression. Since then, a number of about genes and the environment. Epigenetic pro-
papers have replicated this nding (for reviews, cesses are alterations to DNA that activate or
see Shanahan and Hoffer 2011; Uher 2014). silence the DNA sequence. In lay terms, epigene-
These ndings have set in motion a series of stud- tic processes switch genes off or on. Most impor-
ies and discussions about the relative inuences tantly, these processes seem to be affected by
on mental health of genes, environments, their environmental exposures such as toxins, pollut-
interactions and correlations. The complexity of ants, diet, and perhaps even social experience.
Mental Health 423

Although research on epigenetic processes is that ultimately produce symptoms of mental


almost exclusively based on experiments with health problems.
animal models, one of the most provocative nd- Whether one chooses to investigate the social
ings has been that rat pups raised by more nur- and social psychological complexities that
turant mothers are less responsive to stressors in describe mental health across the life course or
later life and that these patterns may be heritable whether one incorporates features from biosci-
(Cameron et al. 2005; Meaney and Szyf 2005). ence into such studies ought to depend on the
The heritability of this lower responsiveness to questions asked. Consistent with Shanahan and
stress implies that the nurturing behaviors may Hofer (2011), the review of the literature pre-
have produced epigenetic changes that inuence sented here clearly suggests that strictly socio-
subsequent generations. behavioral studies of mental health and the life
Uher (2011) has proposed a life course model course will answer questions that are just as
of the epidemiology and treatment of depression pressing and important as those asked by molecu-
that hypothesizes the concurrent processes of lar geneticists or physiological psychologists
genetic DNA sequences, epigenetic modica- with training in genomics or cell biology.
tions, and social experience. For life course For social scientists, the focus remains on
researchers, this offers a range on intriguing understanding the complexities underling the
research questions to be pursued. links among social disparities, exposure to stress,
At the same time, some caution seems war- and the experience of mental health across the life
ranted in any uncritical race to incorporate the course. These questions can be answered without
Barker hypothesis, inammatory response mod- reference to molecular genetics or physiology.
els, social genomics, or genetic and epigenetics For some social scientists and most bioscien-
into life course studies of mental health. As tists, the pressing questions concern the micro-
Schnittker (2014) has noted with regard to social processes that translate social experience into
genomics, the eld has been far more genomic physiological processes and how these processes
that social. The ways in which social environ- affect mental health across the life course. To
ments inuence physiological processes may be answer these questions, a full array of scientists
extremely nuanced and may require considerably from diverse disciplines will need to collaborate
more input from social and behavioral scientists to tell the full story about mental health across
than has been the case thus far. the life course.
Regardless of the particular focus that
researchers select, there is a rich set of opportuni-
4 Which Way Forward? ties for life course research on mental health. All
of this work will require careful conceptualiza-
It seems clear that there is an enormous number tions of the research questions, rigorous yet
of opportunities for life course researchers inter- inventive research designs, and sophisticated data
ested in studying mental health. The rst half of analytic strategies.
this chapter identied a series of important
research questions that has arisen out of the syn-
thesis of the stress process paradigm with the life References
course perspective. These questions have as their
primary focus the relationship between social Aneshensel, C. S. (2013). Mental illness as a career:
Sociological perspectives. In C. S. Aneshensel, J. C.
disparities and mental health outcomes.
Phelan, & A. Bierman (Eds.), Handbook of the sociol-
The second half of this chapter has briey ogy of mental health (2nd ed., pp. 603620).
reviewed new developments in the biological and New York: Springer.
life sciences that enable us to understand the Aneshensel, C. S., & Sucoff, C. A. (1996). The neighbor-
hood context of adolescent mental health. Journal of
physiological pathways and processes that trans-
Health and Social Behavior, 37, 293301.
late social experiences into biological processes
424 W.R. Avison

Aneshensel, C. S., Frerichs, R. R., & Huba, G. J. (1984). Awad, A. G., & Voruganti, L. N. P. (2008). The burden of
Depression and physical illness A multiwave, nonre- schizophrenia on caregivers: A review.
cursive causal model. Journal of Health and Social PharmacoEconomics, 26, 149162.
Behavior, 25, 350371. Barker, D. J. P. (1995). Fetal origins of coronary heart dis-
Aneshensel, C. S., Rutter, C. M., & Lachenbruch, P. A. ease. British Medical Journal, 310, 171174.
(1991). Competing conceptual and analytic models: Barker, D. J. P. (1998). Mothers, babies and health in life
Social structure, stress, and mental health. American (2nd ed.). Edinburgh: Churchill Livingstone.
Sociological Review, 56, 166178. Barker, D. J. P. (2006). Birth weight and hypertension.
Aneshensel, C. S., Pearlin, L. I., & Schuler, R. H. (1993). Hypertension, 48, 357358.
Stress, role captivity, and the cessation of caregiving. Barker, D. J. P., Osmond, C., Rodin, I., Fall, C. H. D., &
Journal of Health and Social Behavior, 34, 5470. Winter, P. D. (1995). Low-weight gain in infancy and
Aneshensel, C. S., Pearlin, L. I., Mullan, J. T., Zarit, S. H., suicide in adult life. British Medical Journal, 311,
& Whitlach, C. J. (1995). Proles in caregiving: The 1203.
unexpected career. San Diego: Academic. Barker, D. J. P., Osmond, C., Forsn, T. J., Kajantie, E., &
Aneshensel, C. S., Botticello, A. L., & Yamamoto-Mitani, Eriksson, J. G. (2005). Trajectories of growth among
N. (2005). When caregiving ends: The course of children who have coronary events as adults. New
depressive symptoms after bereavement. Journal of England Journal of Medicine, 353, 18021809.
Health and Social Behavior, 45, 422440. Barrett, A. E., & White, H. R. (2002). Trajectories of gen-
Aneshensel, C. S., Phelan, J. C., & Bierman, A. (Eds.). der role orientations in adolescence and early adult-
(2013). Handbook of the sociology of mental health hood: A prospective study of the mental health effects
(2nd ed.). New York: Springer. of masculinity and femininity. Journal of Health and
Avison, W. R. (2001). Unemployment and its conse- Social Behavior, 43, 451468.
quences for mental health. In V. W. Marshall, Barrett, A. E., & Turner, R. J. (2005). Family structure and
W. Heinz, H. Krueger, & A. Verma (Eds.), mental health: The mediating effects of socioeco-
Restructuring work and the life course (pp. 177200). nomic status, family process, and social stress. Journal
Toronto: University of Toronto Press. of Health and Social Behavior, 46, 156169.
Avison, W. R. (2010). Family structure and womens Bierman, A., & Pearlin, L. I. (2011). SES, trajectories of
lives: A life course perspective. In W. R. Avison, C. S. physical limitations, and change in depression in late
Aneshensel, S. Schieman, & B. Wheaton (Eds.), life. Society and Mental Health, 1, 139152.
Advances in the conceptualization of the stress pro- Broidy, L. M., Nagin, D. S., Tremblay, R. E., et al. (2003).
cess: Essays in honor of Leonard I. Pearlin Developmental trajectories of childhood disruptive
(pp. 7192). New York: Springer. behaviors and adolescent delinquency: A six-site,
Avison, W. R., & Cairney, J. (2003). Social structure, cross-national study. Developmental Psychology, 39,
stress, and personal control. In S. H. Zarit, L. I. Pearlin, 222245.
& K. W. Schaie (Eds.), Personal control in social and Brown, G. W., & Harris, T. O. (1978). Social origins of
life contexts (pp. 127164). New York: Springer. depression. New York: Free Press.
Avison, W. R., & Comeau, J. (2013). The impact of men- Brown, G. W., & Harris, T. O. (1989). Life events and ill-
tal illness on the family. In C. S. Ansehensel, J. C. ness. New York: Guilford Press.
Phelan, & A. Bierman (Eds.), Handbook of the sociol- Brown, A. S., Susser, E. S., Lin, S. P., Neugebauer, R., &
ogy of mental health (2nd ed., pp. 543564). Gorman, J. M. (1995). Increased risk of affective dis-
New York: Springer. orders in males after second trimester prenatal expo-
Avison, W. R., & Davies, L. (2005). Family structure, sure to the Dutch Hunger Winter of 194445. British
gender, and health in the context of the life course. Journal of Psychiatry, 161, 601606.
Journals of Gerontology: Series B Psychological Brown, A. S., van Os, J., Driessens, C., Hoek, H. W., &
Sciences and Social Sciences, 60B, S113S116. Susser, E. S. (2002). Further evidence of relation
Avison, W. R., & McAlpine, D. D. (1992). Gender differ- between prenatal famine and major affective disorder.
ences in symptoms of depression among adolescents. American Journal of Psychiatry, 157, 190195.
Journal of Health and Social Behavior, 33, 7796. Cameron, N. M., Champagne, F. A., Parent, C., Fish,
Avison, W. R., & Turner, R. J. (1988). Stressful life events E. W., Ozaki-Kuroda, K., & Meaney, M. J. (2005).
and depressive symptoms: Disaggregating the effects The programming of individual differences in
of chronic strains and eventful stressors. Journal of defensive responses and reproductive strategies in
Health and Social Behavior, 29, 253264. the rat through variations in maternal care.
Avison, W. R., Davies, L., Willson, A., & Shuey, K. Neuroscience and Biobehavioral Reviews, 29,
(2008). Family structure and mothers mental health: 843865.
A life course perspective on stability and change. In Caspi, A., Sugden, K., Moftt, T. E., et al. (2003).
H. A. Turner & S. Schieman (Eds.), Stress processes Inuence of life stress on depression: Moderation by a
across the life course: Advances in life course research polymorphism in the 5-HTT gene. Science, 391,
(Vol. 13, pp. 233255). New York: Elsevier. 386389.
Mental Health 425

Chen, Z. Y., & Kaplan, H. B. (2003). School failure in Eaton, W. W., Shao, H., Nestadt, G., Lee, B. H., Bienvenu,
early adolescence and status attainment in middle O. J., & Zandi, P. (2008). Population-based study of
adulthood: A longitudinal study. Sociology of rst onset and chronicity in major depressive disorder.
Education, 76, 110127. Archives of General Psychiatry, 65, 513520.
Chen, E., & Miller, G. E. (2013). Socioeconomic status Elder, G. H., Jr., George, L. K., & Shanahan, M. J. (1996).
and health mediating and moderating factors. Annual Psychosocial stress over the life course. In H. B.
Review of Clinical Psychology, 9, 723749. Kapan (Ed.), Psychosocial stress: Perspectives on
Clausen, J. A., & Yarrow, M. R. (1955). Introduction: structure, theory, life- course, and methods (pp. 247
Mental illness and the family. Journal of Social Issues, 292). New York: Academic.
11, 35. Forthofer, M. S., Kessler, R. C., Story, A. L., & Gotlib,
Cohen, S., Janicki-Deverts, D., Chen, E., & Matthews, I. H. (1996). The effects of psychiatric disorders on the
K. A. (2010). Childhood socioeconomic status and probability and timing of rst marriage. Journal of
adult health. Annals of the New York Academy of Health and Social Behavior, 37, 121132.
Sciences, 1186, 3755. Gage, T. B., Fang, F., ONeill, E., & Stratton, H. (2009).
Cole, S. W. (2009). Social regulation of human gene Maternal age and infant mortality: A test of the
expression. Current Directions in Psychological Wilcox-Russell hypothesis. American Journal of
Science, 18, 132137. Epidemiology, 169, 294303.
Cole, S. W. (2014). Human social genomics. PLoS Gayman, M. D., Turner, R. J., & Cui, M. (2008). Physical
Genetics, 10, 17. limitations and depressive symptoms. Journal of
Colman, I., Ploubidis, G. B., Wadsworth, M. E. J., Jones, Gerontology: Social Sciences, 63B, S219S228.
P. B., & Croudace, T. J. (2007). A longitudinal typol- George, L. K. (2007). Life course perspectives on social
ogy of symptoms of depression and anxiety over the factors and mental illness. In W. R. Avison, J. D.
life course. Biological Psychiatry, 62, 12651271. McLeod, & B. A. Pescosolido (Eds.), Mental health,
Colman, I., Jones, P. B., Kuh, D., Weeks, M., Naicker, K., social mirror (pp. 191218). New York: Springer.
Richards, M., & Croudace, T. J. (2014). Early devel- George, L. K. (2013). Life-course perspectives on mental
opment, stress and depression across the life course: health. In C. S. Ansehensel, J. C. Phelan, & A. Bierman
Pathways to depression in a national British cohort. (Eds.), Handbook of the sociology of mental health
Psychological Medicine, 44, 28452854. (2nd ed., pp. 585602). New York: Springer.
Conger, R. D., Conger, K. J., Elder, G. H., Jr., Lorenz, George, L. K. (2014). Taking time seriously: A call to
F. O., Simons, R. L., & Whitbeck, L. B. (1992). A action in mental health research. Journal of Health
family process model of economic hardship and and Social Behavior, 55, 251264.
adjustment of early adolescent boys. Child George, L. K., & Lynch, S. M. (2003). Race differences in
Development, 63, 526541. depressive symptoms: A dynamic perspective on
Costello, E. J., Compton, S. N., Keeler, G., & Angold, A. stress exposure and vulnerability. Journal of Health
(2003). Relationships between poverty and psychopa- and Social Behavior, 44, 353369.
thology: A natural experiment. Journal of the Gibb, S. J., Fergusson, D. M., & Horwood, L. J. (2012).
American Medical Association, 290, 20232029. Childhood family income and life outcomes in adult-
Cummings, E. M., Keller, P. S., & Davies, P. (2005). hood: Findings from a 30-year longitudinal study in
Towards a family process model of maternal and New Zealand. Social Science & Medicine, 74,
paternal depressive symptoms: Exploring multiple 19791986.
relationships with child and family functioning. Goodman, S. H. (2007). Depression in mothers. Annual
Journal of Child Psychology and Psychiatry, 46, Review of Clinical Psychology, 3, 107135.
479489. Goodman, S. H., & Gotlib, I. H. (Eds.). (2002). Children
Danese, A., Moftt, T. E., Pariante, C. M., Ambler, A., of depressed parents: Mechanisms of risk and implica-
Poulton, R., & Caspi, A. (2008). Elevated inamma- tions for treatment. Washington, DC: American
tion levels in depressed adults with a history of child- Psychological Association.
hood maltreatment. Archives of General Psychiatry, Gore, S., & Aseltine, R., Jr. (2003). Race and ethnic dif-
65, 409415. ferences in depressed mood following the transition
Davies, L., Avison, W. R., & McAlpine, D. D. (1997). out of high school. Journal of Health and Social
Signicant life experiences and depression among Behavior, 44, 370389.
single and married mothers. Journal of Marriage and Gotlib, I. H., & Wheaton, B. (Eds.). (1997). Stress and
the Family, 59, 294308. adversity over the life course: Trajectories and
DiPrete, T. A., & Eirich, G. M. (2006). Cumulative advan- turning points. Cambridge: Cambridge University
tage as a mechanism for inequality: A review of theo- Press.
retical and empirical developments. Annual Review of Grant, K. E., Compas, B. E., Stuhlmacher, A. F., Thurm,
Sociology, 32, 271297. A. E., McMahon, S. D., & Halpert, J. A. (2003).
Dix, T., & Meunier, L. N. (2009). Depressive symptoms Stressors and child and adolescent psychopathology:
and parenting competence: An analysis of 13 regula- Moving from markers to mechanisms of risk.
tory processes. Developmental Review, 29, 4568. Psychological Bulletin, 129, 447466.
426 W.R. Avison

Greenberg, J. S., Kim, H. W., & Greenley, J. R. (1997). prevalence and age-of-onset distributions of DSM-IV
Factors associated with subjective burden in siblings disorders in the National Comorbidity Survey
of adults with severe mental illness. American Journal Replication. Archives of General Psychiatry, 62,
of Orthopsychiatry, 67, 231241. 593602.
Hales, N. C., & Barker, D. J. P. (2001). The thrifty pheno- Kramer, M. S., Seguin, L., Lydon, J., & Goulet, L. (2000).
type hypothesis. British Medical Bulletin, 60, 520. Socio-economic disparities in pregnancy outcome:
Hammen, C., Shih, J., Altman, T., & Brennan, P. A. Why do the poor fare so poorly? Paediatric and
(2003). Interpersonal impairment and the prediction of Perinatal Epidemiology, 14, 194210.
depressive symptoms in adolescent children of Kuh, D., Hardy, R., Rodgers, B., & Wadsworth, M. E. J.
depressed and nondepressed mothers. Journal of the (2002). Lifetime risk factors for womens psychologi-
American Academy of Child and Adolescent cal distress in midlife. Social Science & Medicine, 55,
Psychiatry, 42, 571577. 19571973.
Harrington, R., Fudge, H., Rutter, M., Pickles, A., & Hill, Lee, S., Guo, W. J., Tsang, A., He, Y. L., Zhang, M. Y.,
J. (1990). Adult outcomes of child and adolescent Shan, Y. C., & Kessler, R. C. (2011). The prevalence
depression: I. psychiatric status. Archives of General of family childhood adversities and their association
Psychiatry, 47, 465473. with rst onset of DSM-IV disorders in metropolitan
Harris, T. O., Brown, G. W., & Bifulco, A. (1986). Loss of China. Psychological Medicine, 41, 8596.
parent in childhood and adult psychiatric disorder: Lewinsohn, P. M., Clarke, G. N., Seeley, J. R., & Rohde,
Role of lack of adequate parental care. Psychological P. (1994). Major depression in community adoles-
Medicine, 16, 641659. cents Age at onset, episode duration, and time to
Harris, T. O., Brown, G. W., & Bifulco, A. T. (1990). recurrence. Journal of the American Academy of Child
Depression and situational helplessness/mastery in a and Adolescent Psychiatry, 33, 809818.
sample selected to study childhood parental loss. Link, B. G., & Phelan, J. C. (1995). Social conditions as
Journal of Affective Disorders, 20, 2741. fundamental causes of disease. Journal of Health and
Hatch, S. L., & Dohrenwend, B. P. (2007). Distribution of Social Behavior, 36, Extra Issue, 8094.
traumatic and other stressful life events by race/ethnic- Longest, K. C., & Thoits, P. A. (2012). Gender, the stress
ity, gender, SES and age: A review of the research. process, and health: A congurational approach.
American Journal of Community Psychology, 40, Society and Mental Health, 2, 187206.
313332. Luo, Y., & Waite, L. J. (2005). The impact of childhood
Hayward, M. D., & Gorman, B. K. (2004). The long arm and adult SES on physical, mental, and cognitive well-
of childhood: The inuence of early- life social condi- being in later life. Journals of Gerontology Series B
tions on mens mortality. Demography, 41, 87107. Psychological and Social Sciences, 60, S93S101.
Horwitz, A. V., Widom, C. S., McLaughlin, J., & White, Macmillan, R., & Copher, R. (2005). Families in the life
H. R. (2001). The impact of childhood abuse and course: Interdependency of roles, role congurations,
neglect on adult mental health: A prospective study. and pathways. Journal of Marriage and the Family,
Journal of Health and Social Behavior, 42, 184201. 67, 858879.
Huang, C., Phillips, M. R., Zhang, Y., Zhang, J., Shi, Q., Marmot, M. G., & Wilkinson, R. G. (2001). Psychosocial
Song, Z., Ding, Z., Pang, S., & Martorell, R. (2013). and material pathways in the relation between income
Malnutrition in early life and adult mental health: and health: A response to Lynch et al. British Medical
Evidence from a natural experiment. Social Science & Journal, 322, 12331236.
Medicine, 97, 259266. McGue, M., & Iacono, W. G. (2005). The association of
Johnson, W. L., Giordano, P. C., Longmore, M. A., & early adolescent problem behavior with adult psycho-
Manning, W. D. (2014). Intimate partner violence and pathology. American Journal of Psychiatry, 162,
depressive symptoms during adolescence and young 11181124.
adulthood. Journal of Health and Social Behavior, 55, McLeod, J. D. (1991). Childhood parental loss and adult
3955. depression. Journal of Health and Social Behavior, 32,
Kendler, K., Jaffee, S., & Romer, D. (Eds.). (2011). The 205220.
dynamic genome and mental health: The role of genes McLeod, J. D. (2012). The meanings of stress: Expanding
and environments in mental health. New York: Oxford the stress process model. Society and Mental Health,
University Press. 3, 172186.
Kessler, R. C., & Magee, W. J. (1993). Childhood adversi- McLeod, J. D., & Edwards, K. (1995). Contextual deter-
ties and adult depression: Basic patterns of association minants of childrens responses to poverty. Social
in a U. S. national survey. Psychological Medicine, 23, Forces, 73, 14871516.
679690. McLeod, J. R., & Fettes, D. L. (2007). Trajectories of fail-
Kessler, R. C., & Magee, W. J. (1994). Childhood family ure: The educational careers of children with mental
violence and adult recurrent depression. Journal of health problems. American Journal of Sociology, 113,
Health and Social Behavior, 35, 1327. 653701.
Kessler, R. C., Berglund, P., Demier, O., Jin, R., McLeod, J. D., & Pavalko, E. K. (2008). Selection and the
Merikangas, K. R., & Walters, E. E. (2005). Lifetime stress process: What does attention to the life course
Mental Health 427

offer? In H. A. Turner & S. Schieman (Eds.), Stress 3 to 18 years. Developmental Psychopathology, 8,


processes across the life course: Advances in life 399424.
course research (Vol. 13, pp. 75104). Oxford: Moskowitz, J. T., Folkman, S., Collette, L., & Vittinghoff,
Elsevier. E. (1996). Coping and mood during AIDS-related
McLeod, J. D., & Shanahan, M. (1993). Poverty, parent- caregiving and bereavement. Annals of Behavioral
ing, and childrens mental health. American Medicine, 18, 4957.
Sociological Review, 58, 351366. Noh, S., & Avison, W. R. (1988). Spouses of discharged
McLeod, J. D., & Shanahan, M. (1996). Trajectories of psychiatric patients: Factors associated with their
poverty and childrens mental health. Journal of experiences of burden. Journal of Marriage and the
Health and Social Behavior, 37, 207220. Family, 50, 377389.
Meadows, S. O. (2009). Family structure and fathers Noh, S., & Turner, R. J. (1987). Living with psychiatric
well-being: Trajectories of mental health and self- patients: Implications for the mental health of family
rated health. Journal of Health and Social Behavior, members. Social Science and Medicine, 25, 263271.
50, 115131. ORand, A. M., & Hamil-Luker, J. (2005). Processes of
Meaney, M. J., & Szyf, M. (2005). Environmental pro- cumulative adversity: Childhood disadvantage and
gramming of stress responses through DNA methyla- increased risk of heart attack across the life course.
tion: Life at the interface between a dynamic Journals of Gerontology Series B Psychological
environment and a xed genome. Dialogues in Sciences and Social Sciences, 60, S117S124.
Clinical Neurosciences, 7, 103123. Pace, T. W. W., Mletzko, T. C., Algabe, O., et al. (2006).
Meeks, S., Murrell, S. A., & Mehl, R. C. (2000). Increased stress-induced inammatory responses in
Longitudinal relationships between depressive symp- male patients with major depression and increased
toms and health in normal older and middle-aged early life stress. American Journal of Psychiatry, 163,
adults. Psychology and Aging, 15, 100109. 16301633.
Menaghan, E. G. (1997). Intergenerational consequences Papp, L. M. (2010). The course and quality of intimate
of social stressors: Effects of occupational and family relationships among psychologically distressed moth-
conditions on young mothers and their children. In ers. American Journal of Orthopsychiatry, 89, 7179.
I. H. Gotlib & B. Wheaton (Eds.), Stress and adversity Pearlin, L. I. (1983). Role strains and personal stress. In
over the life course: Trajectories and turning points H. B. Kaplan (Ed.), Psychosocial stress: Trends in
(pp. 114196). New York: Cambridge University theory and research (pp. 332). New York: Academic.
Press. Pearlin, L. I. (1989). The sociological study of stress.
Menaghan, E. G. (2009). Stress and distress in childhood Journal of Health and Social Behavior, 30, 241256.
and adolescence. In T. L. Scheid & T. N. Brown (Eds.), Pearlin, L. I. (1999). The stress concept revisited:
A handbook for the study of the sociology of mental Reections on concepts and their interrelationships. In
health (2nd ed., pp. 321333). New York: Cambridge C. S. Aneshensel & J. C. Phelan (Eds.), Handbook of
University Press. the sociology of mental health (pp. 395415).
Menaghan, E. G., & Parcel, T. L. (1994). Parents jobs and New York: Plenum.
childrens lives. New York: Aldine de Gruyter. Pearlin, L. I., & Skaff, M. M. (1996). Stress and the life
Menaghan, E. G., & Parcel, T. L. (1995). Social sources of course: A paradigmatic alliance. Gerontologist, 36,
change in childrens home environments: The effects 239247.
of parental occupational experiences and family con- Pearlin, L. I., Lieberman, M. A., Menaghan, E. G., &
ditions. Journal of Marriage and the Family, 57, Mullan, J. T. (1981). The stress process. Journal of
6984. Health and Social Behavior, 22, 337356.
Miech, R. A., & Shanahan, M. J. (2000). Socioeconomic Pearlin, L. I., Aneshensel, C. S., & LeBlanc, A. J. (1997).
status and depression over the life course. Journal of The forms and mechanisms of stress proliferation: The
Health and Social Behavior, 41, 162176. case of AIDS caregivers. Journal of Health and Social
Miller, G. E., & Chen, E. (2013). The biological residue of Behavior, 38, 223236.
poverty. Child Development Perspectives, 7, 6773. Pearlin, L. I., Pioli, M., & McLaughlin, A. (2001). Role
Miller, G. E., & Cole, S. W. (2012). Clustering of depres- disruption and distress among caregivers to parents
sion and inammation in adolescents previously with Alzheimers disease. In R. H. Binstock & L. K.
exposed to childhood adversity. Biological Psychiatry, George (Eds.), Handbook of aging and the social sci-
72, 3440. ences (5th ed., pp. 238254). New York: Academic.
Miller, G. E., Chen, E., & Parker, K. J. (2011). Pearlin, L. I., Schieman, S., Fazio, E. M., & Meersman,
Psychological stress in childhood and susceptibility to S. C. (2005). Stress, health, and the life course: Some
the chronic diseases of aging: Moving toward a model conceptual perspectives. Journal of Health and Social
of behavioral and biological mechanisms. Behavior, 46, 205219.
Psychological Bulletin, 137, 959997. Phelan, J. C., & Link, B. G. (2005). Controlling disease
Moftt, T. E., Caspi, A., Dickson, N., et al. (1996). and creating disparities: A fundamental cause perspec-
Childhood-onset versus adolescent-onset antisocial tive. Journals of Gerontology Series B Psychological
conduct problems in males: Natural history from ages Sciences and Social Sciences, 60, S27S33.
428 W.R. Avison

Raikkonen, K., Personen, A.-K., Kajantie, E., Heinonen, childhood and adulthood psychopathology in the
K., Forsen, T., Phillips, D. I. W., Osmond, C., Barker, south Africa stress and health study: Associations with
D. J. P., & Eriksson, J. G. (2007). Length of gestation rst-onset DSM-IV disorders. Social Science &
and depressive symptoms at age 60 years. British Medicine, 71, 18471854.
Journal of Psychiatry, 190, 469474. Spence, N. J., Adkins, D. E., & Dupre, M. E. (2011).
Reiss, F. (2013). Socioeconomic inequalities and mental Racial differences in depression trajectories among
health problems in children and adolescents: A sys- older women: Socioeconomic, family, and health
tematic review. Social Science & Medicine, 90, 2431. inuences. Journal of Health and Social Behavior, 52,
Reynolds, J. R., & Turner, R. J. (2008). Major life events: 444459.
Their personal meaning, resolution, and mental health Strohschein, L. (2005). Household income histories and
signicance. Journal of Health and Social Behavior, child mental health trajectories. Journal of Health and
49, 223237. Social Behavior, 46, 359375.
Ross, C. E., & Mirowsky, J. (1999). Parental divorce, life- Thapar, A., & Rutter, M. (2009). Do prenatal risk factors
course disruption, and adult depression. Journal of cause psychiatric disorder? Be wary of causal claims.
Marriage and the Family, 61, 10341045. British Journal of Psychiatry, 195, 100101.
Rutter, M. (2007). Gene-environment interdependence. Thoits, P. A. (2010). Stress and health: Major ndings and
Developmental Science, 10, 1218. policy implications. Journal of Health and Social
Rutter, M., Kim-Cohen, J., & Maughan, B. (2006). Behavior, 51, S41S53.
Continuities and discontinuities in psychopathology Thompson, C., Syddall, H., Rodin, I., Osmond, C., &
between childhood and adult life. Journal of Child Baker, D. J. P. (2001). Birth weight and the risk of
Psychology and Psychiatry, 47, 276295. depressive disorder in late life. British Journal of
Schieman, S., & Plickert, G. (1997). Functional limita- Psychiatry, 179, 150155.
tions and changes in levels of depression among older Townsend, A. L., Biegel, D. E., Ishler, K. J., Wieder, B., &
adults: A multiple-hierarchy stratication perspective. Rini, A. (2006). Families of persons with substance
Journal of Gerontology: Social Sciences, 62B, use and mental disorder: A literature review and con-
S36S42. ceptual model. Family Relations, 55, 473486.
Schieman, S., Van Gundy, K., & Taylor, J. (2001). The Turnbull, J. E., George, L. K., Landerman, R., et al.
relationship between age and depressive symptoms: A (1990). Social outcomes related to age of onset among
test of competing explanatory and suppression inu- psychiatric disorders. Journal of Consulting and
ences. Journal of Health and Aging, 14, 260285. Clinical Psychology, 58, 832839.
Schnittker, J. (2014). The challenge of social genomics: A Turner, R. J., & Avison, W. R. (1992). Innovations in the
commentary on Social genomics and the life course: measurement of life stress: Crisis theory and the sig-
Opportunities and challenges for multilevel popula- nicance of event resolution. Journal of Health and
tion research. In L. J. Waite & T. J. Plewes (Eds.), Social Behavior, 33, 3650.
New directions in the sociology of aging (pp. 277 Turner, R. J., & Avison, W. R. (2003). Status variations in
283). Washington, DC: National Academies Press. stress exposure: Implications for the interpretation of
Seabrook, J. A., & Avison, W. R. (2010). Genotype- research on race, socioeconomic status, and gender.
environment interaction and sociology: Contributions Journal of Health and Social Behavior, 44, 488505.
and complexities. Social Science & Medicine, 70, Turner, H. A., & Catania, J. A. (1997). Informal caregiv-
12771284. ing to persons with AIDS in the United States:
Selye, H. (1956). The stress of life. New York: Caregiver burden among central cities residents eigh-
McGraw-Hill. teen to forty-nine years old. American Journal of
Shanahan, M. J. (2014). Social genomics and the life Community Psychology, 25, 3559.
course: Opportunities and challenges for multilevel Turner, R. J., & Lloyd, D. A. (1995). Lifetime traumas and
population research. In L. J. Waite & T. J. Plewes mental health: The signicance of cumulative adver-
(Eds.), New directions in the sociology of aging sity. Journal of Health and Social Behavior, 36,
(pp. 255276). Washington, DC: National Academies 360376.
Press. Turner, R. J., & Noh, S. (1988). Physical disability and
Shanahan, M. J., & Hofer, S. M. (2005). Social context in depression: A longitudinal analysis. Journal of Health
gene-environment interactions: Retrospect and pros- and Social Behavior, 29, 2337.
pect. Journals of Gerontology. Series B, Psychological Turner, H. A., & Schieman, S. (Eds.). (2008). Stress pro-
Sciences and Social Sciences, 60, 6576. cesses across the life course (H. A. Turner, &
Shanahan, M. J., & Hoffer, S. M. (2011). Molecular S. Schieman Eds.). New York: Elsevier.
genetics, aging, and well-being: Sensitive period, Turner, H. A., & Turner, R. J. (2005). Understanding vari-
accumulation, and pathway models. In R. H. Binstock ations in exposure to social stress. Health, 9,
& L. K. George (Eds.), Handbook of aging and the 209240.
social sciences (pp. 135147). New York: Elsevier. Turner, J. B., & Turner, R. J. (2013). Social relations,
Slopan, N., Williams, D. R., Seedat, S., Moomal, H., social integration, and social support. In C. S.
Herman, A., & Stein, D. J. (2010). Adversities in Aneshensel, J. C. Phelan, & A. Bierman (Eds.),
Mental Health 429

Handbook of the sociology of mental health (pp. 341 mental health in early adulthood. American
356). New York: Springer. Sociological Review, 68, 680706.
Turner, R. J., Wheaton, B., & Lloyd, D. A. (1995). The Wheaton, B., Roszell, P., & Hall, K. (1997). The impact of
epidemiology of social stress. American Sociological twenty childhood and adult traumatic stressors on the
Review, 60, 104125. risk of psychiatric disorder. In I. H. Gotlib &
Turner, R. J., Lloyd, D. A., & Taylor, J. (2006a). Stress B. Wheaton (Eds.), Stress and adversity over the life
burden, drug dependence and the nativity paradox course: Trajectories and turning points (pp. 5072).
among U.S. Hispanics. Drug and Alcohol Dependence, Cambridge: Cambridge University Press.
83, 7989. Whisman, M. A., & Kaiser, R. (2008). Marriage and rela-
Turner, R. J., Lloyd, D. A., & Taylor, J. (2006b). Physical tionship issues. In K. S. Dobson & D. J. A. Dozois
disability and mental health: An epidemiology of psy- (Eds.), Risk factors in depression (pp. 363384).
chiatric and substance disorders. Rehabilitation Oxford: Elsevier.
Psychology, 51, 214233. Wickrama, K. A. S., Conger, R. D., & Abraham, T. W.
Uher, R. (2011). Genes, environment, and individual dif- (2008a). Early family adversity, youth depressive
ferences in responding to treatment for depression. symptom trajectories, and young adult socioeconomic
Harvard Review of Psychiatry, 19, 109124. attainment: A latent trajectory class analysis. In H. A.
Uher, R. (2014). Gene-environment interactions in com- Turner & S. Schieman (Eds.), Stress processes across
mon mental disorders: An update and strategy for a the life course: Advances in life course research (Vol.
genome-wide search. Social Psychiatry and 13, pp. 161192). Oxford: Elsevier.
Psychiatric Epidemiology, 49, 314. Wickrama, K. A. S., Conger, R. D., Lorenz, F. O., & Jung,
Umberson, D., Williams, K., Powers, D. A., Liu, H., & T. (2008b). Family antecedents and consequences of
Needham, B. (2005). Stress in childhood and adult- trajectories of depressive symptoms from adolescence
hood: Effects on marital quality over time. Journal of to young adulthood: A life course investigation.
Marriage and the Family, 67, 13321347. Journal of Health and Social Behavior, 49, 468483.
Wade, T. J., & Pevalin. (2004). Marital transitions and Wilcox, A. J. (2001). On the importance and unimpor-
mental health. Journal of Health and Social Behavior, tance of birthweight. International Journal of
45, 155170. Epidemiology, 30, 12331241.
Walsemann, K. M., Gee, G. C., & Geronimus, A. T. Wilcox, A. J., & Russell, I. T. (1983a). Birthweight and
(2009). Ethnic differences in trajectories of depressive perinatal mortality: I. On the frequency distribution of
symptoms: Disadvantage in family background, high birthweight. International Journal of Epidemiology,
school experiences, and adult characteristics. Journal 12, 314318.
of Health and Social Behavior, 50, 8298. Wilcox, A. J., & Russell, I. T. (1983b). Perinatal mortal-
Wheaton, B. (1985). Models for the stress-buffering func- ity: Standardizing for birthweight is biased. American
tions of coping resources. Journal of Health and Journal of Epidemiology, 118, 857864.
Social Behavior, 26, 352365. Wineld, J., George, L. K., Swartz, M., & Blazer, D. G.
Wheaton, B. (1994). Sampling the stress universe. In (1990). Sexual assault and psychiatric disorders
W. R. Avison & I. H. Gotlib (Eds.), Stress and among a community sample of women. American
mental health: Contemporary issues and prospects Journal of Psychiatry, 147, 335341.
for the future (pp. 77114). New York: Plenum Yang, Y. (2004). How does functional disability affect
Press. depression in late life: The roles of social support and
Wheaton, B. (2010). The stress process as a successful personal resources. Gerontologist, 44, 373384.
paradigm. In W. R. Avison, C. S. Aneshensel, Yang, Y. (2006). How does functional disability affective
S. Schieman, & B. Wheaton (Eds.), Advances in the depressive symptoms in late life? The role of per-
conceptualization of the stress process: Essays in ceived social support and psychological distress.
honor of Leonard I. Pearlin (pp. 231252). New York: Journal of Health and Social Behavior, 47, 355372.
Springer. Yang, Y., & George, L. K. (2005). Functional disability,
Wheaton, B., & Clarke, P. (2003). Space meets time: disability transitions, and depressive symptoms in late
Integrating temporal and contextual inuences on life. Journal of Aging and Health, 17, 263292.
Agency Across the Life Course

Steven Hitlin and Hye Won Kwon

Life course studies have moved beyond the stances having a sense of agency explains behav-
abstract agency vs. structure debates that were ior and life course outcomes. Within the study of
commonplace in sociology over the past decades, the life course, agency is fundamental to under-
moving the focus away from the timeless realm standing structurally bounded human beings who
of the abstract (Nisbet 1969) and into the empiri- age within structural and cohort inuences. Its
cal domain. Accordingly, agency, is less a matter existence is not problematic within life course
of philosophical exploration and moreso a con- studies, rather the focus is on its constitution and
struct needing scientic specication and explora- relative inuence on social outcomes.
tion across societal forms and human development. While many aspects of agency-and-structure
Although agency is not a problematic construct in have seen advancements, there are some notable
many social science elds, it is contested in sociol- directions for future inquiry that we will discuss,
ogy, a discipline that largely focuses on the nature below. Scholars have largely measured agency as
of social forces that shape, direct, constrain, and a subjective sense of control, and this conception
enable individual lives. In sociology unlike much has proven fruitful across a number of domains.
of psychology, economics, and political science Yet, we argue, there are additional aspects of
individual volition is often of secondary concern agency that need to be conceptualized and mea-
and even the subject of doubt (Fuchs 2001; Loyal sured, and its links to structure have largely been
and Barnes 2001; Meyer and Jepperson 2000). established within Western cohorts and societies,
Developing a sense of agency appears to be but with little attention to other times and places.
part of human development across cultures The bulk of research on agency equates it with
(Scholz et al. 2002; You et al. 2011) and the concepts like efcacy and personal control. We
capacity for people to inuence their own lives is suggest, however, that other aspects, as captured
a fundamental tenet of life course research (Elder in constructs like optimism, expectations, and
1998). For life course scholars, debating agencys self-identication boundaries, are less theorized
existential nature is less important than exploring and empirically linked with agency. We suggest
its facets and mechanisms, how cultural and that many of the pieces of an inuential life course
structural factors shape it, and in what circum- understanding of agency exist across disciplines
and subdisciplines, though not enough work has
explored its constituent parts across cultures. Life
S. Hitlin (*) H.W. Kwon
course studies hold many of the pieces for an
Department of Sociology, University of Iowa,
Iowa City, IA, USA eventual synthesis, given its fundamentally inter-
e-mail: steven-hitlin@uiowa.edu disciplinary approach (Elder et al. 2003).

Springer International Publishing Switzerland 2016 431


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_20
432 S. Hitlin and H.W. Kwon

This Handbook comes at a propitious time for Hitlin and Long (2009) suggest that this dis-
the eld, and our contribution attempts a global tinction between subjective and objective agency
view of agency in order to identify past successes is important; children are socialized to learn a
and future opportunities for linking individuals sense of control over their surroundings, but
with a proper longitudinal understanding of their likely their subjective sense outweighs (for a
development and societal change. We suggest that time) their actual capacities to inuence the
exploring agency with a cross-cultural lens helps world. That subjective sense likely reects
to identify those theoretical claims about the topic broader structural forces and social locations
that are most useful for motivating future research. beyond the childs control: they cannot affect
We build on the tenet (Elder 1998) that individu- their parents occupation, their location, or the
als make choices in life, though such choices are legal systems in which the family is embedded.
quite bounded (Shanahan and Hood 1998). While Recent psychological work suggests that aspects
some of these debates range into the abstract of this subjective sense, as well as some atten-
(Hitlin and Elder 2007b), this chapter promotes dant skills like delaying gratication, can be
an empirically anchored discussion, drawing on taught (Duckworth et al. 2011); it is easier for
important social psychological treatments of this educators to inuence subjective agency, rather
core, multifaceted construct and offering some than larger social or cultural resources. The
direction for the next stage of exploration. study of the life course is, in part, the study of
how these subjective and objective factors inter-
act to shape that childs future opportunities and
1 Agency: Inuential outcomes.
Denitions An early statement on the nature of agency
was issued by Sewell (1992) in his attempt to
Any discussion of agency should begin with a key restore a notion of agency as contrasted with
distinction between its objective and subjective inuential theories of social structure. Responding
dimensions (Hitlin and Long 2009). Sociological to Giddens (1984) notion that structure and
treatments focus less directly on objective agency presuppose each other and that structure
agency, which would consist of actual skills and both constrains and enables the individual actor,
resources (Clausen 1991; Sewell 1992) that an Sewell highlights that being an agent means to
individual possesses. Some people have more be capable of exerting some degree of control
economic, social, or psychological resources for over the social relations in which one is
encountering lifes vicissitudes and thus possess enmeshed. Agency arises from the actors
more options to deal with potential setbacks. A knowledge of schemas, which means the ability
greater amount of empirical work expressly on to apply them to new contexts (p. 20). This
agency highlights its subjective aspects, peo- enterprise was part of a general trend to counter
ples internal sense that they can inuence their sociological notions that reduced human actors to
lives. In theory, ones subjective sense of agency unrecognizable simplications (Kohn 1989), and
is linked to the objective skills and resources they even those aspects of sociology aimed at linking
possess, but in practice this is not a perfect corre- the individual to social structure the Social
lation. Men, for example, appear to develop attri- Structure and Personality tradition (McLeod and
bution patterns that motivate persistence in the Lively 2003) which were seen as overly deter-
face of failure more than women (Correll 2001), ministic (Elder and Johnson 2002). There is a
meaning that failures are perceived as surmount- theoretical debate over whether agency and struc-
able obstacles, rather than evidence of personal ture should be considered as discrete, de-coupled
failings. This subjective sense of agency has a concepts (Archer 2006; Emirbayer and Mische
number of overlapping conceptualizations we 1998), or largely as two intertwined sides of the
will discuss, in constructs such as personal con- same social process (Cockerham 2005; Hays
trol, mastery, and self-efcacy. 1994).
Agency Across the Life Course 433

Perhaps the preeminent theoretical statement used theoretical tradition that conceptualizes the
of the nature of agency (Emirbayer and Mische self as transcending situations and thus includes a
1998) holds that agency is a temporally embed- temporal element (Flaherty and Fine 2001; Mead
ded process of social engagement, informed by 1932). Recently, Hitlin and Johnson (2015) dem-
the past (in its habitual aspect), but also oriented onstrated that empirical models incorporating a
toward the future (as a capacity to imagine alter- future-oriented aspect of agentic beliefs provide
native possibilities) and toward the present (as a additional purchase in predicting life course out-
capacity to contextualize past habits and future comes beyond earlier, atemporal, measures.
projects within the contingencies of the moment) Thus, our understanding of agency would benet
(p. 963). This model is especially important for from empirical exploration that incorporates
life course theory, given the importance of tempo- actors who utilize beliefs and feelings about the
rality as a basic component of agency. Hitlin and future while acting in the present, especially for
Elder (2007a, b) build on this by anchoring agency actions with potential life course inuences.
within social psychological understandings of the Ultimately, we suggest that agency has served
self, suggesting that different temporal foci lead as a popular and useful theoretical construct due
to four analytical types of agency (existential, in part to its slippery nature (see discussions of
identity, pragmatic and life course), the last other complicated terms like identity (Brubaker
being the most anchored within empirical studies and Cooper 2000) or dignity (Misztal 2013),
of agentic beliefs across the life course, discussed which allows it to serve as a placeholder for
below. They suggest that only the last three types scholars interested in carving out room for indi-
of agency are sociologically meaningful. The vidual volition within a range of social forces
rst, existential agency, involves the most basic (see the discussion by Marshall 2000). We focus
sense of free will and is thus sociologically banal. largely on agency as a variable, not an existential
Identity agency captures the large literature on capacity to act (e.g., Hitlin and Elder 2007b). As
situational action (e.g., Burke and Stets 2009) that a variable, agency captures an individuals ability
focuses on how people interpret and enact impor- to shape their behavior in novel, routine, or life
tant identities and roles. Even this set of behav- course relevant situations. The subjective sense
iors, which sociologists often consider as a xed, of agency changes, though to what extent remains
structural pattern, involves agentic, self-directed an empirical question, in light of individual
action. Relatedly, pragmatic agency deals with development, structural location, cultural beliefs,
those aspects of the person that necessarily act and individual outcomes. Some work, discussed
when the previous form of routine situations below, has attempted to instantiate some of these
break down (e.g., Gross 2009). Life course inuential notions of agency from Sewell
agency draws on the large empirical literature we (1992) and Giddens (1984) through testing
discuss in this chapter, the notion that subjective claims with relevant social psychological mea-
beliefs at one point in time inuence later trajec- sures and samples. We encourage more of this
tories and life course outcomes. within the eld.
This conceptualization of temporality is gain-
ing currency in current theorizing, notably with
Misches (2009) idea of the importance of futures 2 Agency as an Empirical
for determining social action, and Fryes (2012) Construct: Synonyms
discussion of how optimistic views of the future and Future Opportunities
can help determine positive social outcomes
within challenging circumstances. This body of A range of constructs have appeared in the litera-
work is especially relevant to life course theoriz- ture in the last half century that purports to deal
ing, contrasting with classic models of the self with individual volition, broadly understood. In
that focus only on its existence within situations. addition to the constructs most inuential to the
Instead, this work draws on a less commonly study of the life course, discussed in this section,
434 S. Hitlin and H.W. Kwon

aspects of agency appear in ideas like self- associated with declines in mastery (Pearlin et al.
direction (Kohn and Schooler 1982), personal 2007; Schieman and Turner 1998). For some
autonomy (Seeman and Seeman 1983), and inter- people, mastery beliefs in older age are linked to
nal locus of control (Rotter 1966; see Haidt and the perception that one has managed their life
Rodin 1999 for a psychologically focused over- effectively, suggesting that the internal standards
view of some of these constructs). Our discussion for establishing a sense of agency may shift
will focus more depth on those constructs most across the life course (Pearlin et al. 2007).
tied to life course issues, primarily mastery Self-Efcacy. Self-efcacy is one of the most
(Pearlin et al. 1981), self-efcacy (Bandura 1982, commonly utilized constructs referring to agency.
1997; Gecas 1989), personal control (Mirowsky Developed by Albert Bandura (1986, 1992,
and Ross 1998), and planful competence (Clausen 1997), self-efcacy beliefs focus on the notion a
1991, 1993). In general, differences in these mea- person possesses that she has an ability to take
sures involve methodological tweaks, that we actions that generate expected outcomes. It mea-
discuss in turn, but many reviews (Gecas 2003; sures domain-specic cognitive beliefs about
Haidt and Rodin 1999) treat them as largely ones ability to organize and execute courses of
equivalent. action required to produce certain achievements
Work on agency often has a cognitive focus, in (Bandura 1982). Bandura argues that self-efcacy
part due to its measurement through survey items is best measured in specic domains (e.g., sing-
that involve conscious reporting of ones subjec- ing self-efcacy, academic self-efcacy), though
tive beliefs. It is likely that not all aspects of sub- there are later general self-efcacy scales that
jective agency occur consciously. Inuential profess to measure this as a global orientation
notions of how non-conscious processes guide (Chen et al. 2001; Sherer et al. 1982). Evidence
interaction, whether linked to structure (Bourdieu suggests that developing a greater sense of eco-
1977, 1984) or to cultural repertoires (Vaisey nomic self-efcacy, a core self-domain within the
2009), are undoubtedly at play. This is an arena stratication system, predicts educational attain-
for future inquiry, the extent that measures of ment (Grabowski et al. 2001) as well as socioeco-
agency capture non-conscious outlooks, expecta- nomic achievement into adulthood (Lee and
tions, and feelings. Mortimer 2009). Yet it is less often a focus of life
Mastery. Mastery is one of the older socio- course study (though see Gecas 2003 for a
logical attempts to render subjective agency as a review).
measurable construct. It refers to the extent that Personal Control. Personal control is perhaps
an individual believes she can control her life the most commonly used construct within life
chances, rather than having them predetermined course studies. The sense of control involves the
(Pearlin and Schooler 1978). Most often belief that a person can control her life outcomes
employed within the stress literature, mastery is (Mirowsky and Ross 1998; Ross and Mirowsky
an important buffer between external stressors 2013). Individuals with a sense of control feel
and individual health (Pearlin et al. 1981). responsible for their successes and failures and do
Mastery generally increases across the life not attribute those outcomes to external factors.
course, until older age (Pearlin and Schooler The sense of control indicates a generalized belief
1978). With age, people achieve more physical by an individual about her power over life out-
and social mastery over their lives, net of social comes, and is more global than the largely domain-
class. Mastery is inuenced by both social struc- specic construct of self-efcacy. Personal control
tural realities, and the social psychological per- is benecial to the individual, helping her address
ception of those realities (Christie-Mizell and difcult life situations (Conger et al. 2009;
Erickson 2007). In later life, the decline of physi- Wheaton 1980; Berrenberg 1987).
cal skills, cognitive functioning, as well as the Personal control is a generalized, learned trait,
loss of social status and signicant others, is measured through an 8 item-scale, comprising
Agency Across the Life Course 435

positive and negative subscales asking about feel- Lachman and Weaver (1998) nd that those with
ing control over good and bad outcomes. The lower incomes report a lower sense of control.
scale includes items such as I am responsible for Most commonly, a sense of control is strongly
my own successes, I can do just about anything positively linked with education (Mirowsky and
I really set my mind to, My misfortunes are the Ross 2007; Ross and Mirowsky 2013; Schieman
result of mistakes I have made, and I am 2001; Schieman and Plickert 2007). The ratio-
responsible for my failures. This construct was nale for this lies in the content of education,
developed to empirically improve upon the self- which helps people develop useful skills and
efcacy scale as well as the locus of control mea- resources, also leading to better job positions that
sure. Those other measures possessed agreement can, in turn, increase the sense of personal con-
bias, and were more likely to obtain positive trol (Ross 2000). Obtaining more education, due
responses from older adults and the less educated to its strong links with income, means that people
(Mirowsky and Ross 2007). Personal control is have more resources with which to face adversity
acquired and becomes part of ones self-concept, and potentially bad outcomes that would, in turn,
but it reects life circumstances; repeated dif- harm a sense of self control (Mirowsky and Ross
cult life events, for example, can wear away at 1998; Ross and Van Willigen 1997). The litera-
this sense of self (Wolinsky et al. 2003). This ture strongly suggests that this belief is one factor
sense helps people persevere in the face of hard- in reproducing the class structure over time, as
ships (Trommsdorff 1994), especially when individual agentic decisions reect greater or
linked to positive views of the future. lesser beliefs in ones capacity to make poten-
In practice, the sense of control is the con- tially advantageous changes in ones life trajecto-
struct most commonly linked with stratication ries. Future work should focus on disentangling
and life course outcomes; mastery, for example, the sense of control from related issues actual
is most commonly found within the aging litera- capacities, beliefs about the future that also
ture. Its popularity stems from its empirical util- comprise agency as it affects individual and
ity. Having a sense of control allows individuals structural stratication.
to build different life paths compared with others Future Orientations and Planful Competences.
with a lack of personal control: individuals with a Theoretical treatments of agency (Emirbayer and
higher sense of control typically evidence higher Mische 1998; Hitlin and Elder 2007b) have begun
levels of achievement, like higher high school to incorporate temporality as fundamental to a
and college GPAs (Gifford et al. 2006; Perry notion of agency (Pickering 1993), while Hitlin
et al. 2001; You et al. 2011), and better educa- and Johnson (2015) have begun to test these ideas
tional and occupational attainment (Wang et al. empirically. Future orientation (Lewin 1948;
1999). The sense of control is especially linked to Yowell 2000), in general, has recently been high-
education, and this in turn contributes to better lighted as important for understanding social
health as the educated pursue healthier lifestyles actors (Frye 2012; Mische 2009; Tavory and
and seek information that improves life chances Eliasoph 2013), in that it motivates and guides
(Mirowsky and Ross 1998). In addition, a sense action. This orientation can be linked to the notion
of control motivates effort and work engagement, of life projects (Archer 2003; McAdams 2013;
which also translate into better life outcomes Taylor 1989), the ongoing stories that people
(Mirowsky and Ross 2007). develop to make sense of their lives that, in many
The sense of control has a well-established ways, guide interpretation and intentions for future
strong relationship with social class position: activity. Forethought is a core aspect of human
higher class people tend to have a stronger sense agency, understood through a psychological lens
of control than those in lower classes (Kraus (Bandura 2001), with the notion of looking ahead
et al. 2009; Mirowsky and Ross 1998; Ross and being an important force in human cognition
Van Willigen 1997; Ross and Mirowsky 2002). (Seligman et al. 2013). This forward-looking
436 S. Hitlin and H.W. Kwon

orientation contains an emotional element and a Classically, the primary life course rendering
set of cognitive appraisals. Having a positive out- of agency-as-forward-oriented was found in
look, a sense of optimism for the future, has ben- Clausens (1991, 1993) planful competence.
ecial effects on ones life course (Frye 2012; Clausen drew on a set of classic longitudinal
Peterson 2000), and can also serve to buffer indi- studies (including the Berkeley Growth Study
viduals from setbacks (Oyserman et al. 2004). and the Oakland study), developing a measure
In sociology, aspects of looking forward have based on scales and Q-sort techniques to develop
been captured somewhat in the literature on three aspects of planful competence: dependabil-
expectations and aspirations (Morgan 2006), ity, self-reexivity, and self-condence. These
though this literature is rarely linked to life course external evaluations of abilities and skills, once
notions of agency (but see Rudd and Evans 1998). the gold-standard for psychological measure-
Having higher educational expectations in ado- ment, differ from the self-reported beliefs that
lescence is associated with advantaged life course comprise the bulk of current research; personal-
outcomes (Andrew and Hauser 2011; Bozick ity psychologists suggest that we do not know
et al. 2010). These and other expectations involve ourselves as well as others know us. Planful com-
cognitive appraisals of the conditions that a per- petence is not fully an objective measure, but it
son grew up within (Hallerod 2011; Reynolds foregoes the subjectivity most common in those
and Johnson 2011), yet shift in line with life studies linking individual functioning with larger
course achievements or failures (Mortimer et al. social structures.
2002; Reynolds and Baird 2010). Mentally What is notable is that this measure performs
healthy people tend to be positively biased in well as a predictor of later life outcomes (Clausen
their expectations (Taylor and Brown 1988), and 1991, 1993). An adolescent with more planful
this optimism has benecial effects on social life competence makes more advantageous life-
(Andersson 2012b), mental and physical well- decisions and sticks to advantageous trajectories.
being (Carver et al. 2010), and labor market suc- This construct anchors the capacity for making
cesses (Vuolo et al. 2012). Optimism in the health advantageous long-term plans (Shanahan et al.
domain, what Hitlin et al. (forthcoming) term 2003), and was a good predictor of adolescents
health agency, captures a sense of ones capac- who had more stable life courses (e.g., marriages,
ity to handle and recover from illness, and has occupational stability). Historical circumstances
longitudinally advantageous life course outcomes are important, however; planfulness was an
across a range of mental and physical health important life outcome predictor for men who
domains. came of age during the Great Depression, as
Beliefs about ones future are theoretically compared to men who grew up a few years earlier
and empirically distinct from beliefs about per- where this skill had little consequence for out-
sonal control (Epel et al. 1999; Trommsdorff comes (Shanahan et al. 1997).
1994). Structural conditions, for example, might Psychological Cognates. Agency is largely
foster a sense of hopelessness even as an indi- problematic only in structurally oriented disci-
vidual feels a personal sense of efcacy (Young plines. In those that begin at the level of the indi-
2004). Aspirations are likely to change in vidual, notably psychology, agentic capacity has
response to experiences during life transitions as largely been taken for granted, though recent
people assess opportunities and results, with a discoveries of dual-system psychology (Gilbert
general trend toward increasing realism (Heinz 2006; Kahneman 2011; Wilson 2002) have ren-
2014). Recent work (Hitlin and Johnson 2015) dered conscious knowledge of our actions more
suggests that adolescent life expectations are as problematic. Work on agency has not directly
strong predictors of a range of later life outcomes engaged dual-system psychology, the idea that
as the more commonly employed subjective we have a conscious and automatic pair of related
notion of mastery, demonstrating the utility of an brain systems. This view, currently ascendant
expanded notion of agency. within psychology, holds that we have more
Agency Across the Life Course 437

capacity to control the conscious system, while McRae 1985) suggests that one core component,
we are also motivated by an implicit set of under- conscientiousness, might be an individual-level
standings, beliefs, fears, and emotions that often aspect of agency given its focus on goal-direction,
shape our conscious perception (Haidt 2001). focus, persistence and industriousness (Reed
Work in this vein is ltering into sociology (Firat et al. 2013).
and Hitlin 2012; Hitlin 2008; Massey 2002;
Vaisey 2009), and has implications for under-
standings of agency that have not been fully 3 Subjective Agency
articulated. and the Life Course: What
That said, a few psychological constructs We Know and What We Need
overlap with the general sense of agency we have to Know
discussed. Grit is a newly developed construct
used to predict achievement (Duckworth et al. Understanding agency allows life course scholars
2007; Duckworth and Quinn 2009; Reed et al. the capacity to circumscribe conditions under
2013; Singh and Jha 2008; Von Culin et al. 2014). which individual-level factors shape meaningful
Qualitative and quantitative studies have exam- social outcomes. Traditionally, life course studies
ined this trait that is found in high achieving indi- explore issues of mental and physical health,
viduals, dened as passion and perseverance for occupational and educational attainment, family
a long term goal (Duckworth et al. 2007, formation and the intergenerational transmission
p. 1087). Grit contains two components, consis- of advantage or disadvantage. Versions of agency
tency of interest and perseverance of effort; grit are thought to be core for reproducing and gener-
encourages people to maintain a constant focus ating inequality (Grabowski et al. 2001; Shanahan
and work hard to complete a goal, what its proge- and Bauer 2004; Wolinsky et al. 2003). Some
netors argue is distinct from the self-regulatory aspect of the trajectories that connect early life
aspects of the previously discussed life course with these later outcomes is thought to be cap-
constructs (Duckworth and Gross 2014). tured under the umbrella of agency; how much
Individuals high in grit can maintain their deter- can we credit people with successful outcomes,
mination, interest, and effort despite failures or or hold them accountable for less advantageous
adversities, something that seems to develop as outcomes? Some people feel a greater sense of
people age (Duckworth et al. 2007), and is linked control or efcacy, and this sense may simply
to achievement through the amount of deliberate reproduce class advantages, or it may represent
practice that people engage in (Duckworth et al. an individuals motivation for shifting away from
2011). disadvantaged social locations. We know a decent
The practice of goal-setting, itself, has an amount about how subjective agency (as captured
extensive literature and is relevant to the actual in efcacy, personal control and mastery) is
capacity to achieve life aspirations. Carver and shaped and inuences later outcomes within
Scheier (1998) conceptualize this as self- Western contexts; we know much less about
regulation, the capacity to maintain focus on par- other subjective aspects of agency, such as opti-
ticular goals. They argue that human beings are mism and expectations, or how wider cultural
motivated toward goal representations captured forces moderate the relationship between struc-
within mental constructs, and are learned from ture and individual social psychology.
ones culture and social networks. Motivation Social class inuences normative expectations
(Turner 1987) is something captured implicitly for what agentic choices mean to people mak-
with the focus on future orientation, discussed ing choices. For example, Stephens et al. (2007)
above, but needs a great deal of work before it is nd that people from middle class backgrounds
a useful life course construct. Finally, work on make choices that distinguish themselves from
the Big Five personality typology (Costa and others more often than those from working class
438 S. Hitlin and H.W. Kwon

backgrounds, who exhibit a preference for high- efcacy, along with parenting behaviors have
lighting similarity with those around them. been proposed as crucial factors affecting chil-
Classic social psychological results suggest that drens agency development. Family size has a
parents occupational position leads to value negative effect on developing self-efcacy.
socialization in line with these patterns, with Socioeconomic resources are less available and
working class children taught to value confor- events are more unpredictable for children from
mity and middle class children taught a focus on large families compared to children from smaller
autonomy (Kasser et al. 2002; Kohn 1969, 1977; families (Gecas 1989). Rodin (1976) nds that
Kohn and Schooler 1983; Schooler et al. 2004), children from large families tend to report lower
even into adolescence (Kasser et al. 2002). Social levels of feeling of choice and perceived control.
class inuences intertwine with education, the Parents education also inuences childrens
primary mechanism through which class advan- levels of sense of control. Higher class parents
tages translate into individual senses of control. transmit dispositions, competences, and knowl-
Obtaining higher socioeconomic status, along edge that help children to obtain advantages
with higher education, may offer a smoother life (Calarco 2011; Lareau 2003). Parents with a
path, reducing a persons exposure to adversities higher education also encourage children to
in her life that may hurt her beliefs of personal develop abilities and habits (Mirowsky and Ross
control (Mirowsky and Ross 2007). 1998). Lewis et al. (1999) nd that adolescent
Gender and race, core sociological areas of sense of control is positively associated with par-
study, have less global inuence on agency than ents levels of education. Well-educated parents
one might suspect. Females have a lower sense of have a greater sense of personal control and their
control than males (Ross and Mirowsky 2013; children are affected by such beliefs. Parental
Slagsvold and Srensen 2008) although often efcacy about inuencing childrens academic
this difference turns out to be statistically insig- performance increases academic aspirations and
nicant (Mirowsky and Ross 2013). Womens self-efcacy (Bandura et al. 2001). Parents
sense of control declines more than mens as they beliefs that they can successfully carry out child-
age, showing a gender gap in the sense of control rearing activities can also affect childrens self-
that increases with age (Ross and Mirowsky efcacy by shaping actual parenting practices
2002). Further, this growing gender gap in the (Schneewind 1995). In addition, better-educated
sense of control is greater for older cohorts than parents tend to focus more on childrens indepen-
for younger cohorts (Ross and Mirowsky 2002), dence, self-direction, and personal responsibility
suggesting that cohort differences in work expe- while less-educated working class parents tend to
rience contribute to this observed gap. If so, we emphasize conformity and obedience (Lewis
may continue to see change in this pattern in et al. 1999; Pearlin and Kohn 1966). In accor-
future cohorts. Blacks in the United States report dance with this tendency, well-educated parents
a lower sense of control than their white counter- transmit these attitudes regarding agency to their
parts in every age group, net of other socioeco- children (Lewis et al. 1999). At least within a
nomic factors (Bruce and Thornton 2004; Shaw Western context, the family, as a prime socializa-
and Krause 2001). Pearlin et al. (2007) suggest tion agent, is a fundamental link in shaping sub-
that race effects disappear, however, when con- jective agency that partially serves to reproduce
trolling for education. This suggests that agency, social structural advantage.
as currently measured, is not a key factor explain- Close friends also have a signicant inuence
ing some of sociologys major culprits for trans- on ones development of agency (e.g., Adler and
mitting inequality across the life course. Alder 1998). Adolescents tend to be friends with
Family is one of the most important develop- those who are similar to them (Ryan 2000), and
mental loci for developing ones volition, agency, students who observe similar friends achieve-
or self-direction (Gecas 1989). Family size, as ment have stronger beliefs that they can also
well as parents educational level and self- achieve (Deci et al. 2006; Schunk and Meece
Agency Across the Life Course 439

2006). In contrast, experiences of peer pressure Work and occupational characteristics, often
are negatively associated with self-efcacy correlated with education, provide crucial set-
among adolescents (Kiran-Esen 2012). tings that inuence ones sense of agency.
School experiences help adolescents establish Building on the well-established positive rela-
beliefs about their agency (Schunk and Meece tionship between education and agency, research-
2006). Formal educational institutions are places ers have also focused on employment as the
where socially desirable and legitimate cultural underlying mechanism that connects education
knowledge about attitudes and behaviors are and personal control (Ross and Van Willigen
transmitted (Bourdieu 1984; Calarco 2014; 1997; Schieman and Plickert 2007). Individuals
Gordon 1989; Pollack and Thoits 1989). with higher education levels tend to participate in
Children not only learn how to critically and the labor force and obtain higher status occupa-
logically think and solve problems, but also tions (Blau and Duncan 1967; Grusky and
develop attitudes and skills to face problems DiPrete 1990; Schieman and Plickert 2007;
with condence and perseverance through Warren et al. 2002), that, in turn, tend to provide
diverse activities in schools (Mirowsky and Ross greater job authority, employment security, and
1998). The learning sequence of the school envi- exibility in schedule, creativity and autonomy
ronment makes children face progressively chal- (Ross 2000; Ross and Mirowsky 1992; Ross and
lenging sets of problems and, step-by-step, Van Willigen 1997; Schieman and Plickert 2007).
children learn that things that look initially dif- Drawing on Marxs notion of workers alienation,
cult can be construed and learned via practic- researchers nd that individuals with non-routine,
ing. This training and experience of autonomous, creative, and problem-solving work
accomplishment help them to build a sense of develop greater senses of agency over their lives
control over their lives (Ross and Mirowsky (Ross 2000; Ross and Mirowsky 1992; Schieman
2007). Lewis, Ross, and Mirowsky (1999) report and Plickert 2007). Causality is difcult to disen-
that dropping out of high school has a signi- tangle, as these higher status occupations also
cant, negative effect on a persons sense of con- offer better economic rewards, thereby reducing
trol, controlling for their initial levels, exposure to adversities that can challenge per-
demonstrating the negative importance of this sonal control beliefs (Mirowsky and Ross 2007).
potential life course transition. Class inuences on the intergenerational shaping
Schooling further encourages people to of attitudes, as mentioned previously, lead work-
acquire strong beliefs that they can achieve better ing class parents to be more likely to teach con-
outcomes by using their capacities and resources formity as opposed to autonomy taught by
(Ross and Mirowsky 2013; Schieman and parents who themselves have greater occupa-
Plickert 2007). School is not the only force tional autonomy, a nding with cross-cultural
increasing a sense of control, but it is vital. evidence (Kohn and Schooler 1983; Kohn and
Mirowsky (2013) modied his original notion of Schoenbach 1993; Kohn et al. 1986).
agency as built up in school, used up in old age Culture is a crucial factor that inuences indi-
(p. 415) to encompass a growing set of ndings viduals senses of agency. The majority of
that personal control increases across early adult- agency studies have focused on the United States
hood and middle-age, regardless of additional specically, or Western nations generally,
schooling. Educational institutions teach valu- extrapolating issues of the nature of the person
able skills and abilities benecial for yielding from an admittedly narrow, privileged sample
better achievement in later life, but aging, itself, (Henrich et al. 2010a, b). Building on the classic
contributes to people obtaining a sense of control distinction between collectivistic and individual-
in life until older age. Schooling, however, pro- istic culture, cross-psychological approaches
vides a focused location enabling individuals to typically suggest understanding Western cul-
obtain useful resources and socioeconomic posi- tures as individual-focused cultures, with a focus
tions in society. on the self as independent from others (Kondo
440 S. Hitlin and H.W. Kwon

1990). East Asian cultures are generally consid- culture, obscuring important variation by class,
ered collectivistic cultures, in which individuals race, religion, and other factors that might plau-
see themselves as interdependent with other peo- sibly contribute to people within a particular
ple (Cross and Gore 2012; Markus and Kitayama society developing different subjective orienta-
1991; Matsumoto et al. 2008; Snibbe and Markus tions. Relying on the East/West simple dualism
2005; Triandis 1995). A sense of agency is much may obscure within-culture variation, and thus
more intertwined with what Elder (1994, 1998) lead researchers to focus too much on between-
refers to as linked lives; to be an agent is to society differences and less on within-society
take important others into account as an aspect patterns (see Kwon 2013). Complicating this
of the self. Because of the emphasis on collectiv- dichotomy, Vandello and Cohen (1999) show
istic culture in East Asian nations, the concept of the United States, perhaps the paradigmatic
individual agency has received less attention example of an individualistic nation and thus
among researchers; delimiting the possibility lauding agencys importance, has extensive
that agency is fundamentally experienced intra-national variations in individualistic ori-
differently. entations. Yamawaki (2012) also nds that indi-
A great deal of cross-cultural research in viduals who reside in larger Japanese cities
East Asian countries has reafrmed the link such as Tokyo report a lower collectivistic ori-
between Western cultures and individualistic entation when compared to residents in rural
orientations versus East Asian cultures and col- Japanese regions. Researchers nd that younger
lectivistic orientations. Several studies have Koreans diverge from collectivistic orientations
found that East Asians have a lower sense of of the older generation, reporting more individ-
control than Westerners, and therefore linked ualistic orientations (Cha 1994; Hyun 2001).
low sense of control among Asians with collec- These studies suggest that overlooking varia-
tivistic cultures generally (OConner and tion within cultures to conform to one index
Shimizu 2002; Smith et al. 1995; You et al. within individualism or collectivism may
2011). In a review of cross-cultural studies of obscure within-cultural variation critical to
the locus of control, Smith et al. (1995) con- understanding the role of agency, and how
clude that East Asians, particularly Japanese, changing global forces contribute to changing
tend to report lower sense of control compared cohort beliefs in the importance of subjective
to others. You et al. (2011) report Asian- control over ones life. We do not know enough
American students show lower senses of con- about how much subjective agencys contribu-
trol compared to American students from tion to later life outcomes is moderated by the
non-Asian groups. cultural systems of belief surrounding those
Although these cross-cultural studies have individual subjective appraisals, nor how sub-
illuminated the important role of culture in jective agency, itself, differs across nations and
shaping agency, they have not yet provided a cultural forms. Nor do we have a great deal of
useful correction to the western bias that suf- condence that our measures capture agency
fuses much sociological discussion of the topic. well for those who conceptualize the self differ-
Cross-cultural research often accepts this ently. Future research would benet life course
individualistic-collectivistic dichotomy at face studies by exploring longitudinal data on the
value, suggesting that it is a universal categori- sense of control in different countries, and
zation of societies, and thus becomes a source specifying measurement to include issues like
for explaining cultural differences in perceived expectations and aspirations as core agency
agency. This simple cultural dichotomy, how- components. The study of such cultural differ-
ever, has come up against criticism in recent ences might draw on improved methodology
years. Cross-cultural research tends to largely that can better distinguish age, period, and
presuppose an internally consistent societal cohort effects (Zheng et al. 2011).
Agency Across the Life Course 441

While cultures differ on the relative impor- factors that contribute to the generally accepted
tance they extend toward the construct of agency, inverted U-curve, with agentic beliefs increasing
it appears that even collectivist cultures still through adulthood and declining in older age.
produce people who claim a sense of agency (Li Inuential longitudinal work suggests that much
2006). Counter to prevalent notions about a low of the cross-sectional association between age
sense of agency among East Asian population, Li and the sense of personal control is, in fact, due to
(2006) nds Chinese adolescents have particu- cohort differences in education levels (Ross and
larly high senses of agency in the domain of Mirowsky 2013; Schieman 2001; Wolinsky and
learning. Using Pearlins 7-item Mastery Scale, Stump 1996). That is, lower levels of personal
Jang et al. (2009) explore a sense of mastery in control in older age are in part due to lower levels
Korean American elders and nd nancial inse- of education in earlier cohorts (Wolinsky and
curity hurts their feelings of mastery. While East Stump 1996; see Ross and Mirowsky 2013 for
Asian populations have received less attention by further discussion). Slagsvold and Srensen
researchers, even these so-called collectivistic (2013) also distinguish between aging and cohort
cultures in which individuals independence and effects using data from two waves (collected in
autonomy are assumed to be less socially valued, 20022003 and 20072008) of The Norwegian
ndings are in line with studies on non-Asian study of Life course, Aging, and Generation
populations. Culture, age and historical cohort (NorLAG), and conclude that the observed distri-
intersect to shape subjective senses of agency, yet bution of the sense of control by age should be
research on these linkages is in its infancy. attributed to cohort effects and not to aging
In addition, structure interacts with culture effects (p. 302). Beyond education, however,
even in Western nations to shape agency; analyz- there are several important factors that show
ing a Norwegian adult sample, Slagsvold and aging inuencing the sense of control: impact of
Srensen (2013) suggest that age-related losses physical impairment, retirement, loss of partner,
in mastery are inuenced by cultural and struc- etc (Mirowsky and Ross 1998, 2007; Rodin 1986;
tural differences across countries. Whereas previ- Schieman 2001). Future work should attempt to
ous studies in the United States found that the further disentangle these issues, and adjudicate
sense of control declines around age 60 how much of the general inverted U-shaped curve
(Mirowsky and Ross 2007), Slagsvold and of increasing agency through adulthood until a
Srensen show that the decline in the sense of noticeable drop in older age is shaped by cultural
control begins around age 75 in Norway. They and structural factors, and how much simply
argue that differences can be attributed to differ- occurs through the biological aging process.
ent welfare systems between the two countries,
implying that more generous welfare provisions
in Norway reduce elders uncertainty and provide 4 Conclusion: General
greater predictability about their future lives. Conclusions and Some
This in turn protects a sense of mastery. A pro- Continuing Holes
vocative thesis worth pursuing would be that in Knowledge
culture is becoming less inuential in shaping
subjective agency as national economic systems Agency is considered a core sociological con-
develop, such that structural inuences might struct and one of the main principles of life course
become the dominant forces shaping perceptions studies (Elder 1994). Its existence is likely more
of agency. problematic within sociology than within most
Aging, itself, offers complication in under- other academic disciplines, or the wider popular
standing general principles about the develop- consciousness, as American society is largely
ment and loss of agency. We know less about how dened by the ideology that hard work will pay
agentic beliefs are shaped in adolescence (Conger off, and that everybody has a fair chance to
et al. 2009), and whether it is age, itself, or other achieve the American Dream. Specifying agency
442 S. Hitlin and H.W. Kwon

has fueled decades of theoretical discussions and and socially shaped skill-sets, that are linked with
arguments, with various positions assigning more higher achievement, better mental and physical
or less agency on the part of individuals acting health, and optimal functioning (Vaillant 2012).
within a web of social structures and forces. Sociologists are often more comfortable discuss-
Understanding the relationship between an indi- ing social problems than making prescriptions
vidual and the structured life course pathways for what the alternative might be, including posi-
that are presented to her within a given society tive aspects of human functioning that psychol-
makes getting an empirical handle on agency ogy has embraced much more fully (Dahlsgaard
important, if not fully successful, within the eld et al. 2005; Haidt 2006) in the recent past. As a
of life course studies. eld, we might focus more directly on optimal
We focus here primarily on the subjective levels of agency, both subjective and objective,
sense of agency that people develop, and how that might serve to inform policy makers to a
that inuences their life courses. In general, hav- greater extent.
ing more of a sense of agency appears to be ben- Beyond this general thesis, that having more
ecial for mental and physical health (Gecas of a sense of agency is associated with positive
1989; Mirowsky and Ross 1989; Wheaton 1980). outcomes we do not know if, like self-esteem
Having a sense of control over ones life is a posi- (Baumeister et al. 1996), there can be a dark
tive force on well-being, though, as we suggest in side to having too much agency. There are a
this chapter, having that subjective sense does not great deal of questions remaining specifying
necessarily mean that a person actually has more agencys contours, development, and specic
agentic capacity. In this sense, agency can be a contribution across social domains.
positive illusion (Taylor and Brown 1988, 1994), Contentiousness, perhaps related to aspects of
contributing to optimistic outlooks that, in turn, subjective agency, are largely positive for life
are positive for life outcomes (Andersson 2012a, course outcomes (Shanahan et al. 2014), but also
b; Hitlin and Elder 2007a; Peterson 2000). can have a dark side when people high in this
Subjective agency, measured with a range of con- orientation encounter failure. This is an underex-
structs reviewed here, appears to buffer the inu- plored aspect; Alexander (1993) suggests we too
ence of various life course stressors. often focus on agencys heroic aspects.
Measuring objective agency, the actual skills A primary issue to be resolved involves the
and capacities that contribute to advantaged nature of the subjective construct, itself. For the
social achievement, points to a potentially impor- purposes of this review (see also Gecas 2003), we
tant area for future inquiry. Value judgments collapse different traditions into the general sense
about particular traits and skills that might that people have that they can exert inuence on
improve outcomes can quickly bump up against their lives. This treats the construct as a global,
issues that were once captured within the culture cross-situational sense that is in many ways at
of poverty debate, with a concern about blaming odds with Banduras (1997) domain-specic
the victim. The concern is that pointing out dis- notion of self-efcacy. Some scholars suggest
tinct skills that lead to advantageous outcomes, that it makes sense to talk about agency, or their
such as Clausens underutilized treatment of preferred indicator of this concept, as something
planful competence, holds those who achieve that transcends the various roles and identities
lower stratication outcomes as responsible for they play. The psychological cognate, self-
their own disadvantage. Sometimes discussions efcacy, is most accurately assigned to specic
of the ways lower SES leads to different skills/ tasks, roles, and outcomes. Just because some-
orientations has been interpreted as holding peo- body has an exceedingly high sense of
ple in these circumstances causally culpable for self-efcacy as a singer or a programmer does
their stratication position. We advocate cautious not necessarily mean they will have condence in
engagement with those psychological proles, other domains.
Agency Across the Life Course 443

In addition, agency is sometimes discussed as stage, general structure, and facets of individual
a motivational force, other times as a dependent functioning can start to be developed. For exam-
indicator of status, other times as a buffer from ple, Shanahan et al. (2014) offer the Life Course
social stresses; across studies, then, agency oper- of Personality Model, offering theoretical speci-
ates like another multi-faceted construct, self- city about when and how personality traits
esteem (Cast and Burke 2002). As a result, agency interact with structural advantages at different
has been used variously as a cause, an effect, a periods of the life course, intersecting with linked
mediator and a moderator. While there may never lives and situational contexts to inuence particu-
be exact agreement over its properties, we do sug- lar outcomes. Future work can move beyond their
gest researchers be clear about how they are using focus on health to a range of stratication out-
their indicators, a problem made even more com- comes to demonstrate the conditions under which
plicated by the fact that many of the scholars cited agency is particularly inuential in helping indi-
here might be surprised to know their work has viduals shape their life courses.
been subsumed under the umbrella of agency.
We follow Ross and Mirowskys (2013) sugges-
tions that more work using representative samples Bibliography
is needed to disentangle a number of issues relat-
Adler, P. A., & Adler, P. (1998). Peer power: Preadolescent
ing to aging, gender, and the sense of control, as
culture and identity. New Brunswick: Rutgers
well as issues of racial difference. They also call University Press.
for more exploration on the relationship of per- Alexander, J. C. (1993). More notes on the problem of
sonal control, their measure of subjective agency, agency: A reply. Schweizerische Zeitschrift fr
Soziologie/Revue suisse de sociologie, 19, 501506.
and anxiety and anger.
Andersson, M. (2012a). Dispositional optimism and the
Its subjective nature, as we outline here, sug- emergence of social network diversity. Sociological
gests a variety of questions about how agency is Quarterly, 53(1), 92115.
formed and maintained across situations, rela- Andersson, M. A. (2012b). Identity crises in love and at
work dispositional optimism as a durable personal
tionships, and the life course. Sociological treat-
resource. Social Psychology Quarterly, 75(4),
ments of cognition (Howard and Renfrow 2003) 290309.
and emotion (Stets 2003; Turner and Stets 2006) Andrew, M., & Hauser, R. M. (2011). Adoption?
are likely implicated, as well as the psychology Adaptation? Evaluating the formation of educational
expectations. Social Forces, 90(2), 497520.
of cognitive processing, biases, and illusions. To
Archer, M. S. (2003). Structure, agency, and the internal
the extent that ones beliefs about themselves and conversation. Cambridge: Cambridge University
their futures are important, the nature of these Press.
beliefs, their cultural content, and their inuence Archer, M. (2006). Making our way through the world:
Human reexivity and social mobility. Cambridge:
over situated action is an open question. Hitlin
Cambridge University Press.
and Elder (2007b) suggest, for example, that the Bandura, A. (1982). The self and mechanisms of agency.
form of agency present in maintaining an estab- In J. Suls (Ed.), Psychological perspectives on the self
lished social identity differs from the sort of (pp. 340). Hillsdale: Lawrence Erlbaum.
Bandura, A. (1986). Social foundations of thought and
agency referred to at key potential turning points
action. Englewood Cliffs: Prentice Hall.
in the life course. While life course researchers Bandura, A. (1992). Exercise of personal agency through
have accomplished a good deal explicating vari- the self-efcacy mechanism. In R. Schwarzer (Ed.),
ous facets of this construct that inuence indi- Self-efcacy: Thought control of action (pp. 338).
Washington: Hemisphere Publishing Corporation.
vidual life outcomes, there is a good deal of
Bandura, A. (1997). Self-efcacy: The exercise of control.
bridge-building to be accomplished with psycho- New York: Freeman.
logical theories that might spell out more com- Bandura, A. (2001). Social cognitive theory: An agentic
pletely how and when these senses develop and perspective. Annual Review of Psychology, 52, 126.
Bandura, A., Barbaranelli, C., Caprara, G. V., & Pastorelli,
inuence action. A large enough body of work
C. (2001). Self-efcacy beliefs as shapers of childrens
has been developed such that stronger claims aspirations and career trajectories. Child Development,
about the interaction of situation, life course 72(1), 187206.
444 S. Hitlin and H.W. Kwon

Baumeister, R. F., Smart, L., & Boden, J. M. (1996). The Clausen, J. A. (1991). Adolescent competence and the
relation of threatened egotism to violence and aggres- shaping of the life course. American Journal of
sion: The dark side of self-esteem. Psychological Sociology, 96, 805842.
Review, 103(1), 533. Clausen, J. A. (1993). American lives: Looking back at
Berrenberg, J. L. (1987). The belief in personal control children of the great depression. New York: Free
scale: A measure of God-mediated and exaggerated Press.
control. Journal of Personality Assessment, 51(2), Cockerham, W. C. (2005). Health lifestyle theory and the
194206. convergence of agency and structure. Journal of
Blau, P. M., & Duncan, O. D. (1967). The American occu- Health and Social Behavior, 46, 5167.
pational structure. New York: Wiley. Collins, R. (2004). Interaction ritual chains. Princeton:
Bourdieu, P. (Ed.). (1977). Outline of a theory of practice. Princeton University Press.
Cambridge: Cambridge University Press. Conger, K. J., Williams, S. T., Little, W. M., Masyn, K. E.,
Bourdieu, P. (Ed.). (1984). Distinction: A social critique & Shebloski, B. (2009). Development of mastery dur-
of the judgement of taste. London: Routledge. ing adolescence: The role of family problem-solving.
Boyce, C. J., Wood, A. M., & Brown, G. D. A. (2010). Journal of Health and Social Behavior, 50(1), 99114.
The dark side of conscientiousness: Conscientious Correll, S. J. (2001). Gender and the career choice pro-
people experience greater drops in life satisfaction fol- cess: The role of biased self-assessments. American
lowing employment. Journal of Research in Journal of Sociology, 106(6), 16911730.
Personality, 44, 535539. Costa, P. T., & McRae R. R. (1985). NEO personality
Bozick, R., Alexander, K., Entwisle, D., Dauber, S., & inventory Form R.
Kerr, K. (2010). Framing the future: Revisiting the Cross, S. E., & Gore, J. S. (2012). Cultural models of the
place of educational expectations in status attainment. self. In M. R. Leary & J. P. Tangney (Eds.), Handbook
Social Forces, 88(5), 20272052. of self and identity (2nd ed., pp. 587614). New York:
Brubaker, R., & Cooper, F. (2000). Beyond Identity. Guilford Press.
Theory and Society, 29(1), 147. Dahlsgaard, K., Peterson, C., & Seligman, M. E. P. (2005).
Bruce, M. A., & Thornton, M. C. (2004). Its my world? Shared virtue: The convergence of valued human
Exploring black and white perceptions of personal strengths across culture and history. Review of General
control. The Sociological Quarterly, 45(3), Psychology, 9(3), 203213.
597612. Deci, E. L., La Guardia, J. G., Moller, A. C., Scheiner,
Burke, P. J., & Stets, J. E. (2009). Identity theory. M. J., & Ryan, R. M. (2006). On the benets of giving
New York: Oxford University Press. as well as receiving autonomy support: Mutuality in
Cahill, S. E. (1998). Toward a sociology of the person. close friendships. Personality and Social Psychology
Sociological Theory, 16(2), 131148. Bulletin, 32(3), 313327.
Calarco, J. M. C. (2011). I Need Help! social class and Duckworth, A., & Gross, J. J. (2014). Self-control and
childrens help-seeking in elementary school. grit: Related but separable determinants of success.
American Sociological Review, 76(6), 862882. Current Directions in Psychological Science, 23(5),
Calarco, J. M. C. (2014). Coached for the classroom par- 319325.
ents cultural transmission and childrens reproduction Duckworth, A. L., & Quinn, P. D. (2009). Development
of educational inequalities. American Sociological and validation of the Short Grit Scale (GRITS).
Review, 79(5), 10151037. Journal of Personality Assessment, 91(2), 166174.
Carver, C. S., & Scheier, M. (1998). On the self- Duckworth, A. L., Peterson, C., Matthews, M. D., &
regulation of behavior. Cambridge: Cambridge Kelly, D. R. (2007). Grit: Perseverance and passion for
University Press. long-term goals. Journal of Personality and Social
Carver, C. S., Scheier, M. F., & Segerstrom, S. C. (2010). Psychology, 92(6), 10871101.
Optimism. Clinical Psychology Review, 30(7), Duckworth, A. L., Kirby, T. A., Tsukayama, E., Berstein,
879889. H., & Ericsson, K. A. (2011). Deliberate practice
Cast, A. D., & Burke, P. (2002). A theory of self-esteem. spells success why grittier competitors triumph at the
Social Forces, 80(3), 10411068. national spelling bee. Social Psychological and
Cha, J.-H. (1994). Aspects of individualism and collectiv- Personality Science, 2(2), 174181.
ism in Korea. In U. Kim, H. C. Triandis, C. Kagitcibasi, Dunn, R. G. (1997). Self, identity, and difference: Mead
S.-C. Choi, & G. Yoon (Eds.), Individualism and col- and the poststructuralists. The Sociological Quarterly,
lectivism: Theory, method, and applications (pp. 157 38(4), 687705.
174). Thousand Oaks: Sage. Elder, G. H., Jr. (1994). Time, human agency, and social
Chen, G., Gully, S. M., & Eden, D. (2001). Validation of a change: Perspectives on the life course. Social
new general self-efcacy scale. Organizational Psychology Quarterly, 57(1), 415.
Research Methods, 4(1), 6283. Elder, G. H., Jr. (1998). The life course and human devel-
Christie-Mizell, C. A., & Erickson, R. J. (2007). Mothers opment. In R. M. Lerner & W. Damon (Eds.),
and mastery: The consequences of perceived neigh- Handbook of child psychology (Theoretical Models of
borhood disorder. Social Psychology Quarterly, 70(4), Human Development, Vol. 1, pp. 939991). New York:
340365. Wiley.
Agency Across the Life Course 445

Elder, G. H., Jr., & Johnson, M. K. (2002). The life course Haidt, J. (2001). The emotional dog and its rational tail: A
and aging: Challenges, lessons, and new directions. In social intuitionist approach to moral judgement.
R. A. Settersten Jr. (Ed.), Invitation to the life course: Psychological Review, 108(4), 814834.
Toward new understandings of later life (pp. 4981). Haidt, J. (2006). The happiness hypothesis: Finding mod-
Amityville: Baywood. ern truth in ancient wisdom. New York: Basic Books.
Elder, G. H., Jr., Johnson, M. K., & Crosnoe, R. (2003). Haidt, J., & Rodin, J. (1999). Control and efcacy as
The emergence and development of life course theory. interdisciplinary bridges. Review of General
In J. T. Mortimer & M. J. Shanahan (Eds.), Handbook Psychology, 3(4), 317337.
of the life course (pp. 319). New York: Kluwer. Hallerod, B. (2011). What do children know about their
Emirbayer, M., & Mische, A. (1998). What is agency? futures: Do childrens expectations predict outcomes
American Journal of Sociology, 103(4), 9621023. in middle age? Social Forces, 90(1), 6584.
Epel, E. S., Bandura, A., & Zimbardo, P. G. (1999). Hays, S. (1994). Structure and agency and the sticky prob-
Escaping homelessness: The inuences of self- lem of culture. Sociological Theory, 12(1), 5772.
efcacy and time perspective on coping with Heinz, W. (2014). Did the great recession affect young
Homelessness. Journal of Applied Social Psychology, peoples aspirations and reinforce social inequality?
29(3), 575596. Longitudinal and Life Course Studies, 5(2), 189198.
Firat, R., & Hitlin, S. (2012). Morally bonded and Henrich, J., Heine, S. J., & Norenzayan, A. (2010a). The
bounded: A sociological introduction to neurology. weirdest people in the world? Behavioral and Brain
Advances in Group Processes, 29, 165199. Sciences, 33(2/3), 175.
Flaherty, M., & Fine, G. A. (2001). Present, past, and Henrich, J., Heine, S. J., & Norenzayan, A. (2010b). Most
future: Conjugating George Herbert Meads perspec- people are not WEIRD. Nature, 466(7302), 2929.
tive on time. Time & Society, 10(23), 147161. Hitlin, S. (2008). Moral selves, evil selves: The social psy-
Franzese, A. T. (2013). Motivation, motives, and individ- chology of conscience. New York: Palgrave Macmillan.
ual agency. In J. Delamater & A. Ward (Eds.), Hitlin, S., & Elder, G. H., Jr. (2007a). Agency: An empiri-
Handbook of social psychology (pp. 281318). cal model of an abstract concept. In R. Macmillan
New York: Springer. (Ed.), Advances in life course research: Constructing
Frye, M. T. (2012). Bright futures in Malawis New dawn: adulthood: Agency and subjectivity in adolescence
Educational aspirations as assertions of identity. and adulthood (pp. 3367). Amsterdam: Elsevier/JAI
American Journal of Sociology, 117(6), 15651624. Press.
Fuchs, S. (2001). Beyond agency. Sociological Theory, Hitlin, S., & Elder, G. H., Jr. (2007b). Understanding
19(1), 2440. Agency: Clarifying a curiously abstract concept.
Gecas, V. (1989). The social psychology of self-efcacy. Sociological Theory, 25(2), 170191.
Annual Review of Sociology, 15, 291316. Hitlin, S. & Johnson, M. K. (2015). Reconceptualizing
Gecas, V. (2003). Self-agency and the life course. In J. T. agency within the life course: The power of looking
Mortimer & M. Shanahan (Eds.), Handbook of the life ahead. American Journal of Sociology 120(5),
course (pp. 369388). New York: Kluwer. 14291472.
Giddens, A. (Ed.). (1984). The constitution of society: Hitlin, S., & Long, C. (2009). Agency as a sociological
Introduction of the theory of structuration. Berkeley: variable: A preliminary model of individuals, situa-
University of California Press. tions, and the life course. Sociology Compass, 3(1),
Gifford, D. D., Briceno-Perriott, J., & Mianzo, F. (2006). 137160.
Locus of control: Academic achievement and reten- Hitlin, S., Erickson L., & Brown J. S. (forthcoming).
tion in a sample of university rst-year students. Agency and mental health: A transition to adulthood
Journal of College Admission, 191, 1825. paradox. Society and Mental Health.
Gilbert, D. (2006). Stumbling on happiness. New York: Holstein, J. A., & Gubrium, J. F. (Eds.). (2000). The self
Alfred A. Knopf. we live by narrative identity in a postmodern world.
Gordon, S. (1989). Socialization of childrens emotions: New York: Oxford University Press.
Toward a social constructionist theory. In C. Saarni & Howard, J. A., & Renfrow, D. G. (2003). Social cognition.
P. Harris (Eds.), Childrens understanding of emotion In J. Delamater (Ed.), Handbook of social psychology
(pp. 3857). Cambridge: Cambridge University Press. (pp. 259282). New York: Kluwer.
Grabowski, L. J. S., Call, K. T., & Mortimer, J. T. (2001). Hyun, K. J. (2001). Sociocultural change and traditional
Global and economic self-efcacy in the educational values: Confucian values among Koreans and Korean
attainment process. Social Psychology Quarterly, American. International Journal of Intercultural
64(2), 164179. Relations, 25, 203229.
Gross, N. (2009). A pragmatist theory of social mecha- Jang, Y., Chiriboga, D. A., Lee, J., & Cho, S. (2009).
nisms. American Sociological Review, 74(3), Determinants of a sense of mastery in Korean
358379. American elders: A longitudinal assessment. Aging &
Grusky, D. B., & DiPrete, T. A. (1990). Recent trends in Mental Health, 13(1), 99105.
the process of stratication. Demography, 27(4), Kahneman, D. (2011). Thinking, fast and slow. New York:
617637. Farrar, Straus, and Giroux.
446 S. Hitlin and H.W. Kwon

Kasser, T., Koestner, R., & Lekes, N. (2002). Early family Lewis, S. K., Ross, C. E., & Mirowsky, J. (1999).
experiences and adult values: A 26 year prospective Establishing a sense of personal control in the transi-
longitudinal study. Personality and Social Psychology tion to adulthood. Social Forces, 77(4), 15731599.
Bulletin, 28(6), 826835. Li, J. (2006). Self in learning: Chinese adolescents goals
Kiran-Esen, B. (2012). Analyzing peer pressure and self- and sense of agency. Child Development, 77(2),
efcacy expectations among adolescents. Social 482501.
Behavior and Personality: An International Journal, Loyal, S., & Barnes, B. (2001). Agency as a Red
40(8), 13011309. Herring. Philosophy of the Social Sciences, 31(4),
Kohn, M. L. (1969). Class and conformity. Homewood: 507524.
Dorsey. Markus, H., & Kitayama, S. (1991). Culture and the self:
Kohn, M. L. (1977). Class and conformity: A study in val- Implications for cognition, emotion, and motivation.
ues, with a reassessment. Chicago: University of Psychological Review, 98, 224253.
Chicago Press. Marshall, V. W. (2000). Agency, structure, and the life
Kohn, M. L. (1989). Social structure and personality: A course in the era of reexive modernization. In The life
quintessentially sociological approach to social psy- course in the 21st century. Washington, DC.
chology. Social Forces, 68(1), 2633. Marshall, V. W. (2005). Agency, events, and structure at
Kohn, M. L., & Schoenbach, C. (1993). Social stratica- the end of the life course. Advances in Life Course
tion, parents values and childrens values. In D. Krebs Research, 10, 5791.
& P. Schmidt (Eds.), New directions in attitude mea- Massey, D. S. (2002). A brief history of human society:
surement (pp. 118151). New York: Walter de Gruyter. The origin and role of emotion in social life. American
Kohn, M. L., & Schooler, C. (1982). Job conditions and Sociological Review, 67(1), 129.
personality: A longitudinal assessment of their recip- Matsumoto, D., Yoo, S. H., & Fontaine, J. (2008).
rocal effects. American Journal of Sociology, 87(6), Mapping expressive differences around the world: The
12571286. relationship between emotional display rules and indi-
Kohn, M. L., & Schooler, C. (Eds.). (1983). Work and per- vidualism versus collectivism. Journal of Cross-
sonality: An inquiry into the impact of social strati- Cultural Psychology, 39(1), 5574.
cation. Norwood: Ablex Pub. Corp. McAdams, D. P. (2013). The psychological self as actor,
Kohn, M. L., Slomczynski, K. M., & Schoenbach, C. agent, and author. Perspectives on Psychological
(1986). Social stratication and the transmission of Science, 8(3), 272295.
values in the family: A cross-national assessment. McLeod, J. D., & Lively, K. J. (2003). Social structure and
Sociological Forum, 1, 73102. personality. In J. DeLamater (Ed.), Handbook of social
Kondo, D. (1990). Crafting selves: Power, gender, and psychology (pp. 77102). New York: Kluwer.
discourses of identity in a Japanese workplace. Mead, G. H. (Ed.). (1932). The philosophy of the present.
Chicago: University of Chicago Press. Chicago: Open Court Publishing.
Kraus, M. W., Piff, P. K., & Keltner, D. (2009). Social Meyer, J. W., & Jepperson, R. L. (2000). The Actors of
class, sense of control, and social explanation. Journal modern society: The cultural construction of social
of Personality and Social Psychology, 97(6), agency. Sociological Theory, 18(1), 100120.
9921004. Mirowsky, J. (2013). Depression and the sense of control:
Kwon, H. W. (2013). Bridging social stratication and Aging vectors, trajectories, and trends. Journal of
cross-cultural approaches to the sense of control: Health and Social Behavior, 54(4), 407425.
Effect of age and social class in the United States and Mirowsky, J., & Ross, C. E. (1989). Social causes of
South Korea. Masters thesis, Department of psychological distress. New York: Aldine de
Sociology, University of Iowa, IA. Gruyter.
Lachman, M. E., & Weaver, S. L. (1998). The sense of Mirowsky, J., & Ross, C. E. (1998). Education, personal
control as a moderator of social class differences in control, lifestyle and health: A human capital hypoth-
health and well-being. Journal of Personality and esis. Research on Aging, 20(4), 415449.
Social Psychology, 74(3), 763773. Mirowsky, J., & Ross, C. E. (2007). Life course trajecto-
Lareau, A. (2003). Unequal childhoods: Class, race, and ries of perceived control and their relationship to edu-
family life. Berkeley: University of California Press. cation. American Journal of Sociology, 112(5),
Laub, J. H., & Sampson, R. J. (2005). Human agency in 13391382.
the criminal careers of 500 Boston Men, circa 1925 Mische, A. (2009). Projects and possibilities: Researching
1995. In Agency and human development under condi- futures in action. Sociological Forum, 24(3),
tions of social change. Jena, Germany 694704.
Lee, J. C., & Mortimer, J. T. (2009). Family socialization, Misztal, B. A. (2013). The idea of dignity: Its modern sig-
economic self-efcacy, and the attainment of nancial nicance. European Journal of Social Theory, 16(1),
independence in early adulthood. Longitudinal and 101121.
Life Course Studies, 1(1), 4562. Morgan, S. L. (2006). Expectations and aspirations. In
Lewin, K. (1948). Resolving social conicts; selected G. Ritzer (Ed.), The Blackwell encyclopedia of sociol-
papers on group dynamics. Oxford: Harper. ogy (pp. 15281531). New York: Wiley-Blackwell.
Agency Across the Life Course 447

Mortimer, J. T., Zimmer-Gembeck, M. J., Holmes, M., & Ross, C. E., & Mirowsky, J. (2002). Age and the gender
Shanahan, M. J. (2002). The process of occupational gap in the sense of personal control. Social Psychology
decision making: Patterns during the transition to Quarterly, 65(2), 125145.
adulthood. Journal of Vocational Behavior, 61(3), Ross, C. E., & Mirowsky, J. (2013). The sense of personal
439465. control: Social structural causes and emotional conse-
Nisbet, R. A. (1969). Social change and history: Aspects quences. In C. S. Aneshensel, J. C. Phelan, &
of the western theory of development. New York: A. Bierman (Eds.), The handbook of the sociology of men-
Oxford University Press. tal health (2nd ed., pp. 379402). New York: Springer.
OConnor, D. B., & Shimizu, M. (2002). Sense of per- Ross, C. E., & Van Willigen, M. (1997). Education and the
sonal control, stress, and coping style: A cross-cultural subjective quality of life. Journal of Health and Social
study. Stress and Health, 18, 173183. Behavior, 38, 275297.
Oyserman, D., Bybee, D., Terry, K., & Hart-Johnson, T. Rotter, J. B. (1966). Generalized expectancies for internal
(2004). Possible selves as roadmaps. Journal of versus external control of reinforcement. Psychological
Research in Personality, 38(2), 130149. Monographs, 80(1), 128.
Pearlin, L. I., & Kohn, M. L. (1966). Social class, occupa- Rudd, P., & Evans, K. (1998). Structure and agency in
tion, and parental values: A cross-national study. youth transitions: Student experiences of vocational
American Sociological Review, 31(4), 466479. further education. Journal of Youth Studies, 1(1),
Pearlin, L. I., & Schooler, C. (1978). The structure of cop- 3962.
ing. Journal of Health and Social Behavior, 19(1), Ryan, A. M. (2001). The peer group as a context for the
221. development of young adolescent motivation and
Pearlin, L. I., Menaghan, E. G., Lieberman, M. A., & achievement. Child Development, 72(4), 11351150.
Mullan, J. T. (1981). The stress process. Journal of Schieman, S. (2001). Age, education, and the sense of
Health and Social Behavior, 22(4), 337356. control a test of the cumulative advantage hypothesis.
Pearlin, L. I., Nguyen, K. B., Schieman, S., & Milkie, Research on Aging, 23(2), 153178.
M. A. (2007). The life-course origins of mastery Schieman, S., & Plickert, G. (2007). Functional limita-
among older people. Journal of Health and Social tions and changes in levels of depression among older
Behavior, 48(2), 164179. adults: A multiple-hierarchy stratication perspective.
Perry, R. P., Hladkyj, S., Pekrun, R. H., & Pelletier, S. T. The Journals of Gerontology. Series B, Psychological
(2001). Academic control and action control in the Sciences and Social Sciences, 62(1), S36S42.
achievement of college students: A longitudinal eld Schieman, S., & Turner, H. A. (1998). Age, disability, and
study. Journal of Educational Psychology, 93(4), the sense of mastery. Journal of Health and Social
776789. Behavior, 39, 169186.
Peterson, C. (2000). The future of optimism. American Schneewind, K. A. (1995). Impact of family processes on
Psychologist, 55(1), 4455. control beliefs. In A. Bandura (Ed.), Self-efcacy in
Pickering, A. (1993). The mangle of practice: Agency and changing societies (pp. 114148). NY: Cambridge
emergence in the sociology of science. American University Press.
Journal of Sociology, 99(3), 559589. Scholz, U., Doa, B. G., Sud, S., & Schwarzer, R. (2002).
Pollack, L. H., & Thoits, P. (1989). Processes in emotional Is general self-efcacy a universal construct?
socialization. Social Psychology Quarterly, 54, 2234. Psychometric ndings from 25 countries. European
Reed, J., Pritschet, B. L., & Cutton, D. M. (2013). Grit, Journal of Psychological Assessment, 18(3),
conscientiousness, and the transtheoretical model of 242251.
change for exercise behavior. Journal of Health Schooler, C., Mulatu, M. S., & Oates, G. (2004).
Psychology, 18(5), 612619. Occupational self-direction, intellectual functioning,
Reynolds, J. R., & Baird, C. L. (2010). Is there a downside and self-directed orientation in older workers:
to shooting for the stars? Unrealized educational Findings and implications for individuals and societ-
expectations and symptoms of depression. American ies. American Journal of Sociology, 110(1), 161197.
Sociological Review, 75(1), 151172. Schunk, D. H., & Meece, J. L. (2006). Self-efcacy devel-
Reynolds, J. R., & Johnson, M. K. (2011). Change in the opment in adolescences. In F. Pajares & T. C. Urdan
stratication of educational expectations and their (Eds.), Adolescence and education: Self-efcacy and
realization. Social Forces, 90(1), 85110. adolescents (pp. 7196). Charlotte: Information Age
Rodin, J. (1976). Density, perceived choice, and response Publishing.
to controllable and uncontrollable outcomes. Journal Seeman, M., & Seeman, T. E. (1983). Health behavior and
of Experimental Social Psychology, 12(6), 564578. personal autonomy: A longitudinal study of the sense
Rodin, J. (1986). Aging and health: Effects of the sense of of control in illness. Journal of Health and Social
control. Science, 233(4770), 12711276. Behavior, 24, 144160.
Ross, C. E. (2000). Occupations, jobs, and the sense of Seligman, M. E. P., Railton, P., Baumeister, R. F., &
control. Sociological Focus, 33(4), 409420. Sripada, C. (2013). Navigating into the future or
Ross, C. E., & Mirowsky, J. (1992). Households, employ- driven by the past. Perspectives on Psychological
ment, and the sense of control. Social Psychology Science, 8(2), 119141.
Quarterly, 55(3), 217235.
448 S. Hitlin and H.W. Kwon

Sewell, W. H., Jr. (1992). A theory of structure: Duality, Stets, J. E. (2003). In J. Delamater (Ed.), Handbook of
agency, and transformation. American Journal of social psychology (pp. 309335). New York: Kluwer.
Sociology, 98, 129. Tavory, I., & Eliasoph, N. (2013). Coordinating futures:
Shanahan, M. J., & Bauer, D. J. (2004). Developmental Toward a theory of anticipation. American Journal of
properties of transactional models: The case of life Sociology, 118(4), 908942.
events and mastery from adolescence to young adult- Taylor, C. (Ed.). (1989). Sources of the self: The making of
hood. Development and Psychopathology, 16, the modern identity. Cambridge, MA: Harvard
10951117. University Press.
Shanahan, M. J., & Hood, K. E. (1998). Adolescents in Taylor, S. E., & Brown, J. D. (1988). Illusion and well-
changing social structures: Bounded agency in life being: A social psychological perspective on mental
course perspective. In R. Silbereisen & E. Crockett health. Psychological Bulletin, 103(2), 193210.
(Eds.), Negotiating adolescence in times of social Taylor, S. E., & Brown, J. D. (1994). Positive illusions and
change: Cross-national perspectives on developmen- well-being revisited: Separating fact from ction.
tal processes and social intervention (pp. 123134). Psychological Bulletin, 116(1), 2127.
Cambridge: Cambridge University Press. Triandis, H. C. (1995). Individualism and collectivism.
Shanahan, M. J., Elder, G. H., Jr., & Miech, R. A. (1997). Boulder: Westview.
History and agency in mens lives: Pathways to Trommsdorff, G. (1994). Future time perspective and con-
achievement in cohort perspective. Sociology of trol orientation: Social conditions and consequences.
Education, 70(1), 5467. In Psychology of future orientation. 20-616
Shanahan, M. J., Hofer, S. M., & Miech, R. A. (2003). (pp. 3962). Lublin: Wydawnictwo Towarzystwa
Planful competence, the life course, and aging: Naukowego Katolickiego Uniwersytetu Lubelskiego.
Retrospect and prospect. In S. H. Zarit, L. I. Pearlin, & Turner, J. H. (1987). Toward a sociological theory of moti-
K. W. Schaie (Eds.), Personal control in social and life vation. American Journal of Sociology, 52, 1527.
course contexts (pp. 189211). New York: Springer. Turner, J. H., & Stets, J. E. (2006). Sociological theories
Shanahan, M. J., Hill, P. L., Roberts, B. W., Eccles, J., & of human emotions. Annual Review of Sociology, 32,
Friedman, H. S. (2014). Conscientiousness, health and 2552.
aging: The life course of personality model. Vaillant, G. E. (2012). Triumphs of experience. Cambridge,
Developmental Psychology, 50(5), 14071425. MA: Harvard University Press.
Shaw, B. A., & Krause, N. (2001). Exploring race varia- Vaisey, S. (2009). Motivation and justication: A dual-
tions in aging and personal control. Journal of process model of culture in action. American Journal
Gerontology: Social Science, 56B(2), 119124. of Sociology, 114(6), 16751715.
Sherer, M., Maddux, J. E., Mercandante, B., Prentice- Vandello, J. A., & Cohen, D. (1999). Patterns of individu-
Dunn, S., Jacobs, B., & Rogers, R. W. (1982). The alism and collectivism across the United States.
self-efcacy scale: Construction and validation. Journal of Personality and Social Psychology, 77(2),
Psychological Reports, 51(2), 663671. 279292.
Singh, K., & Duggal Jha, S. (2008). Positive and negative Von Culin, K. R., Tsukayama, E., & Duckworth, A. L.
affect, and grit as predictors of happiness and life sat- (2014). Unpacking grit: Motivational correlates of
isfaction. Journal of the Indian Academy of Applied perseverance and passion for long-term goals. The
Psychology, 34(4045), 4045. Journal of Positive Psychology, 9(4), 306312.
Slagsvold, B., & Srensen, A. (2008). Age, education, and Vuolo, M., Staff, J., & Mortimer, J. T. (2012). Weathering
the gender gap in the sense of control. The International the great recession: Psychological and behavioral tra-
Journal of Aging and Human Development, 67(1), jectories in the transition from school to work.
2542. Developmental Psychology, 48(6), 17591773.
Slagsvold, B., & Srensen, A. (2013). Changes in sense of Wang, L.-Y., Kick, E., Fraser, J., & Burns, T. J. (1999).
control in the second half of life: Results from a 5-year Status attainment in America: The roles of locus of con-
panel study. The International Journal of Aging and trol and self-esteem in educational and occupational
Human Development, 77(4), 289308. outcomes. Sociological Spectrum, 19(3), 281298.
Smith, P. B., Trompenaars, F., & Dugan, S. (1995). The Warren, R., Hauser, R., & Sheridan, J. (2002).
rotter locus of control scale in 43 countries: A test of Occupational stratication across the life course:
cultural relativity. International Journal of Psychology, Evidence from the Wisconsin longitudinal study.
30(3), 377400. American Sociological Review, 67, 323346.
Snibbe, A. C., & Markus, H. R. (2005). You cant always Wheaton, B. (1980). The sociogenesis of psychological
get what you want: Educational attainment, agency, disorder: An attributional theory. Journal of Health
and choice. Journal of Personality and Social and Social Behavior, 21, 100124.
Psychology, 88, 703720. Wilson, T. D. (2002). Strangers to ourselves: Discovering
Stephens, N. M., Markus, H. R., & Townsend, S. S. M. the adaptive unconscious. Cambridge, MA: Belknap
(2007). Choice as an act of meaning: The case of Press of Harvard University Press.
social class. Journal of Personality and Social
Psychology, 93(5), 814830.
Agency Across the Life Course 449

Wolinsky, F. D., & Stump, T. E. (1996). Age and the sense ment. The Journal of Educational Research, 104(4),
of control among older adults. The Journals of 253266.
Gerontology, 51B(4), S217S220. Young, A. A. (2004). The minds of marginalized black
Wolinsky, F. D., Wyrwich, K. W., Kroenke, K., Babu, men: Making sense of mobility, opportunity, and
A. N., & Tierney, W. M. (2003). 9-11, personal stress, future life chances. Princeton: Princeton University
mental health, and sense of control among older Press.
adults. The Journals of Gerontology Series B: Yowell, C. M. (2000). Possible selves and future orien-
Psychological Sciences and Social Sciences, 58(3), tation: Exploring hopes and fears of Latino boys
S146S150. and girls. The Journal of Early Adolescence, 20(3),
Yamawaki, N. (2012). Within-culture variations of collec- 245280.
tivism in Japan. Journal of Cross-Cultural Psychology, Zheng, H., Yang, Y., & Land, K. C. (2011). Variance
43(8), 11911204. function regression in hierarchical age-period-
You, S., Hong, S., & Ho, H.-z. (2011). Longitudinal cohort models. American Sociological Review,
effects of perceived control on academic achieve- 76(6), 955983.
Cognitive Development
and the Life Course: Growth,
Stability and Decline

Duane F. Alwin, Jason R. Thomas,


and Linda A. Wray

1 Introduction izing the interplay between these developmental


processes and what we take to be the institution-
Life course theorists have persuasively argued alized life course, that is, the normative structures
that the life course perspective is useful for the that move human lives through a sequence of
study of human development over the entire life positions, activities, and roles, around which peo-
span (e.g., Elder 1997a, b; Elder and Shanahan ples lives are constructed (Kohli 2007). Although
2006). Yet, there are actually very few studies life course theorists often speak about human
that take the long view and look at human development in the abstract, they rarely focus on
development over the entire life span using the the ways in which life course events, transitions,
concepts, premises and principles of the life and trajectories actually impinge on and are
course perspective.1 This chapter focuses on a set affected by developmental outcomes over the life
of key human factors important in the study of span. In this chapter we focus specically on
human livesthe domain of cognitive function cognitive function (CF)also known as intelli-
(CF)across the entire life span, from early to gence, intellectual abilities, cognitive develop-
later life experiences, using a life course perspec- ment, or just cognitiona multidimensional
tive. Life span theorists have focused on the array of abilities that are examined at various life
development, maintenance, and decline in CF stages in diverse literatures.
over the entire life span, and have made substan- A signicant degree of recent attention has
tial contributions (e.g., Baltes et al. 1999). Here been paid to CF by the National Research
we add importantly to this work by conceptual- Council, among other government agencies,
attending equally to cognitive development in
1 early life(see the NRC report titled Neurons
One exception is Newcombs Bennington studies (see
Alwin et al. 1991), which studied political attitudes over and Neighborhoods, edited by Shonkoff and
nearly 50 years in the lives of their respondents. Another Phillips 2000)and cognitive aging in later life
exception is the Terman study of gifted children (Friedman (see the NRC report titled The Aging Mind, edited
and Kern 2014; Kern and Friedman 2008).
by Stern and Carstensen 2000). The purpose of
D.F. Alwin (*) J.R. Thomas this chapter is to provide a conceptual framework
Sociology and Demography, The Pennsylvania State
based in the life course perspective that integrates
University, University Park, PA, USA
e-mail: dfa2@psu.edu the consideration of within-person change in CF
over the entire life span with an understanding of
L.A. Wray
Biobehavioral Health, The Pennsylvania State the role of the institutionalized life course in
University, University Park, PA, USA affecting that change. This exercise necessitates

Springer International Publishing Switzerland 2016 451


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_21
452 D.F. Alwin et al.

knowledge of multiple literatures that speak to meaning as the concepts of life span, life cycle, or
these issues, including child development, educa- life history, and should be kept distinct (Elder
tion, social stratication, demography, social ger- 2000; Alwin 2012). The life course consists of a
ontology, and aging, in order to illuminate the complex set of interlocking trajectories, or
social pathways for cognitive development across pathways, over the life span that are marked by
the entire life span. sequences of events, transitions and exposures
The chapter is organized into three parts. First, across several biologically- and socially-dened
we introduce our understanding of the life course life stages (or phases) that impact upon the devel-
perspective and how it may improve on the life- opment of individual lives (see Featherman 1983;
span developmental approach to the study of Elder et al. 2003; Alwin 2012). Such life course
CF. Second, we provide a foundational frame- processes are dened by institutional frame-
work for conceptualizing the concept of within- worksreferred to here as the institutionalized
person change using a latent variable approach, life course (see Kohli 2007)that dene the
and how the life course perspective can be used transition from one life stage to the next, e.g. the
to develop event-centered models for understand- transition to full-time school attendance (e.g.,
ing the heterogeneity of within-person change. Entwistle et al. 2003), or the transition from
Specically, we emphasize the relevance of latent school to work (e.g., Hogan 1981; Hogan and
difference score models as an embodiment of the Astone 1986; Shanahan 2000), or the transition
study of the growth, maintenance, and decline in to retirement (e.g., Henretta 2003). Life course
CF over the life span, and discuss how this analysis typically focuses on the nature and
approach can be integrated with the life course determinants of those transitions, their timing,
analysts interest in events, transitions, and links to events and exposures in other life stages,
changes in the social environment. The third and consequences of specic pathways for out-
major section of the chapter considers four major comes of human development (e.g. CF). The life
areas where our approach to the study of within- course perspective assumes that the social path-
person change can be applied in understanding ways marked by these transitions and the social
the nature of the social processes surrounding life processes that surround them occur across spe-
course transitions and human development: (1) cic life stages that dene the entire life span and
early child cognitive development and the transi- are embedded in social institutions and subject to
tion to school, focusing on factors that contribute historical variation and change.
to successful transitions, and how transitions As argued by Elder (1975, 1985) and others,
themselves contribute to further change; (2) the an understanding of the life course builds on (but
transition to adulthood and midlife, in which CF does not replace) several strong theoretical tradi-
in adolescence leads to major inuences on edu- tions that help give it meaning: (1) the life-span
cational achievement, occupational success and developmental approach (e.g., Baltes 1987,
CF in adulthood; (3) the potential for changes in 1997); (2) the understanding of the human life
CF during midlife; and (4) the applicability of a cycle, and its biological underpinnings (e.g.
life course framework in the study of transitions Hogan 2000; Carey 2003); and (3) the sociologi-
in later life and how theoretically CF has a role to cal study of age stratication and the institution-
play as both consequence and cause of these alized life course (Riley 1987; Kohli 2007). We
transitions. take full advantage of the theoretical richness
offered by these foundational literatures and rely
on an integrated framework that permits us to
2 The Life Course acknowledge the roles of both ontogenesis (bio-
logical development) and sociogenesis (environ-
We begin by clarifying what the life course is ment inuences) in cognitive change over the life
(and is not). First, it is important to note that the span (Featherman and Lerner 1985).
concept of life course does not have the same
Cognitive Development and the Life Course: Growth, Stability and Decline 453

A key element offered by the life course per- graphical and historical time are completely con-
spective, which gives it a unique place in the founded, and are inseparable, i.e. within-person
theoretical panoply of developmental theories, is change is both a reection of history and biogra-
its emphasis on institutionalized age-graded tran- phy.] We focus on the analysis of within-person
sitions, which shape sequences of roles, activi- change and on the processes that determine its
ties, and social pathways across the life span. The direction and its rate, and how these serve as both
focus on social pathways allows research to con- causes and consequences of life course transi-
centrate on certain transition points between life tions. Specically, we raise the issue of whether
cycle stages, such as the transition to school, the research has effectively addressed these matters.
transition to adulthood, or the transition to retire- On the latter point, we believe there is too little
ment in older age. The study of the interplay research that addresses cognitive development
between the institutionalization of life stages and within a life course framework. Baltes (1987,
human development of cognitive capacities offers p. 613) does make an effort to address contextual
a unique opportunity to take seriously a compre- and structural factors in CF, in his consideration
hension of the life course that ignores neither the of contextual and historical inuences; unfortu-
ontogeny of development nor the institutionaliza- nately, these attempts do not adequately address
tion of life stages (aka the age grading of expe- the conceptual and methodological frameworks
riences), as articulated in the work of life course needed to analyze within-person change over
theorists. The life course conception of an age- time.
graded sequence of statuses implies the existence We focus on cognitive function in a broad and
of resources and opportunities that inhere in multidimensional manner, based on Cattells
making the transitions, and one important indi- (1963, 1971a) and Wechslers (1952, 1958) work
vidual resource worthy of the serious attention from the 1950s and 1960s that distinguished two
we intend to give it, is cognitive or intellectual inter-related components of CFuid and crys-
functioning. We focus here on the interplay tallized abilities. This distinction is important for
between life course transitions and the continuity two key reasons: (1) each has a potentially differ-
and change in CF. The approach we propose for ent relationship to age; and (2) each has a differ-
the study of CF can be applied in the study of ent connection to the schooling process. Fluid
other developmental trajectories where social intelligence is conceptualized as the capacity
context is relevant. for insight into complex relations independent
of the sensory or cultural area in which the tests
are expressed (emphasis in the original).
3 The Development Crystallized intelligence has its origins in experi-
of Cognitive Function ence but is not expected to be independent of
other capacities because it arises as the result of
The primary focus of this chapter is on cognitive the investment of uid intelligence, over the
function (CF), considered broadly to refer to the years, in whatever higher-level cultural skills the
human ability to manipulate the environment in individual is exposed to (Cattell 1971a, p. 13;
such ways as to solve both simple and complex see also Cattell 1971b; Denny 1982; Horn 1968,
problems posed by that (primarily external) envi- 1976, 1994; Horn and Cattell 1967; Horn and
ronment, and specically within-person change Donaldson 1980). The distinctions drawn by
in CF across biographical/historical time. This Cattell have subsequently been mentioned by
concept of within-person change is called by a others using somewhat different terminology,
variety of namesaging, human development, which attests in part to their importance. Baltes
maturation, gains/lossesbut the essential fea- (1987) used the terms mechanics and prag-
ture is that changes occur within persons over matics to refer to uid and crystallized abilities
biographical/historical time. [Note that for a (see also Salthouse 1991, p. 34; see also Salthouse
given person, the demographic reality is that bio- 1999). Contemporary models of intelligence now
454 D.F. Alwin et al.

include a more complex set of domains (see Schaie 1988), and that the declines increase even
Carroll 1996, 1998; Woodcock 1994), but the more rapidly after age 80 (Scherr et al. 1988). In
fundamental distinction between uid and crys- contrast, measures of crystallized or education-
tallized intelligence continues to guide current based abilities, such as vocabulary recognition or
thinking (see Flanagan et al. 2000; McGrew verbal reasoning, decline later and less predictably
1997; McGrew and Flanagan 1998). For this rea- (Park 1999; Park et al. 1996; Alwin and
son, any argument concerning the development McCammon 1999, 2001).
of cognitive function must take seriously the role Most people agree that there are no pure mea-
of education and schooling, which is largely sures of either CF component and that any given
ignored in Baltes (1987, 1997), but which a life measure may constitute some combination of
course perspective can add. both. To the extent that available cognitive mea-
A life course perspective on cognitive devel- sures depend on education, there may be a con-
opment across the life spanin early childhood, founding of age-related trajectories with
late adolescence, midlife, retirement and older schooling, given the well-known pattern of inter-
agemust consider the master processes of cohort differences in levels of schooling. Note
educational achievement, work and retirement, that Cattells classic prediction about the age-
about which much has been written. For our pres- related patterns of uid and crystallized abilities
ent purposes, we arbitrarily break the institution- is about levels and trajectories of cognitive func-
alized life course into four distinct age-graded tion; however, it does not address the question of
periods of life: (1) preparation for school (across the relationship between environmental change
all stages of schooling); (2) preparation for work and individual change.
life, which includes the transition to adulthood
(Shanahan 2000); (3) development of a work
career (Schooler 1987); and (4) retirement and 4 Principles of Life Span
older age (Henretta 2003). The consideration of Development
these transitions is reinforced by the technologi-
cal and institutional factors that govern the edu- There are several highly-inuential and widely-
cational and occupational systems, shape the cited theoretical statements about human cogni-
pursuits of individuals, and constrain the oppor- tive development in adulthood that stress the
tunities and resources inhering in the connections multi-causal nature of development (e.g., Baltes
of individual lives to these organizational struc- 1987, 1997; Baltes et al. 1999; Baltes and Mayer
tures. With the premises of the life course per- 1999; Featherman and Lerner 1985; Schaie
spective in hand (see Alwin 2012), we believe we 2005). Development is embedded in multiple
can develop a useful framework for investigating contexts in which the ontogeny of development
the interplay between human abilities (CF) and interacts with the social environment, within a set
the life course. of interconnected social and cultural settings
With regard to the average trajectory of cogni- (e.g., Bronfenbrenner 1979). In the past 30 years,
tive abilities, early life is marked by a period of the sub-discipline of human development became
growth, up until growth peaks, and patterns then aware of the importance of the life span devel-
level off (Cattell 1971a, b). Subsequently, uid opmental perspective, and this work has guided
abilities decline systematically with age, and substantial amounts of research on cognitive
crystallized abilities increase slightly or other- change in the older years. Thus, human develop-
wise remains relatively stable with age (see Horn ment across the entire life span is conceptualized
and Cattell 1967). The available literature on as multi-causal, multi-dimensional, and multi-
cognitive aging generally conrms Cattells directional change.
model. Research shows that there are signicant The main theoretical argument resulting from
cognitive declines in uid or process-based abili- this perspective can be found in Baltes 1996
ties (e.g., memory) well past age 65 (Hertzog and address to the American Psychological Association,
Cognitive Development and the Life Course: Growth, Stability and Decline 455

in which he draws upon both evolutionary and need or demand for culture (Baltes 1997,
ontogenetic ideas about development, and theo- p. 368). By culture, Baltes refers to the entirety
rizes about the contribution of biological-genetic of psychological, social, material, and symbolic
factors and the social-cultural arrangements to (knowledge-based) resources that promote
human development (Baltes 1997). He posits human development. In the case of cognitive
three foundational (constraining) principles of the functioning, this refers to the set of factors we
life span architecture of human ontogeny, which discuss below under the headings of social envi-
can be applied to the phenomenon of cognitive ronment and social structure and all of the
development. He rst argues that there is a negative factors contained therein (e.g. social resources,
relationship between the benets resulting from social capital, and social status). The argument is
evolutionary selection and chronological age. that with age, largely due to principles of evolu-
Evolutionary pressures have emphasized reproduc- tionary and biological ontogeny, there is an
tion rather than longevity (e.g., Finch and Kirkwood increased demand or need for culture to play a
2000). Humans have evolved to reproduce them- role. Social environmental resources play in
selves, and experience the greatest protection from increasingly important role in cognitive function-
their biological and genetic architecture in the ing in older age (see, e.g. Rowe and Kahn 1998;
younger years. Consequently, the human genome Baltes and Baltes 1990). As evidence of this
in older ages contains an increasingly larger num- need, one can point to improvements in life
ber of deleterious genes and dysfunctional gene expectancy in industrialized societies over the
expressions (Baltes 1997, p. 367) compared to twentieth century.
younger ages. In short, reproduction tness Finally, the third foundational principle of life
trumps longevity tness. The existence of the span development is that over the course of time,
dementias, such as Alzheimers Disease, illustrate there is an age-related loss in the effectiveness
his point (see Martin et al. 1996), where the chronic or efciency of cultural factors and resources,
condition does not manifest itself until age 70 and largely conditioned by the negative biological
prevalence rates increase exponentially thereafter trajectory of the life course (Baltes 1997,
(see Brookmeyer and Gray 2000; Brookmeyer p. 368). Further, there continues to be a great deal
et al. 1998; Kawas and Brookmeyer 2001). of exibility or plasticity in development in older
Reproductive tness, in the broad evolutionary age, and the extent of this potential for change
schema, is neutral with respect to the diseases of may in fact be greater than normally believed, but
old age. the scope of plasticity of the human organism
In addition to the adaptive conditions afforded declines with age (Baltes 1997, p. 368). He uses
humans by the realities occasioned by their evo- cognitive learning in old age as an example of
lutionary circumstances, humans also experience this phenomenon. He cites the common nding
the phenomena associated with biological (or of declining speed of processing in older age,
ontological) processes of aging, which produce a (e.g. Salthouse 1996), which strongly suggests
number of age-associated mechanisms involving that when it comes to high levels of performance,
biological loss. As Baltes noted, evolution and older adults may never be able to reach the same
biology are not good friends of old age (Baltes levels of functioning as younger adults even after
(1997, p. 368)biological losses amplify the extensive training (Baltes 1997, pp. 368369).
evolutionary neglect of old age. Further, an We discuss the differing degrees of stability or
unnished architecture of life span develop- plasticity in cognitive functioning in a later
ment is created by the course of evolutionary section.
selection and biological ontogeny. These pro- These considerations lead to three major
cesses work together to produce a parallel phe- issues drawn from the life-span developmental
nomenon, the second cornerstone of a life span perspective: (1) the treatment of CF across the
architecture of human ontogenesis (or develop- life span in terms of successive periods of growth,
ment), namely an age-related increase in the maintenance, and decline; (2) the role of the
456 D.F. Alwin et al.

environment in promoting cognitive change dur- Our interest here is in developing this model
ing these periods; and (3) the degree of stability within a framework of within-person change and
or plasticity of CF over these periods. We take up specify it as a growth model that can be theoreti-
each one of these topics in the following cally formulated across the entire life span within
sections. a life course perspective. At the early stages of the
life span, the environment promotes change, and
differences in the experience with the environ-
4.1 The Architecture ment are most potent in terms of producing indi-
of Development: vidual differences during the period of rapid
From the Ground Up changes at the maturational level. In Fig. 2 we
depict Blooms (1964) empirically-based devel-
Baltes (1997) architecture of the life span dynam- opmental curve that species the nature of aggre-
ics between biology and culture forms a general gate growth trajectories, but the limits of
framework within which human development is environmental impact as well. This goes beyond
embedded and the contours for his selective opti- the baseline model shown earlier (see Fig. 1) in
mization with compensation (SOC) model (see that it explicitly includes the role of the environ-
Baltes and Baltes 1980, 1990; Staudinger et al. ment. The gradually rising curve of the level of
1995). SOC involves three general functions or development that changes as a function of age in
outcomes of development: (1) growth, (2) mainte- Fig. 2 is intended to represent the average trajec-
nance, and (3) regulation of loss. According to tory of the development of cognitive function.
Baltes and his colleagues, a systematic shift in the This curve exhibits its highest rates of change at
relative allocation of resources to these three the youngest ages and as depicted here the rate of
functions occurs over the lifespan. In his words: change slows as age increases. The shaded area
In childhood, the primary allocation is directed surrounding this curve is intended to represent the
toward growth; during adulthood, the predominant potential for differences in environments to affect
allocation is toward maintenance and recovery the rate of change. The theoretical potential for
(resilience). In old age, more and more resources the inuence of environmental factors is repre-
are directed toward regulation or management of
loss. (Note) that the reallocation of resources sented by the wideness of the shaded area in the
toward maintenance of functioning and regulation gure, which in this case narrows with increasing
of loss is facilitated by the tendency of individuals age. The greater the potential for environmental
to prefer avoidance of loss over enhancement of effects, the less stable are the individual differ-
gains. (Baltes 1997, p. 370)
ences in the attributes of individuals; whereas the
We do not dwell on the intricacies of Baltes less potential for environmental change, the more
SOC framework here, as that is stated eloquently highly stable are individual differences.
in the above citations, except to note that the the- In Blooms (1964, p. vii) words, (v)ariations
ory is not accompanied by the specication of in the environment have greatest quantitative
methods for studying the joint inuences of mat- effect on a characteristic at its most rapid period
urational and environmental factors in cognitive of change and least effect on the characteristic
function across the life span. At a general level, during the least rapid period of change. This pic-
the SOC framework is useful in terms of orient- ture of development in the early years is assumed
ing one to the general considerations necessary in Blooms scheme to be very general, in the
for studying cognitive aging as an adaptive pro- sense that it applies to aspects of physical devel-
cess involving cognitive, neural, and environ- opment, e.g. height and weight, as well as cogni-
mental resources. It lacks specicity, however, tive abilities and personality. Blooms curve is
regarding research strategies for studying the also consistent with the models of growth pre-
changing role of the social environment in the sented in Fig. 1, including Cattells (1963) two
development, maintenance and decline in cogni- inter-related components of cognitive function
tive functioning across the life span. uid and crystallized abilities.
Cognitive Development and the Life Course: Growth, Stability and Decline 457

Fluid (Mechanics)
Crystallized (Pragmatics)
Performance

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
Age

Fig. 1 Aggregate patterns of change in uid (mechanics) and crystallized (pragmatics) cognitive function over the life
span (Adapted from Cattell 1971b, p. 12)

Fig. 2 Developmental curve


and the limits of environmental
inuence (From Bloom 1964)

Blooms (1964) empirical work does not go development is possible if the forces reecting
beyond age 18, but he speculated that for most environmental change are powerful enough to
models of cognitive development, in terms an produce cognitive change, although he surmised
absolute scale, it is assumed that intelligence from the available evidence that during most of
remains constant after about age 20, but he con- adult life environments are decidedly static and
ceded that at the time he did his study there was that massive levels of intellectual stability is typi-
very little evidence available on adult intelligence cal (pp. 8090). Ultimately, this pattern is not
(1964, p. 80). He speculated that the environ- necessarily normal, or inevitable, since the level
ments in which people live after young adulthood of stability is a reection of the relationship of
would likely determine the nature of further intel- the person and the environment (Ryder 1965;
lectual development. He suggested that further Alwin 1994, pp. 146147).
458 D.F. Alwin et al.

There are three ways in which environments reect the net effects of within-person change,
can inuence cognitive development via life but they do not directly measure within-person
course processes: via its inuence on levels of change (see Alwin and McCammon 1999, 2001).
cognitive functioning, trajectories of change, and In the case of longitudinal designs, the focus is on
its inuence on rates of within-person cognitive gross age-related changes, where the hypothesis
change (see Fig. 2). A trajectory of development of aging may be only one possible interpretation
is involved as one traces levels of functioning of net levels of age-related within-person change.
through time, and it is therefore linked to within- Generally speaking, neither design can by itself
person change. The vast majority of the research address the foundational assumption of an inter-
literature on the life course and cognitive func- action between the environment and human
tion has focused on levels rather than on trajecto- development, but most people agree that there is
ries and rates of change. In the present essay, we a substantial degree of heterogeneity in the nature
argue that in order to understand the nature of the and degree of cognitive change which is better
relationship between social structure, and life- studied in a longitudinal design. If there were no
span changes in cognitive function, it is impor- heterogeneity in age-related levels, trajectories
tant to be highly sensitive to this distinction and levels of stability in cognitive functioning
between average trajectories and average rates of over most of the life span, there would be little
change over the life span. Indeed, we argue that it need to provide the careful investigation of the
is because of the earliest inuences of social inuences of environmental change on changes
structural factors on cognitive functioning, there in individual growth trajectories.
may be considerably less of a role on cognitive As depicted in Fig. 3, at the early stages of the
change later in life (see below). life span, the environment promotes change.
From a descriptive point of view, cognitive Rates of change are highest early in life, and fol-
change is typically inferred from either repeated lowing Blooms (1964) line of argument, differ-
cross-sectional (between-persons) or panel ences in the experience with the environment are
(within person) data. It is important to bear in most potent during rapid changes at the matura-
mind that cross-sectional age-related differences tional or growth level. Variations in the environ-

Fig. 3 Developmental curve and changes in the limits of environmental inuence across the life span (From Alwin
2010)
Cognitive Development and the Life Course: Growth, Stability and Decline 459

ment have greatest quantitative effect on a the family, and there is a great deal of agreement
characteristic at its most rapid period of change among social scientists concerning the impor-
and least effect on the characteristic during the tance of the family of origin for a multitude of
least rapid period of change (Bloom 1964, p. vii). individual outcomes, from academic performance
Later on in midlife, the environment may act to and achievement test scores to social and eco-
promote stability rather than change, and due to nomic attainments in adulthood. The systematic
the increasingly stable properties of the environ- and persistent advantages enjoyed by individuals
ment (Ryder 1965), efforts aimed at changing born to socioeconomically successful parents
individuals through environment change may be have been thoroughly documented in the litera-
doomed to failure (see Musgrove 1977). These ture over more than a half-century using a variety
ideas are consistent with Baltes (1997) ideas that of data sets representative of the U.S. population
the challenges of midlife focus on maintenance, (e.g., Alwin and Thornton 1984; Blau and Duncan
rather than gains or losses. Finally, in older age 1967; Bowles et al. 2005; Duncan et al. 1972,
the experience of the individual may again paral- 2005; Lareau and Conley 2008; Morgan et al.
lel that of childhood and youthand the environ- 2006; Sewell and Hauser 1975; Smith et al. 1997).
mental inuences may regain their importance as Although achievement and other child outcomes
factors linked to cognitive change, in part due to are malleable over the childs life span, it is impor-
increasing rates of change/decline in older age as tant to recognize that the childs early years are
in Fig. 3 (see Alwin 2010). almost entirely monopolized by his or her family,
and that it is during these early years that basic
language, academic and social skills are devel-
4.2 The Role of the Environment: oped. These initial opportunities and constraints
Advantage vs. Disadvantage serve as important factors in the development of
the child (e.g., Hart and Risley 1995).
The complexities of the environment become The idea that parental inputs to early child-
incorporated into the cognitive function of the indi- hood experiences are important to later develop-
vidual through a sequence of increasingly complex ment is not new; and despite the widely shared
activities, interpersonal structures, and linked view that parents are important for development,
micro-settings (i.e., meso-environments) (see some have criticized that thesis. In a popular
Bronfenbrenner 1979, pp. 5665; Bronfenbrenner book, The Nurture Assumption: Why Children
and Ceci 1994). Clearly, the study of the linkage Turn Out the Way They Do, Judith Harris (1998)
between these complexities of mind and environ- argued that very, very bad parents can cause
mental structures is necessary if we are to under- irremediable harm to their children (p. 390) but
stand the nature of cognitive development across in the main how parents raise their children
the entire life span. It is the role of the family (and their child-rearing practiceshave little if any
other social institutions) in structuring the nature of effect on how their children turn out. She bases
the complexity of the environment that has been of her argument on a number of powerful pieces of
interest to child developmentalists, especially those evidence, including a famous review of the paren-
interested in documenting the role in early environ- tal socialization literature from the early 1980s
ments in producing individual differences in cogni- by Maccoby and Martin (1983) that reviewed
tive outcomes. intra-familial correlations in personality charac-
It has long been held that the early years of a teristics among both biologically-related and
childs development are the most important, lay- unrelated siblings. Their conclusion was that
ing the groundwork for later experiences, and that parental behaviors on the whole had no bearing
effective parenting associated with class or socio- on child outcomes (see also Maccoby 1992; Scarr
economic differences is an important component 1992, 1995; Harris 1995; Rowe 1994; Loehlin
of those critical experiences. Social stratication 1997; Pinker 2002).
has always been an important area in research on
460 D.F. Alwin et al.

Contrasting evidence exists in the behavioral The question raised in this chapter is whether
genetics literature for a role of both environment the development of cognitive resources and the
and genetics (see Plomin 1994, 1999; Rowe accrual of cognitive reserve (see Stern 2007)
1994; Rowe and Teachman 2001; Scarr 1995; follows the same pattern, as Merton suggests it
Shanahan et al. 2003). The claim is that individ- does in other realms. Few doubt that the socio-
ual differences in developmental outcomes arise, economic environment is an important element in
not from family differences in access to opportu- cognitive developmentit is a well-accepted fact
nities, nor from differences in socialization that cognitive function is shaped by differences in
approaches, but from genetically-variable attri- structural opportunities and other factors; theo-
butes of families. The behavioral genetics litera- retically, greater opportunities promote the
ture on cognitive abilities is extensive, and we do investment in activities that lead to greater cogni-
not attempt to review it here. We limit ourselves tive development. In fact, one of the prevailing
to Rowes (1994) assessment with respect to cog- assumptions is that early and later environments
nitive function, in which he reviewed the avail- during childhood contribute independently to the
able behavioral genetics studies of cognitive development of intellectual resources later on
abilities, indicating that heritability estimates (see Alwin and Thornton 1984).
ranged from 40 % to 70 %, indicating that sub- The argument is typically extended to suggest
stantial variation in intellectual ability has sub- there is a further compounding or accentuation,
stantial genetic basis (p. 105). There is also of the inuences of the social environment over
some initial support for the possibility that child- time, but this has not been closely examined in
rearing experiences may combine additively with the realm of cognitive functioning. Not only do
inherited advantage to account for individual dif- socio-environmental inequalities impact upon
ference; after further examination of a range of individual differences at multiple time points
studies; however, he concluded that the existing over the life span, there is considerable theory
data fail to demonstrate that child rearing envi- suggesting that the residues of these inuences in
ronments inuence cognitive function once chil- individual differences cumulate over time.
dren are older (Rowe 1994, p. 113). Hence, there is a literature that has developed
One of the timeworn hypotheses in the literature under the topic of cumulative advantage / disad-
on social stratication, but one that is not often vantage theory (ORand 1996; Dannefer 2003;
tested empirically, is that the consequences of early DiPrete and Eirich 2006; Ferraro and Shippee
socioeconomic differences in the lives of individu- 2009; see also Ferraro, chapter Life Course
als are accentuated over time. Using the metaphor Lens on Aging and Health this volume, for a
of the parable of the talents, dubbed the Matthew review).
Effect in Robert Mertons (1968) famous paper in It is widely believed that children from differ-
Science about inequalities in the reward systems ent family backgrounds enter schooling with dif-
governing credit in scientic authorship, quoted the fering levels of cognitive skills that differentially
following Biblical passage: unto everyone that prepare them for the development of academic
hath shall be given, and he shall have abundance; skills and that these differences grow over time.
but from him that hath not shall be taken away even Potter and Roksas (2013) recent research using
that which he hath. In short, Merton suggested that reading and math test scores from the Early
the rich get richer and the poor get poorer, or in Childhood Longitudinal StudyKindergarten
other words, that advantage and disadvantage cohort (ECLS-K) shows that cumulative family
cumulate over time. Few would doubt Mertons experiences account for most of the growing
observation that the social environment is struc- inequality in academic achievement between
tured in such a way as to promote the accrual of children from different social class backgrounds
greater resources to those who already have them over time (Baumert et al. 2012; Potter and Roksa
or, cumulative advantageand the withholding of 2013; see also Potter et al. 2013). These ndings
resources from those who begin with lessor support claims from the cumulative advantage
cumulative disadvantage. perspective that cognitive score differences
Cognitive Development and the Life Course: Growth, Stability and Decline 461

among children from different family back- individual or within-person malleability in


grounds enter schooling with different levels of development over time (see, e.g., Baltes et al.
academic skills, and such test score differences 1999, pp. 480481; Baltes and Baltes 1980). The
grow over time. strong concern of lifespan researchers with intra-
Few researchers have applied the hypothesis of individual plasticity (malleability) highlights the
cumulative advantage/disadvantage to cognitive search for the potentialities of development,
functioning in adulthood (but see Alwin 2010). including its upper and lower boundary condi-
One related body of work concerns research on tions. Implied by the idea of plasticity is that any
the relationship of social status factors and physi- given developmental outcome is but one of numer-
cal health, which consistently nds inequalities in ous possible outcomes, and that the search for the
health across the life span. Cognitive function and conditions and range of ontogenetic plasticity,
indicators of physical health would presumably including age-associated changes, is fundamental
operate in similar ways; despite this persistence to the study of development. At the same time,
and posited increasing strength of the relationship there is a great deal of evidence that cognitive test
between social status and health with increasing scores are quite stable (in contrast to plastic or
age; however, a growing body of research indi- exible) over rather lengthy periods of the life
cates that the association is generally strongest at span (see Baltes et al. 1999; see also, Alwin 1994).
the older working ages, and subsequently This is a direct consequence of declining
diminishes later in life (e.g. Crimmins 2001, amounts of change in growth patterns in young
2005; Crimmins et al. 2004; Hayward et al. 2000; persons after the early 20s. Growing evidence
House et al. 1992, 1994, 2005; Kunst and from MRI studies of children and adolescents
Mackenbach 1994; Lynch 2003; Marmot and suggests that the growth in the development of
Shipley 1996; Molla et al. 2004; Robert and structural changes in the brain peaks in late ado-
House 1996). These researches have consistently lescence (e.g., Lebel and Beaulieu 2011; Giedd
found a declining role of social status factors in 2008; Giedd et al. 1999; Johnson et al. 2009;
older age, despite the parallel ndings of the Paus et al. 1999). As is the case with many human
declining signicance of genetic differences in traits that reect both maturational and environ-
affecting health outcomes and cognition in older mental inuences, individual differences in test
age (see Rowe and Kahn 1998; Pedersen and scores are relatively less stable in childhood and
Lichtenstein 1997). Moreover, unless one factors early adolescence, but with age the differences
in the effects of selective mortality, such ndings among persons tend to stabilize, at least through
are hard to account for based on the premises of midlife (see Alwin 1994, 2008, 2009, 2010).
cumulative advantage/disadvantage theory, in that Here we examine how much exibility (or plas-
one would expect the socio-economic gradient to ticity) there is in cognitive abilities.
become steeper with timei.e., the rich get There is no question that environment plays an
richer and the poor get poorer. One would expect important role in cognitive development, and it
that those higher in status (i.e., more highly edu- also has a role in the maintenance of individual
cated), who are also more cognitively advantaged, differences over the life span. No one would
would actually increase their advantage over time, likely dispute the possibility that major environ-
although as far as we can tell, there is hardly any mental inputs can contribute to a exibility and
evidence for it (see Alwin 2010). change in older age in a range of abilities, but the
ndings we introduce here strongly suggests a
picture of high degrees of stability in cognitive
4.3 Stability and Change functioning from midlife into old age. It is well-
in Cognitive Function over known that individual differences in cognitive
the Life Span abilities are one of the most stable components of
human behavior that has been studied. Where the
The developmental literature on this subject uses stability of individual differences in human abili-
the term plasticity to refer to this intra- ties has been taken as problematic, measures of
462 D.F. Alwin et al.

Fig. 4 Molar stability estimates for cognitive function from 22 longitudinal studies spanning various ages (From
Alwin, unpublished ndings)

intellective variables are highly stable over most posed a best tting curve that is described by a
of the adult life span. In Fig. 4, we present a sum- polynomial function of the second order.
mary of estimates of molar stability (using a Stability results for the adult years show there
molar index of 8 years) from 24 longitudinal is very little change in the distributional place-
studies of cognitive functioning spanning a vari- ment of individuals relative to others after the age
ety of ages (from Alwin and others n.d.).2 These of 20. For example, the Seattle Longitudinal
results show that, as predicted on the basis of Study reects the typical patternHertzog and
Blooms (1964) work, the levels of molar stabil- Schaies (1986) estimate of the stability of the
ity are relatively low in the childhood years and common factor underlying individual differences
increase with age. In Fig. 4 we have superim- in a version of Thurstones Primary Mental
Abilities is .92 over a 14-year period. They also
showed that levels of stability increase with age.
2
The molar stability coefcient is an estimate of the per- Similarly, Kohn and Schooler (1978) nd a nor-
sistence of behavior or behavioral orientations as
mative stability value of .93 for their concept of
expressed in age-homogenous rates of change over speci-
ed periods of time. The concept was introduced as a intellectual exibility assessed over a 10-year
means of organizing empirical information on human period. Across the entire range of studies
constancy and change, and of comparing raw stability indicates that stability grows in magnitude from
estimates across studies having different re-measurement
adolescence onward and from the age of 40 the
intervals and across different concepts (see Alwin 1994,
pp. 155158; Alwin 1995, pp. 233238). As distinct from typical molar stability of intellectual ability is
other concepts of stability (see Alwin 1994), molar stabil- roughly .900.3
ity is dened as j/k where is the cohort-specic or age-
homogenous stability estimate observed empirically, k is
3
the number of years over which raw stability is assessed, In this discussion normative stability refers to the preser-
and j is the number of years selected to express molar sta- vation of individual differences in a quality within a con-
bility. In the examples used in this chapter, j = 8. stant population over a specied amount of time, whereas
Cognitive Development and the Life Course: Growth, Stability and Decline 463

These results may seem to pose a serious Cronbach and Furby 1970). More recently, mod-
threat to some interpretations arguing that social- ern statistical models for the analysis of growth
ization or learning affects basic intellectual abili- (and decline) were stimulated in part by early
ties well into adulthood. Obviously, some applications of growth curve analysis to the mea-
openness to change is possible during adulthood, surement of change (Rogosa et al. 1982) and later
but if we understand that stability is a function of developments (see Curran and Bollen 2001;
constancies of person-situation or person- Bollen and Curran 2006). A somewhat parallel
structural linkages, then continuities over time literature involving the analysis of change using
may be viewed as reections of the stability of causal models of change in panel data (Heise
socially structured experience, and the proper 1970; Jreskog 1974; Alwin 1988), and present-
adjudication of the issue of whether a socio- day models incorporate both inter-individual dif-
environmental interpretation exists for the stabil- ferences and intra-individual (or within-person)
ity of cognitive abilities would have to focus on change (see Curran and Bollen 2001).
the segment of the population that experiences
change in social locations at different points in
the life cycle, e.g. persons who make the transi- 5.1 Latent Difference Scores
tion to retirement, so that life-course linkages to
changes in socially-structured experiences could The latent difference score model is an advanta-
be determined. We return to a discussion of this geous approach to the study of within-person
issue when we take up the topic of cognitive (intra-individual) change (see Alwin 1988,
change in adulthood and older age. pp. 137139; Steyer et al. 1997; McArdle 2001;
McArdle and Hamagami 2001), and it ts nicely
with current latent growth models. Building upon
5 Latent Change Models the idea of latent difference scores, latent growth
curve (LGC) models are well-suited to the study
Contemporary growth models for CF are rooted of CF because they focus on continuous pro-
in the historical concerns of educational and psy- cesses of change and can allow for measurement
chological researchers with the measurement and errors in the variables assessed over time.4 There
analysis of change (Harris 1963). Early xed- are some difculties with the interpretation of
effects ANOVA models were formulated for measurement errors in these models because the
repeated measures designs involving two or more errors on the indicators represent both measure-
groups of different subjects in which each subject ment error and the lack of t of the growth func-
had measurements on two or more occasions tion to the individuals observed score at time t,
(Lindquist 1953). These early statistical models but we do not dwell on these matters here (see
focused explicitly on within-person change, Willett and Sayer 1994; Singer and Willett 2003).
although notions of individual differences in All growth models begin with the specica-
growth or development were fairly primitive in tion of a within-person (i.e. intra-individual)
these early statistical treatments. Early method- model for individual change over the period of
ological contributions concentrated primarily on measurementthese are often referred to as
the problems and limitations of difference
scores for assessing change (Lord 1963;
4
Because of the limitations on space, we do not discuss
these models in detail here and introduce them primarily
molar stability refers to the persistence of a quality as to provide a conceptual orientation to the study of within-
expressed in the rate of change for an age-homogenous person change, the central concept used throughout this
cohort over a specied period of time. As used here, the chapter. Further discussion of these models for those
term normative stability simply refers to molar stability in uninitiated into the study of within-person change and the
the entire population, not broken down by cohort (see SEM approach may wish to consult introductory material
Alwin 1994, p. 139). on change models (see, e.g. Alwin n.d.).
464 D.F. Alwin et al.

time-based or occasion-based models, although Models based on time in study will include
for a given individual measured longitudinally, covariates such as chronological age and other
there is a perfect correlation between time of covariates to account for the heterogeneity in
measurement and age. In the multi-level frame- individual differences in level and change. Age-
work, the intra-individual growth model is some- based models, on the other hand, model age-
times called a level-1 model. Once this model is differences and age-changes as average or xed
formulated and estimated, one can examine a effects but will often include other covariates as
level-2 model for explaining inter-individual dif- well. Alternatively, event-centered time struc-
ferences in levels (intercepts) and rates of change tures can be used to account for heterogeneity in
(slopes) in the level-1 model (see Willett and initial status and rate of change in terms of com-
Sayer 1994, pp. 36364). Although we do not mon patterns of change that occur prior to or fol-
represent them here, one of the purposes of the lowing a discrete time-based structure or event
level-2 model is to include covariates to account (see Alwin et al. 2006). Heterogeneity in rates of
for variation in the intercepts and slopes. These age-based decline could be explained in terms of
are often referred to as random coefcients mod- the misalignment of individuals with respect to
els, since the groups (individuals) are conceptual- the causal process producing their cognitive gains
ized as, and in many cases, are in fact, a sample or losses (dementia). These types of alternative
from some larger population. Of particular inter- time specications will lead to different interpre-
est here is the inclusion of covariates that repre- tations of the same data. In non-informative time
sent some function of timeeither historical or metrics (e.g., time in study), the heterogeneity of
biographical timeto account for levels (inter- initial status and change (random effects) is
cepts) and the rates of change of the dependent accounted for by covariates whereas in models
variable of interest. Such a latent growth curve with informative time metrics, the heterogeneity
model is statistically equivalent to a random is accounted for by the alignment of individuals
coefcients model for change over time when with the explanation of the time-based process
time values are discrete across occasions for all focused on the average or xed effects in account-
or most individuals. The multilevel approach per- ing for heterogeneity in change.
mits more exible specication of time in that It is not possible to summarize all of the cur-
values can be different for each individual at each rent research studies of CF that have formulated
occasion (see Hox 2002). developmental change in terms of latent curve
There are several alternate ways to construct models. We provide one example from the work
individual latent growth models (e.g., Willett of Farkas and Beron (2004), who analyzed
and Sayer 1994; McArdle and Bell 2000; vocabulary data (the Peabody Picture Vocabulary
Meredith and Tisak 1990). One common Score) from the Children of the NLSY79
approach is to conceptualize the models within (CNLSY) collected at two or more observations
a SEM framework, but there is an equally com- at different time points for many children in the
mon (and formally equivalent) approach in the sample. As Farkas and Beron (2004, p. 466) note,
multilevel regression analysis tradition (see most researchers analyzing the CNLSY treat the
Macmillan and Furstenberg, chapter The Logic data in what is essentially cross-sectional form,
and Practice of Growth Curve Analysis: taking a particular year and an age-homogenous
Modeling Strategies for Life Course Dynamics set of children, and analyzing the determinants of
this volume). The growth model for within-per- CF scores. By contrast, Farkas and Beron (2004)
son change in the generic case expresses change estimate multilevel random coefcient growth
as a function of time, where an individuals curve models in which intercept and slope
score at a given point in time is a function of an (including a quadratic specication) were permit-
intercept (or level) parameter, a linear slope ted to depend on race, class and gender. The data
parameter, and possibly a parameter expressing analyzed by Farkas and Beron (2004, p. 447) are
some curvature. shown in Fig. 5, which depicts the average oral
Cognitive Development and the Life Course: Growth, Stability and Decline 465

160

140

120

100
PPVT

80

60

40 White PPVT Score

Black PPVT Score


20

0
36
40
44
48
52
56
60
64
68
72
76
80
84
88
92
96
100
104
108
112
116
120
124
128
132
136
140
144
148
152
Age in Months

Fig. 5 Raw PPVT average scores by race (3 month moving average)from Farkas and Beron (2004) (Reproduced
with permission)

vocabulary (PPVT) scores by single month of year olds. An event-centered approach is a natu-
age, separately for African-Americans and ral model for life course analysts who study the
Whites. These data show that over the span of 10 interplay between human development and life
years, from 36 to 156 months of age (ages 313) course events. Given an event of substantive
childrens oral vocabulary scores grew from interest it may be of value to describe processes
approximately 20 words identied correctly to of change that precede or follow the event,
approximately 120 words identied correctly, regardless of the age of the individual. These
and for each group the trend was generally event-centered strategies permit a potentially
upward over these ages (Farkas and Beron 2004, more nuanced version of within-person change
p. 477). than that which might be captured by relying on
time in study or age as the metric used for time
(see discussion in Alwin et al. 2006, pp. 3132).
5.2 Event-Centered Latent For example, one might dene time of entry to
Change Models school as a transition for pre-school children and
study within-person change before and/or after
The model we wish to draw attention to in the the transition. In this case, growth is assessed
present context is the class of event-centered from the time of the event rather than from birth.
growth models, where within-person change is Or, taking another example, one might dene
assessed relative so some event or transition, and widowhood as an event, studying patterns of
we illustrate these models with respect to the depression or other measures of psychological
example given above from the Farkas and Beron well- being as a function of time since the event.
(2004) analysis of vocabulary knowledge in 313 There are a number of process-based time
466 D.F. Alwin et al.

structures and event-centered approaches that In the hypothetical shown in Figs. 7 and 8, we
have been investigated which provide exemplars depict what a set of generic event-centered
of the fruitfulness of the approach (e.g., time to growth curve models where there is an effect of
death; time to dementia diagnosis). the event considered. Again time-t denotes the
In the Farkas and Beron (2004) model above, time of the event (e.g. entry to school, or entry to
childrens test scores are centered on chronologi- preschool), and the groups may be dened as
cal age, that is, time is measured from birth. In a above. In the case of Fig. 9b, there is a divergence
life course event-centered approach, the analysis in the growth trajectories among the three groups,
might be centered instead on time of entry into whereas in Fig. 9c there is a convergence in
school. In this case, the units of time would be growth trajectories.
months prior to entry into school and/or months One example of divergence conditioned by an
following entry into school. This would necessi- event is the interesting discovery by Canadian
tate measuring cognitive scores at time before/ psychologist Roger Barnsley, who rst drew
since the event, rather than at time since birth. In attention to the phenomenon called relative age
Fig. 6, we depict what a set of generic event- (Musch and Grondin 2001). While watching a
centered growth curve models would look like hockey match with his wife and two sons, they
for three groups, A, B and C, were t denotes the noticed from the roster that a large majority of the
time of the event (e.g. entry to school, or entry to players from both teams were born in January,
preschool), and the groups may be racial, class or February or March. After reecting on the social
gender groups. In this case, the event at time-t has circumstances that might have produced this cor-
no apparent effect of the trajectories of growth relation, the explanation was obvious: the junior
for the three groups, although as in the Farkas league cut-off for accepting young players in
and Beron (2004) example, there is a main effect Canada was January 1, which means that selec-
separating three groups. tion favors size advantages, which are very

0.9

0.8

0.7

0.6

0.5

0.4
C
0.3
B
0.2 A

0.1

0
t-3 t-2 t-1 t t+2 t+3 t+3

Fig. 6 Depiction of event-centered growth curve models for three groups where there are no differences in the growth
functions for the groups
Cognitive Development and the Life Course: Growth, Stability and Decline 467

0.9

0.8

0.7

0.6

0.5

0.4
C
0.3
B

0.2 A

0.1

0
t-3 t-2 t-1 t t+2 t+3 t+3

Fig. 7 Depiction of event-centered growth curve models for three groups showing an example of a divergence in
growth patterns following the event

0.9

0.8

0.7

0.6

0.5

0.4
C
0.3
B
0.2 A

0.1

0
t-3 t-2 t-1 t t+2 t+3 t+3

Fig. 8 Depiction of event-centered growth curve models for three groups showing a convergence in growth patterns
following the event
468 D.F. Alwin et al.

0.5

EVENT
-0.5

-1

-1.5
-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9 10

Fig. 9 A framework for studying growth trajectories using event-centered latent change models, where the time-metric
is centered on the event

important among young boys. To those born such a question, but it could be answered using
immediately after the cut-off there were a num- the test score data from the CNLSY employed by
ber of advantages related to their relative size. An Farkas and Beron (2004). We know of no exam-
imaginative social psychologist can specify the ples of the convergence hypothesis either, but it,
processes implicated by Barnsley and others in too, could be addressed using the CNLSY data.
accounting these effects: involving selection,
streaming and differentiated experience (see
Barnsley et al. 1985; Barnsley and Thompson 6 Cognitive Development
1988). We do not go into the details of these pro- in Childhood
cesses herethey are written about extensively
in the literaturebut note that the phenomenon is Evolutionary anthropologists and bio-
not unique to hockey. Malcolm Gladwell (2008) demographers have argued that humans have
describes a similar phenomenon with soccer in evolved a life cycle that requires a prolonged
Englandthe best players are those born imme- period of child dependence on adults in order to
diately after the cut-off date. insure post-natal brain development (e.g., Kaplan
The conclusion is obvious: if one wants to et al. 2000; Carey 2003). At the same time, devel-
raise a star hockey or soccer player, one would be opmental psychologists have emphasized both
smart to nd out the junior league eligibility cut- the innate genetic capacity for learning, as well
off dates, and engage in some family planning. as the prenatal environment during infancy. One
The question for the development of CF with example of the latter is the hypothesized effect of
respect to the relative age effect might be posed low birth weight on subsequent brain develop-
in terms of the event-centered models proposed ment (e.g., Barker 1998; de Kieviet et al. 2012).
above, that is, are there advantages that derive Mechanisms connected with innate structures
from school entry (or pre-school entry) when and abilities raise the issue of the relevance of
prospective students are older relative to time of other developmental factors that occur before the
entry. We know of no research that has posed child begins interacting with the social world.
Cognitive Development and the Life Course: Growth, Stability and Decline 469

Still, it is during these early phases of social life with the physical environment, the mechanism
that cognitive abilities are developed; and while through which economic deprivation operates
social scientists are primarily concerned with dif- (e.g., fewer toys, books, and stimulating objects
ferences in the social environment, it should be present in the household); (3) Social transmis-
clear that species-level genetic material and, per- sionlanguage, education, and socialization that
haps more importantly, gene-environment inter- structures how a child benets from teach-
actions, also play an indispensable role in early ing,thus, children who have acquired early
development. skills may get more out of the exact same situa-
The important advances in the science of early tion, than children lacking a particular skill
childhood development over the last century which is an inuential idea similar to the
were recently documented by the National economic models of human capital accumula-
Research Council and Institute of Medicine tion; and (4) Equilibrationwhen faced with a
report titled From Neurons to Neighborhoods new piece of evidence, or something that the
(Shonkoff and Phillips 2000). As indicated in that child is not able to accommodate within their
report, much of the contemporary eld of the existing structure of knowledge (Piaget 1985).
study of early childhood cognitive development More recent advances in developmental the-
has been shaped by the revolutionary thinking of ory make an argument that has some parallels to
Jean Piaget (1952) and Noam Chomsky (1965). Piagets equilibration mechanism, but the propo-
These ideas celebrated the active role of young nents criticize Piagets understanding of infants
children in their own development and attributed as having very limited cognitive capacities.
early skill acquisition to the universal emergence Through innovations in experimental techniques
of innate cognitive and linguistic structures with infants, there is increasing evidence that
(Shonkoff and Phillips 2000, p. 58). Piagets they lead much more cognitively complex lives
work stressed the importance of development than previously thought (Gopnik and Meltzoff
(ontogeny) versus learningthe former is a more 1998; Mandler 2004). The work of Gopnik
general process that explains or causes the latter, (2012) and others suggest that very early on chil-
as opposed to the idea that the accumulation of dren have probabilistic models of the world that
learning leads to developmental progress. A sec- allow them to predict, infer, and use their new
ond key idea of Piaget is the idea of cognitive experiences to update their models of how the
operationshaving an understanding of an world works. In other words, cognitive develop-
object that enables one to perform action (e.g., ment is similar to the Bayesian process of updat-
classifying, organizing/ordering, counting, mea- ing prior beliefs with new data to arrive at a new
suring) on the object. In Piagets theory, opera- (posterior) set of beliefs. While this view of cog-
tions link together to form a larger complex of nitive development does involve the acquisition
knowledge, referred to as structures or the total of new experiences through interaction with the
structure (Piaget 1965, 1970). Development is environment, the mechanisms driving change are
the process of building operations, linking them innate cognitive structures and capacities for
together to form structures, and reconstructing change (Gopnik and Meltzoff 1998; Mandler
them to account for new information. 2004).
Piagets work emphasized the basic idea of Similar arguments point to the complexity of
building complexes of understanding and recon- the social world and the need for (domain-
structing them, involving four basic factors, specic) constraints to guide cognitive develop-
which are useful for identifying potential sources ment (Gelman 1990). Evolution has produced
of social differentiation and the beginning of these constraints (or internal biases) that enable
inequality: (1) Maturation, the mechanism infants and children to become efcient learners
through which biological processes (and the in a world so rich with stimuli. As pointed out by
social dimensions along which these biological Goldstone and Landy (2010), it is interesting to
processes are stratied) operate; (2) Experiences note how the evolution of thinking on this topic
470 D.F. Alwin et al.

has shifted to include the idea that individuals social roles in which performance is variable,
can acquire constraints through their interaction with the quality of cognitive performance being
with the social world. the basis of the assignment of social status
Despite the importance of considering the (Duncan et al. 1972, pp. 7779). The criteria
innate potential of infants to learn, early cogni- being used to test for role performance in the
tive development obviously does not occur in a schools at an early age are often linked abstractly
social vacuumit happens within a congura- to notions of mental or cognitive function associ-
tion of opportunities and constraints linked to ated with the concept of intelligence (or CF). This
environmental and institutional arrangements, as concept, as well as those that are highly linked to
well as interpersonal relationships. A key to it, such as Schoolers (1987) concept of intellec-
understanding early cognitive functioning is tual exibility, or others, are an important com-
identifying the mechanisms for change, includ- ponent of what students of cognitive aging
ing life course events and transitions (Siegler investigate, although the aging years are often
1989; McClelland and Siegler 2001). We focus studied separately from child development.
here on causes of change in cognitive ability
because they are sources of disparities between
individuals (some individuals grow more, or at a 6.2 How Do Families Affect
faster rate, than others), and they are crucial for Cognitive Growth?
designing teaching strategies and interventions
aimed at facilitating cognitive growth. Identifying Some clarication of this may be useful. In order
the mechanisms at work also help illustrate the to understand developmental outcomes in chil-
social context of cognitive change, a topic we dren, it is no longer possible to phrase the ques-
address at the end of this section. tion of environmental effects as an either-or
question. The question is not one of nature vs.
nurture, but one of how genes and environment
6.1 Environmental Factors interact to shape developmentone of nature
in Early Cognitive and nurture. There are three broad categories of
Development possible ways that differences among families
or between-family differencescan produce
Environments are essential for development to individual differences in child outcomes. By
take placegenes cannot nd their phenotypic child outcomes, we refer not only to any differ-
expression without ecological or environmental ences in cognitive function, but also to personal-
settings within which to do so. And environments ity differences, values, preferences, interests,
cannot function as facilitators of human behavior skills, accomplishments, ways of behaving, and
without the genetic potential to work with. In a the like. These family-specic mechanisms for
variety of behavioral domains, research on nature cognitive change are: (1) families can differ in
(genes) and nurture (environments) is converging the opportunities for development of particular
(see Shanahan et al. 2003). In focusing on envi- outcomes, e.g., in the complexity of the learning
ronmental inputs, a distinction is typically made environments they provide their children; (2)
between primary and secondary socialization. families can differ in genetic endowments that
Primary socialization happens early in life, mainly contribute to the development of particular out-
in the family and school, whereas secondary comes; or (3) families can differ in the way in
socialization involves learning roles relevant to which they nurture or socialize their children
occupation and family. There are obvious connec- (Scarr 1995). These three explanations do not
tions between cognitive skills captured in early address the possibility that there are within-
parts of the life span and those required for adult family differences in parenting practices that
roles, and later in the chapter we explore these contribute to child outcomes (e.g., Frijters et al.
linkages. Some sociologists even go so far as to 2013), a topic which we leave for future
say that the occupational structure denes the key discussion.
Cognitive Development and the Life Course: Growth, Stability and Decline 471

The rst explanation of between family dif- disadvantage (Sewell and Hauser 1975; Jencks
ferences is the standard sociological explanation et al. 1979; Todd and Wolpin 2007).
for why some children do better in school, or Those who are skeptical that parental inputs to
achieve more socio-economic status, than oth- early childhood experiences are important to
erstheir families provide them with specic later development (e.g., Harris 1998, discussed
advantages or different opportunities for learning above) are presumably not disputing that such
or success. In practice, these inputs are com- inequalities in opportunity exist, although some
monly represented in analytic models through the of what has been written could be mistakenly
inclusion of family income, a proxy for economic interpreted this way, e.g. when Harris says things
resources family income, a proxy for economic like: the evidence indicates that differences
resources, educational attainment of the parents, between one home and another, between one set
or some combination. More specicity is advan- of parents and another, do not have long-term
tageous, since the presence of economic resources effects on the children who grow up in those
does not always translate into using those homes (1999, p. 391), this gives the impression
resources for making child investments. that she is including any and all differences
Complicating the task even further is the possi- among families and not just those having to do
bility that investment decisions depend on the with socialization practices. In any event there
demonstrated or perceived ability of the child, as are a number of other examples, besides those
well investments made by teachers, schools, or having to do with educational and socio-
other sources (Todd and Wolpin 2003). Several economic outcomes, which can be given for dif-
studies have sharpened our understanding of this ferential opportunities and child outcomes.
general mechanism by looking at specic exam- The problem with sorting out these various
plesincluding nutrition, stimulating objects explanations is that the key explanatory factors
(e.g., books, toys, musical instruments), experi- highlighted by each category are correlated with
ences (e.g., visiting museums, traveling to other one another. If genetic differences occur between
countries, working with tutors or taking piano families that have educational and socioeco-
lessons), and access to physical environments nomic consequences for children, then they are
(e.g., dwellings and neighborhoods) free of likely to be related, for example, to the factors
developmental stressors or insults (Berger et al. that shape differential opportunity structures, and
2009; Bradley and Corwyn 2002; Guo and Harris complicating matters further is the possibility of
2000). Such efforts are useful, though it is worth gene-environment interactions (Turkheimer et al.
noting the potential pitfalls of the coveted 2003). Or, as we shall see below, parental social-
kitchen sink approach when it comes to inter- ization practices are linked to family differences
preting results and confounded effects (Todd and in opportunities. They may not be adding any-
Wolpin 2003).5 Despite these complications, it thing independent to the explanation of individ-
seems that the playing eld is clearly not level, ual differences in developmental outcomes, but
and children from different families turn out dif- their role may be one of mediating the effects of
ferently as a consequence differential advantage/ other (genetic and environmental) differences
among families. Perhaps the most common
examples involve the overlap of family socioeco-
5
Todd and Wolpin (2003) identify a useful example using nomic status with positive parental involvement
a model in which child achievement is regressed on family and stress, and the subsequent connections to
income, the number of books in the home, and additional
cognitive development (e.g., Bodovski and
covariates. The issue arises with the interpretation of the
estimated effectsincreasing the number of books, while Farkas 2008; Davis-Kean 2005; Gershoff et al.
holding family income constant, implies a reduction in 2007). We cannot resolve the intricacies for sort-
some other area of family consumption which may ing out the truth or falsity of these various expla-
include child investment (e.g., fewer educational toys;
nations here, but sufce it to say that the
assuming these inputs are purchased). Thus, the estimated
effect is potentially confounded with a change in the level complexity of the issues should not prevent us
of an additional (unobserved) input. from confronting them head on.
472 D.F. Alwin et al.

6.3 Event-Centered Models There is also evidence that this experience


for Educational Transitions increases the initial level of academic skills and
the subsequent upward trajectories (Aikens and
One of the earliest life course milestones to which Barbarin 2008; Cheadle 2008; Magnuson et al.
investigators have paid attention is the transition 2004, 2007), as well as other longer-term out-
to school. Indeed, in recent years, life course comes indicative of cognitive ability (Blau and
researchers have established a beachhead in a Currie 2006). Structural characteristics of child
new understanding of schooling and educational care providers (e.g. child-caregiver ratio and
attainment by studying the developmental pro- caregivers training) affect the quality of the
cesses that lead up to school entry. In the rst edi- child-caregiver interactions (process features)
tion of the Handbook of the Life Course, Doris which, in turn, inuence cognitive outcomes for
Entwisle and her colleagues articulated a core children (NICHD Early Child Care Research
assumption of the life course frameworkthat Network 2002a). While there is little evidence
developmental processes and outcomes are suggesting that the quantity of non-maternal
shaped by the life trajectories children follow. child care affects cognitive outcomes (NICHD
We would add a corollary to this, which reverses ECCRN 2000), there is support for an effect on
the causal ordering to include the principle that behavioral measures (NICHD ECCRN 2003).
the life trajectories children follow are shaped The provision of non-maternal child care may be
by developmental processes and the outcomes an important remedy for reducing socioeco-
they bring to the setting. Social scientists have nomic, racial, and ethnic disparities in school
long appreciated the dual nature of the interplay readiness and cognitive trajectories. It is worth
between developmental inputs to social settings, noting, however, that the effect sizes of early
otherwise referred to using some notion of the child-care experiences on cognitive outcomes are
concept of selection, while at the same time smaller than the effects of family factors, such as
postulating a socialization effect of the life parenting quality and family socioeconomic sta-
paths followed by individuals. We suggest that tus (NICHD ECCRN 2002b). There are, how-
this principle is no less important in early school ever, concerns that the initial benets may fade
transitions, and in keeping with these ideas will out over time (e.g., Votruba-Drzal et al. 2008),
review both (a) the available literature that has which may be due to characteristics of the school
covered the inuence of preschool and kindergar- and classrooms attended after child care (Currie
ten experiences on rst-grade performance, as and Thomas 1995; Magnusson et al. 2007).
well as (b) literature on some of the longer-term Turning now to longer-term outcomes, an
outcomes related to the rst-grade transition. ideal study would measure growth from an early
Non-maternal child care is an increasingly age, and track the nature of the post-event transi-
common experience among children and has tions. Such a model would allow a test of the
served as an experimental setting for investigat- proposition that education is the great equalizer,
ing the potential of early interventions (Blau and allowing post-event convergence in growth tra-
Currie 2006; Diamond et al. 2007; jectories. On the other hand, post-event expo-
U.S. Department of Health and Human Services sures may simply increase the advantage that
2010). Early child care has also been the focus of exists at the time of the transition. This discus-
extensive research through the NICHD Study of sion underscores the importance of considering
Early Child Care and Youth Development event-centered latent change models for school
(NICHD Early Child Care Research Network transitions we considered earlier in the chapter,
2005). This research suggests that early child where we provided examples of latent curve
care provides several benets for children that models centered on educational events, for exam-
include cognitive development and school ple, on school entry.
readiness. The focus of such a model is to examine het-
erogeneity in growth trajectories (intercepts and
Cognitive Development and the Life Course: Growth, Stability and Decline 473

slopes) with respect to the occurrence of an insti- (e.g., health, approaches to learning, and reading
tutionalized event, such as entry to rst grade. outside of school) drive socioeconomic dispari-
Unfortunately, we rarely have change data on ties. While a considerable amount has been
children prior to the age of 5, when they typically learned about diverging trajectories in cognitive
enter school, and we can only predict intercepts abilities, a substantial proportion of the variabil-
at the time of school entry. In the absence of ity remains unexplained (Downey et al. 2004).
growth data prior to the event of school entry, one Diverging trajectories in cognitive develop-
can think of post-event within-person change as ment are consistent with an economic model of
an extension of earlier developmental patterns. In skill acquisition. Through their own work and
such a situation, an interesting approach would reliance on previous literature, Cunha and
be to t a latent growth model, and allow the Heckman (2007, 2008) build this model of early
intercept to correlate with the slope of the devel- development upon two important stilts. The rst
opmental trajectorya positive correlation sup- is the concept of self-productivity, which refers to
porting the cumulative advantage hypothesis the notion that capabilities are self-reinforcing in
described earlier (e.g., Downey et al. 2004). that the development of early abilities serves as
An important resource for studying early cog- the building block for the development of later
nitive development which employs such an (more advanced) skills. Learning at young ages
event-centered approach is the Early Childhood (and the successful application of newly acquired
Longitudinal Study, Kindergarten Class of 1998 skills) embeds in the child the value of learning,
1999 (ECLS-K), a study of a nationally represen- which serves as motivation for future skill devel-
tative sample with information from a on reading opment. Furthermore, the early mastery of skills
and math scores for children measured in kinder- increases the efciency with which skills are
garten, as well as grades 1, 3, 5, and 8. (See acquired at older ages (Knudsen et al. 2006;
Duncan et al. 2007 for an investigation into sev- Heckman 2007). (Footnote: A related concept is
eral additional data sets well-suited for the study cross-fertilization, the idea that capabilities in a
of child development.) Downey and colleagues certain domain (e.g., communication and lan-
(2004) use these data to analyze cognitive trajec- guage skills) enhance the childs ability to acquire
tories by relying on models similar to the growth skills from a different domain (say, cognitive
models described earlier. They show that while skills).) The second leg of the model developed
initial levels of cognitive ability (i.e., intercepts) by Cunha and Heckman (2007) is dynamic com-
are negatively correlated with growth (i.e., plementarity, a concept suggesting that the abili-
slopes) in ability during the school year, there is a ties developed at an earlier age increase the
positive correlation between the intercept and productivity of investments made at an older age.
rate of growth during the summer months Together, these concepts simply imply that skill
between the end of kindergarten and the begin- begets skill (Heckman 2007). The implication of
ning of rst grade. this model is that we should make investments in
Furthermore, socioeconomic inequality in children early and often to ensure that they are
cognitive growth increases more slowly during equipped with early cognitive skills that blossom
the school year compared to summer months, into future cognitive abilities, and so that they
highlighting the importance of this transition. It (and the rest of the society) can reap the benets
is important to note, however, that equalizing of these early investments (Heckman 2007).
effects of early schooling do not extend to racial
disparities in cognitive growth. Research by
Condron (2009) suggests that school factors 7 The Transition to Adulthood
(e.g., racially segregated schools, differences in
ability grouping, and private vs. public school Given our culture and its history, childhood and
differences) exacerbate black-white differences youth involve life stages that are relatively dense
in cognitive growth, while non-school factors with respect to the occurrence and frequency of
474 D.F. Alwin et al.

life course events in the family, school, commu- Test of Intelligencesimilar to other standard-
nity, and workplace that shape early outcomes ized measures of general ability. The WLS
(see, e.g. Rindfuss 1991). These early stages are Henmon-Nelson measure has been widely used
considered particularly important periods for the in predicting early socioeconomic attainments
formation of cognitive capacities. Based on the (Sewell and Hauser 1975; Hauser et al. 1983).
foregoing conceptual framework, in this section For our purposes, we assume such tests contain
we argue that (1) differences in individual experi- some combination of both uid and crystallized
ences organized on the basis of institutionalized components. We employed a model, using path
social structures (experiences involving analysis conventions, which species the roles of
environmentally-linked access to opportunities family background and pre-adult cognitive levels
and resources) contribute to individual differ- in shaping educational level and high school aca-
ences in cognitive functioning throughout the life demic experiences, all of which, in turn, affect
span; (2) the potential for differences in orga- occupational attainment (either directly or indi-
nized experiences to affect individual differences rectly) across the life-span. These early factors
in cognitive functioning is greatest during peri- all help shape socioeconomic attainments and
ods in which change at a developmental level is cognitive developments later in life.6
the greatest; (3) social structural factors not only The results of our analysis show that pre-adult
help put individual differences into place, they cognitive functioning is the most powerful pre-
also help maintain individual differences through- dictor of the WAIS score at age 52. Roughly one-
out most of the adult life span; and (4) the poten- third of the variance in the WAIS score is
tial for change in cognitive functioning is greatest attributable to the WLS Henmon-Nelson test
in the early and late periods of the life span. score, some 35 years later. This is remarkable
One of the longstanding assumptions of the given the differences between the two tests and
life-span developmental perspective is that indi- the possibility that they are tapping different
vidual intellectual development peaks in young dimensions of intellectual functioning. In addi-
adulthood and has major consequences therefrom tion, if one considers the fact that only 40 % of
for the choices of life pathways. Consistent with the variance in WAIS is predictable variance,
these notions are the studies of status attainment, given our model, the bulk of the prediction is
e.g., the Wisconsin Model of Status Attainment, coming from the Henmon-Nelson score. Even
based on the Wisconsin Longitudinal Study after school success, years of schooling and
(WLS), among others (Hauser et al. 1992, 1993; occupational position variables are included in
Sewell et al. 2003). Such models posit a strong the model as potential mediating mechanisms,
role played by individual differences in adoles- the pre-adult test score continues to have a sub-
cent CF for later educational and occupational stantial direct effect; and indeed, the largest direct
achievement. In this section of the chapter we effect. The direct effect accounts for more than
briey review the relative importance of such 70 % of the total effect of pre-adult levels on sub-
cognitive factors relative to non-cognitive and sequent levels of intellectual ability. The indirect
social background factors in educational and effect operates primarily through school perfor-
occupational pathways, as well as the impact of mance and years of schooling.
adolescent cognitive factors on later assessments The ndings are consistent an emerging litera-
of CF. ture on early life experiences and later adult out-
Our own reanalysis of the WLS data permits comes, which suggests that early cognitive
the assessment of the long-term effects of pre- development is predictive of cognitive function
adult intellectual levels and socioeconomic back-
ground on cognitive functioning at midlife
6
(Alwin et al. 2008a). A WAIS score involving ten We employ structural equation methods in this analysis,
survey measures of verbal meaning is available in and focus on the total rather than direct effects of variables
in a causal model (see Alwin 1988). The details of our
the WLS at age 52, and there is a pre-adult mea- treatment of the WLS data, including the measures and
sure of cognitive abilitythe Henmon-Nelson modeling strategies used, are given in Alwin et al. (2008a).
Cognitive Development and the Life Course: Growth, Stability and Decline 475

and other health outcomes later on in life (e.g., the life cycle, such that life-course linkages to
Snowdon et al. 1996; Kuh et al. 2007; Richards changes in socially-structured experiences could
and Deary 2005; Gottfredson and Deary 2004; be used as event markers around which to study
Whalley and Deary 2001). Another important cognitive change. Schoolers (1987) analysis of
nding is reected in the fact that family socio- the development of intellectual exibility over
economic background has an important effect on the life span through exposure to changes in the
the WAIS score, its total effect accounting for 10 complexity of the environment represents one
% of the variance in individual differences in attempt to apply theories of social structure and
cognitive functioning. These results are consis- personality to human development. However,
tent with recent research involving nationally data from this research program have not been
representative studies of children, which indicate adequately organized to reect life-span varia-
that oral language development is strongly inu- tions in stability, nor have these analyses concep-
enced by family socioeconomic factors from tualized within-person change with respect to
birth through age 3 (Farkas and Beron 2004). measured events that have theoretically caused
Childhood socioeconomic position is known to the change. It may be that the greatest change in
exercise an effect on later cognitive function in CF occurs in persons whose environments change
adulthood (see, e.g., Kaplan et al. 2001). most rapidly, rather than in those with stable
Two additional ndings of interest from these environments (Musgrove 1977; Alwin et al.
results concern the effects of occupational levels 1991). There is no question that environment
on intellectual functioning. This involves, rst, plays an important role in cognitive development,
the effect of occupational position at age 35, and and it also has a role in the maintenance of indi-
second, net of this, the effect of occupational vidual differences over the life span despite high
position at age 52. The latter coefcient reects degrees of stability in CF into older age.
the effects of occupational change and speaks The weakness of the argument that changes in
directly to the learning generalization hypothesis adult intellectual functioning are associated with
advanced by Schooler (1987). The rst thing to occupational level and change in adulthood is
note here is that for males (and not for females) further addressed by the results from the WLS
there is a small independent effect of occupa- (see Alwin et al. 2008a). Our analysis further
tional position at age 35 on the WAIS score at age examined the extent to which these results are
52, net of pre-adult ability scores, family back- affected by explicitly controlling for pre-adult
ground, and school experiences. There is no levels of intellectual functioning (results not
detectable signicant effect of occupational level shown). Our results demonstrate that Kohn and
at age 52 on the WAIS score, net of earlier occu- his colleagues may have reached an incorrect
pation. This nding undermines the argument of conclusion about the magnitude of these effects
Schooler (1987) that changes in the substantive by not having controlled for pre-adult intellectual
complexity of the job will result in cognitive background in estimating the reciprocal effects of
shifts. There is evidence for a small positive job experiences and cognitive factors (Kohn and
effect, but it does not reach statistical Schooler 1978, 1983; Kohn and Slomczynski
signicance. 1990; Schooler 1984, 1987, 1998). Specically,
if we omit the measure of pre-adult intellectual
functioning from the estimated models, we see
8 Cognitive Change that a much larger effectalmost twice the mag-
in Adulthood nitudeemerges from the analysis. In this case,
both occupational level at age 35, as well as the
Despite the high level of stability in CF in adult- net effect of occupation at age 52, are highly sig-
hood (i.e. lower levels of change), it is possible nicant, suggesting to us that Kohn, Schooler and
that portions of the population that experience their colleagues may have isolated an effect that
change in social locations at different points in was instead partly attributable to pre-adult levels
476 D.F. Alwin et al.

of intellectual functioning, a variable which was (see Hofer and Alwin 2008, pp. 122258, for a
not taken into account in their analysis. To the review). Although the differences among persons
extent that current job conditions reect earlier tend to remain stable through midlife, there is a
selection processes of the type described above, great deal more cognitive change in older age and
they may have overestimated the role of changes therefore more room for environmental inu-
in environmental conditions in current intellec- ences (see Fig. 3 above). The systematic effects
tual functioning by not taking into account selec- of environmental change on aspects of CF under-
tion processes captured by the measure of scores the plasticity of the phenomenon, includ-
pre-adult intellectual functioning). In their ing its age-related changes. There is hardly any
results, the effects of education are stronger than disagreement among students of aging and
those of job experiences, clearly reecting pre- cognition that with age there are systematic
adult differences in cognitive skills: The impact declines in cognitive function, but there is an
of education may be much more inuential in increasing demand placed on research to identify
peoples cognitive development than those of the environmental events and transitions that
occupations. accompany these declines.

9 Cognitive Change 9.1 Event-Centered Models


in Older Age for Transitions in Older Age

The literature on cognition and aging is enor- Event-centered models are appropriate for the
mous. Cognitive function involves many abili- interplay between cognitive change and the events
ties; some, but not all of which, decline in older of older agehealth events, retirement and other
agee.g., attention, reaction time, processing transitions in and out of the labor market. Figure
speed, learning, memory, language, sensory 10 presents a general framework for conceptual-
function, inductive reasoning, and knowledge izing studying growth trajectories using event-

0.5

-0.5
50 55 60 65

Fig. 10 Event-based growth model formulations for the study of cognitive aging and labor-force transitions
Cognitive Development and the Life Course: Growth, Stability and Decline 477

centered latent change models. The purpose of researchers proceeding from other perspectives
such a framework is to be able to (a) conceptual- (Stern and Carstensen 2000, p. 9).
ize patterns of change before and after a life Consistent with the conclusions of The Aging
course event, such as a health event, or labor force Mind, there is a growing body of research that has
transitions, (b) to formulate models for the predic- raised several hypotheses to explain the apparent
tion of event occurrences from information con- decline in cognitive functioning with increased
tained in the growth/change trajectory, and (c) to aging, which are often bounded by disciplinary
formulate models for the prediction of rates of biases. The rst of these, which is popular among
change from knowledge of the prior growth/ biological aging specialists, posits that cognitive
change trajectory and the event of interest. decline is the result of a more general process of
The model depicted in Fig. 10 reveals one way neurological decline (a common cause) that affects
to think about event-based growth models using both cognitive and sensory functioning. Cognitive
the example of retirement. The pattern of growth decline is, thus, simply one manifestation of a
(or decline) on the y-axis is intended to represent more general neurological decline (Stern and
some predictor of a life course event, say health or Carstensen 2000, p. 10). A second set of explana-
some other measure of well-being. For our pres- tory factors involves overall age-related declines
ent purposes, consider the y-dimension to be cog- in specic organ function that lead to increases in
nitive function, and the x-dimension is time from the onset of illnesses or impairments, and risks of
the event at t = 0. We return later to a discussion of experiencing multiple illnesses or impairments.
one study that used this approach in the study of Research drawing on this perspective focuses on
retirement, calibrating changes in cognitive func- disease processes per se (e.g. the onset and man-
tioning prior to retirement, centering the measure- agement of diabetes), and their consequences for
ment of change on the timing of the event. Other CF (Stern and Carstensen 2000, p. 10). Third, cog-
examples are possible, but we will not go into nitive decline may be related directly to other age-
them here. In the following brief sections, we related functional limitations, particularly in
review the available literature that has focused on sensory function and physical health.
two major life course eventschanges in health In contrast to the common cause theories of
trajectories and transitions in and out of the labor cognitive aging (mentioned above), there is a
force, specically on retirement. great deal more evidence supporting the role of
biological factors in cognitive aging. Cognitive
change has been linked to a number of chronic
9.2 Cognition and Health disease conditions, including cardiovascular dis-
ease (Geroldi et al. 2003; Pavlik et al. 2003), pul-
There is a great deal of literature on the connec- monary function (Anstey et al. 2003), type 2
tions between cognition, health and aging (see diabetes (Hassing et al. 2003, 2004a, b; Cosway
Alwin and Hofer 2011); and there is an emerging et al. 2001; Crooks et al. 2003), stroke (Ostir et al.
consensus that a multidisciplinary theoretical 2003), and depressive symptoms (Wilson et al.
approach is necessary to understand the nature of 2004). Chronic health conditions reect serious
processes of cognitive aging; however, the imple- limitations that may inhibit normal activities that
mentation of this objective has not been fully promote the maintenance of cognitive function.
achieved. The National Research Council report Longitudinal data are needed in order to assess
The Aging Mind (Stern and Carstensen 2000), for the extent to which change in health conditions
example, suggested that while progress is being and disability status are linked to age-related cog-
made by behavioral science, cognitive science nitive change. These issues are best investigated
and neuroscience researchers in understanding using event-centered models of the types dis-
cognitive changes during the aging process cussed here, in which onset of disease is used to
what is being learned from each research per- investigate pre- and post-event declines in CF.
spective has not fully penetrated the work of
478 D.F. Alwin et al.

A 2014 Alzheimers Disease International 9.3 Cognition and the Transition


Report (2014) provides an excellent summary of to Retirement
meta-analytic evidence to date from cohort stud-
ies of the association of cardiovascular factors For many years, we have known that the onset of
and cognitive aging. The summary provides health problems and the adequacy of pension
information on whether meaningful conclusions benets were key factors in the decision to retire
could be based on existing cohort studies and dif- (e.g., Bound et al. 1999; Cahill et al. 2006), but
ferentiates between both distal and proximal that little is known about the non-pecuniary
early-life, midlife, and later life factors. The con- determinants of retirement decisions (Hanushek
ceptual framework for this report coincides with and Maritato 1996). The lack of information in
that presented in an intriguing recent paper by these spheres has persisted into more recent
Muller and colleagues (2014), in which they pro- years. Although these studies have provided
posed a life course model to explain the role of important pieces of the health-retirement rela-
fetal and other early origin factors on later dis- tionship, few have examined whether other
eases implicated in cognitive aging. In particular, domains of healthin particular, cognitive func-
they examined factors related to at the intrauter- tioning and psychological well-beingmight
ine environment, placental function, and material also affect the retirement decision.
nutrition during pregnancy, socioeconomic status Given the shift toward greater technological
and nutrition in childhood and adolescence, and sophistication in most industries and the probable
lifestyle factors and CVD risks and disease and, need for greater numbers of older workers to
in turn, to brain structure and function later in remain in their jobs in the future, a broader under-
life. Their analyses showed that small birth standing of the health (and other) factors that
sizeas an indicator of an adverse intrauterine push workers into retirement may help inform
environmenthas lifelong consequences on later both employers and policy makers. To date there
exposure to health risks and opportunities, as are no studies assessing the role of cognitive
well as subsequent brain tissues and cognitive functioning in transitions to retirement; instead,
function, and that this relationship is particularly retirement is examined as a precursor to cogni-
strong in people with lower educational levels tive change. Although these studies have pro-
who may have fewer other social capital resources vided important pieces of the health-retirement
available to ameliorate the earlier adverse effects. relationship, none have examined whether other
Expanding on the Muller et al. (2014) frame- domains of functioningin particular, cognitive
work, the ADI (2014) report used a life course performancemight also affect the retirement
approach to examine early life, midlife, and later decision. however, to our knowledge, no studies
life factors that have been hypothesized to affect to date have examined the relationship between
cognitive decline. Although they found insuf- cognitive performance and changes in work sta-
cient evidence for meaningful conclusions on tus. One can imagine that different cognitive
many early-life factors and psychological factors skills might either help or hinder the ability of a
in midlife and later life, moderate or robust and worker to perform his/her job at the desired or
consistent evidence was found for some of those required level. For example, better memory is
factors. Higher levels of education, for example, likely to help workers manage quickly changing
were consistently associated with lower risk for or analytic demands associated with their jobs,
cognitive aging; and occupational status effects and may also compensate for declines in work-
produced similar protective effects, largely work- ers physical health by providing ideas of alterna-
ing through education. Depression was also tive ways to perform their jobs. In contrast,
strongly associated with cognitive agingbut declines in such skills may foster psychological
only in later lifebut the causal ordering remains or physiological vulnerabilities, and, in turn,
unclear. reduce the effects of job performance that may
Cognitive Development and the Life Course: Growth, Stability and Decline 479

impel workers to retire earlier than they might retirement. The study results were clear and
otherwise want or need. strong: The levels of subjective memory predict
One possible approach to the study of retire- retirement in older workers, net of other docu-
ment as a developmental process is depicted in mented social, health, and economic correlates of
Fig. 10. In this case there are three points along retirement. We also hypothesized that increasing
the age metric where retirement may occur, age rates of decline in either objective or subjective
55, age 60 and age 65. These are completely arbi- memory performance would impel workers to
trary choices, as retirement can obviously happen leave the work force. The study results did not
at any agethese are simply examples. Hence, support this hypothesis. Instead, as noted, level of
the model suggests that retirement at any age subjective memory appeared to be more impor-
may be accompanied by the prior within-person tant than rates of change in either measure. These
pattern of change. If so, then one might think of results are somewhat surprising given a strong
the event of retirement to structure the analysis, reason for the belief that changes in cognitive sta-
so that rather than chronological age, or time in tus are linked to a disengagement from the labor
study, one would measure change on the metric force. Finally, we posited that the effects of sub-
of time to retirement. In an effort to broaden our jective memory would mediate the effects of
understanding of the range of factors known to objective memory performance, and we found
have a bearing in retirement, Wray and col- that it does not.
leagues (Wray et al. 2010) investigated the poten- One possible explanation for these results may
tial impact of one aspect of cognitive be that the effects on work transitions of levels of
functioningmemory performanceon the or changes in memory (or other cognitive) per-
decision to retire, using such a model. Guided by formance depend on the type of job a person
the event-centered model discussed here, esti- holds and, in particular, on the level of cognitive
mated using a combination of latent curve and demands placed by that job. For example, Park
probit regression models, they examined the (1994) posited that cognitive performance may
effects of the level and change in memory perfor- not readily affect job performance (and, by exten-
mance as predictors of retirement in the Health sion, retirement decisions) in most older workers.
and Retirement Study (HRS), where detailed They may hold maintenance jobs which require
employment histories are available along with no new learning; their job experience, age, and/or
overtime measures of cognitive performance. tenure may protect them against declines in the
They used six waves of data from the Health and abilities needed for their jobs; or they may use
Retirement Study on a sample of men and women environmental supports (e.g., younger col-
age 5161 in 1992 in order to estimate how two leagues, teamwork, shifting to jobs with different
reports of memory performanceobjective skill sets) to compensate for actual or perceived
assessments and perceptions of performance declines. In contrast, workers whose jobs are
between 1996 and 2000are linked to the transi- characterized by high levels of complexity (e.g.,
tion from working in 1992 and 1994 to retirement memory skills, abstract thinking) may well be
between 2000 and 2002. Their research employed affected by either actual or perceived decits in
both an objective measure of memory perfor- their cognition, such that the complexity may
mance and a subjective assessment of memory, slow declines and keep workers performing satis-
within a framework that accounts for their rela- factorily, or the complexity may outweigh work-
tionship, in our examination of these effects. In ers actual or perceived abilities.
sum, the study results relating to memory perfor- The results indicated that memory perfor-
mance and perceptions of memory performance mance appears to play a role in transitions to
are intriguing and potentially policy-relevant. retirement. In particular, high levels of subjective
The study results augmented current knowl- memory (but not objective performance) are
edge by answering questions on the effects of associated with a lower likelihood of retirement in
memory performance on exits from paid work to this population, net of other relevant predictors.
480 D.F. Alwin et al.

However, in our results rates of change in mem- events, transitions and changes in the social
ory performance over 4 years do not alter the environment.
likelihood of retirement. Poorer healthboth Finally, our review discussed four major areas
physical and cognitivenegatively impacts indi- where our approach to the study of within-person
viduals and society via lost productivity, impaired change can be effectively used in broadening our
family and social functioning, and economic understanding the nature of the social processes
well-being. The results of this and future studies surrounding life course transitions and human
may point to areas for policy or workplace inter- development. We described how event-centered
ventions that enable workers to remain in their latent change models can be applied to early
jobs longer than they might otherwise by com- child cognitive development and the transition to
pensating for either perceived or actual memory school, specically focusing on the factors that
decits. contribute to successful transitions, and how
transitions themselves contribute to further
change. We then discussed the transition to adult-
10 Conclusions hood and midlife, in which CF in adolescence
leads to major inuences on educational achieve-
In this chapter, we focused on the importance of ment, occupational success and CF in adulthood.
considering a life course perspective in under- On the basis of several empirical examples, we
standing the life-span development of cognitive concluded that potential changes in CF during
function (CF), considered broadly to refer to the midlife were minimal. And nally, we consid-
human ability to manipulate the environment in ered the applicability of our life course frame-
such ways as to solve both simple and complex work to the study of transitions in later life and
problems posed by that (primarily external) envi- how theoretically CF has a role to play both as
ronment. We argued that these processes condi- consequence and cause of health-related and
tion the formation, expression, maintenance and labor-force transitions in older age. The approach
loss of CF across the entire life span. We began we propose can be applied in the study of other
by reviewing the best available theoretical mod- developmental trajectories where social context
els for the life-span development of CF, which is relevant.
stressed the fact that childhood and adolescence
are periods of growth, adulthood is a period of
maintenance/stability, and older age is a period of References
decline. In contrast to the dominant literatures on
CF, our approach emphasizes the fundamental Aikens, N. L., & Barbarin, O. (2008). Socioeconomic dif-
ferences in reading trajectories: The contribution of
nature of the structure, sequence and dynamics of
family, neighborhood, and school contexts. Journal of
life course events and social pathways that are Educational Psychology, 100(2), 235251.
relevant at various life stage phases with respect Alwin, D. F. (1988). Structural equation models in
to cognitive development. research on human development and aging. In K. W.
Schaie, R. T. Campbell, W. Meredith, & S. C. Rawlings
We provide a foundational framework for con-
(Eds.), Methodological issues in aging research
ceptualizing the concept of within-person change (pp. 71170). New York: Springer.
using a latent variable approach, and show how Alwin, D. F. (1994). Aging, personality and social change:
the life course perspective can be used to develop The stability of individual differences over the adult
life-span. In D. L. Featherman, R. M. Lerner, &
event-centered models for understanding hetero-
M. Perlmuter (Eds.), Life-span development and
geneity of within-person change. Specically, we behavior (Vol. 12, pp. 135185). Hillsdale: Lawrence
introduced the concept of the latent difference Erlbaum Associates.
score as an embodiment of the study of the Alwin, D. F. (2008). Social structure and cognitive
change. In S. M. Hofer & D. F. Alwin (Eds.),
growth, maintenance, and decline in CF over the
Handbook of cognitive aging: Interdisciplinary per-
life span, and discuss how this approach can be spectives (pp. 418444). Thousand Oaks: Sage
integrated with the life course analysts interest in Publications.
Cognitive Development and the Life Course: Growth, Stability and Decline 481

Alwin, D. F. (2009). History, Cohorts, and Cognitive Alzheimers Disease International (ADI). (2014). World
Aging. In H. B. Bosworth & C. Hertzog (Eds.), Aging Alzheimers report 2014: Dementia and risk reduc-
and cognition: Research methodologies and empirical tion: An analysis of protective and modiable risk fac-
advances [Lead chapter in a Festschrift for K. Warner tors. London: ADI.
Schaie] (pp. 928). Washington, DC: American Anstey, K. J., Hofer, S. M., & Luszcz, M. A. (2003). A
Psychological Association. latent growth curve analysis of late-life sensory and
Alwin, D. F. (2010). Social structure and cognitive ageing. cognitive function over 8 years: Evidence for specic
In D. Dannefer & C. Phillipson (Eds.), International and common factors underlying change. Psychology
handbook of social gerontology (pp. 265279). and Aging, 18, 714726.
London: Sage Publications International. Baltes, P. B. (1987). Theoretical propositions of life-span
Alwin, D. F. (2012). Integrating varieties of life course developmental psychology: On the dynamics between
concepts. Symposium on the Life Course. Journals growth and decline. Developmental Psychology, 23,
of Gerontology: Psychological Sciences and Social 611626.
Sciences, 67B, 115. Baltes, P. B. (1997). On the incomplete architecture of
Alwin, D. F. (n.d). A primer in the use of growth models to human ontogeny: Selection, optimization and com-
study within-person change. Unpublished manuscript. pensation as the foundation of developmental theory.
Center for Life Course and Longitudinal Studies, American Psychologist, 52, 366380.
Pennsylvania State University. Baltes, P. B., & Baltes, M. (1980). Plasticity and variabil-
Alwin, D. F., & Hofer, S. M. (2011). Health and cognition ity in psychological aging: Methodological and theo-
in aging research. In D. F., Alwin, & S. M., Hofer retical issues. In G. E. Gurski (Ed.), Determining the
(Eds.), Cognition, health and aging: Integrating per- effects of aging on the central nervous system
spectives across disciplines. Supplement to The (pp. 4166). Berlin: Schering.
Journals of Gerontology: Psychological Sciences and Baltes, P. B., & Baltes, M. M. (1990). Psychological per-
Social Sciences, 66, 916. spectives on successful aging: The model of selective
Alwin, D. F., & McCammon, R. J. (1999). Aging vs. optimization with compensation. In P. B. Baltes & M.
cohort interpretations of intercohort differences in M. Baltes (Eds.), Successful aging: Perspectives from
GSS verbal scores. American Sociological Review, 64, the behavioral sciences (pp. 134). New York:
272286. Cambridge University Press.
Alwin, D. F., & McCammon, R. J. (2001). Aging, cohorts, Baltes, P. B., & Mayer, K. U. (1999). The Berlin aging
and verbal ability. Journal of Gerontology: Social study: Aging from 70 to 100. Cambridge, UK:
Sciences, 56B, S1S11. Cambridge University Press.
Alwin, D. F., & Thornton, A. (1984). Family origins and Baltes, P. B., Staudinger, U. M., & Lindenberger, U.
the schooling process: Early vs. late inuence of (1999). Lifespan psychology: Theory and application
parental characteristics. American Sociological to intellectual functioning. In Annual review of psy-
Review, 49, 784802. chology, 50 (pp. 471507). Palo Alto: Annual Reviews.
Alwin, D. F., Cohen, R. L., & Newcomb, T. M. (1991). Barker, D. J. P. (1998). Mothers, babies, and health in
Political attitudes over the life span: The Bennington later life. London: Churchill Livingstone.
women after fty years. Madison: University of Barnsley, R. H., & Thompson, A. H. (1988). Birthdate and
Wisconsin Press. success in minor hockey: The key to the NHL.
Alwin, D. F., Hofer, S. M., & McCammon, R. J. (2006). Canadian Journal of Behavioral Science, 20,
Modeling the effects of time: Integrating demographic 167176.
and developmental perspectives. In R. H. Binstock & Barnsley, R. H., Thompson, A. H., & Barnsley, P. E.
L. K. George (Eds.), Handbook of aging and the social (1985). Hockey success and birthdate: The relative age
sciences (pp. 2038). New York: Academic. effect. Canadian Association for Health, Physical
Alwin, D. F., Wray, L. A., & McCammon, R. J. (2008a). Education, and Recreation, 51, 2328.
What determines individual differences in intellectual Baumert, J., Nagy, G., & Lehmann, R. (2012). Cumulative
functioning in midlife. Center for Life Course and advantages and the emergence of social and ethnic
Longitudinal Studies, Penn State University. inequality: Matthew effects in reading and mathemat-
Unpublished paper. ics development within elementary schools? Child
Alwin, D. F., McCammon, R. J., Wray, L. A., & Rodgers, Development, 83(4), 13471367.
W. L. (2008b). Population processes and cognitive Berger, L. M., Paxson, C., & Waldfogel, J. (2009). Income
aging. In S. M. Hofer & D. F. Alwin (Eds.), Handbook and child development. Child Youth Services Review,
of cognitive aging: Interdisciplinary perspectives 31(9), 978989.
(pp. 6989). Thousand Oaks: Sage Publications. Blau, D., & Currie, J. (2006). Pre-school, day care, and
Alwin, D. F., & others (n.d). The stability of individual after-school care: Whos minding the kids? In E. A.
differences in cognitive functioning over the life span. Hanushek & F. Welch (Eds.), Handbook on the eco-
Unpublished manuscript. Center for Life Course and nomics of education (Vol. 2). Amsterdam: North
Longitudinal Studies, Pennsylvania State University. Holland.
482 D.F. Alwin et al.

Blau, P. M., & Duncan, O. D. (1967). The American occu- kindergarten through the third grade. Sociology of
pational structure. New York: Wiley. Education, 81, 131.
Bloom, B. S. (1964). Stability and change in human char- Chomsky, N. (1965). Aspects of the theory of syntax.
acteristics. New York: Wiley. Cambridge, MA: MIT Press.
Bodovski, K., & Farkas, G. (2008). Concerted cultiva- Condron, D. J. (2009). Social class, school and non-
tion and unequal achievement in elementary school. school environments, and black/white inequalities in
Social Science Research, 37, 903919. childrens learning. American Sociological Review,
Bollen, K. A., & Curran, P. J. (2006). Latent curve mod- 74(5), 683708.
els: A structural equation perspective. Hoboken: Cosway, R., Strachan, M. W. J., Dougall, A., Frier, B. M.,
Wiley. & Deary, I. J. (2001). Cognitive function and informa-
Bound, J., Schoenbaum, M., Stinebrickner, T. R., & tion processing in type 2 diabetes. Diabetic Medicine,
Waidmann, T. (1999). The dynamic effects of health 18, 803810.
on the labor force transitions of older workers. Labour Crimmins, E. M. (2001). Mortality and health in human
Economics, 6, 179202. life spans. Experimental Gerontology, 36, 885897.
Bowles, S., Gintis, H., & Groves, M. O. (2005). Unequal Crimmins, E. M. (2005). Socioeconomic differentials in
chances: Family background and economic success. mortality and health at the older ages. Genus, 61,
Princeton: Princeton University Press. 163176.
Bradley, R. H., & Corwyn, R. F. (2002). Socioeconomic Crimmins, E. M., Hayward, M. D., & Seeman, T. (2004).
status and child development. Annual Review of Race/ethnicity, socioeconomic status, and health. In
Psychology, 53, 371399. N. Anderson, R. Bulatao, & B. Cohen (Eds.), Critical
Bronfenbrenner, U. (1979). The ecology of human devel- perspectives on race and ethnic differences in health
opment: Experiments by nature and design. in later life (pp. 310352). Washington, DC: National
Cambridge, MA: Harvard University Press. Academies Press.
Bronfenbrenner, U., & Ceci, S. J. (1994). Nature-nurture Cronbach, L. J., & Furby, L. (1970). How should we mea-
reconceptualized in developmental perspective: A bio- sure change Or should we? Psychological Bulletin,
ecological model. Psychological Review, 101, 74, 880.
568586. Crooks, V. C., Buckwalter, J. G., & Petitti, D. B. (2003).
Brookmeyer, R., & Gray, S. (2000). Methods for project- Diabetes mellitus and cognitive performance in older
ing the incidence and prevalence of chronic diseases in women. Annals of Epidemiology, 13, 613619.
ageing populations: Application to Alzheimers dis- Cunha, F., & Heckman, J. J. (2007). The technology of
ease. Statistics in Medicine, 19, 14811493. skill formation. American Economic Review, 97,
Brookmeyer, R., Gray, S., & Kawas, C. (1998). Projections 3147.
of Alzheimers disease in the United States and the Cunha, F., & Heckman, J. J. (2008). Formulating, identi-
public health impact of delaying disease onset. fying and estimating the technology of cognitive and
American Journal of Public Health, 88, 13371342. noncognitive skill formation. Journal of Human
Cahill, K. E., Giandrea, M. D., & Quinn, J. F. (2006). Resources, 43, 738782.
Retirement patterns from career employment. The Curran, P. J., & Bollen, K. A. (2001). The best of both
Gerontologist, 46, 514523. worlds: Combining autoregressive and latent curve
Carey, J. R. (2003). Longevity: The biology and demogra- models. In L. M. Collins & A. G. Sayer (Eds.), New
phy of life span. Princeton: Princeton University Press. methods for the analysis of change (pp. 105135).
Carroll, J. B. (1996). Mathematical abilities: Some results Washington, DC: American Psychological
from factor analysis. In R. J. Sternberg & R. Ben-Zeev Association.
(Eds.), The nature of mathematical thinking (pp. 325). Currie, J., & Thomas, D. (1995). Does head start make a
Mahwah: Erlbaum. difference? American Economic Review, 85,
Carroll, J. B. (1998). Human cognitive abilities. In J. J. 341364.
McArdle & R. W. Woodcock (Eds.), Human cognitive Dannefer, W. D. (2003). Cumulative advantage/disadvan-
abilities in theory and practice (pp. 524). Mahwah: tage and the life course: Cross-fertilizing age and
Erlbaum. social science theory. The Journals of Gerontology
Cattell, R. B. (1963). Theory of uid and crystallized Series B: Psychological Sciences and Social Sciences,
intelligence: A critical experiment. Journal of 58, S327S357.
Educational Psychology, 54, 122. Davis-Kean, P. E. (2005). The inuence of parent educa-
Cattell, R. B. (1971a). The structure of intelligence in tion and family income on child achievement: The
relation to the nature-nurture controversy. In R. Cancro indirect role of parental expectations and the home
(Ed.), Intelligence: Genetic and environmental inu- environment. Journal of Family Psychology, 19,
ences (pp. 330). New York: Grune and Stratton. 294304.
Cattell, R. B. (1971b). Abilities: Their structure, growth De Kieviet, J. F., Zoetebier, L., Van Elburg, R. M.,
and action. Boston: Houghton-Mifin. Vermeulen, R. J., & Oosterlaan, J. (2012). Brain devel-
Cheadle, J. E. (2008). Educational investment, family opment of very preterm and very Low-birthweight
context, and childrens math and reading growth from children in childhood and adolescence: A meta-
Cognitive Development and the Life Course: Growth, Stability and Decline 483

analysis. Developmental Medicine and Child Entwistle, D., Alexander, K. L., & Olson, L. S. (2003).
Neurology, 54, 313323. The rst grade transition in life course perspective. In
Denny, N. W. (1982). Aging and cognitive change. In J. T. Mortimer & M. J. Shanahan (Eds.), Handbook of
B. B. Wolman (Ed.), Handbook of developmental psy- the life course (pp. 229250). New York: Kluwer
chology: Research and theory (pp. 807827). Academic/Plenum Publishers.
Englewood Cliffs: Prentice-Hall. Farkas, G., & Beron, K. (2004). The detailed age trajec-
Diamond, A., Barnett, W. S., Thomas, J., & Munro, S. tory of oral vocabulary knowledge: Differences by
(2007). Preschool program improves cognitive con- race and class. Social Science Research, 33, 464497.
trol. Science, 318, 13871388. Featherman, D. L. (1983). Life-span perspectives in social
DiPrete, T. A., & Eirich, G. M. (2006). Cumulative advan- science research. In P. B. Baltes & O. G. Brim Jr.
tage as a mechanism for inequality: A review of theo- (Eds.), Life-span development and behavior (Vol. 5,
retical and empirical developments. Annual Review of pp. 157). New York: Academic.
Sociology, 32, 271297. Featherman, D. L., & Lerner, R. W. (1985). Ontogenesis
Downey, D. B., von Hippel, P. T., & Broh, B. A. (2004). and sociogenesis: Problematics for theory and research
Are schools the great equalizer? Cognitive inequality about development and socialization across the life
during the summer months and the school year. span. American Sociological Review, 50, 659676.
American Sociological Review, 69(5), 613635. Ferraro, K. F., & Shippee, T. P. (2009). Aging and cumula-
Duncan, O. D., Featherman, D. I., & Duncan, B. (1972). tive inequality: How does inequality get under the
Socioeconomic background and achievement. skin? The Gerontologist, 49, 333343.
New York: Seminar Press. Finch, C. E., & Kirkwood, T. B. L. (2000). Chance, devel-
Duncan, G., Kalil, A., Mayer, S. E., Tepper, R., & Payne, opment and aging. Oxford: Oxford University Press.
M. R. (2005). The apple does not fall far from the tree. Flanagan, D. P., McGrew, K. S., & Ortiz, S. O. (2000).
In S. Bowles, H. Gintis, & M. O. Groves (Eds.), The Wechsler intelligence scales and Gf-Gc theory: A
Unequal chances: Family background and economic contemporary approach to interpretation. Boston:
success (pp. 2379). New York: Russell Sage Allyn and Bacon.
Foundation. Friedman, H. S., & Kern, M. L. (2014). Personality, well-
Duncan, G. J., Dowsett, C. J., Claessens, A., Magnuson, being, and health. Annual Review of Psychology, 65,
K., Huston, A. C., Klebanov, P., Pagani, L. S., 719742.
Feinstein, L., Engel, M., Brooks-Gunn, J., Sexton, H., Frijters, P., Johnston, D. W., Shah, M., & Shields, M. A.
Duckwork, K., & Jape, C. (2007). School readiness (2013). Intrahousehold resource allocation: Do par-
and later achievement. Developmental Psychology, ents reduce or reinforce child ability gaps?
43(6), 14281446. Demography, 50, 21872208.
Elder, G. H., Jr. (1975). Age differentiation and the life Gelman, R. (1990). Structural constraints on cognitive
course. Annual Review of Sociology, 1, 165190. development: Introduction to a special issue of cogni-
Elder, G. H., Jr. (1985). Perspectives on the life course. In tive science. Cognitive Science, 14, 39.
G. H. Elder Jr. (Ed.), Life course dynamics: Geroldi, C., Ferrucci, L., Bendinelli, S., Cavazzini, C.,
Trajectories and transitions, 19681980 (pp. 2349). Zanetti, O., Guralnik, J. M., & Frisoni, G. B. (2003).
Ithaca: Cornell University Press. Journal of the American Geriatrics Society, 51,
Elder, G. H., Jr. (1997). The life course and human devel- 10641071.
opment. In R. M. Lerner (Ed.), Handbook of child psy- Gershoff, E. T., Aber, J. L., Raver, C. C., & Lennon, M. C.
chology: Vol. 1. Theoretical models of human (2007). Income is not enough: Incorporating material
development (pp. 939991). New York: Wiley. hardship into models of income associations with par-
Elder, G. H., Jr. (2000). The life course. In E. F. Borgatta enting and child development. Child Development, 78,
& R. J. V. Montgomery (Eds.), Encyclopedia of soci- 7095.
ology (2nd ed., Vol. 3, pp. 16141622). New York: Giedd, J. N. (2008). The teen brain: Insights from neuro-
Macmillan Reference USA. imaging. Journal of Adolescent Health, 42, 335343.
Elder, G. H., Jr. (1997a, April 5). The life course as devel- Giedd, J. N., Blumenthal, J., Jeffries, N. O., Castellanos,
opmental theory. Presidential address, Society for F. X., Liu, H., Zijenbos, A., Paus, T., Evans, A. C., &
Research on Child Development, Washington, DC. Rapoport, J. L. (1999). Brain development during
Elder, G. H., Jr., & Shanahan, M. J. (2006). The life course childhood and adolescence: A longitudinal MRI study.
and human development. In W. Damon & R. M. Nature Neuroscience, 2, 861863.
Lerner (Eds.), Handbook of child psychology Gladwell, M. (2008). Outliersthe story of success.
(Theoretical models of human development 6th ed., New York: Little Brown & Co.
Vol. 1, pp. 665715). New York: Wiley. Goldstone, R. L., & Landy, D. (2010). Domain-creating
Elder, G. H., Jr., Johnson, M. K., & Crosnoe, R. (2003). constraints. Cognitive Science, 34, 13571377.
The emergence and development of life course theory. Gopnik, A. (2012). Scientic thinking in young children:
In J. T. Mortimer & M. J. Shanahan (Eds.), Handbook Theoretical advances, empirical research, and policy
of the life course (pp. 319). New York: Kluwer implications. Science, 337, 16231627.
Academic/Plenum Publishers.
484 D.F. Alwin et al.

Gopnik, A., & Meltzoff, A. N. (1998). Words, thoughts, and Hayward, M. D., Crimmins, E. M., Miles, T., & Yang, Y.
theories. Cambridge, MA: The MIT University Press. (2000). The signicance of socio-economic status in
Gottfredson, L. S., & Deary, I. J. (2004). Intelligence pre- explaining the race gap in chronic health. American
dicts health and longevity, but why? Current Directions Sociological Review, 65, 910930.
in Psychological Science, 13, 14. Heckman, J. J. (2007). The economics, technology, and
Guo, G., & Harris, K. M. (2000). The mechanisms medi- neuroscience of human capability formation.
ating the effects of poverty on childrens intellectual Proceedings of the National Academy of Sciences,
development. Demography, 37(4), 431447. 104, 1325013255.
Hanushek, E. A., & Maritato, N. L. (Eds.). (1996). Heise, D. R. (1970). Causal inference from panel data. In
Assessing knowledge of retirement behavior. Panel on Sociological methodology 1970 (pp. 327). San
Retirement Income Modeling. Committee on Francisco: Jossey-Bass.
Statistics. Commisson on Behavioral and Social Henretta, J. (2003). The life-course perspective on work
Sciences and Education. National Research Council. and retirement. In R. A. Settersten Jr. (Ed.), Invitation
Washington, D.C.: National Academy Press. to the life course: Toward new understandings of later
Harris, C. W. (1963). Problems in measuring change. life (pp. 85105). Amityville: Baywood Publishing
Madison: The University of Wisconsin Press. Company.
Harris, J. R. (1995). Where is the childs environment? A Hertzog, C., & Schaie, K. W. (1986). Stability and change
group socialization theory of development. in adult intelligence. 1: Analysis of longitudinal cova-
Psychological Review, 102, 458489. riance structures. Psychology and Aging, 1, 159171.
Harris, J. R. (1998). The nurture assumption: Why chil- Hertzog, C., & Schaie, K. W. (1988). Stability and change
dren turn out the way they do. New York: Simon & in adult intelligence: 2. Simultaneous analysis of lon-
Schuster. gitudinal means and covariance structures. Psychology
Hart, B., & Risley, T. R. (1995). Meaningful differences in and Aging, 3, 122130.
the everyday experiences of young American children. Hofer, S. M., & Alwin, D. F. (2008). Handbook on cogni-
Baltimore: Paul H. Brookes. tive aging: Interdisciplinary perspectives. Thousand
Hassing, L. B., Johansson, B., Pedersen, N. L., Nilsson, Oaks: Sage Publications.
S. E., Berg, S., & McClearn, G. (2003). Type 2 diabe- Hogan, D. P. (1981). Transitions and social change: The
tes mellitus and cognitive performance in a population- early lives of American men. New York: Academic.
based sample of the oldest old: Impact of comorbid Hogan, D. P. (2000). Life cycle. In E. F. Borgatta & R. J.
dementia. Aging, Neuropsychology, & Cognition, 10, V. Montgomery (Eds.), Encyclopedia of sociology
99107. (2nd ed., Vol. 3, pp. 16231627). New York:
Hassing, L. B., Grant, M. D., Hofer, S. M., Pedersen, Macmillan Reference USA.
N. L., Nilsson, S. E., Berg, S., McClearn, G. E., & Hogan, D. P., & Astone, N. M. (1986). The transition to
Johansson, B. (2004a). Type 2 diabetes mellitus con- adulthood. Annual Review of Sociology, 12, 109130.
tributes to cognitive change in the oldest old: A longi- Horn, J. L. (1968). Organization of abilities and the devel-
tudinal population-based study. Journal of the opment of intelligence. Psychological Review, 75,
International Neuropsychological Society, 4, 242259.
599607. Horn, J. L. (1976). Human abilities: A review of research
Hassing, L. B., Hofer, S. M., Nilsson, S. E., Berg, S., and theory in the early 1970s. Annual Review of
Pedersen, N. L., McClearn, G., & Johansson, B. Psychology, 27, 437485.
(2004b). Comorbid type 2 diabetes mellitus and Horn, J. L. (1994). Theory of uid and crystallized intelli-
hypertension exacerbates cognitive decline: Evidence gence. In R. J. Sternberg (Ed.), Encyclopedia of human
from a longitudinal study. Age and Aging, 33, intelligence (pp. 443451). New York: Macmillan.
355361. Horn, J. L., & Cattell, R. B. (1967). Age differences in
Hauser, R. M., Tsai, S.-L., & Sewell, W. H. (1983). A uid and crystallized intelligence. Acta Psychologica,
model of stratication with response error in social 26, 107129.
and psychological variables. Sociology of Education, Horn, J. L., & Donaldson, G. (1980). Cognitive develop-
56, 2046. ment in adulthood. In O. G. Brim Jr. & J. Kagan
Hauser, R. M., Sewell, W. H., Logan, J. A., Hauser, T. S., (Eds.), Constancy and change in human development
Ryff, C., Caspi, A., & MacDonald, M. (1992). The (pp. 445529). Cambridge, MA: Harvard University.
Wisconsin longitudinal study: Adults as parents and House, J. S., Kessler, R. C., Herzog, A. R., Mero, R. P.,
children at age 50. IASSIST Quarterly, 16, 2338. Kinney, A. M., & Breslow, M. J. (1992). Social strati-
Hauser, R. M., Carr, D., Hauser, T. S., Krecker, M., Kuo, cation, age, and health. In K. W. Schaie, D. Blazer, &
D., Presit, D., Shinberg, D., Sweeney, M., Thompson- J. S. House (Eds.), Aging, health behaviors and health
Colon, T., & Uhrig, S. C. N. (1993). The class of 1957 outcomes (pp. 132). Hillsdale: Erlbaum.
after 35 years: Overview and preliminary ndings House, J. S., Lepkowski, J. M., Kinney, A. M., Mero,
(CDE Working Paper 93-17). Madison: University of R. P., Kessler, R. C., & Herzog, A. R. (1994). The
Wisconsin-Madison.
Cognitive Development and the Life Course: Growth, Stability and Decline 485

social stratication of aging and health. Journal of Kuh, D., & New Dynamics of Ageing (NDA) Preparatory
Health and Social Behavior, 35, 213234. Network. (2007). A life course approach to healthy
House, J. S., Lantz, P. M., & Herd, P. (2005). Continuity aging, frailty, and capability. Journal of Gerontology:
and change in the social stratication of aging and Medical Sciences, 62A, 717721.
health over the life course: Evidence from a nationally Kunst, A. E., & Mackenbach, J. P. (1994). The size of
representative longitudinal study from 1986 to mortality differences associated with education level
2001/2002 (Americans Changing Lives Study). in nine industrialized countries. American Journal of
Journal of Gerontology: Social Science, 60B(Special Public Health, 84, 932937.
Issue II), 1526. Lareau, A., & Conley, D. (Eds.). (2008). Social class:
Hox, J. J. (2002). Multilevel analysis: Techniques and How does it work? New York: Russell Sage
applications. Mahwah: Lawrence Erlbaum. Foundation.
Jencks, C. (1979). Who gets ahead? The determinants of Lebel, C., & Beaulieu, C. (2011). Longitudinal develop-
economic success in America. New York: Basic. ment of human brain wiring continues from childhood
Johnson, S. B., Blum, R. W., & Giedd, J. N. (2009). into adulthood. The Journal of Neuroscience, 31(30),
Adolescent maturity and the brain: The promise and 1093710947.
pitfalls of neuroscience research in adolescent health Lindquist, E. F. (1953). Design and analysis of experi-
policy. Journal of Adolescent Health, 45, 216221. ments in psychology and education. Boston: Houghton
Jreskog, K. G. (1974). Analyzing psychological data by Mifin.
structural analysis of covariance matrices. In D. H. Loehlin, J. C. (1997). A test of J.R. Harriss theory of peer
Kranz, R. C. Atkinson, R. D. Luce, & P. Suppes (Eds.), inuences on personality. Journal of Personality and
Measurement, psychophysics and neural information Social Psychology, 72, 11971201.
processing. San Francisco: Freeman. Lord, F. M. (1963). Elementary models for measuring
Kaplan, H., Hill, K., Lancaster, J., & Magdalena Hurtado, change. In C. W. Harris (Ed.), Problems in measuring
A. (2000). A theory of human life history evolution: change (pp. 2138). Madison: University of Wisconsin
Diet, intelligence and longevity. Evolutionary Press.
Anthropology, 9, 130. Lynch, S. M. (2003). Cohort and life-course patterns in
Kaplan, G. A., Turrell, G., Lynch, J. W., Everson, S. A., the relationship between education and health: A hier-
Helkala, E. L., & Salonen, J. T. (2001). Childhood archical approach. Demography, 40, 309331.
socioeconomic position and cognitive function in Maccoby, E. E. (1992). The role of parents in the social-
adulthood. International Journal of Epidemiology, 30, ization of children: An historical overview.
256263. Developmental Psychology, 28, 10061017.
Kawas, C. H., & Brookmeyer, R. (2001). Editorialaging Maccoby, E. E., & Martin, J. A. (1983). Socialization and
and the public health: Effects of dementia. The New the context of the family: Parent-child interaction. In
England Journal of Medicine, 344, 11601161. E. M. Hetherington (Ed.), Socialization, personality,
Kern, M. L., & Friedman, H. S. (2008). Early educational and social development. New York: Wiley.
milestones as predictors of lifelong academic achieve- Magnuson, K. A., Meyers, M. K., Ruhm, C. J., &
ment, midlife adjustment, and longevity. Journal of Waldfogel, J. (2004). Inequality in preschool educa-
Applied Developmental Psychology, 30, 419430. tion and school readiness. American Educational
Knudsen, E. I., Heckman, J. J., Cameron, J. L., & Research Journal, 41(1), 115157.
Shonkoff, J. P. (2006). Economic, neurobiological, Magnuson, K. A., Ruhm, C. J., & Waldfogel, J. (2007).
and behavioral perspectives on building Americas The persistence of preschool effects: Do subsequent
future workforce. Proceedings of the National classroom experiences matter? Early Childhood
Academy of Sciences, 103, 1015510162. Research Quarterly, 22, 1838.
Kohli, M. (2007). The institutionalization of the life Mandler, J. M. (2004). The foundations of mind.
course: Looking back to look ahead. Research in New York: Oxford University Press.
Human Development, 4, 253271. Marmot, M. G., & Shipley, M. J. (1996). Do socioeco-
Kohn, M. L., & Schooler, C. (1978). The reciprocal effects nomic differences in mortality persist after retirement?
of the substantive complexity of work and intellectual 25 year follow up of civil servants from the rst
exibility: A longitudinal assessment. American Whitehall Study. British Medical Journal, 313,
Journal of Sociology, 84, 2452. 11771180.
Kohn, M. L., & Schooler, C. (1983). Work and personal- Martin, G. M., Austad, S. N., & Johnson, T. E. (1996).
ity: An inquiry into the impact of social stratication. Genetic analysis of ageing: Role of oxidative damage
Norwood: Ablex. and environmental stresses. Nature Genetics, 13,
Kohn, M. L., & Slomczynski, K. M. (1990). Social struc- 2534.
ture and self-direction: A comparative analysis of the McArdle, J. J. (2001). A latent difference score approach
United States and Poland. Cambridge, MA: Basil to longitudinal dynamic structural analyses. In
Blackwell. R. Cudeck, S. du Toit, & D. Srbom (Eds.), Structural
486 D.F. Alwin et al.

equation modeling: Present and future (pp. 342380). child care. American Educational Research Journal,
Lincolnwood: Scientic Software International. 39, 133164.
McArdle, J. J., & Bell, R. Q. (2000). An introduction to NICHD Early Child Care Research Network. (2003).
latent growth models for developmental data analysis. Does amount of time spent in child care predict socio-
In T. D. Little, K. U. Schnabel, & J. Baumert (Eds.), emotional adjustment during the transition to kinder-
Modeling longitudinal and multilevel data: Practical garten. Child Development, 74, 9761005.
issues, applied approaches and specic examples NICHD Early Child Care Research Network. (2005).
(pp. 69107). Mahwah: Lawrence Erlbaum. Child care and child development: Results from the
McArdle, J. J., & Hamagami, F. (2001). Latent difference NICHD study of early child care and youth develop-
score structural models for linear dynamic analyses ment. New York: Guilford Press.
with incomplete longitudinal data. In L. M. Collins & ORand, A. M. (1996). The precious and the precocious:
A. G. Sayer (Eds.), New methods for the analysis of Understanding cumulative disadvantage and cumula-
change (pp. 137175). Washington, DC: American tive advantage over the life course. The Gerontologist,
Psychological Association. 36, 230238.
McClelland, J. L., & Seigler, R. S. (Eds.). (2001). Ostir, G. V., Maji, M. A., Ottenbacher, K. J., Markides,
Mechanisms of cognitive development: Behavioral K. S., & Goodwin, J. S. (2003). Cognitive function
and neural perspectives. Mahwah: Lawrence Erlbaum and incidence of stroke in older Mexican Americans.
Associates. Journal of Gerontology: Medical Sciences, 58(6),
McGrew, K. S. (1997). Analysis of the major intelligence 531535.
batteries according to a proposed comprehensive Park, D.C. (1994). Aging, cognition, and work. Human
Gf-Gc framework. In D. P. Flanagan, J. L. Genshaft, & Performance, 7, 181205
P. L. Harrison (Eds.), Contemporary intellectual Park, D. C. (1999). Cognitive aging, processing resources,
assessment: Theories, tests, and issues (pp. 151180). and self-report. In D. C. Park & N. Schwarz (Eds.),
New York: Guilford. Cognition, aging, and self-reports (pp. 4569).
McGrew, K. S., & Flanagan, D. P. (1998). The intelligence Philadelphia: Taylor and Francis.
test desk reference (ITDR): Gf-Gc cross-battery Park, D. C., Smith, A. D., Lautenschlager, G., Earles, J.,
assessment. Boston: Allyn and Bacon. Frieske, D., Zwahr, M., & Gaines, C. (1996).
Meredith, W., & Tisak, J. (1990). Latent curve analysis. Mediators of long-term memory performance across
Psychometrika, 55, 107122. the life-span. Psychology and Aging, 11, 621637.
Merton, R. K. (1968). The Matthew effect in science. Paus, T., Zijdenbos, A., Worsley, K., Collins, D. L.,
Science, 159, 5663. Blumenthal, J., Giedd, J. N., Rapoport, J. L., & Evans,
Molla, M. T., Madans, J. H., & Wagener, D. K. (2004). A. C. (1999). Structural maturation of neural pathways
Differentials in adult mortality and activity limitation in children and adolescents: In vivo study. Science,
by years of education in the United States at the end of 283, 19081911.
the 1990s. Population and Development Review, 30, Pavlik, V. N., Alves de Moraes, S., Szklo, M., Knlpman,
625646. D. S., Mosely, T. H., Jr., & Hyman, D. J. (2003).
Morgan, S. L., Grusky, D. B., & Fields, G. S. (2006). Relation between cognitive function and mortality in
Mobility and inequality: Frontiers of research in soci- middle-aged adults in the atherosclerosis risk in com-
ology and economics. Stanford: Stanford University munities study. American Journal of Epidemiology,
Press. 157, 327334.
Muller, M., Sigurdsson, S., Kjartansson, O., et al. (2014). Pedersen, N. L., & Lichtenstein, P. (1997). Biometric
Birth size and brain function 75 years later. Pediatrics, analyses of human abilities. In C. Cooper & V. Varma
134, 761770. (Eds.), Processes in individual differences (pp. 126
Musch, J., & Grondin, S. (2001). Unequal competition as 148). New York: Routledge.
an impediment to personal development: A review of Piaget, J. (1952). The origins of intelligence in children.
the relative age effect in sport. Developmental Review, New York: International Universities Press.
21, 147167. Piaget, J. (1965). Cognitive development in children:
Musgrove, F. (1977). Margins of the mind. London: Piaget. Journal of Research in Science Teaching, 2,
Methuen. 176186.
NICHD Early Child Care Research Network. (2000). The Piaget, J. (1970). Piagets theory. In P. H. Mussen (Ed.),
relation of child care to cognitive and language devel- Carmichaels manual of child psychology (pp. 703
opment. Child Development, 71, 960980. 732). New York: Wiley.
NICHD Early Child Care Research Network. (2002a). Piaget. (1985). The equilibration of cognitive structures:
Child-care structure -> process -> outcome: Direct and The central problem of intellectual development.
indirect effects of child-care quality on young chil- Chicago: University of Chicago Press.
drens development. Psychological Science, 13(3), Pinker, S. (2002). The blank slate: The modern denial of
199206. human nature. New York: Viking.
NICHD Early Child Care Research Network. (2002b). Plomin, R. (1994). Genetics and experience: The inter-
Early child care and childrens development prior to play between nature and nurture. Thousand Oaks:
school entry: Results from the NICHD study of early Sage Publications.
Cognitive Development and the Life Course: Growth, Stability and Decline 487

Plomin, R. (1999). Behavioral genetics. In M. Bennett (Ed.), Scherr, P. A., Albert, M. S., Funkenstein, H. H., Cook,
Developmental psychology: Achievements and pros- N. R., Hennekens, C. H., Branch, L. G., White, L. R.,
pects (pp. 231252). Philadelphia: Taylor & Francis. Taylor, J. O., & Evans, D. A. (1988). Correlates of
Potter, D., & Roksa, J. (2013). Accumulating advantage cognitive function in an elderly community popula-
over time: Family experiences and social class inequal- tion. American Journal of Epidemiology, 128,
ity in academic achievement. Social Science Research, 10841101.
42, 10181032. Schooler, C. (1984). Psychological effects of complex
Potter, D., Mashburn, A., & Grissmer, D. (2013). The environments during the life span: A review and the-
family, neuroscience, and academic skills: An inter- ory. Intelligence, 8, 259281.
disciplinary account of social class gaps in childrens Schooler, C. (1987). Psychological effects of complex
test scores. Social Science Research, 42, 446464. environments during the life span: A review and the-
Richards, M., & Deary, I. J. (2005). A life course approach ory. In C. Schooler & K. W. Schaie (Eds.), Cognitive
to cognitive reserve: A model for cognitive aging and functioning and social structure over the lifecourse
development? Annals of Neurology, 58, 617622. (pp. 2449). Norwood: Ablex.
Riley, M. W. (1987). On the signicance of age in sociol- Schooler, C. (1998). Environmental complexity and the
ogy. American Sociological Review, 52, 114. Flynn effect. In U. Neisser (Ed.), The rising curve:
Rindfuss, R. (1991). The young adult years: Diversity, Long-term gains in IQ and related measures
structural change, and fertility. Demography, 28, (pp. 6779). Washington, DC: American Psychological
493512. Association.
Robert, S., & House, J. S. (1996). SES differentials in Sewell, W. H., Sr., & Hauser, R. M. (1975). Education,
health by age and alternative indicators of SES. occupation and earnings achievement in the early
Journal of Aging and Health, 8, 359388. career. New York: Academic.
Rogosa, D. R., Brandt, D., & Zimowski, M. (1982). A Sewell, W. H., Hauser, R. M., Springer, K. W., & Hauser,
growth curve approach to the measurement of change. T. S. (2003). As we age: The Wisconsin longitudinal
Psychological Bulletin, 90, 726748. study, 19572001. In K. Leicht (Ed.), Social stratica-
Rowe, D. C. (1994). The limits of family inuence: Genes, tion and mobility (Vol. 20, pp. 3111). London: Elsevier.
experience, and behavior. New York: Guilford. Shanahan, M. J. (2000). Pathways to adulthood in chang-
Rowe, J. W., & Kahn, R. L. (1998). Successful aging. ing societies: Variability and mechanisms in life
New York: Pantheon. course perspective. Annual Review of Sociology, 26,
Rowe, D. C., & Teachman, J. (2001). Behavioral genetic 667692.
research designs and social policy studies. In Shanahan, M. J., Hofer, S. M., & Shanahan, L. (2003).
A. Thornton (Ed.), The well-being of children and Biological models of behavior and the life course. In
families: Research and data needs (pp. 157187). Ann J. T. Mortimer & M. J. Shanahan (Eds.), Handbook of
Arbor: University of Michigan Press. the life course. New York: Kluwer Academic/Plenum
Ryder, N. B. (1965). The cohort as a concept in the study Publishers.
of social change. American Sociological Review, 30, Shonkoff, J. P., & Phillips, D. A. (Eds.). (2000). From neu-
843861. rons to neighborhoods: The science of early child
Salthouse, T. A. (1991). Theoretical perspectives on cog- development. Washington, DC: National Academy of
nitive aging. Hillsdale: Lawrence Erlbaum. Sciences.
Salthouse, T. A. (1996). The processing-speed theory of Siegler, R. (1989). Mechanisms of cognitive development.
adult age differences in cognition. Psychological Annual Review of Psychology, 40, 353379.
Review, 103, 403428. Singer, J. D., & Willett, J. B. (2003). Applied longitudinal
Salthouse, T. A. (1999). Pressing issues in cognitive data analysis: Modeling change and event occurrence.
aging. In N. Schwarz, D. Park, B. Knuper, & London: Oxford University Press.
S. Sudman (Eds.), Cognition, aging, and self-reports Smith, J. R., Brooks-Gunn, J., & Klebanov, P. K. (1997).
(pp. 185198). Philadelphia: Psychology Press. Consequences of growing up poor, chap. Consequences
Scarr, S. (1992). Developmental theories for the 1990s: of living in poverty for young childrens cognitive and
Development of individual differences. Child verbal ability and early school achievement.
Development, 63, 119. New York: Russell Sage Foundation.
Scarr, S. (1995). How do families affect intelligence? Snowdon, D. A., Kemper, S. J., Mortimer, J. A., Greiner,
Social environmental and behavior genetic predic- L. H., Wekstein, D. R., & Markesbery, W. R. (1996).
tions. In J. J. Hox, B. F. van der Meulen, J. M. A. M. Linguistic ability in early life and cognitive function
Janssens, J. J. F. ter Laak, & L. W. C. Tavecchio (Eds.), and Alzheimers disease in late life: Findings from the
Advances in family research. Amsterdam: Thesis Nun Study. Journal of the American Medical
Publishers. Association, 275, 528532.
Schaie, K. W. (2005). Developmental inuences on adult Staudinger, U. M., Marsiske, M., & Baltes, P. B. (1995).
intelligence: The Seattle longitudinal study. Oxford: Resilience and reserve capacity in later adulthood:
Oxford University Press. Potentials and limits of development across the life
488 D.F. Alwin et al.

span. In D. Cicchetti & C. Cohen (Eds.), Developmental demic trajectories through fth grade. Child
psychopathology (Risk, Disorder, and Adaptation, Vol. Development, 79(4), 957978.
2, pp. 801847). New York: Wiley. Wechsler, D. (1952). The range of human capacities.
Stern, Y. (Ed.). (2007). Cognitive reserve: Theory and Baltimore: The Williams & Wilkins Company.
applications. New York: Taylor & Francis Group. Wechsler, D. (1958). The measurement and appraisal of
Stern, P. C., & Carstensen, L. L. (2000). The aging mind adult intelligence. Baltimore: The Williams & Wilkins
Opportunities in cognitive research. Washington, DC: Company.
National Academy of Sciences. Whalley, L. J., & Deary, I. J. (2001). Longitudinal cohort
Steyer, R., Eid, M., & Schwenkmezger, P. (1997). Methods study of childhood IQ and survival up to age 76.
of Psychological Research Online, 2(1), 2133. British Medical Journal, 322, 15.
Todd, P. E., & Wolpin, K. I. (2003). On the specica- Willett, J. B., & Sayer, A. G. (1994). Using covariance
tion and estimation of the production function for structure analysis to detect correlates and predictors of
cognitive achievement. Economic Journal, 113, individual change over time. Psychological Bulletin,
F3F33. 116, 363381.
Todd, P. E., & Wolpin, K. I. (2007). The production of cog- Wilson, R. S., de Leon Mendes, C. F., Bennett, C.,
nitive achievement in children: Home, school, and racial Beenias, J. L., & Evans, D. A. (2004). Depressive
test score gaps. Journal of Human Resources, 1, 91136. symptoms and cognitive decline in a community pop-
Turkheimer, E., Haley, A., Waldron, M., DOnofrio, B., & ulation of older persons. Journal of Neurology and
Gottesman, I. I. (2003). Socioeconomic status modi- Neurosurgical Psychiatry, 75, 126129.
es heritability of IQ in young children. Psychological Woodcock, R. W. (1994). Measures of uid and crystal-
Science, 14(6), 623628. lized intelligence. In R. J. Sternberg (Ed.), The ency-
U.S. Department of Health and Human Services, clopedia of intelligence (pp. 452456). New York:
Administration for Children and Families. (2010). Macmillan.
Head start impact study. Final report. Washington, DC. Wray, L. A., Alwin, D. F., Zeiser, K. L., & Hofer, S. M.
Votruba-Drzal, E., Li-Grining, C. P., & Maldonado- (2010). Does memory performance predict retirement
Carreno, C. (2008). A developmental perspective on in older workers. Unpublished Paper. Center for Life
full- versus part-day kindergarten and childrens aca- Course and Longitudinal Studies, Pennsylvania State
University.
Part IV
Life Course Research Methodologies
Longitudinal Qualitative Research

Joseph C. Hermanowicz

Although qualitative methods have always perspectives on the varieties of situations that
availed themselves to longitudinal research, it is characterize their movement in time.
only relatively recently that they have begun to be In studying the same people at more than one
employed in an emerging spectrum of studies. point in time, longitudinal qualitative research is
The wedding of qualitative methods to longitudi- useful in examining developmental change. It is
nal investigation combines hallmark concerns also a key to understanding how people experi-
that can yield novel ways to examine the social ence and respond to developmental change.
world. On the one hand, the use of qualitative Change may be conceived variously at individ-
methods marshals an emphasis on meaning: it ual, group, institutional, or societal levels (see
focuses attention on the perspectives and inter- Cohler and Hostetler 2003). For example, at an
pretations that people develop about experiences individual level, the researcher might show how
and events. The methods open a window through the elderly understand changes in their advanced
which others are able to see how people under- aging that shape interpretations of quality of life.
stand themselves and social situations. On the At a group level, such research might study how
other, a concern with phenomena longitudinally a category among the aged progresses and copes
foregrounds the idea of trajectory: it focuses with a specic disease. At an institutional level, it
attention on the course of experiences and events. may be concerned with how the aged fare with
That is, people and social situations exist, not care and treatment regimens as found in the cul-
statically, but on a path, such that their existence ture and structure of specic assisted-living orga-
comes to be understood as a function of temporal nizations. At a societal level, such research could
passage. Combined, longitudinal qualitative be directed at how passage in advanced years is
research endeavors to understand how people differently understood between national contexts.
successively make meaning about the trajectories The object of study is not, of course, limited to
of their lives, or specic conditions of their lives, old age; longitudinal qualitative research is
by following them through time. This type of employable at all points in the life course. Further,
research offers a way by which to study peoples the methods can address phenomena both that
extend across life phases (e.g., perspectives on
employment experiences across adulthood) and
J.C. Hermanowicz (*)
also phenomena that are delimited within a phase
Department of Sociology, The University of Georgia,
Athens, GA 30602-1611, USA (e.g., the subjective criminal careers of juve-
e-mail: jch1@uga.edu niles). As in the example above, the analytic

Springer International Publishing Switzerland 2016 491


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_22
492 J.C. Hermanowicz

focus shifts from the micro (individual analysis) types of qualitative methods that may be used for
to the meso (group, institutional analyses) and longitudinal research.
then to the macro (societal analysis), but in each
instance attention is centered on change and how
people developmentally understand themselves 1 Theoretic Impetus
in response to shifting social (and biological) of Longitudinal Qualitative
conditions. Research
Since longitudinal qualitative research consti-
tutes a relatively new set of methodological tools The theoretic impetus for longitudinal qualitative
with which to study social life, there is a need to research is derived from life course sociology
outline parameters within which it works. The more generally. In life course sociology, interest
idea in doing so is to sustain the growing interest centers both on patterns and on variation in pro-
in longitudinal qualitative research and to encour- cesses of development and aging. Whereas cross-
age its greater use where conceptual dividends sectional designs in sociological research can
can be large. The maturation and integration of examine developmental change through both ret-
longitudinal qualitative research as a set of tech- rospective and prospective data on life histories,
niques constitutes a methodological advance for prospective longitudinal designs can capitalize on
contemporary social science. In this chapter, I the examination of change by studying phenom-
codify several of the principles and procedures ena in real time. In doing so, longitudinal design
underlying longitudinal qualitative research. I arguably maximizes the likelihood of capturing
explicitly draw and expand upon my prior writ- variation in aging processes. Subjects are followed
ing about longitudinal qualitative research meth- and the multiple, variable courses of their progres-
ods (Hermanowicz 2013) in order to highlight the sion may be observed, together with the nested
essential conditions under which longitudinal conditions that are associated with and/or cause
qualitative data are gathered, analyzed, and pre- the observed variation. As a result, the social
sented. In the course of discussion, I provide underpinnings of human development are poised
illustrations from studies noted for their longitu- to be more fully known and articulated.
dinal qualitative design. I rely centrally on my While these theoretic premises are consistent
research and experience studying scientic with a sociological perspective on lives, they
careers (Hermanowicz 1998, 2009). have, to be certain, not always carried the day,
The chapter proceeds through three parts. nor are they yet as fully utilized as their promise
First, I explain the theoretic impetus and impor- holds. As Dannefer (2013, 794) has commented,
tance of conducting qualitative research longitu- The traditional inclination to regard age-related
dinally. Second, I briey summarize the studies phenomena as largely individual matters gov-
of science careers I conducted, which, combined erned by the imperatives of biological aging on
with examples of work on other topics, form a the one hand and agency or choice-making on
basis of understanding subsequent illustrations. the other has survived largely unscathedmeth-
Finally, in the core of the chapter, I account for odological and analytical progress associated
the conditions essential to conducting longitudi- with longitudinal data and cohort analysis [has
nal qualitative research successfully. In doing so, not been] matched by theoretical advances (see
I address central concerns that form three sets of also Dannefer, this volume). If only to reinforce
issues: design, execution, and analysis. As will the theoretic rationale for what longitudinal qual-
become apparent, many of the parameters within itative design can do for inquiry into the sociol-
which longitudinal qualitative research succeeds ogy of the life course, we may briey re-visit the
are as applicable to one method of collecting reasons that account for its arrival as an analytical
data, such as observational and case study and methodological set of tools.
research, as to another, such as interviews. The Sociological perspectives on the life course
discussion will underscore the similarities among have often taken psychological approaches to
Longitudinal Qualitative Research 493

aging as their counterpoint, largely because the them (Dannefer 1984a, 106). They are, there-
latter so dominated thinking about human devel- fore, of keen interest to the social scientist, for
opment until major sociological contributions to they are the manifestations of other patterns and
the study of lives were made beginning the 1960s pathways by which people age variously.
and 1970s. Psychologically oriented approaches Psychologically oriented views of aging were
to aging conceived of development predomi- challenged by an onset of a sociology of age and
nantly in terms of preprogrammed maturation. the life course, which sought to demonstrate how
The temporally-anchored fact of physical birth age-related patterns were situated in cohort-
is widely assumed to be followed by a set of specic and context-specic experiences. For
xed, organismically driven imperatives of matu- sociologists, aging has a locatedness in contexts,
ration and aging; as a corollary, social life and whether temporally or spatially bound by cohorts
social structure must accommodate such pro- on the one hand and/or environments on the
cesses of individual change over the life course. other. By this view, contexts serve not merely as
Social scientists thinking about aging has devel- a setting for development, as earlier approaches
oped in the context of such assumptions about the held, but as a constituent force of it. Why should
naturalness of aging, pervasive in both popular a universal pattern have been a theoretically
and scientic thought (Dannefer 2013, 793). expected or desired claim to make in the rst
Psychologically-rooted approaches to aging place? What mode of inquiry and what kinds of
have partly achieved their notoriety by promul- assumptions would lead one to assume such
gating stage theories of aging in their various invariance? (Dannefer 1984a, 102103; see also
forms. The popularity of stage theories, espe- Baltes and Nesselroade 1984; Dannefer 1984b).
cially outside of the scientic world, is the likely In this important theoretic sense, different con-
result of rendering complex aging processes into texts entail systematically distinct consequences
simpler and predicable paths on which people are for socialization and development throughout the
able to locate themselves amidst similarity, but life span (see Hermanowicz 2009, 810).
also notable difference from others. Differences Cohort analysis has typically been used by
notwithstanding, people can know that they are sociologists and others to understand how people
not alone, that their situations, challenges, goals, pass through time in socially patterned yet vari-
and crises may be rendered generic. One can thus able ways. In doing so, cohort analysis has sought
recall Eriksons eight stages of man, Levinsons to avoid an ontogenetic fallacypostulating
seasons of a mans and of a womans life, and universality of developmentbut instead to
Sheehys passages, among other stage-like investigate how groups of individuals age differ-
approaches (Erikson 1950; Levinson 1978, 1996; ently. Elders work on the life course, for instance,
Sheehy 1976). Still earlier formulations conned locates individuals in historical times and socio-
development to childhood such that adulthood economic contexts in order to see how develop-
consisted only as a playing-out of developmental ment has transpired differently for cohorts
imprinting that had taken place by the end of proceeding through time under different environ-
childhood. In their various formulations, stage mental conditions. Studying cohorts coming of
theories were asserted to be universal, inherent to age during and just after the Great Depression,
human aging. But differences do stand and these his work has illustrated the differential force of
approaches minimized them. Aging differences time and socioeconomic context on development
gone unexamined represent a serious conceptual in childhood and in ensuing adulthood (Elder
and empirical gap, which creates a signicant 1974, 1981, 1998).
theoretic shortfall in accounting for developmen- In different application, Neugarten used
tal patterns in human aging. Deviations from the cohorts to conduct pioneering studies of the dif-
norm are more than inconvenient life events; they ferential meanings of age. Neugartens work, and
represent patterned variations that can be those in its vein, emphasize normative underpin-
explained by their relation to forces that produce nings of age as a more general component of
494 J.C. Hermanowicz

culture: people operate with shared, if uid, under- intra-cohort variation in the development of a
standings about age and the time and sequence of cohort of men who, with shared beginnings, expe-
age-related events across life domains such as rienced the onset of delinquency as juveniles, but
marriage, family, work, education, and leisure then led divergent lives by proceeding along adult
(Neugarten 1968, 1979, 1996; Neugarten and pathways made different by their variable expo-
Datan 1973; Neugarten et al. 1965; Settersten sure to social controls. The concepts of inter- and
2003; Settersten and Hagestad 1996a, b). The intra-cohort variation have sought to socially situ-
result is an age-graded life course as a durable, but ate the study of human development more rmly.
again elastic and exible, feature of culture. It is against this conceptual backdrop that many
Sociologically, the elasticity of an age-graded sociological studies of the life course have been
life course is key. An age-graded life course is undertaken, including the ones to which the dis-
not tantamount to monolithic stages as character- cussion now turns, which highlight a qualitative
istic of the psychologically oriented approaches methodological means to study the varied condi-
referenced above. Instead, an age-graded life tions and meanings of lives over time.
course refers to the idea of a general conception
and socially desired unfolding of lives through
loosely dened periods of life. Some periods may 2 An Example from the Studies
characterize some individuals and not others. of Careers in Science
Individuals who enter and leave certain periods
may do so at different rates. Some periods may My studies of careers in academic science illus-
be skipped altogether by some subset of people. trate the use, rationale, and utility of longitudinal
The idea of a normative, age-graded life course is qualitative research, and they form a basis of sub-
meant to more fully allow for the possibility of sequent discussion about the conditions under
variation while still attune to the structured, pat- which such research is conducted. By way of
terned ways that people age and interpret their background, the studies began in 19941995
passage through time. when I interviewed a sample of 60 scientists
As a means to emphasize difference in devel- about their professional aspirations and percep-
opmental patterns, life course sociology has uti- tions of their unfolding careers. At that time, the
lized the idea of cohort (Ryder 1965), and has subjects were sampled according to two analytic
come to emphasize the concepts of inter-cohort dimensions: time, indicated by the year in which
and intra-cohort variation (Dannefer and Kelley- the scientists earned their Ph.D.s, and place,
Moore 2009). Inter-cohort differences have indicated by the type of university in which they
emphasized the role of context in shaping aging now worked. The rationale for utilizing these
processes that are historically variable. Elders dimensions was to situate careers temporally and
(1974) study of how the Great Depression differ- contextually in order to see how the experience of
entially affected cohorts coming of age during work and perceptions of careers vary.
and after the event illustrates the signicance of On the time dimension, scientists were sam-
inter-cohort variation in development. Intra- pled and grouped by three cohorts: scientists who
cohort differences have also emphasized the role received their Ph.D.s after 1980 which, at the
of context, but also inequality and the ways in time of the interviews, placed the subjects in
which processes of cumulative advantage and dis- early career phases; scientists who received their
advantage characterize members of the same Ph.D.s between 1970 and 1980, which drew
cohort as they age. Laub and Sampsons (2003) upon scientists in middle phases of their careers;
study of criminal careers demonstrates how per- and nally scientists who received their Ph.D.s
sistence in and desistance from crime operates in prior to 1970, which captured scientists in late
conjunction with the socially controlling effects career phases. Hence, temporally, the design
of marriage, family, employment, and military operationalized a study of scientists at early, mid-
service. It thereby illustrates the signicance of dle, and late career.
Longitudinal Qualitative Research 495

On the place dimension, scientists were sam- sectional data collection, is a standard of life his-
pled at a range of university types. The rationale tory research (Scott and Alwin 1998).
was to maximize the types of academic environ- I interviewed 55 of the same scientists again in
ments in which scientists work, and to thereby be 20042005, creating a longitudinal design from
inclusive of the types of careers found in aca- which to study how academics, working in a vari-
demic science. The universities, which form a ety of institutions, age in relation to their work. Up
representative continuum, consist of those stress- to this point, no prior study of the academic pro-
ing research in the presence of teaching and other fession had followed the same people over time. A
roles, which were termed elite. Examples include unique opportunity was thus created to study the
Harvard University, Cal Tech, and the University experience, meaning, and interpretation of work in
of California-Berkeley. Institutions that stress one of the main professions in modern society
research and teaching as well as other roles were (Ben-David 1972; Gustin 1973). The 10-year time
termed pluralist. Examples include the University interval is conceptually signicant for the specic
of Kansas, the University of Missouri, and Purdue occupation studied: it advances the subjects into a
University. Institutions that stress teaching in the subsequent career phase. Consequently, the sequel
presence of research and other roles were termed allowed one to see how academics perceptions of
communitarian. Examples include the University work evolve with felt costs and rewards, from
of Tulsa, the University of Louisville, and the early to mid career, from mid to late career, and
University of Toledo. from late to post career, including the stage of
The design of the study is captured in Table 1. retirement and exit from the career.
The number of respondents are arrayed by cohort In the follow-up work, age and institutional
and by the type of institution in which they location provide the structure to analyze individ-
worked. The full results of this study and broader ual, subjective careers through diachronic
discussion appear in The Stars Are Not Enough: change. Diachronic refers to change between
ScientistsTheir Passions and Professions successive points in time. Longitudinal data add
(Hermanowicz 1998). spatial and temporal dimensions to synchronic
In this baseline work, the subjects discussed study. Synchronic refers to characteristics and
their evolving careers, including their past and conditions existing at one point in time. In mov-
hoped-for future. In these respects the work con- ing from a synchronic to a diachronic perspec-
stituted a cross-sectional design in which tempo- tive, we are consequently in a position to answer
rality arose by two means: by the three cohorts the following research questions that were central
employed in the design, which enabled a study of to the longitudinal study:
meanings assigned to experiences by career
phase, and by interview questions that sought How do academics account for the unfolding
data on retrospective and prospective percep- of their careers in light of the goals and aspira-
tions. This type of design, where retrospection tions that socially situate their profession?
and prospection form key components of cross- What continuities and changesin aspiration,
satisfaction, motivation, commitment, and
identication with workmark the careers of
Table 1 Research design and number of scientists in academics?
baseline study, by type of institution and cohort What knowledge have academics acquired
Cohort (by year of Ph.D.) about themselves, their institutions, and the
Pre- 1970 Post- academic profession in 10 years?
Institution 1970 1980 1980 Total How does this knowledge vary by individual
Elite 9 6 8 23 age and type of university?
Pluralist 6 5 7 18
Communitarian 7 5 7 19 Table 2 illustrates the evolved design guiding
Total 22 16 22 60 the research in the longitudinal study. Time and
496 J.C. Hermanowicz

Table 2 Research design of longitudinal study, by type of institution and cohort


Cohort/career phase Elite Pluralist Communitarian Total
Early to mid 8 6 7 21
Mid to late 6 4 5 15
Late to post 9 5 5 19
Total 23 15 17 55

place remain key analytic dimensions, but dia- the same opportunities and constraints in their
chronic analysis is added to what was previously occupational (and personal) lives. As analytic
synchronic study. Thus in the follow-up work, resources, cohort and contexttime and place
time is captured as it was in the baseline work by are utilized in these studies, as they may else-
differences in cohort meanings and by retrospec- where, to socially situate the study of
tion and prospection, but also by change from developmental change.
one point in time to another. The full results and
broader discussion of the longitudinal study
appear in Lives in Science: How Institutions 3 Conditions of Longitudinal
Affect Academic Careers (Hermanowicz 2009). Qualitative Research
Professors from one academic eldphys-
icscomposed the studies, but the results are not We may proceed by posing the question, How
limited to them. I discuss issues of generalizabil- does one do longitudinal qualitative research?
ity elsewhere (Hermanowicz 2009, 252268). There are three clusters of issues that, in tandem,
Physicists were selected because in the wider cul- uniquely situate longitudinal qualitative research,
ture they are perceived to embody the scientic whether involving interviews, observation, case
discipline par excellence. They possess a recog- study methods, or other approaches, and which
nizable genealogy of immortals, such as Kepler, consequently one must take into account when
Newton, and Einstein, who promote a heroism undertaking this type of research in its variety of
and dene a paradigmatic life course. Thus if one forms. These issues encompass matters that arise
is interested in seeing how academic aspirations and range in time from the inception of a study to
develop and evolve and how careers play out over the dissemination of ndings. They include:
time, particularly against the backdrop of a eld issues of design, execution, and analysis.
that imposes a paradigmatic template on the pas-
sage of time, the eld of physics made for an
ideal setting. 3.1 Issues of Design
By virtue of the research designs in both the
baseline and longitudinal work, emphasis is Three main issues pertain to the design of longi-
placed on inter- and intra-cohort variation and on tudinal qualitative research: origination, which
context, ideas central to life course analysis as refers to the point at which longitudinal inquiry is
discussed in the prior section of the chapter. The conceived; number and frequency, which
incorporation of cohorts allow for a view and involves the points of contact with research sub-
treatment of aging in a career as variable, not jects or the setting under study; and protocol for-
monolithic or universal to stages. The incorpora- mat, which is key to the use of interviews in
tion of contexts, indicated by the different institu- longitudinal qualitative research. Each are dis-
tions in which the subjects have worked, opens a cussed in turn.
way in which to examine how they operate as
constituent forces of, as opposed to merely set- 3.1.1 Origination
tings for, development. Clearly not all careers in Longitudinal qualitative research may proceed
a eld are of one kind, nor are people exposed to from two different starting points. It may be
Longitudinal Qualitative Research 497

factored into the design at the outset of a study, and exibility in some data gathering by virtue of
where researchers anticipate and plan in advance established relationships that, by turn, are not
the use of serial contact. Alternatively, longitudi- brought about in cross-sectional inquiry.
nal qualitative research may be formulated into a In sharp contrast, the baseline study of scien-
design after a study has been completed, when tists was conceived and completed in the absence
the research is conducive to longitudinal of entertaining a possibility of longitudinal work.
formulation and where the benets of such work The subsequent discovery that such an undertak-
are likely large. ing would be the rst longitudinal study of the
Armstrong and Hamiltons (2013), Paying for academic profession suggested that this might be
the Party: How College Maintains Inequality, a protable new way to examine careers. I began
exemplies the former approach. That is, the contacting respondents of the original sample in
researchers sought at the outset of their work to Spring, 2004. I did so by sending them a letter,
follow their subjects, who were women college which for illustration is presented in Fig. 1 (see
students at a major public university in the United also Hermanowicz 2009, appendix B). Indicative
States. The goal of the work was to understand of the task of contacting respondents after an
contemporary college culture and its variable elapsed period of time and in which there was no
inuence on the practices, goals, and aspirations indication at the rst point of contact that they
of female undergraduates. The researchers stud- would be solicited again, the letter sought to
ied their sample through a combination of obser- accomplish several objectives. It attempted to
vational and interview work over the course of 5 place the longitudinal study in context by remind-
years. In the rst year, alternating members of a ing the respondents of their previous participa-
research team took turns occupying a room in tion in the foundational work. It explained what
one of the universitys dormitories, which enabled the longitudinal study would seek to accomplish.
observational work through varied hours of the It informed them about what their continued par-
day. Following the year of observational study, ticipation would involve (i.e., the basic subject
interviews were conducted serially, across the matter of the interview, the estimated interview
remaining college years and then shortly after length, where interviews would be held, and the
college, with a large subset of the 53 women who like). Finally, the letter made clear some proce-
had lived on the residence hall oor. Armstrong dures (e.g., the use of a recorder to retain data)
and Hamilton note that the initial year of obser- and rights of and assurances to the respondents
vational work enabled rapport, which facilitated (e.g., voluntariness, anonymity, the use of a
students subsequent cooperation in interviews human subjects review).
that yielded high rates of response. After approximately 10 days from sending the
Many of the nal interviews were conducted letter, I contacted the respondents by telephone.
around the country as students fanned out from At this time, I re-introduced myself and the study,
the university to nd and take jobs. A portion of asked if they had questions, and attempted to
these interviews were conducted face-to-face, schedule interviews. Any success of the longitu-
others via telephone owing to travel and budget dinal study hinged, of course, on respondents
constraints. Telephone contact is often not the electing to participate again in the follow-up
method of choice in semi-structured interviews work. In this simple regard, as is true in much
because it can compromise rapport and data qual- longitudinal research in which the same subjects
ity (Hermanowicz 2002). Armstrong and are studied, the stakes were high for the ways in
Hamilton noted that telephone interviews, in this which communication was handled with the
instance, went relatively unhampered given the respondents. The procedures described above
rapport that had already been established over the generated a response rate of 93 %. (This rate of
prior years of contact with the subjects. In ways response also generated a sense of relief for the
that may seem ironic, then, longitudinal qualita- researcher, in that it meant the work could go for-
tive research might introduce a degree of latitude ward, and a project could proceed into fruition).
498 J.C. Hermanowicz

Sample Re-Contact Letter, Study of Scientists Careers

April 7, 2004

Dr. ----------------------
Department of Physics
--------------------------
--------------------------
--------------------------

Dear Professor ------------:

It has been 10 years since we met. In 1994, I interviewed you for a study of careers in
science. Funded by the National Science Foundation, and conducted under the auspices of the
University of Chicago, that study explored scientists aspirations and identities related to their
work. The study was based on interviews with scientists across the United States. As the
principal investigator of that study, I know very well that you formed a critical part of the
sample, and I remember very well how much your participation contributed to the work.

I write to ask for your help. A 10-year follow-up study is being conducted entitled Lives
of Learning: Continuity and Change in Science Careers. The study design calls for
interviewing the same participants who composed the original work. This is both substantively
and historically significant: the study will be the first of its kind to follow professors in their
careers. It therefore holds real potential to generate important findings about how careers are
experienced and understood by scientists themselves. The study presents the unique opportunity
for you as a scientist to convey knowledge about careers and science acquired over the years of
your extensive experience in physics.

Your participation would involve an interview, conducted again by myself, that would
last approximately an hour. As before, interviews would customarily take place in your office,
and I would meet you at an agreed upon time. (If you happen to be one of the several scientists
in the sample who has retired, we would make alternative meeting arrangements as necessary
and as agreeable to you.) The interview would consist generally of questions about changes and
continuities in your career over the past 10 years. Like before, the interview would normally be
tape-recorded simply to keep accurate track of information, and subsequently the tape would be
destroyed once the study is completed. Participation and all interview material will be strictly
confidential. Both personal and institutional identities will be concealed in published work,
following standard conventions of work of this kind. Participation is voluntary. Nevertheless, I
very much hope you can participate; the success of the work depends on you. All aspects of this
project have passed the usual human subjects reviews at the University of Georgia.

I will call you shortly to invite your participation and answer any questions you might
have. Please know how greatly I appreciate your time and help with this request.

Yours sincerely,

Joseph C. Hermanowicz
Assistant Professor

Fig. 1 Sample re-contact letter, study of scientists careers


Longitudinal Qualitative Research 499

An implication posed by origination involves submit things to published journals anymoreIm


disgusted by the whole thingI got tired of getting
opportunity. Many opportunities have been
referee reports [on manuscripts submitted to jour-
missed by failing to see how once cross-sectional nals for peer review] that spend a page talking
studies may be adapted for longitudinal inquiry. about the bibliography; they were entirely con-
Might we have imagined The Social Order of the cerned with whether I cited their work or their
friends work, and they hadnt read the paper. I got
Slum (Suttles 1968), Habits of the Heart (Bellah
to the point where at [national] meetings I was tell-
et al. 1985), or Coming of Age in New Jersey ing people, Please dont reference my paper, if
(Moffatt 1989) 10, 15, or 20 years later? It is not you dont read it, dont reference it. Its a game to
an exaggeration to assert that many, if not most, so many people, and there are many fools. I didnt
do this [go into an academic career] to deal with
cross-sectional qualitative studies have been and
fools. They dont understand basic thingsI went
are amenable to longitudinal work, and that the from not having tenure to slowly being delighted
introduction of diachronic analysis to these stud- with tenure because I can do the right thing
ies would altogether change, in a powerful and There are more important things in life than getting
grants from the National Science Foundation, get-
robust way, the face of social science research.
ting Nobel Prizes even or any of that stuff. Thats
Where opportunity was seized and longitudi- all just a game. Im interested in solving prob-
nal work accomplished, a different implication lemsI am at a crossroad.
arises: what may call emergence. In qualitative Interviewer: Do you see yourself getting back to
research?
work, meaning and signicance of data emerges
Respondent: If you mean publishing papers and
over time as researchers grapple with formulating going to conferences and advising graduate stu-
interpretation. This is true in both cross-sectional dents, no I dontWhat do I care for refereed
and longitudinal work. In longitudinal research, publications?Im not angry about it anymore, I
just dont care about it
however, more time is introduced for ideas to
Interviewer: How would you complete the sen-
emerge. Specic research emphases, questions, tence, I am more X and less Y compared to a
and themes may change over time. In my base- decade ago?
line study of scientists, the research emphasized Respondent: I would like to say I am wiser and I
am less nave. But it could be just the opposite as
ambition and its role in constructing careers,
far as I know. I really am in a very transitional
whereas the longitudinal study emphasized aging stage. Im questioning whether I want to be in
and adaptation to work. In the baseline study, physics. Ive gone a little bit even beyond that. Im
acceleration and deceleration were the major thinking I probably will not stay in professional
physics. I want to do something very different.
themes in interpreting interview data, whereas in
(Hermanowicz 2009, 105106)
the longitudinal study they were continuity and
change. Thus the baseline study could not have Because people change over time, newly
anticipated, or grappled theoretically, with the posed research questions, ideas and themes
changed outlooks that prove to mark many emerge to grapple analytically with the qualita-
careers years later. A conceptual re-tooling was tive complexity in the data. Themes of accelera-
necessary in order to understand lives diachronic- tion and deceleration as initially used are not
ally. Such change is illustrated in the account pro- suitable for an accounting of data 10 years later
vided by a respondent who had passed from early that speak, for example, of disengagement, frus-
to mid career phases: tration, and even exit. This reality imposes
I would saymy research career, ten years ago, signicant limitations on the extent to which lon-
was at a peak. I was working with two or three gitudinal qualitative research can be designed at
graduate students continuously and two or three the outset as fully operational. By circumstance
post-docs continuouslyMy attitudes about the alone, designs will change in such work, and
job, about me, and about the university have under-
gone tremendous changes in the last ten years. Ive even the best-planned project will not, at the out-
gone from having a fairly large amount of [grant] set, be able to anticipate and accommodate what
money, especially for the stage of my career, to arises subsequently as newly emphasized areas of
having my name on a grant, but not taking any interest.
money out of it at all. Im not sure I want to even
500 J.C. Hermanowicz

The intensity of emergence as a factor in the sufcient amount of time needed to pass in order
research process likely increases as the time to track change and continuity in the respondents
interval increases between episodes of data col- professional and personal lives.
lection. While emergence is always a condition Theoretically, 10 years of time accomplished
of qualitative work, and while it was by their own a major outcome: it placed all of the original
accounting a feature in Armstrong and Hamiltons respondents at different phases of their careers.
college student study, one can surmise it would Because the respondents were originally sampled
gure even more prominently had their inter- at early, middle, and late career phases, the
views been separated by greater lengths of time, 10-year interval advanced all of them into the
giving the objects of interest time to change all next set of three parallel phases, enabling longi-
that more substantially. Even when their subjects tudinal work to capture how people make transi-
were studied over consecutive years, the research- tions throughout a career. In using this specic
ers state that Although we followed a general time interval, data combined from the two studies
interview schedule each yearthe ow of a typi- enabled a consideration of careers that spanned
cal interview was highly informal, indicating from their beginning, for members of the young-
that change over time required a exibility in the est cohort, to their end, for members of the eldest
researchers to adapt their focus (Armstrong and cohort.
Hamilton 2013, 271). In these respects, one can Other research problems, however, present
see why emergence gures prominently in those different time considerations. It studying how
studies, such as the ones of scientists, where the college culture conditions behavior, attitudes,
time between contact is that much greater. and aspirations, it was incumbent upon Armstrong
and Hamilton (2013) to identify and begin study-
3.1.2 Number and Frequency ing their respondents upon college entry and then
Number refers to the total amount of research re-visit them across and shortly after the college
episodes, frequency to the periodicity at which years, thus creating four waves of contact beyond
they occur. The number and frequency of serial the initial year of research. Electing to study the
episodes that compose a longitudinal study will students at the beginning and then only at the end
depend on how a given research problem is of college (i.e., two total points of contact) would
posed, and will thus vary from study to study. in all likelihood have obscured numerous qualita-
Another way of stating this consideration is via tive details necessary to understand variability in
the question of how much time should pass identity formation. What is more, another option
before a subsequent round of data are collected remains in play: whether to continue following
(Saldana 2003). The answer is that it should be an the subjects, and if so, with what frequency, as
amount of time sufcient to examine relevant they navigate the worlds of employment and
change from one point to another. eventually marriage and family, in order to draw
In the follow-up study of scientists, a 10 year longer arcs of inuence between college and
interval was used for just a second point of con- adult life. Recalling the discussion above, the
tact. The 10-year interval had both practical and option presents an exciting theoretic opportunity
theoretic importance. Practically, the 10-year in sociological research.
mark represented a point at which the greatest Still other studies illustrate the permutations
number of respondents from the original sample in the number and frequency of research episodes
would have been available for longitudinal study. in longitudinal work. In medical studies with
A longer time interval would have posed risk of patients, the time interval may be relatively short,
involuntary attrition; one of the respondents had and the number of iterations of contact relatively
already passed away, others would of course fol- large. In their studies of people with chronic ill-
low in time. A shorter time interval would not nesses, Murray et al. (2009) used 3 month inter-
have accomplished the works theoretic objec- vals to study patients with lung cancer, but
tiveto study aging and adaptation in work. A 6 month intervals to study patients with
Longitudinal Qualitative Research 501

obstructive pulmonary disease, which develops identical protocols; instead the questions at sub-
less quickly. Corden and Nice (2007) tracked sequent contact are likely different.
individuals who participated in employment pro- The issue presents two means by which to
grams on their journey to employment. The structure interview protocols that undergird lon-
determination of time intervals depended on spe- gitudinal qualitative work. One consists of pos-
cic events occurring in respondents lives, which ing the same questions on the same themes. The
did not occur at the same rate. The timing of second consists of posing different questions on
serial contact was contingent upon an event selected same and newly emergent themes. The
sequence in the research subjects varied lives in rst means is perhaps more recognizable because
order to establish appropriate parameters in it parallels quantitative longitudinal design
which to assess change. (Menard 2002; Ruspini 2002; Scott and Alwin
In studies such as Shaws (1930) portrayal of 1998). The second means is not found readily in
Stanley, the delinquent boy whose life story com- quantitative longitudinal design; it is a strength of
prises the illustrious sociological work The Jack- qualitative study. It arises in large part because
Roller, subjects may need to be interviewed lives change and qualitative methods are attune to
multiple times in a given iteration. Through inter- emergence, discussed under origination,
viewing and diary techniques, Shaw obtained wherein the researcher strives to best characterize
data on nearly all facets of Stanleys life: his per- thematic patterns in development.
ceptions about criminal involvement, street life, In the study of scientists, a preponderance of
schooling, work, friends, and courtship. But to do new questions were used in the longitudinal
so, Shaw had to meet up and work with Stanley work. (To compare the longitudinal and baseline
many times for each iteration of study across the interview protocols, see Hermanowicz 2009,
6 years that Shaw followed him. Wide-ranging appendix A and D.) Many of the questions asked
data on peoples lives obtained at numerous of scientists in the baseline study were time-
points will often compel a multiplicity of research bound. Consider the following: How did you
visits. Time intervals vary from one type of group come to arrive at this university? or What aspi-
to another, but nevertheless must be considered to rations did you have as a graduate student? or
design longitudinal inquiry with a theoretic logic. You were a graduate student at ______. Is this
the type of university where you wanted to end
3.1.3 Protocol Format up? These types of questions had low utility
When interviews are a chief form of data collec- value in being asked again in longitudinal work.
tion in longitudinal qualitative research, origina- Instead in the longitudinal work, one sees ques-
tion, discussed previously, affects the design of tions such as: In light of the past 10 years, what
protocols. Longitudinal qualitative research are your current aspirations? or Looking back
designed in advance lends itself to protocols con- over the past 10 years, has your career progressed
taining identical questions posed to respondents as you expected? or If you were starting all
at different times in order to assess change. But over again, what would you do differently, know-
longitudinal qualitative research, unlike much ing what you know now about your line of work?
longitudinal quantitative research, is not restricted How does one think about comparison of data
to the use of identical questions. We again note when different interview questions are posed
Armstrong and Hamilton (2013) whose longitu- between research episodes? The conditions of
dinal study of college students was projected and research subjects are established from base round
well-planned, but whose serial interviews data, which may include an account of the setting
evolved, even across the span of 4 years, such in which the subjects operate, usually through
that questions changed from one interview epi- heavy contextualization. Both qualitative and
sode to another. What is more, longitudinal quali- quantitative evidence may be used to establish
tative research designed after an initial and inquire about these conditions, as often found
investigation is even less likely to avail itself to in ethnographic work (e.g., Suttles 1968).
502 J.C. Hermanowicz

Longitudinal work is then used to examine 3.2.1 Attrition and Retention


change, not against responses to identical ques- Because samples in qualitative studies tend to be
tions, but against the themes that emerged to comparatively small, subject attrition is an espe-
characterize the conditions of people and their cially prominent consideration in longitudinal
social settings. qualitative work. In those studies where longitu-
Shaw (1930) did not ask Stanley identical dinal inquiry is designed in advance, it is likely
questions serially. Instead, Shaw crafted charac- advantageous to factor attrition into the size of
terizations of Stanleys condition, in conjunction the sample as part of the study design. Thus, in
with his milieu, which were then compared and such designs, comparatively small samples will
contrasted with characterizations of Stanley at become larger by necessity. In all longitudinal
subsequent points in time. studies, whether implemented in advance of or
In this format, questions do not in and of them- after initial study, thought ought to be given to the
selves serve as the baseline of comparison, but care of respondents, not only to enhance the qual-
rather characterizations of people and their situa- ity of data collected but also to promote subject
tions. Perceptions of crime, work, and courtship, retention.
for example, are compared across time, not spe- I relied exclusively on the rapport established
cic questions about crime, work, or courtship. with my scientist respondents 10 years prior to
This logic is made still more evident when the follow-up contact. Careful and considered
researchers change the method of data collection framing of written correspondence at the time of
among episodes of research. Armstrong and follow-up (see Fig. 1) further aided subject reten-
Hamilton (2013) began by collecting observa- tion, including explicit explanation that the suc-
tional data, and then proceeded in all of their sub- cess of the work depended crucially on subjects
sequent research rounds to semi-structured continued participation, all while also having to
interviewing. Likewise Lareau (2011) conducted make clear that such participation was voluntary.
her baseline study through mainly observational In addition, thought and care should be given to
work, but in the longitudinal study switched to correspondence immediately following contact,
interviews. It is noteworthy that such exibility such as in the use of thank you letters. Thank you
of method is possible in qualitative research letters should always be sent to participants in a
designs. It also reinforces the understanding that research study when their identities and addresses
qualitative researchers undertaking longitudinal are known to the researcher. These letters should
work typically craft characterizations and themes have personalized salutations (i.e., not Dear
to represent the conditions of their subjects at Research Participant), their wording and con-
given points in time in order to spell out continu- tent should reect genuine (not perfunctory) grat-
ities and changes across time. The result is often itude, conveying in essence that, were it not for
a vivid depiction of how lives unfold. them, there would be no work to be writing about.
Such letters might be handwritten as an extended
note or on a card; they also can be typed, as were
3.2 Issues of Execution mine, owing mostly to the illegibility of my pen-
manship. I strongly support mailing such corre-
Three sets of issues especially situate the execu- spondence, not sending it electronically. Taking
tion of longitudinal qualitative research: attrition time to send thankful messages received by mail
and retention, or the capacity to keep original are more demonstrative of the gratitude they
subjects in subsequent research rounds; respon- intend to convey. (It is ne, for example, to
dent reaction, which bears on rapport, relation- include ones email address and/or telephone
ships, and continuing access to research subjects, number in the thank you message for purposes of
and; ethics, which involve moral considerations continued contact should respondents desire to
in re-contacting and researching the same people be in touch. What is more, I negotiated with my
over time. universitys human subjects ofce that their
Longitudinal Qualitative Research 503

Thank you Letter, Study of Scientists Careers

May 15, 2004

Dr. ----------------------
Department of Physics
--------------------------
--------------------------
--------------------------

Dear Professor ------------:

Having recently completed our interview, I want to take the opportunity to thank you for
all your help. You are most kind and gracious not only in your time, but most especially with
your insight about careers in science, and your capacity to communicate some of the meaningful
aspects of your life in physics. This means more to me, and to the work it forms, than I can tell.

If you have any questions or want to get in touch with me, you should feel free to do so at
any time. My departmental address and telephone number are on this letterhead,and my e-mail
address is: jch1@uga.edu.

As you know, this project has passed customary human subjects reviews. Should you
have any questions regarding your rights as a research participant, you may contact me or --------
-------, Human Subjects Office, University of Georgia, 606A Boyd Graduate Studies Research
Center, Athens, GA 30602-7411; Tel: (706) -------------; E-mail: --------.

Please accept my many thanks for all the help you have given, and my very best wishes
for the months and years to come.

Yours sincerely,

Joseph C. Hermanowicz
Assistant Professor

Fig. 2 Thank you letter, study of scientists careers

contact information be included in the thank you tudinal study of scientists; it appears in Fig. 2
letter, since (1) it had to be communicated to the (see also Hermanowicz 2009, appendix C).
research subjects somehow; (2) this format was It is likely that ancillary characteristics of the
less invasive and disruptive to the organic ow of subjects aided retention in my study. In this case,
exchange that I sought to establish at the time of the subjects lives were oriented to research and
the interviews, and; (3) the particular study con- teaching, and thus broadly to helping and inform-
tained risks that were low so that a rationale could ing others. This occupational characteristic may
be made in providing this information after con- have conditioned the subjects proclivity to assist
tact with subjects. Other studies, involving other someone whose own work depended on their
types of risks, will, of course, warrant different involvement.
procedures in communicating the human subjects Other types of studies using different types of
information that is essential to transmit to participants may not be ordered on such auspi-
research participants.) As an example, I have cious grounds. Researchers may consider send-
included the thank you letter I used in the longi- ing birthday cards and holiday notes to
504 J.C. Hermanowicz

participants, and messages of congratulation on issues are readily apparent. First, the success of
celebratory occasions, such as weddings and replacement is contingent on the number and fre-
births, and messages of condolence in occasions quency of research episodes. Replacement may
of grief or stress. In her study of families and be more feasible when the number and frequency
childrearing, Lareau (2011, 313) notes that she of points of contact is relatively high, though this,
sent holiday cards to the children in the families too, will depend on the specic research prob-
she studied with a ve dollar bill tucked inside. lem. College students under study who come into
These practices likely help to maintain rapport contact with researchers regularly may be more
and cooperation with respondents. These types of readily substituted in the event of attrition than
practices may carry the further consequence of career criminals who come into contact with
preserving, if not deepening, relationships such researchers decades apart. In the former instance,
that participants at times of follow-up feel a key data may be more readily brought current
warmth, security, and openness with a researcher. by new participants, whereas in the later instance
In other instances it may be difcult to locate more time has elapsed for data to be comparable
former participants, particularly when a long across cases. Second, the success of qualitative
period of time has elapsed between episodes of work is contingent on contextualization. Even if
contact. This was the case for Laub and Sampson replacements can be matched approximately with
(2003) who attempted to locate and contact male the characteristics of lost cases (the procedure
offenders last studied by a different team of followed in quantitative panel studies), the indi-
researchers decades earlier. The last addresses for vidual narrative will still differ between the case
the men were 35 years old, few of them had tele- lost and added. Thus, in qualitative work, replace-
phone numbers in their case les, and only about ment would only seem feasible when the subse-
1 in 20 had established a social security number, quent number and frequency of points of contact
which otherwise is a key means of tracking peo- would allow for sufcient intra-comparison of
ple in modern databases (Laub and Sampson the substituted case over time.
2003, 62). They proceeded by undertaking crimi-
nal records searchers through both the state of 3.2.2 Respondent Reaction
Massachusetts (where the offenders were last An issue that can interfere with subject retention
located) and the Federal Bureau of Investigation at its extremes and with their cooperation at its
for national records, a process that itself took 18 minimum involves the reactions that respondents
months, but which yielded information on the have to prior ndings from the research in which
vast majority of their sample. In addition, they they have participated. The researcher nds him
conducted death record searches through state or herself confronted by strong emotions.
and federal registries as well as searches through Respondents may express negative views about
the Boston Globe, the newspaper most likely to the research. They may be offended by particular
have published death notices for these particular interpretations or representations of themselves
men. Still other resources were utilized: tele- or others. Some may object to particular conclu-
phone directories (paper and electronic), motor sions. The feelings arise for various reasons: in a
vehicles searches, voter lists, and, as a last resort, researchers decision not to reveal plans to pub-
even the Cold Case Squad of the Boston Police lish a piece of work in which subjects are por-
Department. These efforts, extensive and time- trayed; in a researchers decision to conceal
consuming, yielded a locate rate of 79 % (Laub information that subjects believe is important
and Sampson 2003, 72). (Lareau 2011); in feeling used by a researcher
In quantitative designs replacement of missing (ten Have 2004); and even in having to leave the
subjects is a customary practice. The technique eld and conclude relationships (Reiss 2005).
may be less feasible in qualitative research; there From her study of mental illness in rural
is little in the way of precedent in accounting for Ireland, Scheper-Hughes (2000) recounts how
the possibility in the research literature. Two the work was promptly made a classic of
Longitudinal Qualitative Research 505

anthropology yet simultaneously criticized in the ticipants with an eventual copy of the publication,
Irish press as an extreme breach of privacy. Her she advocates instead devising an informative
return years later to the village in which she com- brochure that identies key points, themes, and
pleted the original work resulted in her expul- conclusions. A letter summarizing results, per-
sion. Ellis (1995) tells the tale of the remote haps including charts and tables, may also be uti-
shing communities she studied, where upon her lized. In this way, the researcher fullls any
return, residents reacted angrily toward her prior obligation to inform participants of results while
work. From his study of street life, Whyte (1996) also creating an opportunity for feedback in a
reported tense reaction by community members way that protects both the role of the researched
to Street Corner Society. Studying class-based and that of the researcher.
parenting patterns, Lareau (2011) reported highly Even this approach has limitations. Some par-
negative responses from several participants, ticipants, in some studies, will be curious enough
which damaged relationships in some instances to nd their rightful way to the more complete
and ended others altogether. work. The internet and ready access to it via cel-
One might conclude that a way around this lular telephones and mobile devices makes this
potential dilemma is to decide not to share the especially feasible. Studies have yet to document
published work with participants. This is a pos- this occurrence and any consequences for the
sibility, but it does not address fully the issue of research. Following Lareau, letters and brochures
informing respondents what will come of the could include statements that channel more
research. Such disclosurethe explicit statement inquisitive subjects to fuller accounts and to
that the researcher intends to publish work based thereby guard against subjects routing them-
on the research while protecting the anonymity of selves to less responsibly gathered or less
participantsis often treated as an ethical condi- informed documents that litter various media. In
tion of conducting the work. Nor is a solution addition, researchers can always make them-
necessarily found in giving respondents pre- selves available out of the eld via mail or phone
published work and amending it for publication to questions from research subjects, and defuse
according to respondent wishes. This is a version potentially volatile situations (and the possibility
of the practice of respondent validation, of attrition) by personalized communication
wherein participants are given pieces of writing regarding facets of the work that prove to gener-
to afrm or disconrm the validity of written ate continued interest in participants. In fact, pro-
material and/or to establish the veracity of par- viding contact information to research subjects is
ticular points. The less standard practice of enlist- a customary, and often mandatory, procedure.
ing participants to edit, re-write, or change Such exchange can benet all parties: the research
representation presents perhaps more problems subject, who has questions or concerns addressed
than it resolves. Publication goals of researchers if not always resolved, and the researcher, who
and requirements of publishers for content and may use feedback to hone analyses, interpreta-
style will often diverge from the desires of par- tions, or conclusions (Rupp and Taylor 2011).
ticipants. What is more, researchers surrender Upon publication of the baseline study, I sent
their license and mandate as trained professionals participants a letter describing the outcome of the
while bestowing expertise upon others far out- work and where the results were published. In
side the researchers community of professional returning to the eld for the longitudinal study,
peer-judges. one of my interviewees took umbrage at the prior
Scrutinizing her own experience, Lareau publication. That this had occurred 10 years earlier
(2011) concluded that there is not an easy solu- conveys the depth of the respondents sentiment.
tion to this pitfall. By her account, one must The respondent disputed a specic way in which
clearly inform participants at the outset of their I had constructed a set of tables (in which he
participation about the goals of the work, includ- had been able to infer that he was treated as an
ing its publication plans. Rather than furnish par- outlier and excluded from selected computations).
506 J.C. Hermanowicz

I attempted to explain at the interview the deliberately vigilant in how they go about their
distortions that would arise were this procedure recruitment and re-recruitment of subjects. As in
not followed. I explained that, under the specic all research, nancial incentives may be some-
conditions, this was a standard methodological times used to constitute samples, but they must
procedure in my eld and practiced widely by be used only as long as the incentives are neither
others. coercive nor binding. That is, all respondents,
I had allotted 2 h for our interview; despite my regardless of any incentive, must voluntarily
efforts to address the matter and dispense with it, choose to participate in any and in each succes-
the discussion consumed 40 min of our time. The sive round of longitudinal research, and their
respondent, a highly accomplished researcher right to withdraw at any time without prejudice to
and teacher, said angrily that he expected me to them must be communicated clearly by
change the data presentation and if I did not do researchers.
so, he would consider the present interview a To create and maintain ethical standards of
waste of his time, unfair, a discredit to his longitudinal research, it should be incumbent
university, and would not participate in an inter- upon authors to account in their published work
view with me again. I told him that I would con- for their research and eld procedures. This is
fer with my colleagues upon my return home. notably lacking in much of the published work
But my responses and gestures were to no avail. to-date. Researchers should be able to live up
A pall clouded the entire meeting. The interview before a community of their professional peers to
was irreparably marred by the respondents open- an accounting and justication of the procedures
ing hostility; it proceeded perfunctorily, and they followed in handling human subjects and
ended sooner than it should have. collecting data just as they are expected to
In the thank you letter I subsequently sent to account and justify their data analysis, interpreta-
this respondent, I might have tailored it to further tion, research conclusions, and pertinent policy
acknowledge his concerns. But I also knew that I insights or recommendations. By the same token,
would be unable to accept his conditions and that professional peer reviewers, of article and book
the change he requested would not be made in the manuscripts and of grants proposing longitudinal
longitudinal work. I elected not to blow air over research, should insist on accounts from authors
a smoldering re. This type of experience who can state clearly the procedures followed or
occurred only once across 55 respondents. But it to be followed in a piece of work. Such guide-
was surely memorable. I came to the conclusion, lines, for producers and for gate-keepers of longi-
however, that I had done all I could, that the tudinal qualitative research, will help to sustain
respondent had been treated ethically and prop- an ethic for this type of work. The many consid-
erly, and that I was simply paying a price for con- erations discussed above, in attempting to pre-
ducting this type of work. vent subject attrition, in managing respondent
reactions, and in working with research subjects
3.2.3 Ethics ethically, inform the special dynamics that come
Because by its denition longitudinal qualitative into play in entering, leaving, and re-entering the
research depends on respondents continued par- eld in longitudinal qualitative research (cf. Ellis
ticipation, the research tools utilized in follow-up 1995; Gallmeier 1991; Lawton 2001; Reiss
work may be especially susceptible to problems 2005).
in the power dynamics between the researcher
and the researched. The researcher anticipates in
conducting such work that the scholarly pay-off 3.3 Issues of Analysis
can be large, particularly as this family of meth-
ods remains novel. Incentives to complete such Two predominant ways of analyzing longitudinal
work successfully can therefore be substantial. qualitative data characterize this type of work. In
Longitudinal researchers must be recurrently and the rst, what may be called an iterative mode of
Longitudinal Qualitative Research 507

analysis, a researcher emphasizes the characteris- such tables and narrative summaries can be
tics and conditions of subjects at multiple points crafted, see Hermanowicz (2009, tables 21, 22,
of contact. In the second, what may be called a 24, 25, 27, 28).
summative mode of analysis, a researcher empha- To analyze the scientists accounts, specic
sizes the net characteristics and conditions of codes were often adopted to mirror the subjects
subjects seen to be produced over time. Each of the interview questions. For example, Do you
mode contains elements of the other. While they think you are working harder, less hard, or about
are not mutually exclusive, the modes stress dif- as hard as you were 10 years ago? Responses
ferent temporal avenues along which to see and were coded using the same response categories
study subjects. offered in the question. In learning what you
have about academic careers, would you go into
3.3.1 Iterative Mode an academic career if you were starting all over
An iterative mode of analyzing longitudinal qual- again? Responses were coded afrmatively or
itative data emphasizes the representation and negatively, and a probe question was analyzed for
portrayal of the objects of research at each of the the explanation provided in the response, using
points they are studied. Hence if data on a sam- codes such as funding, difculty, lack of
pling of people were collected over ve episodes, reward, and freedom.
there might be ve studies, or ve renderings in a In coding and analyzing the longitudinal data,
single study, each presenting and analyzing data I paid particular attention to how responses
for the corresponding and preceding iterations of coalesced around themes of consistency and
research. Here, too, however, the researcher con- change. Following Saldana (2003, 64), I
fronts important choices. employed a variety of conceptual and thematic
Data at time 1 + time x can be analyzed both questions to help situate data analysis,
cross-sectionally and longitudinally. The data including:
can be analyzed in comparison to time 1 or to any
intervening time period where data has been col- 1. What increases or emerges through time?
lected. In other words, even in the context of lon- 2. What is cumulative through time?
gitudinal design, data may be subject to analysis 3. What kinds of surges occur through time?
synchronically and/or diachronically. This pres- 4. What decreases or ceases through time?
ents signicant conceptual and analytic chal- 5. What remains constant or consistent through
lenges. To contend with this situation in my time?
longitudinal study of scientists, I formulated 6. What is idiosyncratic through time?
tables to characterize for readers the career con- 7. What is missing through time?
ditions obtaining when I rst interviewed my 8. Which changes interrelate through time?
subjects. I then used results generated by the lon- 9. What are participant or conceptual rhythms
gitudinal data to compare characterizations of through time?
careers, both among respondents at time 2 (syn- 10. What is the characterization of across time
chronic analysis) and with respondents between experience, and how do characterizations
times 2 and 1 (diachronic analysis). Thus, in the differ by sub-groups of the sample?
longitudinal work, one table presents Early
Career Patterns as established by the baseline My intent was to formulate understandings of
study and another table presents Early- to Mid- respondents experiences and to derive substan-
Career Patterns as established by the longitudi- tive comparisons and contrasts with respect to the
nal data, and likewise for respondents at all other key dimensions of the research design: the insti-
stages in their careers. This arguably helped to tutional contexts in which scientists worked and
eliminate any need for readers to have read or their career stages. This allowed me to address
been familiar with the rst study in order to the guiding question of how scientists age in
understand the second. For illustration on how their work environments. Thus we are placed in a
508 J.C. Hermanowicz

position to compare how cohorts age in organi- either retired or has gone on to try to do something
else.
zations. For illustration, we can compare a scien-
Interviewer: Since retiring, what do you miss
tists account at an elite research university most about your job?
(scientist 1) with that from a more teaching- Respondent: Not a hell of a lot.
oriented institution (scientist 2), to trace the Interviewer: Is there anything that you miss?
Respondent: No. Not at all. There are very few
effects of institution on the individual. The illus-
people that I really enjoy being around, and none
tration depicts intra-cohort variation. Both scien- of them are my former colleagues. I nd them bor-
tists, in the latter-most phases of their careers at ing. This one guy was a very, is still a good friend.
the time of the longitudinal interviews, were born But, you know, Im around him for fteen, twenty
minutes, and Im thinking, Ive got to get away. He
at approximately the same time (in the early
rattles on and on about the same old things.
1930s) and earned their doctorates in physics a Interviewer: What has been the best part of
year apart from each other. Their professional retirement?
careers were spent in one respective institution; Respondent: Doing whatever the hell I want. I can
get up and go to the [gym] and work out, or ride
they followed signicantly different paths:
my bike [downtown] and have coffee, or even go
Scientist 1: I come in usually around 6:00 a.m., over to the departmentI dont do that very much
6:30 a.m., and leave about 5:30 p.m., 5:15 p.m. Im anymore. (Hermanowicz 2009, 207)
here [at the university] about half the time [of the
year]. December was a light travel month because I utilized an approach to data analysis most
of the holidays. I only went to one foreign coun-
trySweden. In January, I had a really big load: often referred to as constant comparison, a com-
Taiwan, UK, and Japan, in that order. It would have ponent of grounded theorizing (Charmaz 1990,
been nice to have it more continuous. I was sup- 2001; Glaser and Strauss 1967; Strauss and Corbin
posed to go to Chile but couldnt t it in, so it was 1994). In this approach, a researcher simultane-
only a conference call. [Looking ahead], Utah is
the rst week of the month and from Utah I go to ously collects and analyzes data. In the course of
the West CoastI have a panel review Academy doing so, the researcher pursues emergent themes
meeting. I have to do some homework for that, get and begins to discover basic social processes in the
organized. I leave tomorrow. From there, Im sup- data. These themes and processes are elaborated,
posed to go to Brazil. When I come back, I have to
give a plenary talk at a conference in Florida. Right modied, or qualied through further data collec-
after that, I go to Arizona. I come back here for tion and analysis. In time, the researcher constructs
three days or two days. I make many trips to and renes, inductively, abstract conceptual cate-
New YorkIf I never wrote another paper, it gories that explain and synthesize these themes
wouldnt be so bad. But I know Im going to write
many more, because I have many in the pipeline, and processes. The researcher eventually seeks to
things that Im working on. Its hard to imagine a integrate categories into a meaningful theoretic
time I wont be doing this. (Hermanowicz 2009, framework that species conditions and conse-
189) quences of the studied processes (Charmaz 2007;
Scientist 2: I had two or three pretty good ideas Charmaz and Mitchell 2001).
during the course of my career, and I havent had In constant-comparative analysis, typical, pre-
any since. I really dont keep up with the litera-
tureI think early on, even though I did some dominant patterns are gleaned from the data.
fairly decent work, both as a graduate student and Thus, for example, the accounts from the two sci-
in the beginning of my career, I never was satised. entists directly above are indicative of modal
I always thought that I could have done better or career patterns found among scientists employed
sooner or more. In more recent years, I have
become content, not only with what I was doing, at research-oriented versus teaching-oriented
but also how much. I think this is a reection of my universities. The career patterns are differenti-
coming to like myself more. ated with respect to key categories, derived from
Interviewer: What worries or concerns would you analyzing the data comparatively, including:
say you have about your career?
Respondent: None, now. My career as a physicist work/family focus, the attribution of place,
is over I talk to colleagues occasionally. But objects of satisfaction, denition of success, and
theyve gone on to do other things. The faculty has whether scientists would seek an academic career
Longitudinal Qualitative Research 509

again. Scientists at research-oriented institutions For deviating cases, one can attempt to answer
who are in late career phases tend to focus on the questions of why and how they have come to
work as well as leisure, view their institutions as depart from the modal pattern. This type of pro-
a haven for their work, understand research as cedure allows the researcher to strengthen asser-
the principal object of their satisfaction; utilize tions and to qualify suggestive conclusions about
external audiences to dene and characterize patterns indicative of groups and sub-groupings
their success, and would readily pursue an aca- in a sample. This analytical strategy serves the
demic career again were they to start all over. By goals of discovering and accounting for inter-
contrast, scientists at teaching-oriented institu- and intra-cohort differences in development.
tions who are in late career phases tend to focus To these ends, analysis of qualitative data
on just leisure, view their institutions as places enables a researcher to arrive at propositions.
departed (i.e., from which they have disen- Owing to the type of data on which they are based,
gaged), understand retirement as the principal the propositions must be qualied and suggestive
object of their satisfaction, utilize internal means such that they are compatible with observations
to dene and characterize their success through that the data support. Thus, proposition may be
self-crafted measures, and would not pursue an taken to be a form of generalization; the former
academic career again. These are but 5 of 20 ana- is more provisional than the latter.
lytic categories that arose in the larger work In the longitudinal study of scientists careers,
through constant comparison. For the other 15, I made the goal of deriving propositions an
and for illustration in how analytic categories can explicit part of my task. I formed propositions
by turn be displayed in tabular form, see from each of the eight sections that composed my
Hermanowicz (2009, tables 22, 25, 28). concluding chapter, where I brought together the
The task then may turn to deviating cases, or ndings about career patterns presented in the
what others sometimes call negative cases, preceding chapters. Across the eight sections of
which one can dene as those cases departing the chapter, I derived 30 propositions. To empha-
from the typical found in any given sub-grouping size for the reader, I numbered the propositions
(Charmaz 2001)instantiations of inter- and where they arose in the concluding chapter and
intra-cohort variation. Thus, taking the example placed the text in italics. For additional summary
above, some individuals do not conform to the and simplicity, I created an appendix to the book
patterns indicative of scientists in the respective that listed all of the propositions numerically by
career stages and institutional types. Small sub- topic as they appeared in the concluding chapter.
sets of scientists in late career at research-oriented To illustrate the creation and usage of proposi-
institutions more closely resemble late career sci- tions in longitudinal qualitative research, I include
entists in teaching-oriented institutions, and vice 4 of the 30 below, identied with the corresponding
versa. An account from a late career scientist at a topic on careers from which they arose.
teaching-oriented institution illustrates a counter-
vailing case: On Expectations and the Rhythm of Careers:
Without grants, you end up with a nine-month sal- Proposition 1: One observes notable reversals in
ary, and you have to ask the department chair if outlook and identication with the career. In
you can teach a class during the summer, which broad terms, elites enter mid-career highly
takes up your summer. Ive never taught in the satised only to end them with ambivalence.
summer. Ive always been able to fund myself dur-
ing the summer for the past thirty years. Every Communitarians enter mid-career highly dis-
month of the summer Ive been here, Ive been satised and end them with serenity. In the
paid. Not everybody here can say that. There are a middle, pluralists start on a high, proceed to
lot of people who dont have money during the either a low or moderate level of satisfaction,
summer, so they have to teach a course.
(Hermanowicz 2009, 239) and conclude on another high.
510 J.C. Hermanowicz

On Anomie and Adaptation: analysis focuses on variation between and among


Proposition 3A: At the end of their careers, elites points in time. By contrast, in the summative
customarily experience the phenomenon mode, analysis focuses on characteristics and
known as anomie. Communitarians and plu- conditions that result in variation at a nal point
ralists experience anomie also, but typically in in time.
much earlier phases of their careers, when it is Armstrong and Hamilton, in their study of
possible for scientists in these worlds of sci- college women, do not portray them at each point
ence to realize that their career expectations the women are studied, and hence we do not have
cannot be realized. representations of the students as freshmen,
Proposition 3B: The incidence and longevity of sophomores, juniors, seniors, and college gradu-
anomie among elites is greatest because elites ates. Instead, the researchers follow their subjects
are exposed to the greatest potential for and use the accumulation of data from the ve
rewards. points of contact to formulate a discrete study
On Future Cohorts of Scientists and Contexts of that includes a set of representations of how col-
Science: lege ultimately conditions female students. They
Proposition 30: Increased emphases on research found that their subjects possessed different
will be accompanied by increased probabili- resources, embodied different gender styles, and
ties of anomie throughout the system of higher professed different ideas about the purpose of
education. college. Consequently, the women developed,
over the course of the college years, varying ori-
For the remaining propositions, and a more entations that characterized their educational and
extended discussion of the contexts in which they occupational values. Some were primed to
can be made, see Hermanowicz (2009, chapter 5 party, others cultivated for success, and others
and appendix G). still motivated for mobility (Armstrong and
As can be inferred from the examples, Hamilton 2013, 38). In turn, these orientations
Propositions mean that they can be tested in a were found to map onto denable pathways
wider range of subsequent work; they can be through and beyond college: a party pathway
explored using a variety of different empirical built for afuent and socially absorbed students;
methods (and not exclusively qualitative ones); a professional pathway tted to high-aspiring
they can be pursued by any assortment of inter- students from privileged families; and a mobil-
ested researchers; they can be mined by different ity pathway designed for the pragmatic and
disciplinary and cross-disciplinary perspectives. vocationally oriented. The researchers call upon
The goal of formalizing propositions from longi- their longitudinal data to capture the processes by
tudinal research is a step one can take toward which students develop and align themselves
encouraging others to continue the work, to with one of these pathways, and the ways by
extend and build upon it, and to thereby sharpen which institutions correspondingly produce
the empirical lessons and theoretical tools avail- inequalities in aspirations and achievements in
able in the study of lives through time. education and work.
The tables used to codify data in Armstrong
3.3.2 Summative Mode and Hamiltons work do not aim to compare and
In a summative mode of analyzing longitudinal contrast their subjects over time (as depicted in
qualitative data, a researcher places stress on the the use of tables indicative of the iterative mode
net results of having followed subjects over time described above). Instead, tables are used primar-
as opposed to portraying the objects of research ily to summarize (1) how women come to differ
at each research episode. Process, change, and along key dimensions of college experience, and
development remain central ideas to the given (2) outline characteristics of the pathways and
study, but the emphasis is on what is produced by post-collegiate trajectories by which the women
their cumulative effects. In the iterative mode, become aligned (for illustration, see Armstrong
Longitudinal Qualitative Research 511

and Hamilton, tables 5.1, 6.1, 6.2, 7.1, 7.2, 8.1, Variability of the life coursewhich involves
and 8.2). The specic ways by which data are an analytic capacity to discern similarities and dif-
analyzed in a summative mode will often mirror ferences by which people make passages through
those used to describe the iterative analytical timeis a message as central to the discussion of
mode: that is, researchers apply conceptual codes this chapter as to currents in the continuing devel-
to data, constantly compare them, identify and opment of life course sociology. A concern for
evaluate negative cases, and work toward inter- and intra-cohort variation, as well as for con-
inductively deriving an authentic account of their text, offers substantive theoretic means by which
subject matter. Thus, for instance, the three pre- to understand the multiple ways that lives are
dominant pathways that guide an essential under- lived, experienced, and interpreted. This is true as
standing of Armstrong and Hamiltons data do much for lives animating scientic careers as for
not appear out of thin air: we can infer that the educational pathways, involvement in crime, fam-
researchers compared and contrasted the evolv- ily and parenting practices, exposure to serious ill-
ing characteristics of their cases over time to ness, and participation in welfare, recovery, and
gradually codify institutional routes that summa- employment programs, each examples that have
tively characterize womens passage through col- informed the present discussion. In other words,
lege. Going to college, even to the same university, longitudinal qualitative research is amenable to
results in altogether different outcomes by virtue studying a broad spectrum of settings and situa-
of ways students become situated institutionally. tions that characterize contemporary social life.
The result is a rendering of intra-cohort variation Work has begun to reveal its promise and pay-
in the ways a universityoperating as a powerful off in these diverse quarters. But longitudinal qual-
sociocultural context for developmentshapes itative research has only begun. A long-standing,
womens achievements, aspirations, and identities. well-developed body of such work has yet to come
into fruition and a methodological tradition has yet
to fully mature. Varieties of topics await to be pur-
4 Conclusion sued using these techniques. Equally vital, when
we utilize these techniques we need to dedicate
The chapter has discussed the parameters that more explicit attention to explaining what we have
guide the use of qualitative methods in longitudi- done, how we have done it, and why we did it that
nal research. Specic conditions in three clusters way. The articulation of technique will foster
of issuesdesign, execution, and analysis methodological maturation, a basis of empirical
frame this type of social scientic inquiry. Design and theoretic development in future work. An indi-
issues are informed by conditions that include the cation of the empirical reach and theoretic power
points at which longitudinal study originates, the of longitudinal qualitative research will lie in its
number and frequency of research episodes, and inspiration for others to follow.
protocol format. Execution issues are guided by
conditions of attrition and retention of research
subjects, reactions from respondents, and the eth- References
ics of conducting follow-up research with people.
Analysis issues are guided by two related but Armstrong, E. A., & Hamilton, L. T. (2013). Paying for
varying modes of thought, iterative and summa- the party: How college maintains inequality.
Cambridge, MA: Harvard University Press.
tive, which condition both the researchers Baltes, P. B., & Nesselroade, J. R. (1984). Paradigm lost
engagement with longitudinal data and how such and paradigm regained: Critique of Dannefers por-
data are presented to audiences. All longitudinal trayal of life-span developmental psychology.
qualitative research will by necessity confront American Sociological Review, 49, 841847.
Bellah, R. N., Madsen, R., Sullivan, W. M., Swidler, A., &
these conditions, and the success of the research Tipton, S. M. (1985). Habits of the heart: Individualism
will depend signicantly on how methodically and commitment in American life. New York: Harper
researchers consider and work through them. and Row.
512 J.C. Hermanowicz

Ben-David, J. (1972). The profession of science and its work: An inside view of qualitative research. Newbury
powers. Minerva, 10, 351383. Park: Sage.
Charmaz, K. (1990). Discovering chronic illness: Using Glaser, B., & Strauss, A. L. (1967). The discovery of
grounded theory. Social Science and Medicine, 30, grounded theory: Strategies for qualitative research.
11611172. New York: Aldine de Gruyter.
Charmaz, K. (2001). Qualitative interviewing and Gustin, B. H. (1973). Charisma, recognition, and the
grounded theory analysis. In J. F. Gubrium & J. A. motivation of scientists. American Journal of
Holstein (Eds.), Handbook of interview research: Sociology, 78(5), 11191134.
Context and method (pp. 675694). Thousand Oaks: Hermanowicz, J. C. (1998). The stars are not enough:
Sage. ScientistsTheir passions and professions. Chicago:
Charmaz, K. (2007). Grounded theory. In G. Ritzer (Ed.), University of Chicago Press.
Encyclopedia of sociology (pp. 20232027). Oxford: Hermanowicz, J. C. (2002). The great interview: 25 strate-
Blackwell. gies for studying people in bed. Qualitative Sociology,
Charmaz, K., & Mitchell, R. G. (2001). Grounded theory 25(4), 479499.
in ethnography. In P. Atkinson, A. Coffey, S. Delamont, Hermanowicz, J. C. (2009). Lives in science: How institu-
J. Loand, & L. Loand (Eds.), Handbook of ethnog- tions affect academic careers. Chicago: University of
raphy (pp. 160174). Los Angeles: Sage. Chicago Press.
Cohler, B. J., & Hostetler, A. (2003). Linking life course Hermanowicz, J. C. (2013). The longitudinal qualitative
and life story: Social change and the narrative study of interview. Qualitative Sociology, 36, 189208.
lives over time. In J. T. Mortimer & M. J. Shanahan Lareau, A. (2011). Unequal childhoods: Class, race, and
(Eds.), Handbook of the life course (pp. 555576). family life (2nd ed.). Berkeley: University of California
New York: Kluwer/Plenum. Press.
Corden, A., & Nice, K. (2007). Qualitative longitudinal Laub, J. H., & Sampson, R. J. (2003). Shared beginnings,
analysis for policy: Incapacity benets recipients tak- divergent lives: Delinquent boys to age 70. Cambridge,
ing part in pathways to work. Social Policy and MA: Harvard University Press.
Society, 6, 557569. Lawton, J. (2001). Gaining and maintaining consent:
Dannefer, D. (1984a). Adult development and social the- Ethical concerns raised in a study of dying patients.
ory: A paradigmatic reappraisal. American Qualitative Health Research, 11, 693705.
Sociological Review, 49, 100116. Levinson, D. J. (1978). The seasons of a mans life.
Dannefer, D. (1984b). The role of the social in life-span New York: Ballantine.
developmental psychology, past and future: Rejoinder Levinson, D. J. (1996). The seasons of a womans life.
to Baltes and Nesselroade. American Sociological New York: Alfred A. Knopf.
Review, 49, 847850. Menard, S. (Ed.). (2002). Longitudinal research (2nd ed.).
Dannefer, D. (2013). Age and sociological explanation: Thousand Oaks: Sage.
Expanding horizons in the study of aging and the life Moffatt, M. (1989). Coming of age in New Jersey: College
course. Contemporary Sociology, 42(6), 793800. and American culture. New Brunswick: Rutgers
Dannefer, D., & Kelley-Moore, J. A. (2009). Theorizing University Press.
the life course: New twists in the paths. In V. Bengston, Murray, S. A., Kendall, M., Carduff, E., Worth, A., Harris,
M. Silverstein, & N. Putney (Eds.), Handbook of theo- F. M., Lloyd, A., Cavers, D., Grant, L., & Sheikh, A.
ries of aging (pp. 389411). New York: Springer. (2009). Use of serial qualitative interviews to
Elder, G. H., Jr. (1974). Children of the great depression: understand patients evolving experiences. British
Social change in life experience. Chicago: University Medical Journal, 339, b3702.
of Chicago Press. Neugarten, B. L. (Ed.). (1968). Middle age and aging: A
Elder, G. H., Jr. (1981). History and the life course. In reader in social psychology. Chicago: University of
D. Bertaux (Ed.), Biography and society: The life his- Chicago Press.
tory approach in the social sciences (pp. 77115). Neugarten, B. L. (1979). Time, age, and the life cycle.
Newbury Park: Sage. American Journal of Psychiatry, 136, 887894.
Elder, G. H., Jr. (1998). The life course and human devel- Neugarten, B. L. (1996). In D. A. Neugarten (Ed.), The
opment. In R. M. Lerner (Ed.), Handbook of child psy- meanings of age: Selected papers of Bernice
chology: Vol. 1. Theoretical models of human L. Neugarten. Chicago: University of Chicago Press.
development (pp. 939991). New York: Wiley. Neugarten, B. L., & Datan, N. (1973). Sociological per-
Ellis, C. (1995). Emotional and ethical quagmires in spectives on the life cycle. In P. B. Baltes & K. W.
returning to the eld. Journal of Contemporary Schaie (Eds.), Life-span development psychology:
Ethnography, 24, 6898. Personality and socialization (pp. 5379). New York:
Erikson, E. H. (1950). Childhood and society. New York: Academic.
Norton. Neugarten, B. L., Moore, J. W., & Lowe, J. C. (1965). Age
Gallmeier, C. P. (1991). Leaving, revisiting, and staying in norms, age constraints, and adult socialization.
touch: Neglected issues in eld research. In W. B. American Journal of Sociology, 70, 710717.
Shafr & R. A. Stebbins (Eds.), Experiencing eld-
Longitudinal Qualitative Research 513

Reiss, M. J. (2005). Managing endings in a longitudinal (pp. 8198). New York: Kluwer Academic/Plenum
study: Respect for persons. Research in Science Publishers.
Education, 35, 123135. Settersten, R. A., Jr., & Hagestad, G. O. (1996a). Whats
Rupp, L. J., & Taylor, V. (2011). Going back and giving the latest? Cultural age deadlines for family transi-
back: The ethics of staying in the eld. Qualitative tions. Gerontologist, 36(2), 178188.
Sociology, 34, 483496. Settersten, R. A., & Hagestad, G. O. (1996b). Whats the
Ruspini, E. (2002). Introduction to longitudinal research. latest? II. Cultural age deadlines for educational and
London: Routledge. work transitions. Gerontologist, 36(5), 602613.
Ryder, N. B. (1965). The cohort as a concept in the study Shaw, C. R. ([1930] 1966). The Jack-Roller: A delinquent
of social change. American Sociological Review, 30, boys own story. Chicago: University of Chicago Press.
843861. Sheehy, G. (1976). Passages: Predictable crises of adult
Saldana, J. (2003). Longitudinal qualitative research: life. New York: Bantam.
Analyzing change through time. Walnut Creek: Strauss, A. L., & Corbin, J. (1994). Grounded theory
AltaMira. methodology: An overview. In N. K. Denzin & Y. S.
Scheper-Hughes, N. (2000). Ire in Ireland. Ethnography, Lincoln (Eds.), Handbook of qualitative research
1, 117140. (pp. 273285). Thousand Oaks: Sage.
Scott, J., & Alwin, D. (1998). Retrospective versus pro- Suttles, G. D. (1968). The social order of the slum:
spective measurement of life histories in longitudinal Ethnicity and territory in the inner city. Chicago:
research. In J. Z. Giele & G. H. Elder Jr. (Eds.), University of Chicago Press.
Methods of life course research: Qualitative and ten Have, P. (2004). Understanding qualitative research
quantitative approaches (pp. 98127). Thousand and ethnomethodology. Thousand Oaks: Sage.
Oaks: Sage. Whyte, W. F. (1996). On the evolution of street corner
Settersten, R. A., Jr. (2003). Age structuring and the society. In A. Lareau & J. Schultz (Eds.), Journeys
rhythm of the life course. In J. Mortimer & through ethnography: Realistic accounts of eldwork
M. Shanahan (Eds.), Handbook of the life course (pp. 973). Boulder: Westview.
Causality in Life Course Studies

Ravaris Moore and Jennie E. Brand

1 Introduction A CHANGE IN y INITIATING A CHANGE IN x x AND y


CAUSING CHANGES IN EACH OTHER SIMULTANEOUSLY OR
4HIS CHAPTER SURVEYS COMMON METHODS EMPLOYED x AND y BOTH RESPONDING TO CHANGES IN A THIRD
FOR CAUSAL INFERENCE IN LIFE COURSE RESEARCH MEASURE z AND  'ENERATE AN UNBIASED ESTIMATE
#ROSS SECTIONAL DATA CAN ESTABLISH ASSOCIATIONS OF THE CHANGE IN y ASSOCIATED WITH A CHANGE IN x.
BETWEEN VARIABLES BUT ALONE CANNOT ESTABLISH 4HE PRIMARY METHODOLOGICAL EMPHASIS IN THIS
CAUSALITY 0EARL   )N THE METHODS DESCRIBED CHAPTER AND INDEED IN THE LITERATURE IS ON ACHIEV
BELOW CAUSAL INFERENCE IS THE RESULT OF ESTIMATING ING OBJECTIVE   OBJECTIVE  HOWEVER DESERVES
THE CONDITIONAL CHANGE IN AN OUTCOME ASSOCIATED DISCUSSION AS WELL
WITH CHANGES IN AN INDEPENDENT VARIABLE IN A THE 4HE MOST CONVINCING FRAMEWORKS FOR FACILITAT
ORETICAL FRAMEWORK WHERE THE IDENTIlED RELATION ING CAUSAL INFERENCE TEND TO ARGUE THAT A PROCESS
SHIP CAN BE PLAUSIBLY INTERPRETED AS CAUSAL EXOGENOUS TO THE OUTCOME OF INTEREST y GOVERNED
)NFERRING CAUSALITY GENERALLY RESTS UPON THE THE INDEPENDENT EVENT OF INTEREST x AND THAT x
RESEARCHERS ABILITY TO COMPLETE TWO OBJECTIVES OCCURRED CHRONOLOGICALLY PRIOR TO y !N INDEPEN
 %STABLISH THAT THE RELATIONSHIP BETWEEN TWO DENT VARIABLE x CAN BE UNDERSTOOD AS EXOGENOUS
MEASURES x AND y SHOULD BE INTERPRETED AS A TO y IF THE PROCESS GOVERNING x IS INDEPENDENT OF
CHANGE IN x INITIATING A CHANGE IN y AS OPPOSED TO y %XOGENEITY OF x IMPLIES THE ABSENCE OF A
MEASURE z THAT INmUENCES BOTH x AND y WHILE THE
CHRONOLOGICAL TIMING IMPLIES THAT y COULD NOT
HAVE CAUSED x 4HIS LEAVES A CAUSAL EFFECT OF x ON
y AS THE PLAUSIBLE INTERPRETATION "EYOND EXPERI
MENTAL OR QUASI EXPERIMENTAL SETTINGS IN WHICH
4HIS RESEARCH MADE USE OF FACILITIES AND RESOURCES AT THE
#ALIFORNIA #ENTER FOR 0OPULATION 2ESEARCH 5#,! WHICH THESE CONDITIONS MAY BE MORE CONVINCINGLY SATIS
RECEIVES CORE SUPPORT FROM THE .ATIONAL )NSTITUTE OF #HILD lED WE OFTEN AIM TO EMPLOY CONDITIONAL INDE
(EALTH AND (UMAN $EVELOPMENT 'RANT 2($ PENDENCE AS A WAY OF LIMITING POTENTIAL PATHWAYS
4HE .ATIONAL )NSTITUTES OF (EALTH 'RANT  2
OF CAUSALITY #ONDITIONAL INDEPENDENCE IMPLIES
($! PROVIDED lNANCIAL SUPPORT FOR THIS
RESEARCH 4HE IDEAS EXPRESSED HEREIN ARE THOSE OF THE THAT ALTHOUGH TWO FACTORS DO NOT IN GENERAL OCCUR
AUTHORS INDEPENDENTLY THEY DO OCCUR INDEPENDENTLY
R. Moore (* s *% "RAND UNDER CERTAIN CONDITIONS 4HAT IS WE ASSUME THAT
$EPARTMENT OF 3OCIOLOGY 5NIVERSITY OF #ALIFORNIA n AFTER CONTROLLING FOR DIFFERENCES IN SOME SET OF
,OS !NGELES  (AINES (ALL ,OS !NGELES OBSERVABLES w x IS GOVERNED BY A PROCESS
#!   53!
E MAIL RAVARISM UCLAEDU

Springer International Publishing Switzerland 2016 515


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_23
516 R. Moore and J.E. Brand

EXOGENOUS TO y. 4HIS ASSUMPTION OF CONDITIONAL DESCRIBES OUR SIMULATED DATA AND THE RELATION
INDEPENDENCE WITH THE CONTINUED ASSUMPTION SHIP BETWEEN TREATMENT AND OUTCOME 3ECTION 
THAT x OCCURRED CHRONOLOGICALLY PRIOR TO y CAN COMPARES ESTIMATES FROM VARIOUS CAUSAL INFER
ALSO BE SUFlCIENT FOR CAUSAL INFERENCE )N PRAC ENCE METHODS TO THE TRUE EFFECT ESTIMATES AND
TICE THIS TYPE OF ARGUMENT IS EMPLOYED VERY OFTEN DISCUSSES THE STRENGTHS AND WEAKNESSES OF EACH
WITH THE CAVEAT THAT THERE MAY BE ADDITIONAL METHOD 3ECTION  CONCLUDES 7ITH THE METHODS
EXCLUDED OBSERVABLE OR UNOBSERVABLE MEASURES IN THIS CHAPTER AND THE EXERCISE OF STUDYING
THAT RESULT IN A VIOLATION OF CONDITIONAL INDEPEN SIMULATED DATA WE ULTIMATELY lND EVIDENCE THAT
DENCE AND COMPROMISE GROUNDS FOR CAUSAL OUR MOST COMMONLY USED CAUSAL INFERENCES
INFERENCE TOOLS ARE SUFlCIENT FOR UNDERSTANDING KEY RELA
"ELOW WE DISCUSS A NUMBER OF METHODS COM TIONSHIPS UNDERLYING COMPLEX DATA GENERATING
MONLY EMPLOYED TO UNDERSTAND THE CONDITIONAL PROCESSES
CHANGE IN AN OUTCOME ASSOCIATED WITH A CHANGE IN
AN INDEPENDENT VARIABLE 4HESE INCLUDE MULTIVAR
IATE REGRESSION MATCHING MODELS INSTRUMENTAL 2 Methods for Causal
VARIABLE MODELS AND lXED EFFECTS MODELS ALL Inference in Life Course
USEFUL APPROACHES FOR CAUSAL INFERENCE WHEN Research
ESTIMATING EFFECTS WITH DATA OVER THE LIFE COURSE
7E ALSO BRIEmY DISCUSS EXPERIMENTAL AND QUASI 4HIS SECTION DESCRIBES AND ILLUSTRATES THE USE OF
EXPERIMENTAL FRAMEWORKS 4O STREAMLINE THE SEVERAL COMMON CAUSAL INFERENCE APPROACHES
DISCUSSION SEVERAL METHODS WERE OMITTED SUCH AS 7E START EACH SECTION WITH A BRIEF STATISTICAL
STRUCTURAL EQUATION MODELS GROWTH CURVE MODELS PRESENTATION OF EACH MODEL AND ITS NECESSARY
SEE -ACMILLAN AND &URSTENBERG CHAPTER hThe ASSUMPTIONS AND THEN DISCUSS SOME OF THE
,OGIC AND 0RACTICE OF 'ROWTH #URVE !NALYSIS STRENGTHS AND WEAKNESSES OF EACH METHOD
-ODELING 3TRATEGIES FOR ,IFE #OURSE $YNAMICS
THIS VOLUME METHOD OF MOMENT ESTIMATORS AND
A LARGE CLASS OF NON PARAMETRIC ESTIMATORS &OR 2.1 Regression Models
THE MODELS DISCUSSED BELOW WE AIM TO GIVE THE
READER A PRACTICAL INTUITION FOR THE USEFULNESS /RDINARY ,EAST 3QUARES /,3 REGRESSION MODELS
AND APPLICABILITY OF EACH MODEL )N SOME CASES ARE UNDOUBTEDLY THE MOST HEAVILY EMPLOYED TOOL
WE ADD HEURISTIC MATHEMATICAL EXPLANATION TO FOR UNDERSTANDING THE RELATIONSHIP BETWEEN VEC
ILLUSTRATE TECHNICAL POINTS AND THE BENElTS AND TORS /,3 MODELS THE CONDITIONAL EXPECTED VALUE
LIMITATIONS OF MODEL ASSUMPTIONS 7E ALSO CON OF A MEASURE y GIVEN A lXED VALUE OF d. CAN BE
STRUCT SIMULATED DATA AND DEMONSTRATE EMPIRICAL ESTIMATED DIRECTLY AS A LINEAR PROJECTION 'IVEN
APPLICATION OF THE DISCUSSED MODELS N  MATRICES y AND AND N K MATRIX X WITH
4HE CHAPTER PROCEEDS AS FOLLOWS 3ECTION  THE ASSUMED RELATIONSHIP y = d + WE CAN SOLVE
DESCRIBES VARIOUS CAUSAL INFERENCE METHODS AND FOR AN ESTIMATOR GOLS 
DISCUSSES THEIR BENElTS AND LIMITATIONS IN LIFE
COURSE RESEARCH 3ECTION  CONSIDERS CAUSAL GOLS = [ d d ]1 d y
INFERENCE IN THE PRESENCE OF HETEROGENEITY
TIME VARIATION AND MEDIATION 3ECTION  4O OBTAIN CONSISTENT ESTIMATES SUCH THAT THE
PARAMETER ESTIMATE GOLS CONVERGES TO AS SAMPLE

&OR EXAMPLE DISPLACED WORKERS ARE LIKELY TO HAVE LOWER SIZE INCREASES WE MUST MEET THE FOLLOWING
WAGES THAN NON DISPLACED WORKERS EVEN IN THE ABSENCE OF ASSUMPTIONS #AMERON AND 4RIVEDI  
DISPLACEMENT (OWEVER IF WE WERE ABLE TO SPECIFY ALL THE
FACTORS THAT INmUENCE THE PROBABILITY A WORKER WAS DIS
PLACED AND CONDITION UPON THOSE FACTORS WE COULD THEN
I 4HE DATA ARE GENERATED BY A PROCESS THAT CAN
ASSUME CONDITIONAL INDEPENDENCE BETWEEN DISPLACEMENT BE ACCURATELY MODELED AS y = d + WHERE
AND WORKERS WAGES y IS A CONTINUOUS OUTCOME
Causality in Life Course Studies 517

II 4HE K K MATRIX XX IS WELL BEHAVED SUCH )N LIFE COURSE RESEARCH THERE ARE OFTEN A RANGE OF
THAT IT APPROACHES A lNITE EXISTENT NON POTENTIALLY IMPORTANT FACTORS OMITTED FROM A
SINGULAR MATRIX AS N INCREASES MODEL ! NUMBER OF FACTORS MAY ENTER A MODEL
III E diH i = 0 SUCH THAT xi AND ui ARE THROUGH THE ERROR COMPONENT SUCH AS POLITICAL
UNCORRELATED OR HISTORICAL CONTEXTS UNOBSERVED PREFERENCES
IV 4HE DATA ARE INDEPENDENTLY AND IDENTICALLY ANDOR CONSTRAINTS AND OTHER OMITTED VARIABLES
DISTRIBUTED OVER i    x N WITH E[d=   )F THESE FACTORS INmUENCE THE OUTCOME OF INTER
AND Var ( i | d ) = i .
2
EST THEY MAY ALSO INmUENCE VALUES OF X LEADING
TO A CORRELATION THAT VIOLATES THE KEY ASSUMPTION
/,3 IS LIMITED FOR EVALUATING LIFE COURSE OF SELECTION ON OBSERVABLES AND MAY PRODUCE A
CAUSAL EFFECTS &IRST THERE ARE A NUMBER OF WAYS BIASED ESTIMATE OF  .OTHING IN THE STANDARD
IN WHICH LIFE COURSE PROCESSES MAY VIOLATE THE /,3 FRAMEWORK IMPLIES THAT THE ESTIMATED
ASSUMPTIONS SET FORTH ABOVE 7HILE /,3 ASSUMES RELATIONSHIP BETWEEN THE INDEPENDENT AND
A CONSTANT LINEAR RELATIONSHIP BETWEEN THE INDE DEPENDENT VARIABLE IS CAUSAL /,3 OFFERS A WAY
PENDENT VARIABLES AND OUTCOME MANY LIFE PRO OF ESTIMATING THE CONDITIONAL CHANGE IN ONE
CESSES MAY NOT BEHAVE THIS WAY (UNGERFORD AND MEASURE ASSOCIATED WITH A CHANGE IN A RELATED
3OLON  FOR EXAMPLE lND EVIDENCE THAT LOG MEASURE !NY ATTEMPT TO INTERPRET THIS CHANGE AS
WAGES ARE NOT A SMOOTH FUNCTION OF EDUCATION CAUSAL IS BASED ON A THEORY OF THE MECHANISMS AT
4HEY REPORT EVIDENCE OF A hSHEEPSKINv EFFECT PLAY WHICH EXTENDS BEYOND THE MATHEMATICAL
WHERE WAGES INCREASE WITH THE COMPLETION OF PROPERTIES OF /,3
CERTAIN EDUCATIONAL THRESHOLDS ! MODEL THAT &OR BINARY AND CATEGORICAL OUTCOMES WE TURN
INSTEAD ASSUMED A LINEAR RELATIONSHIP BETWEEN TO BINARY AND DISCRETE CHOICE MODELS FOR INFERENCE
YEARS OF COMPLETED EDUCATION AND WAGES WOULD "INARY CHOICE MODELS ARE GENERALLY FORMULATED
BE MIS SPECIlED IF THE DATA GENERATING PROCESS FROM THE CONCEPTUAL FRAMEWORK OF STUDYING A
WERE MORE REmECTIVE OF THE @SHEEPSKIN PROCESS LATENT VARIABLE y* USING OBSERVABLE
0ROCESSES OF CUMULATIVE ADVANTAGE MAY LIKEWISE
HAVE A WELL DElNED NON LINEAR GROWTH PATH ;SEE 1 y* 0
$I0RETE AND %IRICH  FOR A REVIEW OF THE y=
0 else
SOCIOLOGICAL LITERATURE ON CUMULATIVE ADVANTAGE=
3OME SOCIAL PROCESSES MAY FOLLOW A PATTERN SUCH WHERE y* = xi E + . 'IVEN THIS IDENTITY WE CAN
THAT THE RETURN DEPENDS UPON THE STOCK OF A PAR WRITE
TICULAR TYPE OF CAPITAL AT A MOMENT IN TIME &OR
EXAMPLE WE MAY OBSERVE NOTICEABLE DIFFERENCES Pr[ yi = 1 | xi ] = Pr( y* 0)
IN MEAN ACCOUNT BALANCES OVER TIME BETWEEN TWO = Pr( xi E + 0)
SAVINGS ACCOUNTS WITH DIFFERENT LEVELS OF INITIAL
CAPITAL YET STILL SUBJECT TO THE SAME COMPOUND 7E CAN TRANSFORM THIS INTO A PROBIT OR LOGIT ESTI
INTERESTS RATES -ERTON DESCRIBES A PHENOMENON MATOR BY MAKING APPROPRIATE ASSUMPTIONS ON THE
WHERE GREAT CONTRIBUTIONS EARLY IN ONES CAREER DISTRIBUTION OF  !S BINARY OUTCOME MODES ARE
MAY HELP GENERATE RESOURCES AND THAT LEAD TO ESSENTIALLY A TRANSFORMATION AN /,3 MODEL THEY
ADDITIONAL CONTRIBUTIONS AND LONG TERM CAREER INHERIT MANY OF THE SAME LIMITATIONS FACED BY
BENElTS 4HE SAME CONTRIBUTION LATER IN ONES /,3 MODELS FOR CAUSAL INFERENCE OVER THE LIFE
CAREER MAY NOT PRODUCE THE SAME LONG TERM COURSE 7E FACE POTENTIAL UNCERTAINTY ABOUT THE
EFFECT 4HESE TYPES OF CUMULATIVE ADVANTAGE TRUE FUNCTIONAL FORM GOVERNING OUR PROCESS OF
PROCESSES MAY GENERATE RELATIONSHIPS THAT CANNOT INTEREST AND THE UNCERTAINTY MAY INCREASE WITH
BE ACCURATELY CAPTURED WITH A LINEAR MODEL BINARY MODELS BASED ON THE ASSUMED STRUCTURE OF
3ECOND WE HAVE LIMITED ABILITY TO DIFFERENTIATE A LATENT VARIABLE y* AND ITS ASSUMED RELATIONSHIP TO
BETWEEN CAUSAL AND NON CAUSAL RELATIONSHIPS OBSERVED yi -OST IMPORTANTLY WE FACE THE SAME
0OSSIBLE CORRELATION BETWEEN OBSERVABLES AND UNCERTAINTY CONCERNING THE ASSUMED INDEPENDENCE
UNOBSERVABLES THREATENS TO BIAS /,3 ESTIMATES OF OBSERVED xi AND UNOBSERVED i.
518 R. Moore and J.E. Brand

,AGGED DEPENDENT VARIABLE MODELS ARE A VARI INDIVIDUALS HAVE IDENTICAL PROlLES FOR OBSERVABLES
ATION OF /,3 OR LOGISTIC REGRESSION MODELS THAT X WE ATTRIBUTE AVERAGE DIFFERENCES IN VALUES TO
UTILIZE DATA OVER THE LIFE COURSE TO OBTAIN BETTER DIFFERENCES IN TREATMENT ASSIGNMENT 4HIS ASSUMES
ESTIMATES OF CAUSAL EFFECTS 3UCH MODELS INCLUDE THAT CONDITIONING ON X IS SUFlCIENT FOR ACHIEVING
A PRE TREATMENT MEASURE OF THE OUTCOME AS A CONDITIONAL INDEPENDENCE OF THE OUTCOME AND THE
REGRESSOR AND IN SO DOING HELP CONTROL FOR DIFFER TREATMENT SUCH THAT -ATCHING
ENCES IN PRE TREATMENT CHARACTERISTICS THAT MAY METHODS PROVIDE A CONCEPTUAL FRAMEWORK THAT
BIAS EFFECT ESTIMATES (OWEVER WE NOTE A FEW PARALLELS EXPERIMENTAL SETTINGS FOR EVALUATING
LIMITATIONS OF THIS APPROACH &IRST THE APPROACH ESTIMATED EFFECTS IN OBSERVATIONAL SETTINGS WHERE
REQUIRES THE AVAILABILITY OF A PRE TREATMENT ASSIGNMENT TO TREATMENT IS NON RANDOM
OUTCOME MEASURE WHICH MAY NOT BE AVAILABLE IN -ATCHING ESTIMATORS MAY INTRODUCE A DIMEN
SOME RESEARCH SETTINGS 3ECOND CONTROLLING FOR SIONALITY PROBLEM THAT STRAINS AVAILABLE DATA
'IVEN THE NEED TO CONDITION UP X R WHERE Rk
k
PRE TREATMENT DIFFERENCES DOES NOT CONTROL FOR
OTHER CHANGES THAT OCCUR BETWEEN PRE TREATMENT DENOTES THE REAL NUMBERS AND k INDEXES DIMEN
MEASURE AND FOLLOW UP 4HIS IMPLIES A POTENTIAL SIONALITY THE DATA NEED TO HAVE A SUFlCIENT
NEED TO CONTROL FOR CHANGES THAT OCCUR IN THE NUMBER OF K DIMENSIONAL MATCHES TO FACILITATE
INTERIM AND TO LIMIT THE INTERIM TIME PERIOD TO THE ESTIMATION OF AN EXPECTED DIFFERENCE )F K IS RELA
EXTENT POSSIBLE TO LIMIT THE LIKELIHOOD OF UNOB TIVELY LARGE THE DATA MAY NOT HAVE THE AVAILABLE
SERVABLE CHANGES INmUENCING EFFECT ESTIMATES MATCHES %VEN FOR SMALL VALUES OF K WHERE EACH
DIMENSION IS CONTINUOUS HAVING AVAILABLE
MATCHES MAY BE PROBLEMATIC 2OSENBAUM AND
2.2 Matching Models 2UBIN  RECOMMEND PROPENSITY SCORE
MATCHING TO REDUCE THE DIMENSIONALITY OF THE
-ATCHING ESTIMATORS ARE USED TO ESTIMATE TREAT MATCHING PROBLEM 'IVEN THAT ASSIGNMENT TO
MENT EFFECTS BY TAKING THE AVERAGE DIFFERENCE IN A TREATMENT DEPENDS UPON OBSERVABLES W WE
SELECTED OUTCOME BETWEEN INDIVIDUALS WITH THE MODEL THE PROBABILITY TREATMENT AS
SAME PRE TREATMENT OBSERVABLES AND DIFFERENT
TREATMENT ASSIGNMENTS 4HE FUNDAMENTAL DIFl Pr(d = 1 | X ) = F ( X)
CULTY IN ESTIMATING TREATMENT EFFECTS INVOLVES THE
IMPOSSIBILITY OF OBSERVING THE SAME OBSERVATION WHERE Fq IS A CUMULATIVE DISTRIBUTION FUNCTION
UNDER OBSERVED AND COUNTERFACTUAL CONDITIONS AND A BINARY MODEL IS ESTIMATED USUALLY PROBIT OR
'IVEN AN OUTCOME y THAT DEPENDS UPON OBSERV LOGIT %STIMATING SUCH A MODEL ALLOWS THE CAL
ABLES X UNOBSERVABLES AND TREATMENT STATUS d CULATION OF Pr(dn = 1 | X ) A ONE DIMENSIONAL
WE CAN ESTIMATE AN AVERAGE TREATMENT EFFECT BY MEASURE THAT SUMMARIZES ONES LIKELIHOOD OF
TREATMENT GIVEN AVAILABLE OBSERVABLES 7E THEN
= E[ y | Xi , i , d = 1] E[ y | X j , j , d = 0]. MATCH TREATED AND UNTREATED OBSERVATIONS BASED
ON THE ONE DIMENSIONAL ESTIMATED PROBABILITY OF
4HE ABOVE ESTIMATES BY COMPUTING THE EXPECTED TREATMENT INSTEAD OF MATCHING ON K DIMENSIONAL
DIFFERENCE IN OUTCOME VALUE AMONG INDIVIDUALS OBSERVABLES 3EE 2OSENBAUM AND 2UBIN 
WHO DIFFER ONLY BY TREATMENT STATUS 'IVEN THAT  FOR A MORE FORMAL DISCUSSION OF THE METHOD
AND #ALIENDO AND +OPEINIG  ,EUVEN AND
3IANESI  AND -ORGAN AND (ARDING )

4HE COUNTERFACTUAL CONDITION IS THE OUTCOME THAT WOULD
FOR PRACTICAL GUIDANCE ON IMPLEMENTING THE
HAVE RESULTED HAD AN OBSERVATION EXPERIENCED AN ALTERNATIVE
TREATMENT ASSIGNMENT &OR EXAMPLE SUPPOSE WE ARE METHOD 4HERE ARE MANY EMPIRICAL EXAMPLES OF
INTERESTED IN THE LABOR MARKET EFFECTS OF COLLEGE SELECTIVITY PROPENSITY SCORE MATCHING TO ESTIMATE TREATMENT
7E OBSERVE THE WAGES OF STUDENTS WHO ATTEND SELECTIVE EFFECTS IN SOCIAL RESEARCH OVER THE LIFE COURSE 3EE
COLLEGES BUT WE DO NOT OBSERVE THE WAGES OF STUDENTS WHO
"RAND AND (ALABY  AND "RAND ET AL )
ATTEND SELECTIVE COLLEGE HAD THEY NOT ATTENDED THOSE
COLLEGES AND INSTEAD ATTENDED NON SELECTIVE COLLEGES 4HIS FOR A FEW RECENT EXAMPLES OF RESEARCH ON THE
UNOBSERVED OUTCOME IS THE COUNTERFACTUAL EFFECTS OF HIGHER EDUCATION ON LIFE OUTCOMES
Causality in Life Course Studies 519

-ATCHING ESTIMATORS OFFER AN APPROACH FOR 1 N


EFFECT ESTIMATION WHEN PARAMETRIC REGRESSIONS wi yi =1
N i =1

G IV =
ASSUMPTIONS MAY BE VIOLATED AND EXPERIMENTAL
1 N
APPROACHES MAY BE UNAVAILABLE (OWEVER LIKE w d
N i =1 i i =1
REGRESSION APPROACHES MATCHING MODELS ARE
LIMITED 7HEN USING A PROPENSITY SCORE MATCHING 4HE INSTRUMENTAL VARIABLE wi SATISlES THE FOL
APPROACH OR A REGRESSION APPROACH ONE NEEDS A LOWING ASSUMPTIONS
MODEL THAT IS FAIRLY SUCCESSFUL AT PREDICTING
WHETHER AN OBSERVATION WILL BE ASSIGNED TO I wi MUST BE CORRELATED WITH di SUCH THAT
TREATMENT 7ITH SOME PROCESSES OF INTEREST TO LIFE E[xizi] 
COURSE RESEARCH IT MAY BE EXTREMELY DIFlCULT TO II wi MUST BE UNCORRELATED WITH zi SUCH THAT
CONSTRUCT A MODEL OF EXOGENOUS TREATMENT E[wizi=  
ASSIGNMENT 7E MUST AGAIN MAKE ASSUMPTIONS III wi MUST BE UNCORRELATED WITH i SUCH THAT
CONCERNING UNOBSERVABLES  4HOUGH WE OFTEN E[ wi H i ] = 0 .
EXPECT INDIVIDUALS WITH SIMILAR OBSERVABLES TO
HAVE SIMILAR UNOBSERVABLES THIS MAY NOT BE THE 3EE #AMERON AND 4RIVEDI  #HAP  'REENE
CASE 'IVEN A LIFE COURSE PERSPECTIVE WHERE ( CHAP  AND 7OOLDRIDGE   FOR
ENCOMPASES ALL PRIOR CHARACTERISTICS ABOUT AN A MORE GENERAL TREATMENT OF )6 ESTIMATORS
OBSERVATION NOT CAPTURED BY X SIGNIlCANT DIFFER )N LIFE COURSE RESEARCH IT IS EXTREMELY DIFlCULT TO
ENCES BETWEEN i AND j FOR i j ARE TYPICALLY IDENTIFY AN )6 THAT SATISlES THE ASSUMPTIONS ABOVE
EXPECTED )T IS TYPICALLY EASY TO lND A MEASURE THAT IS CORRE
LATED WITH THE TREATMENT OF INTEREST xi BUT DIFlCULT
TO ARGUE THAT A POTENTIAL INSTRUMENT IS UNCORRELATED
2.3 Instrumental Variable Models WITH AN UNOBSERVED CONFOUNDER AND AN INSTRUMENT
WITH NO INDEPENDENT EFFECT ON THE OUTCOME OF INTER
)N MANY RESEARCH SITUATIONS /,3 EFFECT ESTI EST yi BEYOND ITS EFFECT THROUGH xi ,LERAS -UNEY
MATES ARE BIASED DUE TO A VIOLATION OF THE IGNOR ( OFFERS AN EXAMPLE OF IDENTIlCATION OF AN )6
ABILITY OR SELECTION ON OBSERVABLES ASSUMPTION IN LIFE COURSE RESEARCH 3HE AIMS TO ESTIMATE THE
4HE INSTRUMENTAL VARIABLES )6 ESTIMATOR AIMS TO CAUSAL EFFECT OF EDUCATION ON ADULT MORTALITY USING
PRODUCE EFFECT ESTIMATES FREE FROM THE BIAS A METHOD THAT ACCOUNTS FOR THE LIKELY EXISTENCE OF
GENERATED BY VIOLATIONS TO THE IGNORABILITY UNOBSERVABLES THAT MAY LEAD INDIVIDUALS WHO
ASSUMPTION BY IDENTIFYING A SOURCE OF EXOGENOUS CHOOSE TO OBTAIN GREATER EDUCATION TO MAKE OTHER
VARIATION THE INSTRUMENT AND ESTIMATING EFFECTS CHOICES THAT AFFECT THEIR MORTALITY A PHENOMENON
BASED ON THIS EXOGENOUS VARIATION 3UPPOSE WE THAT WOULD LEAD TO A BIAS IN THE ESTIMATED AFFECT OF
WISH TO ESTIMATE THE EFFECT OF di ON yi GIVEN EDUCATION ON MORTALITY ,LERAS -UNEY )
KNOWLEDGE THAT AN UNOBSERVABLE zi EXISTS THAT IS EMPLOYS AN )6 APPROACH USING CHANGES IN COMPUL
CORRELATED WITH di AND AFFECTS yi /,3 HAS NO WAY SORY STATE MANDATED SCHOOLING LEVELS AS AN INSTRU
OF DISTINGUISHING BETWEEN THE EFFECTS ON di ON yi MENTAL VARIABLE #HANGES IN COMPULSORY EDUCATION
AND THE CORRELATED EFFECTS OF zi ON yi )6 ESTIMATION LAWS INCREASE THE EDUCATION LEVEL OF STUDENTS WHO
ADDRESSES THIS LIMITATION OF /,3 BY INTRODUCING WOULD HAVE DISCONTINUED THEIR EDUCATION EARLIER IN
AN INSTRUMENTAL VARIABLE wi THAT IS CORRELATED THE ABSENCE OF THE POLICY CONSTRAINT #HANGES IN
WITH xi BUT UNCORRELATED WITH BOTH wi AND ui. STATE LAWS DISCONTINUITIES IN FEDERAL LAWS GENETIC
-ATHEMATICALLY THIS IMPLIES THAT ASSUMPTIONS i VARIATION AND OTHER VERY INVENTIVE STRATEGIES HAVE
THROUGH iii ABOVE HOLD )NTUITIVELY IT IMPLIES THE BEEN USED TO IDENTIFY INSTRUMENTS EG SEE "LACK
EXISTENCE OF A MEASURE THAT SHIFTS di INDEPEN ET AL FORTHCOMING WHICH USES DISTANCE TO TRAINS
DENTLY OF BOTH zi AND i TO PRODUCE AN ESTIMATE OF AS AN INSTRUMENTAL VARIABLE TO ACCOUNT FOR PROPEN
THAT IS UNBIASED 4HE BIVARIATE )6 ESTIMATOR SITY TO MIGRATE NORTH AMONG SOUTHERN "LACKS IN THE
TAKES THE FORM PREDICTION OF MORTALITY 
520 R. Moore and J.E. Brand

2.4 Fixed and Random Effects PER CHILD CONTROLLING FOR MARITAL STATUS AND
Models HUMAN CAPITAL VARIABLES
3IBLING lXED EFFECTS MODELS ARE MODELS WHERE
&IXED EFFECTS MODELS ARE USEFUL WHEN DATA CON lXED EFFECTS ARE ASSUMED TO BE FAMILY SPECIlC
TAIN REPEATED OBSERVATIONS FROM A GIVEN UNIT INSTEAD OF INDIVIDUAL SPECIlC 3UCH MODELS ARE
LIKELY WHEN WE ESTIMATE EFFECTS OVER THE LIFE WIDELY USED TO UNDERSTAND THE EFFECTS OF LIFE
COURSE %ACH UNIT IS EXPECTED TO HAVE A TIME EVENTS ON CHILDREN IN AN ANALYTICAL FRAMEWORK
INVARIANT VALUE i THAT SHIFTS THE OUTCOME OF THAT CONTROLS FOR DIFFERENCES BETWEEN CHILDREN AND
INTEREST yi IN ADDITION TO CHANGES IN yi THAT ARE FAMILIES THAT MAY OTHERWISE BIAS EFFECT ESTIMATES
CORRELATED WITH X )N A MODEL #URRIE AND 4HOMAS  OFFER AN EXAMPLE OF
SIBLING lXED EFFECTS MODELS APPLIED TO AN ANALY
yit = i + Xit + it
SIS OF THE LONG TERM EFFECTS OF (EAD 3TART ON
WHERE i DENOTES A UNIT OF OBSERVATION AND t DENOTES STUDENTS ACADEMIC ACHIEVEMENT 4HEY lND SIG
THE TIME PERIOD OF THE OBSERVATION /,3 LEADS TO NIlCANT GAINS IN TEST SCORE AND DECLINES IN THE
BIASED ESTIMATES OF OUR PARAMETER OF INTEREST LIKELIHOOD STUDENTS WILL REPEAT A GRADE WHERE THE

IF i IS CORRELATED WITH Xit SUCH THAT E[ ditD i ] 0 . POSITIVE BENElTS ARE CONCENTRATED AMONG WHITE
! lXED EFFECTS MODEL PRODUCES AN ESTIMATE OF CHILDREN 2ANDOM EFFECTS MODELS HAVE THE ADVAN
UNBIASED BY i &IXED EFFECTS MODELS AVOID A BIAS TAGE OF GREATER EFlCIENCY AND GREATER STATISTICAL
FROM i BY ESTIMATING E USING WITHIN PERSON POWER TO DETECT EFFECTS RELATIVE TO lXED EFFECTS
UNIT VARIATION IN Xit AND yit 7E ESTIMATE EFE MODELS AND ALLOW ESTIMATION OF EFFECTS ON TIME
USING AN /,3 REGRESSION OF THE DEVIATIONS IN Xit ON INVARIANT MEASURES 7OOLDRIDGE   2ANDOM
THE DEVIATIONS IN yit 3INCE yit = i + dit + it WE EFFECTS MODELS HOWEVER ASSUME OFTENTIMES
CAN SUBTRACT yit FROM BOTH SIDES OF THE EQUATION UNREALISTICALLY THAT i IS UNCORRELATED WITH Xit.
ABOVE AND REARRANGE TERMS TO GET !DDITIONAL RESTRICTIONS IMPOSED IN THE USUAL
APPLICATION OF BOTH lXED AND RANDOM EFFECTS
yit yit = i + Xit + it yit
MODELS ARE PROBLEMATIC FOR LIFE COURSE RESEARCH
yit yit = i i + ( Xit Xit ) + it it &OR INSTANCE THE MODELS RESEARCHERS TYPICALLY
ESTIMATE ASSUME THAT THE COEFlCIENTS OF THE SAME
yit = X it + it
COVARIATE AND THE ERROR VARIANCES OF EQUATIONS ARE

EQUAL OVER TIME (OWEVER IF INDIVIDUALS ENCOUN

Where y it = yit yit X it = Xit Xit AND TER LIFE COURSE TRANSITIONS OVER TIME THE STABLE
H it = it it . 3INCEi ISTIME INVARIANT D i D i = 0, EFFECTS ASSUMPTIONS COULD BE INVALID "OLLEN AND
WHICH LEAVES OUR lXED EFFECTS ESTIMATE "RAND   ,IKEWISE LAGGED DEPENDENT VARI
EFE i FREE EFE IS AN UNBIASED ESTIMATOR OF ABLES ARE TYPICALLY ASSUMED TO HAVE NO EFFECTS ON

IF X it AND it ARE UNCORRELATED 4IME VARYING CURRENT VALUES ! PRIOR YEARS EMPLOYMENT STATUS
PROCESSES THAT MAY THREATEN THE PLAUSIBILITY OF THE FOR EXAMPLE MAY INmUENCE SUBSEQUENT EMPLOY
IGNORABILITY ASSUMPTION REMAIN A SIGNIlCANT CON MENT STATUS EVEN ADJUSTING FOR CONTROL VARIABLES
CERN "UDIG AND %NGLAND  OFFER AN EXAM 4HESE AND OTHER CONSTRAINTS CAN NEVERTHELESS BE
PLE OF INDIVIDUAL lXED EFFECTS MODELS APPLIED TO MODIlED TO ESTIMATE ALTERNATIVE MODELS THAT RELAX
AN ANALYSIS OF THE WAGE PENALTY OF MOTHERHOOD THESE RESTRICTIONS ;SEE "OLLEN AND "RAND )
4HEY FOUND A WAGE PENALTY OF APPROXIMATELY   FOR ALTERNATIVE SPECIlCATIONS=
-ECHANICAL APPLICATION OF lXED EFFECTS MODELS

WITHOUT ATTENTION TO THE UNDERLYING CAUSAL PROCESS
,AGGED DEPENDENT VARIABLE MODELS AND lXED EFFECTS
MODEL ARE SIMILAR IN THAT THEY BOTH INCORPORATE REPEATED CAN LEAD TO ERRONEOUS CONCLUSIONS 3PECIlCALLY A
MEASURES OF AN OUTCOME 4HEY DIFFER IN THAT WHILE lXED POTENTIAL CONCERN WITH lXED EFFECTS MODELS BOTH
EFFECTS MODELS ASSIGN SPECIAL STATUS TO PRE TREATMENT OUT INDIVIDUAL AND SIBLING MODELS CONCERNS THE
COMES BY INCORPORATING THEM INTO THE DEPENDENT VARIABLE
AMBIGUITY OF EXACTLY WHAT IS SUBSUMED IN THE
LAGGED DEPENDENT VARIABLE MODELS TREAT OUTCOME OBSERVA
TIONS FROM PREVIOUS PERIODS AS SIMPLY ANOTHER REGRESSOR lXED EFFECTS PARAMETER )T IS NOT ALWAYS CLEAR THAT
IN EXPLAINING THE LEVEL OF THE POST TREATMENT OUTCOME A lXED EFFECTS MODEL IS NOT INTRODUCING AN
Causality in Life Course Studies 521

ENDOGENOUS SELECTION BIAS AND POSSIBLY DAMPEN CHANGE FOR LOW INCOME FAMILIES )N THE CONTEXT OF
ING THE TREATMENT EFFECTS OF INTEREST 2ESEARCHERS THIS RESEARCH QUESTION -4/ WAS QUASI EXPERI
SHOULD CAREFULLY CONSIDER WHETHER THE DIFFERENC MENTAL BECAUSE RESEARCHERS ULTIMATELY HAD NO
ING PROCESS UNDERMINES THE VARIABILITY THEY WISH CONTROL OVER WHICH INDIVIDUALS FROM THE VOUCHER
TO EXPLAIN AND INDEED OVER CONTROLS FOR POTEN GROUPS EXERCISED THEIR OPTION TO MOVE %ACH
TIALLY OMITTED VARIABLES TREATMENT FAMILY CHOSE WHETHER OR NOT TO EXERCISE
THEIR OPTION TO MOVE AND THIS LEAD TO SELECTION
ISSUES AND CONTROVERSY CONCERNING THE INTERNAL
2.5 Experimental and Quasi- VALIDITY AND INTERPRETATION OF -4/ EVALUATION
Experimental Designs RESULTS ;SEE 'OERING ET AL  ,UDWIG ET AL
( AND 3ANBONMATSU ET AL  FOR FURTHER
%XPERIMENTAL RESEARCH DESIGNS HELP FACILITATE DISCUSSION OF -4/=
IDENTIlCATION OF TREATMENT EFFECTS BY MAINTAINING ! NATURAL EXPERIMENT IS A TYPE OF QUASI
FULL CONTROL OVER THE TREATMENT ASSIGNMENT PRO EXPERIMENT WHERE THE RESEARCHER HAS NO CONTROL
CESS +NOWLEDGE OF THE TREATMENT ASSIGNMENT OVER THE TREATMENT ASSIGNMENT PROCESS BUT HAS
PROCESS ALLOWS A STRONGER ARGUMENT FOR CLAIMING REASON TO BELIEVE THAT THE PROCESS YIELDS A CAUSAL
THAT TREATMENT IS UNCORRELATED WITH UNOBSERVABLES RELATIONSHIP THAT FACILITATES THE ESTIMATES OF UNBI
THAT MAY LEAD TO BIAS IN PARAMETER ESTIMATES AND ASED TREATMENT EFFECTS 3HARKEY  EXAMINES
THAT THERE ARE NO SELECTION EFFECTS GOVERNING WHO THE ACUTE EFFECTS OF LOCAL HOMICIDES ON COGNITIVE
RECEIVES TREATMENT WHICH IMPLIES NO SYSTEMATIC PERFORMANCE OF CHILDREN (E USES THE EXOGENEITY
DIFFERENCES BETWEEN THE TREATED AND UNTREATED BETWEEN THE TIMING OF LOCAL HOMICIDES AND THE
2ANDOMIZED #ONTROL 4RIALS 2#4S ARE THE MOST TIMING OF VOCABULARY AND READING ASSESSMENTS IN
COMMON TYPE OF EXPERIMENT )N 2#4S EACH UNIT #HICAGO SCHOOLS TO CRAFT A NATURAL EXPERIMENT THAT
OF STUDY IS RANDOMLY ASSIGNED TO EITHER A TREAT GIVES ESTIMATES OF THE SHORT TERM DECREASE IN TEST
MENT OR CONTROL GROUP 7HILE 2#4S ARE HIGHLY SCORES ASSOCIATED WITH EXPOSURE TO LOCAL HOMI
EFFECTIVE AT REACHING CAUSAL EFFECT ESTIMATES THEY CIDE (E lNDS SIGNIlCANT NEGATIVE EFFECTS ON TESTS
ARE OFTEN VERY INFEASIBLE IN LIFE COURSE RESEARCH ADMINISTERED UP TO  DAYS AFTER A LOCAL HOMICIDE
&OR MANY QUESTIONS CONCERNING LIFE PROCESSES AN 4HE ACUTE EFFECT WEAKENS OVER TIME AND BECOMES
EXPERIMENT MANDATING THAT INDIVIDUALS ENDURE ARBITRARILY CLOSE TO ZERO ON AFTER  WEEKS
CERTAIN TREATMENTS IS PRECLUDED FOR BOTH PRACTICAL
AND ETHICAL REASONS
1UASI EXPERIMENTS AIM TO ACHIEVE THE BENElTS 3 Causal Inference
OF A FULL EXPERIMENT IN A SETTING WHERE THE in the Presence
RESEARCHER DOES NOT HAVE FULL CONTROL OVER THE of Heterogeneity, Time-
TREATMENT ASSIGNMENT PROCESS /NE OF THE MOST Variation, and Mediation
WELL KNOWN EXPERIMENTS IN THE STRATIlCATION LIT
ERATURE IS THE -OVING TO /PPORTUNITY -4/ 3.1 Identifying Heterogeneous
EXPERIMENT 4HIS INTERVENTION RANDOMLY ASSIGNED Causal Effects
TWO TYPES OF HOUSING VOUCHERS TO LOW INCOME
FAMILIES IN ORDER TO ASSESS NEIGHBORHOOD EFFECTS )F THERE IS TREATMENT EFFECT HETEROGENEITY AVERAGE
ON FAMILY ECONOMIC HEALTH AND OTHER LIFE COURSE TREATMENT EFFECTS CAN VARY WIDELY DEPENDING ON
OUTCOMES /NE VOUCHER AFFORDED FAMILIES THE THE POPULATION COMPOSITION OF THE TREATED AND
OPTION OF MOVING TO ANOTHER LOW INCOME THUS SIMPLE AVERAGES DO NOT HAVE A STRAIGHTFOR
NEIGHBORHOOD AND THE OTHER OFFERED THE OPTION OF WARD INTERPRETATION )NDEED AN IMPORTANT DEVEL
MOVING WITHOUT CONSTRAINT &OR THE PURPOSE OF OPMENT OF THE CAUSAL INFERENCE LITERATURE IS THE
UNDERSTANDING THE EFFECTS OF GIVING FAMILIES THE RECOGNITION THAT TREATMENT EFFECTS ARE LIKELY TO BE
OPPORTUNITY TO RELOCATE -4/ WAS A RANDOMIZED HETEROGENEOUS !NGRIST AND +RUEGER  "RAND
CONTROL TRIAL 3OME RESEARCHERS INSTEAD WISHED TO AND 3IMON 4HOMAS  (ECKMAN ET AL ;
USE -4/ TO UNDERSTAND THE EFFECTS OF NEIGHBORHOOD 7INSHIP AND -ORGAN  8IE ET AL  8IE
522 R. Moore and J.E. Brand

  4HIS KIND OF HETEROGENEITY DOES NOT MERELY IN THE PRESENCE OF EFFECT HETEROGENEITY MAY BE
REmECT GROUP DIFFERENCES AT THE BASELINE THAT CAN INTERPRETED AS IDENTIFYING LOCAL AVERAGE TREATMENT
BE hCONTROLLED FORv BY COVARIATES IN REGRESSION OR EFFECTS LATE THOSE EFFECTS CORRESPONDING TO
MATCHING MODELS OR lXED EFFECTS 4HE RECOGNI SUBPOPULATIONS ON THE MARGIN OF TREATMENT PAR
TION THAT TREATMENT EFFECTS MAY VARY BY THE PROB TICIPATION INDUCED BY THE PARTICULAR INSTRUMENT
ABILITY OF TREATMENT BEYOND RESPONSE VARIATION UNDER CONSIDERATION
BY SELECTED COVARIATES LIKE GENDER OR RACE HAS LED
TO NEW METHODS OF CAUSAL INFERENCE AND TO RElNED
INTERPRETATIONS OF EFFECT ESTIMATES DERIVED FROM 3.2 Identifying Causal Effects
EXISTING METHODS "RAND AND 8IE  7INSHIP with Time-Varying Treatments
AND %LWERT  -ORGAN AND 4ODD ; and Time-Varying Outcomes
-ORGAN AND 7INSHIP  8IE  8IE ET AL
  $ESPITE WIDESPREAD BELIEF BY PRACTITIO ,IFE COURSE RESEARCH OFTEN INVOLVES THE ANALYSIS
NERS TRADITIONAL REGRESSION ESTIMATES DO NOT REP OF EFFECTS OF EVENTS THAT OCCUR OVER TIME WHICH
RESENT STRAIGHTFORWARD AVERAGES OF INDIVIDUAL LEVEL RAISES COMPLEX ISSUES IN THE ESTIMATION OF CAUSAL
CAUSAL EFFECTS IF INDIVIDUAL LEVEL VARIATION IN THE EFFECTS )NDIVIDUALS WHO EXPERIENCE THE EVENT OF
CAUSAL EFFECT OF INTEREST IS NOT RANDOM )NSTEAD INTEREST EARLY IN LIFE MAY DO SO FOR DIFFERENT
THEY GIVE A WEIGHTED AVERAGE OF THE HETEROGE REASONS THAN THOSE WHO EXPERIENCE THE EVENT
NEOUS INDIVIDUAL LEVEL EFFECTS WHERE POPULATION LATER 2ESEARCHERS MUST CAREFULLY ATTEND TO THE
COMPOSITION WEIGHTS CAN PRODUCE WIDELY DIFFER CONCEPTUAL AND THEORETICAL ISSUES UNDERLYING LIFE
ENT EFFECT ESTIMATES DEPENDING UPON WHO EXPERI COURSE TREATMENTS &OR EXAMPLE "RAND AND
ENCES TREATMENT 3IMON 4HOMAS  LOOK AT THE EFFECTS OF
2EGRESSION AND MATCHING MODELS CAN HOW MATERNAL JOB DISPLACEMENT ON EDUCATIONAL AND
EVER BE USED TO RECOVER SPECIlC SUBPOPULATION SOCIAL PSYCHOLOGICAL OUTCOMES OF CHILDREN USING
TREATMENT EFFECTS OF INTEREST INCLUDING THE TREAT PROPENSITY SCORE MATCHING #ORRECTING FOR SELEC
MENT EFFECT ON THE TREATED TT AND THE TREATED TION INTO DISPLACEMENT REQUIRES A REASONABLY
EFFECT ON THE UNTREATED TUT  ,ET US DElNE THE STRONG MODEL PREDICTING WHICH CHILDREN ARE MOST
AVERAGE DIFFERENCE AMONG THOSE INDIVIDUALS WHO LIKELY TO EXPERIENCE MATERNAL JOB LOSS -ODELS
WERE ACTUALLY TREATED THE TT THAT PREDICT THE LIKELIHOOD OF EXPERIENCING DIS
PLACEMENT USING ONLY COVARIATES AVAILABLE AT OR
G TT = E ( y1 y 0 | d = 1), BEFORE THE CHILDS BIRTH DO A MUCH BETTER JOB OF
PREDICTING DISPLACEMENT EVENTS THAT OCCUR IN EARLY
AND THE AVERAGE DIFFERENCE AMONG THOSE INDIVIDU CHILDHOOD RELATIVE TO DISPLACEMENTS EXPERIENCED
ALS WHO WERE NOT TREATED THE TUT IN MIDDLE CHILDHOOD AND ADOLESCENCE 4HIS ILLUS
TRATES SOME OF THE DIFlCULTY ONE MAY ENCOUNTER
G TUT = E ( y1 y 0 | d = 0). WHEN USING ONE MODEL TO EXPLAIN AN OCCURRENCE
THAT HAPPENS AT DIFFERENT TIMES FOR DIFFERENT REA
3TATISTICAL MODELING TO EXPLORE EMPIRICAL PATTERNS SONS (OWEVER "RAND AND 3IMON 4HOMAS )
OF EFFECT HETEROGENEITY AS A FUNCTION OF THE PRO ALSO PARTITION MATERNAL DISPLACEMENT INTO THREE
PENSITY SCORE HAVE ALSO BEEN EMPLOYED TO RECOVER PERIODS ACROSS CHILDHOOD AND ASSESS EFFECTS THAT
PATTERNS OF TREATMENT EFFECT HETEROGENEITY "RAND OCCUR AT VARYING POINTS IN YOUNG ADULTHOOD
AND 3IMON 4HOMAS  8IE ET AL   ADOPTING THE TIME VARYING CONCEPTUAL FRAMEWORK
3EVERAL RECENT STUDIES HAVE ADOPTED THIS APPROACH OF "RAND AND 8IE  WE DISCUSS BELOW
TO ADDRESS QUESTIONS INVOLVING HETEROGENEOUS "RAND AND 8IE  DISCUSS SOME OF THE
EFFECTS OF HIGHER EDUCATION ON A RANGE OF LIFE CONCEPTUAL CHALLENGES INVOLVED WITH ESTIMATING
COURSE OUTCOMES "RAND  "RAND AND $AVIS CAUSAL EFFECTS OF NON REPEATABLE AND NON
 "RAND ET AL  "RAND AND 8IE ; REVERSIBLE TREATMENTS THAT OCCUR AT DIFFERENT POINTS
-USICK ET AL   )NSTRUMENTAL VARIABLE MODELS IN TIME AND THAT AFFECT OUTCOMES THAT ARE
Causality in Life Course Studies 523

Table 1 4IME VARYING TREATMENT EXPOSURE AND OUTCOME MEASUREMENT


/UTCOME MEASUREMENT
4IME INVARIANT 4IME VARYING
4REATMENT EXPOSURE 4IME INVARIANT #ASE  #ASE 
i = yid = yid > i , v = yid, v=1 yid, v>1
4IME VARYING #ASE  #ASE 
d =t d >T
t,T
i =y
i y
i ivt T = yid v= t yid v>T

MEASURED AT DIFFERENT POINTS IN TIME 4ABLE  AGE  ON EARNINGS AT AGE  AND AGE  RELATIVE
SHOWS THE POSSIBLE COMBINATIONS OF TIME VARYING TO THE EFFECT OF JOB LOSS AT AGE  AND AGE  ON
TREATMENTS AND OUTCOMES 7E LIST THE EFFECT OF EARNINGS AT AGE  AND AGE  )N CASE  THE RELEVANT
INTEREST IN EACH CASE FOR SOME OUTCOME yi MEA COUNTERFACTUAL TO RECEIVING A TREATMENT IN PERIOD
SURED AT TIME v 7E OPERATE IN AN ENVIRONMENT  IS CLEARLY THE SUBSEQUENT OUTCOME THAT WOULD
WHERE NON REPEATABLE AND NON REVERSIBLE TIME OCCUR IN THE ABSENCE OF PERIOD  TREATMENT 4HE
VARYING TREATMENTS MAY OCCUR AT MULTIPLE TIMES SAME HOLDS FOR CASE  EXCEPT THAT WE CAN EMPLOY
BEFORE AND AFTER THE TREATMENT EVENT OF INTEREST )N MEASURES FROM DIFFERENT POST TREATMENT PERIODS
THE TABLE BELOW WE CARE TO STUDY THE TREATMENT AS OUR OUTCOMES OF INTEREST 4HE COUNTERFACTUAL
EVENTS THAT OCCUR AT TIME t BUT THERE MAY ALSO BE BECOMES SLIGHTLY MORE COMPLEX IN CASES  AND 
OBSERVATIONS WHO EXPERIENCE THE EVENT AT TIME t+. 7HEREAS CASES  AND  SEGMENT THE SAMPLE BASED
4ABLE  SHOWS FOUR POSSIBLE WAYS IN WHICH TIME ON SINGLE PERIOD RECEIPT CASES  AND  MUST COM
VARYING TREATMENTS AND OUTCOMES MAY INTERACT PARE A TREATED GROUP TO A COUNTERFACTUAL OF NOT
#ASE  CONSIDERS THE EFFECTS OF A TIME INVARIANT RECEIVING TREATMENT WITHIN THE DATAS OBSERVATION
TREATMENT ON A TIME INVARIANT OUTCOME 4HIS IS PERIOD AND ALL FUTURE PERIODS THROUGH OUTCOME
THE CLASSIC ASSUMED TWO PERIOD FRAMEWORK WHERE MEASUREMENT
TREATMENT OCCURS IN SOME PERIOD AND AN OUTCOME
IS REALIZED THEREAFTER #ASE  CONSIDER THE EFFECT
OF A TIME INVARIANT TREATMENT ON A TIME VARYING 3.3 Identifying Causal Effects
OUTCOME 4HIS WOULD BE USEFUL IF WE AIM TO ASSESS with Mediating Mechanisms
HOW AN EVENT AFFECTS AN OUTCOME MEASURED IN
REPEATED FUTURE PERIODS 4HESE TYPES OF ANALYSES )N SOCIAL RESEARCH AIMED AT CAUSAL INFERENCE OVER
HELP DETERMINE HOW TREATMENT AND TIME SINCE THE LIFE COURSE RESEARCHERS OFTEN AIM TO ASSESS
TREATMENT INTERACT IN THE DETERMINATION OF CAUSAL WHICH MEDIATING MECHANISMS TRANSMIT THE EFFECT
EFFECTS AND OFFER EVIDENCE AS TO WHETHER EFFECTS FROM TREATMENT TO OUTCOME 4OO OFTEN RESEARCHERS
GROW OR WEAKEN OVER TIME #ASE  CONSIDERS A
TIME VARYING TREATMENT AND TIME INVARIANT OUT 
3AMPLE AND CELL SIZE CONSIDERATION AFFECT THE FEASIBILITY
COME 4HIS MODEL CAN SHED FURTHER LIGHT ON HOW OF PRODUCING ESTIMATES #OMPARISONS ARE MADE BETWEEN A
PROCESSES UNFOLD OVER TIME &OR EXAMPLE ONE SUBGROUP THAT EXPERIENCE TREATMENT AT TIME t AND THE
COULD USE THIS APPROACH TO UNDERSTAND WHETHER SUBSAMPLE THAT HAS NOT EXPERIENCED TREATMENT UP TO TIME t.
4HIS IMPLIES THAT ALL INDIVIDUALS WHO EXPERIENCED TREAT
EXPERIENCING MARITAL DISRUPTION HAS DIFFERENT
MENT BEFORE TIME t ARE EXCLUDED FROM THE STATED EFFECT
IMPLICATIONS FOR A WOMENS EARNINGS AT AGE  ESTIMATE 7E THUS NEED ENOUGH INDIVIDUALS EXPERIENCING
DEPENDING UPON WHETHER SHES IN HER S OR S TREATMENT WITHIN THOSE PERIODS TO PRODUCE RELIABLE
WHEN THE DISRUPTION OCCURS #ASE  CONSIDERS ESTIMATES ! WIDER INTERVAL BENElTS FROM THE INCLUSION OF
MORE OBSERVATIONS WHICH CAN YIELD MORE PRECISE ESTI
TIME VARYING TREATMENT AND TIME VARYING OUT
MATES BUT MEANS THAT OUR PRE TREATMENT COVARIATES HAVE
COMES &OR EXAMPLE THIS SET UP COULD ASSESS THE POTENTIALLY REDUCED ABILITY TO PREDICT THE PROBABILITY OF
DIFFERENCES IN THE EFFECTS OF JOB LOSS AT AGE  AND TREATMENT EXPOSURE
524 R. Moore and J.E. Brand

SIMPLY ADD AN ADDITIONAL INTERMEDIARY VARIABLE TO DATA FOR EACH COHORT INCLUDE ROUGHLY  
THE MODEL AND THEN ASSESS THE DEGREE TO WHICH FAMILIES WHO HAVE OVER   HIGH SCHOOL AGED
THE EFFECT OF THE TREATMENT ON THE OUTCOME HAS CHILDREN %ACH FAMILY HAS BETWEEN ONE AND THREE
CHANGED IN RESPONSE TO THE INCLUSION OF THE ADDI CHILDREN 4HE DATA CONTAIN STANDARD DEMOGRAPHIC
TIONAL VARIABLE 4HIS TYPE OF ANALYSIS EVEN WITH CHARACTERISTICS FOR ALL CHILDREN AND FAMILIES
ATTENTION TO SELECTION INTO TREATMENT OFTEN FAILS TO &AMILY STRUCTURE CLASSIlCATIONS INCLUDE A CHILD
ATTEND TO THE CAUSAL PROCESS RELATING THE TREATMENT CHILDREN LIVING WITH BOTH PARENTS MARRIED UNMAR
TO THE MEDIATING VARIABLE OR THE MEDIATING VARI RIED SINGLE MOTHER OR LIVING WITH AN UNMARRIED
ABLE TO THE OUTCOME 7HEN RESEARCHERS AIM TO SINGLE MOTHER WITH A FATHERFATHER lGURE 3INGLE
ASSESS INDIRECT CAUSAL EFFECTS THEY SHOULD DEVOTE FATHER HEADED HOUSEHOLDS AND COMPLEX NON
THE SAME ATTENTION TO CAUSAL PROCESSES LINKING THE TRADITIONAL HOUSEHOLD STRUCTURES WERE EXCLUDED
MEDIATING MECHANISMS AS THEY DO TO THE PRIMARY 0ARENTS EDUCATION INDICATES WHETHER EACH PARENTS
TREATMENT OF INTEREST HIGHEST ACHIEVED EDUCATION IS LESS THAN A HIGH
$IRECTED ACYCLIC GRAPHS $!'S PROVIDE A SCHOOL DEGREE A HIGH SCHOOL DEGREE A COLLEGE
TRACTABLE FRAMEWORK FOR ASSESSING WHETHER A DEGREE OR A GRADUATEPROFESSIONAL DEGREE &AMILIES
MODEL IS IDENTIlED AND THE MECHANISMS THAT MAY LIVE IN ONE OF lVE NEIGHBORHOODS THAT DIFFER IN
MEDIATE EFFECTS OF INTEREST %LWERT  0EARL THEIR DEMOGRAPHIC CHARACTERISTICS AND RISK RATE OF
  #ONVENTIONS INCLUDE THAT AN ARROW INDI EXPOSURE TO THE TREATMENT OF INTEREST &OR EXAMPLE
CATES A DIRECT CAUSAL EFFECT d y WHILE A MISS RESPONDENTS IN SOME NEIGHBORHOODS HAVE A  
ING ARROW INDICATES NO CAUSAL EFFECTS d y  ! LINE CHANCE OF RANDOM EXPOSURE TO VIOLENCE WHEREAS
INDICATES TWO VARIABLES ARE ENDOGENOUSLY CORRE THE PROBABILITY OF RANDOM EXPOSURE IN OTHER
LATED WITH NO CAUSAL DIRECTION x n x  ! CAUSAL NEIGHBORHOODS IS AS HIGH AS  
PATH FROM d TO y CAN BE DEPICTED BY d y y 4HE DATA GENERATING PROCESS BEGINS BY
WHILE d y x y IS A NON CAUSAL PATH FROM SPECIFYING THE DEMOGRAPHIC TRAITS OF A FAMILY &OR
d TO y ! VARIABLE WITH TWO ARROWS ALONG THE PATH EACH FAMILY ONE OF FOUR RACES WAS RANDOMLY
POINTING TO IT IS A COLLIDER &OR EXAMPLE y IS A ASSIGNED TO YIELD A SAMPLE THAT IS APPROXIMATELY
COLLIDER ALONG d y x y 7E SUBJECT OUR   !SIAN   BLACK   (ISPANIC AND 
ESTIMATION TO ENDOGENOUS SELECTION BIAS WHEN  WHITE #ONDITIONAL UPON RACE A HOUSEHOLD
CONDITIONING ON A COLLIDER VARIABLE %NCODED INCOME VALUE IS CHOSEN FROM RACE SPECIlC DISTRI
WITHIN THE $!'S ARE RULES FOR MOVING FROM ASSO BUTIONS 4HE INCOME DISTRIBUTION FOR THE YEAR
CIATION TO CAUSATION #HAINS REPRESENT CAUSAL  COHORT WAS DElNED TO APPROXIMATE THE RACE
ASSOCIATIONS d y y FORKS REPRESENT CON SPECIlC 53 INCOME DISTRIBUTION REmECTED IN
FOUNDING d z y AND INVERTED FORKS REPRE -ARCH  #URRENT 0OPULATION 3URVEY DATA 4HE
SENT ENDOGENOUS SELECTION d y y). We S INCOME DISTRIBUTION IS A TRANSFORMATION OF
REPRESENT THE CAUSAL PROCESS WITH A $!' FOR OUR THE S DISTRIBUTION (OUSEHOLD STRUCTURE IS
EMPIRICAL EXAMPLE BELOW RANDOMLY ASSIGNED ACCORDING TO AN INCOME
QUINTILE SPECIlC DISTRIBUTION !VAILABLE STRUCTURES
INCLUDE  MARRIED PARENT HOUSEHOLD  SINGLE
4 Simulated Data MOTHER HEADED HOUSEHOLD  COHABITING PARENT
HOUSEHOLD )N BOTH COHORTS HOUSEHOLDS WITH
4.1 Description HIGHER INCOME ARE MORE LIKELY TO BE LED BY MARRIED
of the Simulated Data PARENTS WHILE HOUSEHOLDS WITH LOWER INCOMES
ARE MORE LIKELY TO HAVE TO HAVE A SINGLE HEAD OR
&OR THIS CHAPTER WE HAVE CREATED SIMULATED LONGI COHABITING PARENTS  &AMILIES FROM THE S
TUDINAL DATA THAT FOLLOWS ADOLESCENTS THROUGH  COHORT ARE MORE LIKELY TO BE MARRIED WHILE FAMI
YEARS OF SECONDARY EDUCATION 4HE DATA INCLUDE LIES IN THE S HAVE MORE SINGLE MOTHER LED
TWO COHORTS OF YOUNG PEOPLE WHO ENTER HIGH HOUSEHOLDS AND HOUSEHOLDS WITH COHABITING
SCHOOL TWO DECADES APART #OHORT ONE BEGINS IN PARENTS #ONDITIONAL UPON RACE AND HOUSEHOLD
 AND COHORT  BEGINS IN THE YEAR  4HE STRUCTURE EACH FAMILY IS RANDOMLY ASSIGNED
Causality in Life Course Studies 525

BETWEEN ONE AND THREE HIGH SCHOOL AGED CHILDREN GIVEN YEAR WHAT PROPORTION OF THEIR TIME WAS
0ARENTAL EDUCATION TAKES ONE OF FOUR VALUES LESS ALLOCATED TO WAGE LABOR AND THEIR WAGE RATE
THAN HIGH SCHOOL HIGH SCHOOL DEGREE COLLEGE
DEGREE AND GRADUATEPROFESSIONAL DEGREE AND
WAS ASSIGNED BASED ON INCOME DRAWS 4HIS 4.2 Treatment Specication
APPROACH CREATES AN EXPECTED CORRELATION BETWEEN and Mechanisms
PARENTS EDUCATION AND HOUSEHOLD INCOME AS WELL
AS EDUCATIONAL HOMOGAMY BETWEEN PARENTS &IVE 4HE TREATMENT OF INTEREST IS EXPOSURE TO NEIGHBOR
COMMUNITIES ARE DETERMINED THAT DIFFER IN TERMS HOOD GUN VIOLENCE %46 AND THE MAIN OUTCOME
OF RACIAL COMPOSITION AND PROBABILITY OF EXPOSURE OF INTEREST IS HIGH SCHOOL COMPLETION 2ECEIVING
TO VIOLENCE #OMMUNITY  MIRRORS THE SAMPLE THE hTREATMENTv IN THESE DATA IS EQUIVALENT TO
WIDE RACIAL COMPOSITION !LL OTHER COMMUNITIES ANSWERING THE FOLLOWING QUESTION AFlRMATIVELY
HAVE A PROPORTIONALLY DOMINANT RACIAL GROUP )N THE PREVIOUS SCHOOL YEAR HAVE YOU SEEN A
4HE CHARACTERISTICS OF FAMILIES WERE DESIGNED TO SHOOTING OR HAS ANYONE CLOSE TO YOU BEEN SHOT OR
SEE CORRELATIONS BETWEEN EDUCATION INCOME RACE TARGETED BY A GUNMAN !MONG THOSE WHO ANSWER
AND OTHER MEASURES THAT OBFUSCATE THE DIRECT THIS QUESTION AFlRMATIVELY IT IS UNKNOWN WHETHER
RELATIONSHIPS BETWEEN DEMOGRAPHIC MEASURES THE RESPONDENT WAS EXPOSED TO FATAL OR NON FATAL
AND AN OUTCOME OF INTEREST )N COLLECTED DATA WITH SHOOTING )T IS ALSO UNKNOWN WHETHER THE RESPON
UNOBSERVED COUNTERFACTUALS WE HAVE NO WAY OF DENT WAS EXPOSED TO MORE THAN ONE QUALIFYING
KNOWING THE EXTENT TO WHICH THESE CORRELATIONS EVENT )T IS HOWEVER KNOWN THAT FATAL SHOOTINGS
BIAS ESTIMATES OF INTEREST )N THESE SIMULATED DATA ARE RARE AND EXPOSURE TO MULTIPLE SHOOTINGS IN A
WE CAN OBSERVE COUNTERFACTUAL OUTCOMES AND GIVEN YEAR IS UNCOMMON 0ATTERNS OF GUN VIO
THUS CALCULATE THE DIFFERENCE BETWEEN REGRESSION LENCE ACROSS NEIGHBORHOODS HAVE CHANGED VERY
ESTIMATES AND ACTUAL COUNTERFACTUAL OUTCOMES LITTLE OVER THE  YEAR TIME PERIOD )NCIDENTS ARE
4HE CHARACTERISTICS DESCRIBED IN 4ABLE  IN THE RELATIVELY UNCOMMON IN NEIGHBORHOODS ONE AND
TWO SAMPLE PERIODS n AND n TWO WHILE NEIGHBORHOODS THREE THROUGH lVE TEND
HAVE NOTABLE DIFFERENCES &IRST CHILDREN ARE MORE TO EXPERIENCE VIOLENT OCCURRENCES MORE FREQUENTLY
LIKELY TO LIVE WITH TWO MARRIED PARENTS IN THE #OMMUNITY LEVEL GRADUATION RATES ARE NEGATIVELY
S WHILE MOTHER HEADED HOUSEHOLDS AND CORRELATED WITH INCIDENTS OF GUN VIOLENCE 7HILE
HOUSEHOLDS WITH COHABITING PARENTS ARE RELATIVELY THIS IS CONSISTENT WITH AN INVERSE RELATIONSHIP
MORE COMMON IN THE S -OTHERS HOWEVER BETWEEN EXPOSURE TO VIOLENCE %46 AND ACADEMIC
ARE SIGNIlCANTLY MORE LIKELY TO HAVE A GRADUATE OR SUCCESS THERE ARE A RANGE OF OTHER DIFFERENCES
PROFESSIONAL DEGREES IN THE S RELATIVE TO THE BETWEEN COMMUNITIES THAT COULD PARTIALLY OR FULLY
S !LL BUT ONE COMMUNITY BECAME LESS SEGRE EXPLAIN THE DIFFERENCES IN GRADUATION RATES 4HESE
GATED OVER THE  YEAR TIME PERIOD WITH THE LARG INCLUDE OBSERVABLE DIFFERENCES IN PARENTAL EDUCA
EST RACIAL GROUP REPRESENTING A SM ALLER PERCENTAGE TION HOUSEHOLD INCOME SCHOOL QUALITY AND RACIAL
OF THE POPULATION OF EACH COMMUNITY 7E HAVE COMPOSITION AND UNOBSERVABLE DIFFERENCES IN
ACADEMIC DATA FOR STUDENTS FOR  YEARS OF HIGH ABILITY PERSONAL VALUATIONS ON EDUCATION AND
SCHOOL -EASURES INCLUDE WHETHER THE STUDENT VARYING OPPORTUNITY COSTS 7E AIM TO ASSESS
ENROLLED IN HIGH SCHOOL DURING EACH YEAR HOW WHETHER EXPOSURE TO VIOLENCE HAS A CAUSAL EFFECT
MANY CREDITS WERE EARNED TOWARDS GRADUATION ON HIGH SCHOOL COMPLETION AND CREDITS EARNED
AND WHETHER THE STUDENT EARNED ENOUGH CREDITS TO TOWARDS GRADUATION
GRADUATE AFTER  YEARS 4HE DATA ALSO INCLUDES 4HE %46 PROCESS HAS MODEST DIFFERENCES
LABOR MARKET PARTICIPATION FOR STUDENTS INCLUDING BETWEEN THE EARLIER AND LATER COHORTS )N THE S
WHETHER STUDENTS WORKED IN THE LABOR MARKET IN A DATA EXPOSURE RATES ARE NEIGHBORHOOD DEPENDENT
BUT OTHERWISE RANDOM 2ESIDENTS OF EACH NEIGH

3EE !PPENDIX ! FOR FURTHER DETAILS ON THE CONSTRUCTED BORHOOD HAVE A lXED YEAR SPECIlC PROBABILITY OF
DATA COMMUNITY VIOLENCE EXPOSURE AND INDIVIDUAL
526 R. Moore and J.E. Brand

Table 2 $ESCRIPTIVE STATISTICS FOR SIMULATION DATA


 COHORT MEANPROPORTION  COHORT MEANPROPORTION
'ENDER
&EMALE  
-ALE  
2ACEETHNICITY
7HITE  
"LACK  
!SIAN  
(ISPANIC  
-OTHERS EDUCATION
,ESS THAN HIGH SCHOOL  
(IGH SCHOOL DEGREE  
#OLLEGE DEGREE  
'RADUATEPROFESSIONAL DEGREE  
&ATHERS EDUCATION
,ESS THAN HIGH SCHOOL  
(IGH SCHOOL DEGREE  
#OLLEGE DEGREE  
'RADUATEPROFESSIONAL DEGREE  
(OUSEHOLD INCOME
-EAN    
   
 n   
 n   
 n   
   
#OMMUNITY OF RESIDENTS
#OMMUNITY   
#OMMUNITY   
#OMMUNITY   
#OMMUNITY   
#OMMUNITY   
(OUSEHOLD STRUCTURE
"OTH PARENTS MARRIED  
-OTHER HEADED ((  
"OTH PARENTS COHABITING  
-ODEL OUTCOMES
0ERIODS EMPLOYED  
4OTAL WAGE LABOR OVER  PERIODS  
4OTAL ACADEMIC LABOR OVER  PERIODS  
&OUR PERIOD GRADUATION RATE  
4OTAL PERIODS ENROLLED IN SCHOOL  
.
2ESPONDENTS    
&AMILIES    
2ESPONDENTS WITH SIBLINGS    
&AMILIES WITH MULTIPLE KIDS    
Causality in Life Course Studies 527

CHARACTERISTICS PLAY NO PART IN DETERMINING WHO EXPERIENCED %46 EVENTS IN EXACTLY ONE OF THE 
EXPERIENCES THE EVENT OF INTEREST 4HE VIOLENCE YEAR INTERVALS /F THE S S SAMPLE 
EXPOSURE IN THE  COHORT IS SIMILAR EXCEPT    EXPERIENCED %46 EVENTS IN  OF THE 
INDIVIDUALS WITH A HIGH PREFERENCE FOR LEISURE ARE YEARS AND THE REMAINING PERCENTAGE EXPERIENCED
MORE LIKELY TO EXPERIENCE EXPOSURE 4HIS ADDS %46 EVENTS IN  OF THE  YEARS &IGURE 
ENDOGENEITY TO THE PROCESS GOVERNING THE RECEIPT SUMMARIZES THE DISTRIBUTION OF %46 EXPOSURE
OF TREATMENT AND THIS IS AN ISSUE WE OFTEN OBSERVE OVER SIMULATED HIGH SCHOOL YEARS )N BOTH SAM
IN SURVEY DATA PLES WE OBSERVE A SAW TOOTH PATTERN WHERE EXPO
3IMULATED RESPONDENTS HAVE PREFERENCES OVER SURE FALLS IN YEAR  RISES DRAMATICALLY IN YEAR 
CONSUMPTION AS INDICATED BY WAGE LABOR LEI AND THEN FALLS SLIGHTLY IN YEAR  /VERALL EXPOSURE
SURE AND HIGH SCHOOL COMPLETION )N EACH PERIOD RATES ARE UNIFORMLY HIGHER IN THE S DATA
THEY MAKE TIME ALLOCATION DECISIONS THAT DETER 4ABLE  DESCRIBES EXPOSURE PROBABILITIES ACROSS
MINE THEIR CONSUMPTION LEISURE AND ACADEMIC COMMUNITIES
CREDIT EARNED FOR THE PERIOD %XPOSURE TO VIOLENCE 4HE SIMULATED ENVIRONMENT ALLOWS THE ESTIMA
%46 IS A POTENTIALLY TRAUMATIC EXPERIENCE THAT TION OF TRUE EFFECTS BY COMPARING OBSERVED OUT
DEPRESSES PRODUCTIVITY 4HE LOSS OF PRODUCTIVITY COMES WITH OBSERVABLE COUNTERFACTUAL OUTCOMES
CAN LOWER AVAILABLE TIME TO THE POINT THAT SUPPLY &OR OUR SAMPLE OF SIMULATED RESPONDENTS WHO
ING ACADEMIC LABOR IS NO LONGER FEASIBLE OR NO WERE EXPOSED TO VIOLENCE WE CALCULATE THE TRUE
LONGER OPTIMAL %46 ALSO LOWERS THE PERCEIVED AVERAGE EFFECT ON AN OUTCOME AS
GAINS TO A HIGH SCHOOL DEGREE BY RAISING THE
yETV = yETV y ETV .
SUBJECTIVE PROBABILITY THAT ONE MAY NOT SURVIVE TO
REAP THE RETURNS TO ACADEMIC LABOR "OTH MECHA 7E ALSO CALCULATE THE AVERAGE PERCENT CHANGE IN
NISMS LEADING TO %46 EFFECTS CAN LOWER EDUCA THE OUTCOME AS
TIONAL ATTAINMENT
yETV y ETV
%yETV = .
y ETV
5 Empirical Example Effect
Estimates 7E EXAMINE FOUR OUTCOMES  PERIODS EMPLOYED
DURING HIGH SCHOOL  TOTAL TIME DEVOTED TO
5.1 True Causal Effects WAGE LABOR  TOTAL TIME DEVOTED TO ACADEMIC
LABOR AND  HIGH SCHOOL COMPLETION
)N THE S S DATA     OF THE 4ABLES  AND  SUMMARIZE THE CALCULATED
SIMULATED SAMPLE EXPERIENCED SOME EXPOSURE TO EFFECTS FOR BOTH COHORTS )N THE  COHORT DATA
VIOLENCE OVER THE  YEAR PERIOD /F THOSE EXPOSED WE OBSERVE NEGATIVE %46 EFFECTS ON ALL LISTED OUT
    OF THE S S SAMPLE COMES %46 LEADS TO A DECREASE IN YEARS OF WORK

Fig. 1 0ROPORTION OF SAMPLE


EXPERIENCING %46 BY YEAR 6.5
6
5.5
5
1980s Data
4.5
2000s Data
4
3.5
3
1 2 3 4
528 R. Moore and J.E. Brand

AND HOURS WORKED AS WELL AS A DECREASE IN TIME WORKS ALL  YEARS OF HIGH SCHOOL REGARDLESS OF
DEVOTED TO ACADEMIC LABOR AND A LOWER PROPORTION %46 4HIS YIELDS NO EFFECT ON EMPLOYMENT 7E
OF STUDENTS WHO GRADUATED IN  YEARS 7E SEE GEN DO HOWEVER SEE A DECREASE IN HOURS IMPLYING
ERALLY SIMILAR lNDINGS FOR THE S DATA WITH THE SAMPLE CONTINUES TO WORK BUT MAY SUPPLY
SOME DIFFERENCES &IRST THE FULL S COHORT FEWER HOURS OF WAGE LABOR GIVEN %46 EXPOSURE
2ELATIVE TO THE  COHORT %46 LEADS TO A MUCH
Table 3 %46 PROBABILITIES GREATER DECREASE IN ACADEMIC LABOR HOURS AND A
GREATER DECREASE IN GRADUATION RATES
S %46 PROBABILITIES BY COMMUNITY AND PERIOD
3IMULATED INDIVIDUALS COMPARE THE EXPECTED
#OMMUNITY
RETURN OF ENROLLING IN SCHOOL VERSUS NOT ENROLLING
    
IN SCHOOL EACH YEAR AND CHOOSE THE OPTION WITH
0ERIOD      
     
THE HIGHEST VALUE "ORDERLINE CASES WILL NOT ENROLL
   ;= ;= ;=
IF THEY EXPERIENCE %46 IN A GIVEN PERIOD BUT WILL
   ;= ;= ;= ENROLL OTHERWISE 2OUGHLY   OF THE SAMPLE
FALLS IN THE BORDERLINE REGION FOR THE S COHORT
S %46 PROBABILITIES BY COMMUNITY AND PERIOD WHILE ONLY   OF THE SAMPLE IS BORDERLINE FOR
#OMMUNITY THE S COHORT )N ADDITION TO THESE DISTRIBU
    
TIONAL DIFFERENCES THE S COHORT CONTAINS A
0ERIOD      
SELECTION ISSUE SUCH THAT SIMULATED AGENTS WHO
     
ARE ALREADY AT RISK OF DROPPING OUT ARE ROUGHLY
   ;= ;= ;=
  MORE LIKELY TO EXPERIENCE %46 4HESE
   ;= ;= ;=
DIFFERENCES LEAD TO THE NOTICEABLE DIFFERENCES IN
"RACKETS INDICATE PERIODS WHERE A POLICY WAS ENACTED THAT
LEAD TO A SHARP CHANGE IN EXPOSURE RATES EFFECTS BETWEEN COHORTS 4HE ANALYSIS SHOWS

Table 4 $ATA TRUE EFFECT CALCULATIONS S


0ERCENT CHANGE DUE
/BSERVED MEAN #OUNTERFACTUAL MEAN %46 EFFECT TO TREATMENT
4OTAL YEARS EMPLOYED    
DURING (3
4OTAL HOURS DEVOTED TO    
WAGE LABOR
4OTAL HOURS DEVOTED TO    
ACADEMIC LABOR
%NROLLMENT BY PERIOD    
'RADUATED FROM (3 IN    
 YEARS

Table 5 $ATA TRUE EFFECT CALCULATIONS S


0ERCENT CHANGE DUE
/BSERVED MEAN #OUNTERFACTUAL MEAN %46 EFFECT TO TREATMENT
4OTAL YEARS EMPLOYED    
DURING (3
4OTAL HOURS DEVOTED TO    
WAGE LABOR
4OTAL HOURS DEVOTED TO    
ACADEMIC LABOR
%NROLLMENT BY 0ERIOD    
'RADUATED FROM (3 IN    
 YEARS
Causality in Life Course Studies 529

DECREASES IN ALL OUTCOME MEASURES WITH THE COUNTERFACTUAL EFFECTS ON OBSERVABLES DIFFER
EXCEPTION OF PERIODS WORKED FOR THE S COHORT BETWEEN COHORTS AS A RESULT OF DIFFERENCES IN THE
"ELOW WE USE METHODS DISCUSSED ABOVE TO DISTRIBUTIONS OF LATENT AND DEMOGRAPHIC VARIABLES
ATTEMPT TO RECOVER THE SAME INFORMATION DIS 4ABLES  AND  SHOW CONTROLLED AND UNCONTROLLED
CLOSED BY THE CAUSAL ANALYSIS PRESENTED HERE REGRESSION COEFlCIENTS WHETHER THE TRUE EFFECT
FALLS WITHIN THE   CONlDENCE INTERVAL OF THE
EFFECT ESTIMATE AND WHETHER THE EFFECT ESTIMATE
5.2 Causal Effect Estimates HAS THE SAME SIGN AS THE TRUE EFFECT #ONTROLS
INCLUDE MOTHERS EDUCATION FATHERS EDUCATION
4HIS SECTION EMPLOYS METHODS DESCRIBED ABOVE TO RACE RESPONDENT WAGE OFFER HOUSEHOLD STRUCTURE
INFER THE CAUSAL EFFECTS OF %46 4HE TRUE PARENTAL INCOME AND COMMUNITY OF RESIDENCE
COUNTERFACTUAL EFFECTS ARE LISTED AND COMPARED TO 4HERE IS NO TEMPORAL VARIATION IN THE CONTROL
THE EFFECT ESTIMATES FROM VARIOUS METHODS 4RUE VARIABLES

Table 6 #OUNTERFACTUAL AND ESTIMATED EFFECTS S


4RUE %STIMATE
4RUE 5NCONTROLLED #ONTROLLED ESTIMATE HAS THE
COUNTERFACTUAL ESTIMATE 3TD ESTIMATE 3TD WITHIN  CORRECT
/UTCOME -ETHOD EFFECT ESTIMATE error) error)  #) SIGN
0ERIODS EMPLOYED 
/,3 REGRESSION   +
 
0ROPENSITY  +
SCORE MATCHING

4OTAL WAGE LABOR 
/,3 REGRESSION  
 
&IXED EFFECTS 

)NSTRUMENTAL  ***  ***
VARIABLES
 
0ROPENSITY n  ** +
SCORE MATCHING
n 
4OTAL ACADEMIC LABOR 
/,3 REGRESSION  ***  ***
 
&IXED EFFECTS  *** n
 n
)NSTRUMENTAL  ***  ***
VARIABLES
 
0ROPENSITY n  *** +
SCORE MATCHING
n 
%NROLLMENT BY PERIOD 
/,3 REGRESSION  ***  *** +
 
CONTINUED
530 R. Moore and J.E. Brand

Table 6 CONTINUED
4RUE %STIMATE
4RUE 5NCONTROLLED #ONTROLLED ESTIMATE HAS THE
COUNTERFACTUAL ESTIMATE 3TD ESTIMATE 3TD WITHIN  CORRECT
/UTCOME -ETHOD EFFECT ESTIMATE error) error)  #) SIGN
&IXED EFFECTS  *** n
 n
)NSTRUMENTAL  ***  ***
VARIABLES
 
0ROPENSITY n  ***
SCORE MATCHING
n 
'RADUATED IN  YEARS 
/,3 REGRESSION   ***
 
0ROBIT   ***
 
,OGIT   ***
 
)NSTRUMENTAL   +
VARIABLES
 
0ROPENSITY n  *** +
SCORE MATCHED
n 
Notes -ODELS INCLUDE THE FOLLOWING CONTROL VARIABLES MOTHERS EDUCATION FATHERS EDUCATION RACE FAMILY STRUCTURE
HOUSEHOLD INCOME COMMUNITY OF RESIDENCE AND SIMULATED RESPONDENT WAGE OFFERS !STERISKS AND DAGGERS FOR P VALUE
SIGNIlCANCE HAVE THE STANDARD MEANING P   

P   

P   
P    o

%STIMATES FROM THE EMPLOYED METHODS PROPENSITY SCORE MATCHING AND lXED EFFECTS
GENERALLY HAD THE SAME SIGN AS THE TRUE EFFECT 4HE ESTIMATES WERE MOST SUCCESSFUL IN APPROXIMATING
INSTRUMENTAL VARIABLES ESTIMATES WERE THE ONLY TRUE COUNTERFACTUAL EFFECTS "ASED ON THESE
EXCEPTION &OUR OF THE EIGHT INSTRUMENTAL ESTIMATES WE CONCLUDE THAT %46 LEADS TO NEGATIVE
VARIABLES ESTIMATES WHERE THE INSTRUMENT WAS AN SHOCKS TO THE EXAMINED OUTCOMES )N THE YEAR
EXOGENOUS POLICY ENACTED THAT LEAD TO A SHARP FOLLOWING A %46 SIMULATED RESPONDENTS ARE LESS
INCREASE OR DECREASE IN THE LIKELIHOOD OF EXPERI LIKELY TO ENROLL IN SCHOOL /VER THE COURSE OF THE
ENCING %46 WERE POSITIVE WHEN THE TRUE EFFECT HIGH SCHOOL CAREER %46 IS ASSOCIATED WITH LESS
WAS NEGATIVE 4HERE WERE MIXED RESULTS CONCERN OVERALL TIME DEVOTED TO ACADEMIC AND WAGE LABOR
ING THE SUCCESS WITH WHICH A   CONlDENCE AND A LOWER PROBABILITY OF COMPLETING HIGH SCHOOL
INTERVAL HELD THE TRUE EFFECT 2EGRESSION AND IN  YEARS 4HERE IS NO EFFECT OF %46 ON THE NUM
PROPENSITY SCORE MATCHING ESTIMATES APPEAR TO BE BER OF HIGH SCHOOL YEARS IN WHICH A SIMULATED
MOST SUCCESSFUL AT GENERATING A   CONlDENCE RESPONDENT WAS EMPLOYED IN THE S DATA BUT
INTERVAL THAT HELD TRUE EFFECT ESTIMATES 4HE POINT AN APPARENT NEGATIVE EFFECT IS PRESENT IN THE S
ESTIMATES FROM lXED EFFECTS ESTIMATES OFTEN SIMULATED DATA
APPEARED TO BE FAIRLY CLOSE TO THE TRUE PARAMETER
VALUES BUT IN THIS CASE THE SMALLER STANDARD ERRORS
FROM lXED EFFECTS ESTIMATION LEAD TO VERY SMALL 
&OR SIMPLICITY WE DO NOT CONSIDER HETEROGENEITY IN
CONlDENCE INTERVALS THAT OFTEN EXCLUDED THE TRUE EFFECTS HERE 7E ASSUME THAT EFFECTS ARE HOMOGENOUS )N
VALUES )N THIS SIMULATION INSTRUMENTAL VARIABLES APPLIED LIFE COURSE RESEARCH HOWEVER RESEARCHERS SHOULD
ROUTINELY QUESTION THE UNDERLYING HOMOGENEITY ASSUMP
ESTIMATES WERE THE LEAST ACCURATE WHILE REGRESSION TION "RAND AND 3IMON 4HOMAS ).
Table 7 #OUNTERFACTUAL AND ESTIMATED EFFECTS S
4RUE ESTIMATE
4RUE COUNTERFACTUAL 5NCONTROLLED #ONTROLLED ESTIMATE WITHIN   %STIMATE HAS THE
/UTCOME -ETHOD EFFECT ESTIMATE ESTIMATE 3TD ERROR 3TD ERROR #) CORRECT SIGN
0ERIODS EMPLOYED 
/,3 REGRESSION  **  ** +
 
0ROPENSITY SCORE  ** +
MATCHING
Causality in Life Course Studies


4OTAL WAGE LABOR 
/,3 REGRESSION  ***  ***
 
&IXED EFFECTS n  ***
n 
)NSTRUMENTAL VARIABLES  *  ** +
 
0ROPENSITY CORE n  ***
MATCHING
n 
4OTAL ACADEMIC LABOR 
/,3 REGRESSION  ***  ***
 
&IXED EFFECTS n  *** +
n 
)NSTRUMENTAL VARIABLES  ***  +
 
0ROPENSITY SCORE n  * +
MATCHING
n 
%NROLMENT BY PERIOD 
/,3 REGRESSION  ***  ***
 
531

CONTINUED
Table 7 CONTINUED
532

4RUE ESTIMATE
4RUE COUNTERFACTUAL 5NCONTROLLED #ONTROLLED ESTIMATE WITHIN   %STIMATE HAS THE
/UTCOME -ETHOD EFFECT ESTIMATE ESTIMATE 3TD ERROR 3TD ERROR #) CORRECT SIGN
&IXED EFFECTS n  ***
n 
)NSTRUMENTAL VARIABLES  ***  ***
 
0ROPENSITY SCORE n  *** +
MATCHING
n 
'RADUATED IN  YEARS 
/,3 REGRESSION  ***  ***
 
0ROBIT  *** 
 
,OGIT  *** 
 
)NSTRUMENTAL VARIABLE n 
n 
0ROPENSITY SCORE n  o
MATCHING
n 
Notes -ODELS INCLUDE THE FOLLOWING CONTROL VARIABLES MOTHERS EDUCATION FATHERS EDUCATION RACE FAMILY STRUCTURE HOUSEHOLD INCOME COMMUNITY OF RESIDENCE AND SIMULATED
RESPONDENT WAGE OFFERS !STERISKS AND DAGGERS FOR P VALUE SIGNIlCANCE HAVE THE STANDARD MEANING P   

P   

P   
P    o
R. Moore and J.E. Brand
Causality in Life Course Studies 533

)N APPLIED LIFE COURSE RESEARCH WE GENERALLY ENROLLING IN SCHOOL IN ANY PERIOD AND THUS A
DO NOT FULLY UNDERSTAND THE ROLE OF DIFFERENCES IN LOWER LIKELIHOOD OF GRADUATING IN FOUR PERIODS
THE DISTRIBUTION OF OBSERVABLE AND UNOBSERVABLE YEARS 4HE LOST TIME MAY HAVE THE INDIRECT EFFECT
VARIABLES AS WELL AS DIFFERENCES IN THE TREATMENT OF FORCING A SIMULATED RESPONDENT TO CHOOSE
SELECTION PROCESS OVER TIME 4HERE ARE NOTICEABLE BETWEEN EMPLOYMENT AND ENROLLMENT )N THIS
DIFFERENCES IN THE MAGNITUDE OF TRUE EFFECTS ON ALL SIMULATED EXAMPLE EFFECTS ARE MEDIATED BY
OUTCOMES BETWEEN COHORTS )T IS POSSIBLE THAT THE CONSTRAINTS AND LATENT MEASURES THAT ARE UNOB
DISTRIBUTION OF HOUSEHOLD INCOME AND RESPONDENT SERVABLE TO THE RESEARCHER &ULLY RECOVERING THESE
WAGE OFFERS WHICH DIFFER ACROSS THE TWO COHORTS MECHANISMS IS LIKELY IMPOSSIBLE WITH ONLY THE
ARE LARGELY RESPONSIBLE FOR DIFFERENCES IN THE OBSERVABLE MEASURES IN THE DATA )T IS ALSO OFTEN
EFFECT OF %46 ON PERIODS EMPLOYED /THER DIFFER THE CASE IN LIFE COURSE RESEARCH THAT MULTIPLE
ENCES BETWEEN THE COHORTS MAY BE RESPONSIBLE FOR MECHANISMS MAY WORK INDEPENDENTLY OR JOINTLY
THE DIFFERENCE IN ESTIMATES )N THE S DATA TO PRODUCE A NET EFFECT OF INTEREST 4HE CAUSAL
%46 IS A COMPLETELY RANDOM PROCESS THAT IS $!' PRESENTED IN &IG  DEPICTS THE EFFECT OF
INDEPENDENT OF INDIVIDUAL CHARACTERISTICS )N THE %46 ON THE PROBABILITY OF HIGH SCHOOL COMPLE
S DATA %46 IS SLIGHTLY MORE LIKELY TO HAPPEN TION 7E ENCLOSE UNOBSERVED VARIABLES IN OVALS
TO SIMULATED RESPONDENTS WHO HAVE A STRONGER OBSERVED VARIABLES IN ROUNDED BOXES AND
TASTE FOR LEISURE THAN CONSUMPTION IE i > i). TREATMENT AND MAIN OUTCOME IN RECTANGLES 4HERE
4HERE WAS ALSO AN OVERALL SHIFT IN UNOBSERVABLE IS AN ASSOCIATION BETWEEN THE PROBABILITY OF
PREFERENCE PARAMETERS BETWEEN THE PERIODS 4HE EXPERIENCING %46 AND THE LATENT VALUATION OF
S COHORT HAS A STRONGER PREFERENCE FOR LEISURE HIGH SCHOOL COMPLETION BOTH OF WHICH ARE COR
WHILE THE S COHORT HAS A HIGHER PREFERENCE RELATED WITH SOCIO DEMOGRAPHIC CHARACTERISTICS
FOR CONSUMPTION )N THIS MODEL A TASTE FOR CON #ONDITIONING ON SOCIO DEMOGRAPHIC CHARACTERIS
SUMPTION IS MORE CONDUCIVE TO CONTINUED ENROLL TICS IMPLIES CONDITIONAL INDEPENDENCE BETWEEN
MENT AND EVENTUAL GRADUATION 4HE UNOBSERVABLE %46 EXPOSURE AND LATENT VALUATIONS ON HIGH
SHIFT IN PERSONAL PREFERENCES AND THE SELECTION SCHOOL COMPLETION AND ALLOWS US TO IDENTIFY THE
MECHANISM IN THE S DATA MAY EXPLAIN A LARGE CAUSAL EFFECT OF %46 ON ENROLLMENT DECISIONS AND
PART OF THE DIFFERENCE IN SENSITIVITY OF THE EDUCA GRADUATION 4HIS TOTAL EFFECT INCLUDES THE EFFECTS
TIONAL TRACT TO %46 EXPOSURE BETWEEN COHORTS OF THE MEDIATING MECHANISMS )N THESE SIMULATED
DATA WE SPECULATE THAT THE MEDIATING MECHA
NISMS INCLUDE A SHORT TERM DROP IN PRODUCTIVITY
5.3 Mediating Mechanisms THAT RESULTS IN A DECREASE IN THE AMOUNT OF TIME
DEVOTED TO ACADEMIC AND WAGE LABOR AND A LOWER
7E OFTEN CARE ABOUT BOTH THE EFFECT OF A TREATMENT LIKELIHOOD OF ENROLLING IN SCHOOL IN THE COMING
AND THE MECHANISM MEDIATING THAT EFFECT 4HE ACADEMIC YEAR !S A RESULT OF THE PRODUCTIVITY
DATA WERE GENERATED SUCH THAT %46 LEADS TO A DROP AND DECREASED LIKELIHOOD OF ENROLLING YOUTH
LOWER EXPECTED RETURN TO EDUCATION OPERATIONAL FALL BEHIND IN SCHOOL AND MANY ARE UNABLE TO
IZED AS A CHANGE IN A LATENT VARIABLE DENOTING THE MAKE UP THE DIFFERENCE IN TIME TO GRADUATE IN THE
PERSONAL VALUATION PLACED ON GRADUATING AND A  YEAR WINDOW 4HIS LEADS TO A LOWER  YEAR
DECREASE IN THE AMOUNT OF PRODUCTIVE TIME AVAIL GRADUATION RATE AMONG THOSE WHO HAVE EXPERI
ABLE FOR ACADEMIC AND WAGE LABOR ! LOWER VALU ENCED %46 1UANTITATIVE RESEARCHERS ARE OFTEN
ATION ON GRADUATION IMPLIES A LOWER LIKELIHOOD OF LIMITED AS TO WHICH MEDIATING MECHANISMS CAN
BE EMPIRICALLY TESTED )F ALL VARIABLES BELOW WERE
OBSERVABLE WE COULD DECOMPOSE THE MECHA

7E DO NOT ASSESS HOW EXPOSURE TO THE TREATMENT OVER NISMS MEDIATING %46 EFFECTS BY DETERMINING
TIME PRODUCES VARIATION IN INDIVIDUAL EFFECTS )T IS LIKELY
THAT %46 THAT OCCURS EARLY IN CHILDHOOD RELATIVE TO MIDDLE
WHAT PROPORTION OF THE EFFECT WORKS THROUGH THE
CHILDHOOD OR ADOLESCENCE MIGHT INmUENCE THE EFFECTS ON DEVALUATION OF EDUCATION AND WHAT PORTION IS
OUR OUTCOMES OF INTEREST ATTRIBUTABLE TO A LOSS OF PRODUCTIVITY (OWEVER
534 R. Moore and J.E. Brand

Fig. 2 $!' CAUSAL MODEL OF EFFECTS OF %46 ON HIGH SCHOOL GRADUATION

BECAUSE THESE VALUES ARE UNOBSERVED WE DO NOT 4HE TREATMENT HAS TWO SIMULATED EFFECTS &IRST
HAVE THE DATA ONE WOULD NEED TO STUDY THESE IT DECREASES SUBJECTIVE ASSESSMENTS OF PERSONAL
MECHANISMS LONGEVITY WHICH DECREASES THE HORIZON OVER
WHICH ONE WOULD EXPECT TO REAP THE RETURNS TO
EDUCATION THEREFORE LOWERING THE OVERALL
5.4 Summary of the Empirical EXPECTED RETURN TO EDUCATION 3ECOND %46
Example SERVES AS A TRAUMATIC EXPERIENCE THAT DEPRESSES
PRODUCTIVITY IN THE SHORT RUN 4HIS IS OPERATIONAL
7E CONSTRUCT A PROCESS WHERE SIMULATED RESPON IZED AS A LOSS OF TIME THAT COULD HAVE OTHERWISE
DENTS HAVE UNOBSERVED PREFERENCES OVER LABOR BEEN ALLOCATED TO LEISURE ACADEMIC LABOR OR WAGE
LEISURE AND HIGH SCHOOL COMPLETION AND VARYING LABOR 4HE LOSS OF TIME AND DEVALUATION OF
LEVELS OF UNOBSERVED ACADEMIC ABILITY 3IMULATED EDUCATION TOGETHER MAKE THE CHOICE TO ENROLL IN
RESPONDENTS ARE CHARACTERIZED BY SOCIO SCHOOL LESS LIKELY WHICH LEADS TO A LOWER  YEAR
DEMOGRAPHIC CHARACTERISTICS SUCH AS NEIGHBOR SIMULATED GRADUATION RATE FOR THOSE WHO EXPERI
HOOD OF RESIDENCE RACIAL CLASSIlCATION ENCE %46 4HIS PROCESS WAS SIMULATED FOR TWO
HOUSEHOLD INCOME AND PARENTAL EDUCATION LEV COHORTS )N ONE COHORT ASSIGNMENT TO TREATMENT
ELS THAT ARE CORRELATED WITH THEIR PREFERENCES )N HAS AN ENDOGENOUS ASPECT WHERE INDIVIDUALS WHO
EACH OF THE SIMULATED PERIODS RESPONDENTS HAVE A STRONGER TASTE FOR LEISURE THAN CONSUMP
DECIDE WHETHER TO WORK IN THE LABOR MARKET FULL TION ARE A PERCENTAGE POINT MORE LIKELY TO EXPERI
TIME WORK AND GO TO SCHOOL OR GO TO SCHOOL ENCE TREATMENT ABOVE THE %46 EXPOSURE RATES
WITHOUT WORKING 3UPPLYING SUFlCIENT LEVELS OF
ACADEMIC LABOR OVER THE  YEAR PERIOD RESULTS IN 
3EE THE DESCRIPTION OF THE SIMULATION PROCESS IN !PPENDIX
GRADUATION /UR TREATMENT OF INTEREST EXPOSURE ! 3IMULATED RESPONDENTS FOR WHOM I>I PREFER HAVING
MORE CONSUMPTION MEASURED BY LABOR ACTIVITY AND
TO VIOLENCE %46 mAGS SIMULATED RESPONDENTS LESS LEISURE WHILE SIMULATED RESPONDENTS FOR WHOM THE
WHO HAVE SEEN A PERSON SHOT OR SHOT AT IN THE OPPOSITE IS TRUE IE I<I PREFER MORE LEISURE AND LESS
PREVIOUS YEAR CONSUMPTION
Causality in Life Course Studies 535

DElNED BY THEIR NEIGHBORHOOD OF RESIDENTS 4HIS ANALYSIS IN WHICH SOME METHODS OUTPERFORMED
YIELDS A SELECTION EFFECT WHERE INDIVIDUALS WHO OTHERS !LL TECHNIQUES ARE ACCOMPANIED BY A SET
ARE ALREADY MORE LIKELY TO DROPOUT ARE ALSO MORE OF ASSUMPTIONS AND THESE SIMULATED DATA MAY
LIKELY TO EXPERIENCE AN EVENT THAT FURTHER ENCOUR ACCOMMODATE THE ASSUMPTIONS OF SOME TECH
AGES EARLY SCHOOL EXIT 4HE COHORTS ALSO DIFFER NIQUES BETTER THAN OTHERS
ALONG THE DISTRIBUTION OF DEMOGRAPHIC VARIABLES
THAT GOVERN FAMILIAL CHARACTERISTICS
/UR ANALYSIS SUGGESTS THAT %46 LEADS TO A 6 Conclusion
DECREASE IN HIGH SCHOOL GRADUATION RATES 7E ALSO
lND THAT %46 IS ASSOCIATED WITH A DROP IN WAGE 4HIS CHAPTER HAS DISCUSSED SOME OF THE CHALLENGES
AND ACADEMIC LABOR SUPPLY WHICH LEAD TO THE AND STRATEGIES UNDERLYING LIFE COURSE RESEARCH
INFERENCE THAT TIME MAY BE ONE OF THE MEDIATING AIMED AT CAUSAL INFERENCE 2ELATIONSHIPS AMONG
FACTORS TRANSMITTING THE EFFECT OF %46 ON  YEAR VARIABLES IN EMPIRICAL DATA ALONE CANNOT ESTABLISH
HIGH SCHOOL COMPLETION 7E WERE UNABLE TO CAUSALITY #AUSAL INFERENCE REQUIRES SOME KNOWL
UNCOVER THE SECOND MECHANISM WHICH AFFECTS EDGE ABOUT THE DATA GENERATING PROCESS TO SUPPORT
GRADUATION DUE TO A DECREASE IN THE PERSONAL AN ASSUMPTION OF EXOGENOUS VARIABILITY IN OUR
VALUATION OF HIGH SCHOOL COMPLETION 4HE RESPON TREATMENT OF INTEREST 4HIS MAY BE CONTROLLED
DENTS SIMULATED VALUATION ON EDUCATION IS AN RANDOMIZATION OF THE ASSIGNMENT TO TREATMENT
UNOBSERVED VARIABLE AND FOR THAT REASON IT COULD EXOGENEITY OF A NATURAL TREATMENT ASSIGNMENT
NOT BE INCLUDED IN REGRESSION MODELS %STIMATES PROCESS OR CONDITIONAL INDEPENDENCE OF TREAT
FOR THE EFFECTS OF %46 DIFFERED GREATLY BETWEEN MENT ASSIGNMENTS AFTER CONTROLLING FOR SOME SET
COHORTS EVEN THOUGH THE CODED EFFECT OF %46 WAS OF OBSERVABLES 2ESEARCH AIMED AT CAUSAL INFER
THE SAME IN BOTH SETTINGS )N BOTH SIMULATIONS ENCE OFTEN INVOLVES TRYING TO UNDERSTAND COMPLEX
%46 LEAD TO A  POINT DECREASE IN THE VALUATION OF AND DYNAMIC PROCESSES THAT DEPEND UPON BOTH
EDUCATION AND A   DECREASE IN TIME AVAILABLE OBSERVED AND UNOBSERVED FACTORS 7E DISCUSS
TO ALLOCATE BETWEEN LABOR AND LEISURE 4HESE CHALLENGES TO EMPIRICAL ESTIMATION THAT ARISE DUE
DIRECT EFFECTS LEAD TO VERY DIFFERENT INDIRECT EFFECTS TO SELECTION EFFECTS 7E SHOW AND DISCUSS HOW
OF %46 ON LABOR SUPPLIES PER PERIOD ENROLMENT VIOLATIONS OF MODEL ASSUMPTIONS MAY LEAD TO BIAS
AND GRADUATION PROBABILITIES 7E CONCLUDE THAT IN PARAMETER ESTIMATES AND DISCUSS COMMON
DIFFERENCES IN THE DISTRIBUTIONS OF UNDERLYING ESTIMATION TECHNIQUES 4HESE INCLUDE ORDINARY
VARIABLES LEAD %46 TO HAVE VASTLY DIFFERENT EFFECTS LEAST SQUARES REGRESSION DISCRETE CHOICE MODELS
BETWEEN COHORTS PROPENSITY SCORE MATCHING MODELS lXED EFFECTS
4HIS EXERCISE OFFERS AN ENVIRONMENT WHERE WE MODELS AND INSTRUMENTAL VARIABLE MODELS -ANY
CAN ASSESS THE PERFORMANCE OF SOME OF OUR MOST OTHER POTENTIAL ESTIMATION PROCEDURES WERE OMITTED
COMMONLY USED CAUSAL INFERENCE EFFECT ESTIMATORS FROM THIS DISCUSSION FOR SIMPLICITY !DDITIONAL
AND COMPARE THE RESULTING ESTIMATES TO OBSERVED CHALLENGES MAY INCLUDE ESTIMATING EFFECTS WITH
COUNTERFACTUAL OUTCOMES 7E USE A DATA GENERAT COMPLEX COUNTERFACTUALS AND EVALUATING HETERO
ING PROCESS SEE !PPENDIX ! THAT INCLUDES LATENT GENEITY IN EFFECTS ,IFE COURSE RESEARCH ALSO OFTEN
VARIABLES CORRELATED COVARIATES AND ENDOGENOUS INVOLVES ASSESSING HOW EFFECTS VARY OVER TIME AS
SELECTION IN AN ATTEMPT TO CAPTURE THE CHALLENGES WELL AS ASSESSMENT OF MECHANISMS THAT INTERVENE
WE OFTEN FACE WITH OBSERVED DATA /VERALL THE AND HELP EXPLAIN ASSOCIATIONS BETWEEN TREATMENTS
SIMULATION SHOWS THAT THE ANALYTICAL TOOLS USED AND OUTCOMES
WERE SUFlCIENT FOR UNDERSTANDING THE KEY DYNAMICS 7E GENERATE SIMULATED DATA THAT MODELS A
BEHIND OUR DATA GENERATING PROCESS 3OME ESTI COMPLEX SOCIAL PROCESS OVER TIME AND EMPLOY
MATION TECHNIQUES PERFORMED BETTER THAN OTHERS THE DISCUSSED METHODS TO RECOVER THE DYNAMICS
4HIS SHOULD NOT BE INTERPRETED AS EVIDENCE OF AND EFFECTS GENERATED BY OUR CONSTRUCTED PROCESS
GENERAL SUPERIORITY OF SOME TECHNIQUES OVER OTHERS /UR SIMULATED PROCESS CONSIDERS THE EFFECT OF
)NSTEAD IT SHOULD BE UNDERSTOOD AS A PARTICULAR EXPOSURE TO GUN VIOLENCE ON THE OUTCOMES OF
536 R. Moore and J.E. Brand

SIMULATED RESPONDENTS /F THE METHODS EMPLOYED COMPLICATED DATA GENERATING PROCESS GOVERNED BY
/,3 REGRESSION AND PROPENSITY SCORE MATCHING NON LINEAR RELATIONSHIPS UNOBSERVED VARIABLES
MOST CONSISTENTLY PRODUCE RESULTS THAT WERE CLOSE UNKNOWN DISTRIBUTIONS AND AT TIMES ENDOGENE
TO THE TRUE EFFECTS CALCULATED THROUGH COUNTERFAC ITY 4HE FAVORABLE PERFORMANCE OF THESE ESTIMA
TUAL SIMULATION &IXED EFFECTS ESTIMATES WERE ALSO TORS IN THIS SIMULATED SETTING IS INDICATIVE THAT
REASONABLY ACCURATE )NSTRUMENTAL VARIABLES MOD THEY CAN PERFORM WELL TO UNDERSTAND COMPLEX LIFE
ELS HAD THE WEAKEST PERFORMANCE 7E DO NOT TAKE COURSE PROCESSES IN REAL DATA
THIS AS A CONDEMNATION OF )6 METHODOLOGY BUT
RATHER AS EVIDENCE OF SOMETHING WE ALREADY KNEW
THAT THE QUALITY OF AN )6 ESTIMATE DEPENDS HEAVILY Appendix A: Simulation Process
ON THE QUALITY OF THE )6 AND THAT IDENTIlCATION OF
A VALID )6 IS CHALLENGING 4HE )6 MEASURE 4HIS SECTION EXPLAINS THE MATHEMATICAL PROBLEM
EMPLOYED HERE MET THE NECESSARY ORTHOGONALITY SOLVED BY SIMULATED AGENTS IN THE CREATION OF THE
RESTRICTIONS BUT IT WAS AS IS MANY TIMES THE CASE TEST DATA !GENTS HAVE PREFERENCES OVER CONSUMP
A MEASURE THAT GENERATED VERY LITTLE VARIATION IN TION LEISURE AND HIGH SCHOOL COMPLETION 4HEY
THE TREATMENT VARIABLE MAKE CHOICES EACH PERIOD OVER THE ALLOCATION OF
)N OUR SIMULATION WE FOUND THAT SIMPLE LINEAR TIME TOWARDS EDUCATION LABOR OR LEISURE GIVEN
METHODS WERE USEFUL FOR UNDERSTANDING THE BASIC CONSTRAINTS ON TIME AVAILABILITY AND THE CONSUMP
DYNAMICS OF INTEREST IN A HIGHLY NON LINEAR MODEL TION BENElTS OF WAGE LABOR 4HE SEQUENCE OF
THAT INCLUDES A RANGE OF UNOBSERVED VARIABLES CHOICES DETERMINES WHETHER THEY GRADUATE IN FOUR
4HIS LENDS CONlDENCE IN OUR ABILITY TO USE THESE PERIODS 4HIS PROVIDES A NON LINEAR DATA GENERAT
TOOLS TO DISCERN THE EFFECTS OF COMPLEX REAL ING PROCESS WITH COMPUTABLE COUNTERFACTUAL
WORLD CAUSAL RELATIONSHIPS 4HIS EXERCISE MAY BE OUTCOMES WHERE WE CAN ASSESS OUR ABILITY TO
UNDERSTOOD AS A CASE WHERE OUR MOST COMMONLY MAKE CORRECT CAUSAL INFERENCE USING THE LINEAR
EMPLOYED LINEAR METHODS WERE SUFlCIENT FOR MODELS PRESENTED ABOVE
UNDERSTANDING THE KEY COMPONENTS BEHIND A &ORMALLY SIMULATED AGENT i SOLVE

c
( ) ( )
i
V lta,i , ltw,i , xt ,i , et ,i , ct ,i , ai = argmax ( xt ,i )i 1 lta,i ltw,i + Gi ,t 1 lta,i t ,i
{ }
lta,i , ltw,i , xt ,i ai 

subject to : xt ,i ltw,i wi + ei  NENT CAPTURING GAINS FROM GRADUATION #ONSTRAINT


( IS A BUDGET CONSTRAINT LIMITING PRESENT PERIOD
0 1 lta,i ltw,i  CONSUMPTION TO WHAT IS AFFORDABLE GIVEN ENDOW
MENT ei AND EARNED INCOME lt iwwi WHERE wi IS A
xt ,i 0, 0 ltw,i 1, 0 lta,i 1, > 0, > 0  WAGE RATE %NDOWMENTS ARE A SMALL BUT lXED
PERCENTAGE OF HOUSEHOLD INCOME WHILE WAGE
4HE OBJECTIVE FUNCTION  TAKES THE FORM OF A OFFERS ARE RANDOM Gi t IS THE TIME t VALUE THAT
#OBB $OUGLAS UTILITY FUNCTION EXPRESSING PREFER AGENT i PLACES ON EVENTUAL GRADUATION 4HE INDICA
ENCES OVER CONSUMPTION xt i AND LEISURE TOR FUNCTION SHOWS THAT AGENT i ONLY EXPECTS TO
1 lta,i ltw,i , PLUS AN ADDITIONAL ADDITIVE COMPO RECEIVE Gi t IF HE OR SHE SUPPLIES SOME MINIMUM
Causality in Life Course Studies 537

AMOUNT OF ACADEMIC LABOR #ONSTRAINTS  AND 4HE SOLUTION TO THIS PROBLEM TAKES THE FOLLOW
( STIPULATE THAT ALL TIME AND CONSUMPTION ALLO ING FORM GIVEN STATE st ,i = {et ,i , ct ,i , ai ,Gi ,t } :
CATIONS MUST BE NON NEGATIVE

c c c
t ,i if V (lta,i = 0, ltw,i*|l a , xt*,i |l a , st ,i ) V lta,i = t ,i , l w* a ct ,i , x* a ct ,i , st ,i and t ,i 1
ai t ,i =0 t ,i =0
ai t ,i |lt ,i = a t ,i |lt ,i = a ai
lta,i* = i i 

0 otherwise

wi i (1 lta,i* ) ei i
ltw,i* = 
wi ( i + i )

w (1 lta,i* ) + ei
xt*,i = i i 
i + i

4HE INTUITION BEHIND THIS SOLUTION IS AS FOLLOWS LABOR OR LEISURE SINCE i   AND i   )F IT TURNS
4HE RETURN FUNCTION FOR ACADEMIC LABOR IS A NON OUT THAT THE SIMULATED AGENT CANNOT FEASIBLY SUPPLY
DIFFERENTIABLE STEP FUNCTION AND REQUIRES SPECIAL THE DESIRED AMOUNT OF ACADEMIC LABOR SUCH THAT
CARE FOR THAT REASON 4HERE ARE ONLY THREE POSSIBLE
OPTIMAL VALUES FOR ACADEMIC LABOR &IRST ONE ct ,i
ct i >1
COULD SUPPLY UNITS OF ACADEMIC LABOR WHICH ai
ai
IS JUST ENOUGH TO RECEIVE THE EXPECTED RETURN G. OR THAT THE AGENT HAS A HIGHER PRESENT PERIOD GAIN
!NY TIME COMMITTED BEYOND THIS AMOUNT HAS NO IF SHE DEVOTES THE TIME TO WAGE LABOR OR LEISURE
RETURN AND WOULD BE BETTER SPENT TOWARDS WAGE SUCH THAT


(
V lta,i = 0, ltw,i*|l a
t ,i =0
, xt*,i |l a
t ,i = 0
, st , i )
> V


lt
a
, i =
ct ,i w*
, l c , x *
c
ai t ,i |lta,i = at ,i t ,i |lta,i = at ,i
i i
, st , i


THEN lt ,i = 0 MUST BE THE OPTIMAL ACADEMIC LABOR


a*
4HIS MODEL ADDS INTERESTING DYNAMICS TO THE
SUPPLY )N THIS CASE ANY TIME ALLOCATION ABOVE  DATA &IRST INDIVIDUAL ACTIONS ARE HIGHLY SENSITIVE
HAS A HIGHER RETURN AS WAGE LABOR TIME OR LEISURE TO SPECIlCATIONS OF i i AND Gt i ALL OF WHICH ARE
TIME &ROM HERE WE UTILIZE THE CONCAVITY CONTI UNOBSERVED BY THE RESEARCHER !LSO THESE PARAM
NUITY AND DIFFERENTIABILITY OF THE RETURN FUNCTION ETERS ARE CORRELATED WITH FAMILIAL AND COMMUNITY
IN [ ltw i xt i] TO SOLVE FOR  AND  IN TERMS OF CHARACTERISTICS %NDOWMENTS ei ARE A FUNCTION OF
PARAMETERS AND lta,i* . 4HESE FORMULAS ARE SUFlCIENT HOUSEHOLD INCOME 4HE DEPENDENCE OF THESE
FOR CALCULATING THE CURRENT PERIOD RETURN TO CON PARAMETERS ON FAMILY CHARACTERISTICS SHOULD LEAD
TINUING EDUCATION AND THE CURRENT PERIOD RETURN TO AN ESTIMABLE DEGREE OF INTERGENERATIONAL TRANS
TO DROPPING OUT 3IMULATED AGENTS CHOOSE THE MISSION OF ADVANTAGE
OPTION WITH THE HIGHEST PRESENT PERIOD RETURN
538 R. Moore and J.E. Brand

References THEORETICAL AND EMPIRICAL DEVELOPMENTS Annual


Review of Sociology, 32 n
%LWERT &   'RAPHICAL CAUSAL MODELS )N Handbook
!NGRIST * $  +RUEGER ! "   %MPIRICAL
of causal analysis for social research PP n 
STRATEGIES IN LABOR ECONOMICS )N Handbook of labor
.ETHERLANDS 3PRINGER
economics  PP n 
Goering, J., Feins, J. D., & Richardson, T. M. (2003). What
"LACK $ ! 3ANDERS 3  4AYLOR % * FORTHCOMING 
have we learned about housing mobility and poverty
4HE GREAT MIGRATION AND !FRICAN !MERICAN MORTALITY
deconcentration. In Choosing a better life? Evaluating
%VIDENCE FROM THE DEEP SOUTH American Economic
the moving to opportunity social experiment (pp. 3n36).
Review.
Washington, D.C.: Urban Institute Press.
"OLLEN +  "RAND * %   ! GENERAL PANEL MODEL
'REENE 7 (   Econometric analysis TH ED 
WITH RANDOM AND lXED EFFECTS ! STRUCTURAL EQUATIONS
5PPER 3ADDLE 2IVER 0RENTICE (ALL
APPROACH Social Forces, 89 n
(ECKMAN * * 5RZUA 3  6YTLACIL %  
"RAND * %   #IVIC RETURNS TO HIGHER EDUCATION !
5NDERSTANDING INSTRUMENTAL VARIABLES IN MODELS WITH
NOTE ON HETEROGENEOUS EFFECTS Social Forces, 89
ESSENTIAL HETEROGENEITY The Review of Economics and
n
Statistics, 88 n
"RAND * %  $AVIS $   4HE IMPACT OF COLLEGE
(UNGERFORD 4  3OLON '   3HEEPSKIN EFFECTS IN
EDUCATION ON FERTILITY %VIDENCE FOR HETEROGENEOUS
THE RETURNS TO EDUCATION The Review of Economics and
EFFECTS Demography, 48 n
Statistics, 69 n
"RAND * %  3IMON 4HOMAS *   #AUSAL EFFECT
,EUVEN %  3IANESI "   03-!4#( 3TATA
HETEROGENEITY )N Handbook of causal analysis for
MODULE TO PERFORM FULL -AHALANOBIS AND PROPENSITY
social research PP n  .ETHERLANDS 3PRINGER
SCORE MATCHING COMMON SUPPORT GRAPHING AND
"RAND * %  4HOMAS * 3   *OB DISPLACEMENT
COVARIATE IMBALANCE TESTING Statistical Software
AMONG SINGLE MOTHERS %FFECTS ON CHILDRENS OUTCOMES
Components.
IN YOUNG ADULTHOOD American Journal of Sociology,
,LERAS -UNEY !   4HE RELATIONSHIP BETWEEN EDU
119 n
CATION AND ADULT MORTALITY IN THE 5NITED 3TATES The
Brand, J. E., & Halaby, C. N. (2006). Regression and
Review of Economic Studies, 72 n
matching estimates of the effects of elite college attend-
,UDWIG * ET AL   7HAT CAN WE LEARN ABOUT NEIGHBOR
ance on educational and career achievement. Social
HOOD EFFECTS FROM THE MOVING TO OPPORTUNITY EXPERI
Science Research, 35(3), 749n770.
MENT American Journal of Sociology, 114 n
"RAND * %  8IE 9   )DENTIlCATION AND ESTIMA
-ORGAN 3  (ARDING $   -ATCHING ESTIMATORS
TION OF CAUSAL EFFECTS WITH TIME VARYING TREATMENTS AND
OF CAUSAL EFFECTS 0ROSPECTS AND PITFALLS IN THEORY AND
TIME VARYING OUTCOMES Sociological Methodology,
PRACTICE Sociological Methods and Research, 35
37 n
n
"RAND * %  8IE 9   7HO BENElTS MOST FROM
-ORGAN 3 ,  4ODD * *   ! DIAGNOSTIC ROUTINE
COLLEGE %VIDENCE FOR NEGATIVE SELECTION IN HETEROGE
FOR THE DETECTION OF CONSEQUENTIAL HETEROGENEITY OF
NEOUS ECONOMIC RETURNS TO HIGHER EDUCATION American
CAUSAL EFFECTS Sociological Methodology, 38
Sociological Review, 75 n
n
"RAND * % 0FEFFER & 4  'OLDRICK 2AB 3   4HE
-ORGAN 3  7INSHIP #   "RINING CONTEXT AND
COMMUNITY COLLEGE EFFECT REVISITED 4HE IMPORTANCE OF
VARIABILITY BACK INTO CAUSAL ANALYSIS )N Oxford hand-
ATTENDING TO HETEROGENEITY AND COMPLEX COUNTERFACTU
book of the philosophy of the social sciences
ALS Sociological Science, 1 n
PP n 
"UDIG  %NGLAND   4HE WAGE PENALTY OF MOTHER
-USICK + "RAND * %  $AVIS $   6ARIATION IN
HOOD American Sociological Review, 66 n
THE RELATIONSHIP BETWEEN EDUCATION AND MARRIAGE
#ALIENDO -  +OPEINIG 3   3OME PRACTICAL
-ARRIAGE MARKET MISMATCH Journal of Marriage and
GUIDANCE FOR THE IMPLEMENTATION OF PROPENSITY
Family, 74 n
SCORE MATCHING Journal of Economic Surveys, 22
0EARL *   Causality: Models reasoning, and infer-
n
ence #AMBRIDGE #AMBRIDGE 5NIVERSITY 0RESS
#AMERON ! #  4RIVEDI 0 +  
2OSENBAUM 0 2  2UBIN $ "   4HE CENTRAL ROLE
Microeconometrics: Methods and applications.
OF THE PROPENSITY SCORE IN OBSERVATIONAL STUDIES FOR
.EW 9ORK #AMBRIDGE 5NIVERSITY 0RESS
CAUSAL EFFECTS Biometrika, 70 n
#URRIE *  4HOMAS $   $OES HEAD START MAKE A
2OSENBAUM 0 2  2UBIN $ "   #ONSTRUCTING A
DIFFERENCE American Economic Review, 85
CONTROL GROUP USING MULTIVARIATE MATCHED SAMPLING
n
METHODS THAT INCORPORATE THE PROPENSITY SCORE The
$I0RETE 4 !  %IRICH ' -   #UMULATIVE ADVAN
American Statistician, 39 n
TAGE AS A MECHANISM FOR INEQUALITY ! REVIEW OF
Causality in Life Course Studies 539

3ANBONMATSU , ET AL   Moving to opportunity for 7OOLDRIDGE * -   Econometric analysis of cross sec-
fair housing demonstration program Final impacts tion and panel data ND ED  #AMBRIDGE -! -)4 0RESS
evaluation 7ASHINGTON $# 53 $EPARTMENT OF 7OOLDRIDGE *   Introductory econometrics: A
(OUSING AND 5RBAN $EVELOPMENT /FlCE OF 0OLICY modern approach -ASON #ENGAGE ,EARNING
$EVELOPMENT AND 2ESEARCH Xie, Y. (2011). Population heterogeneity and causal infer-
3HARKEY 0   4HE ACUTE EFFECT OF LOCAL HOMICIDES ON ence (pp. 11731). University of Michigan Population
CHILDRENS COGNITIVE PERFORMANCE Proceedings of the Studies Center Research Report.
National Academy of Sciences, 107 n 8IE 9   Population heterogeneity and causal infer-
7INSHIP #  %LWERT &   %FFECT HETEROGENEITY AND ence PP n  5NIVERSITY OF -ICHIGAN 0OPULATION
BIAS IN MAIN EFFECTS ONLY REGRESSION MODELS )N 3TUDIES #ENTER 2ESEARCH 2EPORT
Heuristics, probability and causality: A tribute to Xie, Y. (2013). Population heterogeneity and causal infer-
Judea Pear PP n  ,ONDON #OLLEGE 0UBLICATIONS ence. Proceedings of the National Academy of Sciences,
7INSHIP #  -ORGAN 3   "RINGING CONTEXT AND 110(16), 6262n6268.
VARIABILITY BACK IN TO CAUSAL ANALYSIS )N ( +INCAID 9U 8 "RAND * %  *ANN "   %STIMATING HETERO
%D Oxford handbook of the philosophy of the social GENEOUS TREATMENT EFFECTS WITH OBSERVATIONAL DATA
sciences /XFORD /XFORD 5NIVERSITY 0RESS Sociological Methodology, 42 n
The Logic and Practice of Growth
Curve Analysis: Modeling
Strategies for Life Course
Dynamics

Ross Macmillan and Frank Furstenberg

1 Introduction Following early attempts to discover and


devise unique methods for describing and ana-
A life course perspective is particularly attentive lyzing the timing of life course events (Elder
to the dynamics of time and how time is critical 1999; Modell et al. 1976), recent decades have
to understanding the meaning of events, experi- seen the development of several methodological
ences, and attainments over the life span. Its roots strategies that focused explicitly on time. In
can be traced to the path breaking work of demography, life tables were elaborated in a
W.I. Thomas and Znanieckis (19181920) number of important ways (Coale and Demeny
The Polish Peasant and C. Wright Mills (1959) 1966). In economics, there were signicant
The Sociological Imagination. A life course per- advances in time-series analysis (Lutkepohl
spective is increasingly the dominant framework 2007). In epidemiology and biostatistics, event
for examining the temporal dynamics of human history models became increasingly sophisti-
lives: age/aging, period/history, and cohort cated in their ability to capture time-to-failure
(unique exposure to historical time), and how (Petersen 1993). And in psychology, criminol-
these produce unique developmental trajectories ogy, and educational studies, there were
for individuals and populations (Elder 1998). signicant developments in the modeling of
Indeed, the key concepts of a life course perspec- within-person change on scalar outcomes
tive including lives in historical times, the tim- (Meredith and Tisak 1990; Muthen and Muthen
ing of lives, linked lives, and human agency all 2000; Nagin 2009; Singer and Willett 2003).
reect an underlying concern with time and tem- This chapter is organized around the intersec-
porality in the shaping of human experience tion of one such technique and life course
(Elder 1994; Elder et al. 2003). conceptions of time and timing. Specically, it
considers and elaborates the expression of key
life course concepts in statistical terms and trans-
lates them into an approach that has become
R. Macmillan (*) popular in many of the social sciences growth
Department of Policy Analysis & Public curve analysis. Growth curve analyses have par-
Management, Bocconi University, Milano, Italy
ticular afnity with a life course perspective:
e-mail: ross.macmillan@unibocconi.it
both emphasize stability and change in the
F. Furstenberg
unfolding life course with the aim of identifying
Department of Sociology, University of Pennsylvania,
Philadelphia, PA, USA the origins of developmental pathways and events
e-mail: fff@sas.upenn.edu or conditions that alter probabilistic routes. This

Springer International Publishing Switzerland 2016 541


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_24
542 R. Macmillan and F. Furstenberg

chapter has ve sections. Section 2 discusses a including excess weight. Also known as the
prime example of the utility of growth curve Quetelet index, the simple formula is weight in
analysis, specically developmental trends in kilograms divided by height in meters squared
body mass index (BMI) that operationalize the (or weight in pounds divided height in inches
risk of obesity and consequent health problems squared multiplied by 703). In the United States
across the life course (Harris 2010). We use tra- for example, general trends in body mass went
jectories of BMI to illustrate how core life course virtually unchanged from 1960 to 1980 with 11
principles can be operationalized and tested. % deemed to be obese. In recent decades how-
Section 3 identies four key principles of a life ever, rates of obesity increased markedly, to 23 %
course perspective and describes theoretical through 1990 and to just under 30 % by 2000 (US
expectations within a growth curve analysis con- Department of Health and Human Services
text. Section 4 describes the underlying model in 2011). In 2014, it is estimated that almost 35 % of
formal statistical terms with each conceptual the adult population in the US is obese (Ogden
elaboration associated with a statistical elabora- et al. 2014). Ng and colleagues (2014) report
tion. Section 5 uses data from the National similar increases in overweight and obesity across
Longitudinal Survey of Youth 97 to show the a number of countries over the last 30 years.
practical application of life course concepts Obesity as a social problem reects its asso-
within growth curve analysis of BMI trajectories ciation with a wide array of health limitations
in the transition to adulthood. Section 6 con- (Khaodhiar et al. 1999). Obesity is associated
cludes with discussion of issues of estimation with increased risk of metabolic disorders,
and parameterization, as well as evaluation of including insulin resistance and type 2 diabetes
effect sizes and statistical signicance. mellitus. There is also increased risk of cardio-
vascular problems, such as hypertension, dyslip-
idemia, cardiovascular disease, and stroke.
2 Body Mass Index, Obesity, Certain cancers are also associated with obesity,
and Life Chances including colorectal and prostate cancer in men
in American Society and endometrial, breast, and gallbladder cancer
in women. Some evidence suggests that obesity
The latter part of the twentieth century saw a is also associated with increased rates of psycho-
striking change in the weight of populations logical distress, perhaps given the associated
(Stearns 2002). Beginning rst in wealthy north- stigma of being overweight (Dong et al. 2004).
ern hemisphere countries but rapidly diffusing Given the range of comorbidities, there is evi-
across the globe (Swinburn et al. 2011), popula- dence that excess body weight is also associated
tions started getting heavier. Why this happened with substantial increases in early mortality
is the subject of much debate, but the actual phys- (Masters et al. 2013). In short, understanding the
iology of weight gain is well understood and developmental features of obesity is critical to
quite simple. Weight gain occurs when the intake contemporary public health and a life course per-
of calories through food consumption outpaces spective offers a uniquely powerful framework
the burning of calories through activity. A com- for understanding its epidemiology.
plicating factor is recent evidence of a fat gene,
also known as FTO, which inuences the rate at
which calories are transformed into fat and is 3 Life Course Concepts
variable across individuals and groups. Yet, even and Implications for Growth
with this complicating factor, the central mecha- Curve Analysis of BMI
nism of weight gain is a caloric imbalance.
Particularly for large sample or population Although there are numerous life course con-
comparisons, researchers use the body mass cepts that are relevant to an analysis of develop-
index (BMI) as an indicator of weight status, ment (see discussion in Shanahan and Macmillan
The Logic and Practice of Growth Curve Analysis: Modeling Strategies for Life Course Dynamics 543

2008), we highlight four that have unique rele- lar how schema and resources are implicated in
vance to both elaborating the value of a growth caloric imbalance.
curve analytic strategy and shedding light on the On the resource side of things, race and social
developmental structure of body mass. These class, for example, imply variation in economic
include (1) social structure and developmental resources that might impact upon ones ability to
processes, (2) longitudinal associations, (3) tra- nd and purchase healthier foods. Here, one
jectories, transitions, and turning points, and (4) could imagine that the choice set of foods is sim-
accumulation, accentuation, and leveling over ply narrower when one lacks resources and those
the life course. with less would be more oriented to low cost,
higher calorie foods, even when alternatives are
available. Yet, alternatives may not actually be
3.1 Social Structure available. Nutritional epidemiologists talk of
and Development food deserts that describe particular neighbor-
hoods or geographies, typically lower income,
A central, organizing principle of life course the- where available foods are typically fast food
ory and research is the idea that developmental and pre-prepared, packaged foods that have high
processes are situated within and shaped by sugar contents and where natural or unprocessed
social structure. Although notoriously difcult to foods are rare or difcult to nd (Cummins and
dene, Sewell (1992) denes social structures as Macintyre 1999). Similarly, there is an increasing
the manifestations of interconnected schema need to buy exercise in the form of clubs and
(cultural blueprints, values, and norms) and gym memberships given the increasingly seden-
resources (access to material goods, social capi- tary character of occupations. Such costs may be
tal, and power resulting from social position). In outside the option set for low-income people or
a life course perspective, social structures com- may not even be available in lower income
monly index positions in status hierarchies, eco- communities.
logical context, and historical period (Elder 1994; On the schema side, one could think of a num-
Elder and Rockwell 1979). Human development ber of weight related models that might come
over time depends in large measure upon the into play. In the realm of diet, there would likely
unique congurations of social structures that are be variation in what constitutes good and bad
rooted in the social position of family and child- food and some people, by virtue of biography
hood origins as well as social conditions occur- and group history, might be more oriented
ring over the passage of time as reected in the towards bad foods. A similar principle could
aging of individuals (Elder 1998). apply to the perceived value or utility of exercise.
Using this conceptual underpinning, BMI can Anthropologists have also drawn attention to cul-
be viewed as a probabilistic variation in an indi- tural variation in ideal body types with some cul-
viduals starting point at birth (or later on) and tures expressing a preference for heavier types.
changes with advancing age. Both quantities, Thus, one could further think that reference
including the key developmental aspect of change points for optimal weight would be inuenced by
over time, can be understood in reference to those in ones immediate environment and this
social structures that denote positions in status would connote different ideas about body shape
hierarchies (e.g., race, class, gender), ecological and weight. In short, there are a variety of ways
context (i.e., neighborhood characteristics, in which socio-structural position could connect
regional location, country), and historical context to the proximal mechanisms of weight gain and
(i.e., year of birth, period). In recognizing that ultimately produce variable population weight
social structural location implies variation in and variation in trajectories of weight and weight
both schema and resources, one could think about gain over the life course.
how this might inuence mechanisms in the Given this framework, one could focus on race
dynamics of BMI over the life course, in particu- as a particularly unique social position that brings
544 R. Macmillan and F. Furstenberg

together many of these themes and that has a a sample of Boston children to show continuity
unique relationship to excess weight. In particu- and change in life circumstances over several
lar, African-Americans in the US have relatively decades. Numerous factors, associated with family
high levels of poverty that are uniquely concen- disadvantage in the childhood, were either durable
trated and coupled with high levels of segregation predictors of later adult outcomes or were impli-
(Massey and Denton 1993; Wilson 1987). As cated in well-dened causal pathways that pro-
such, one could expect a profound concentration duced such outcomes. Elaborating such themes
of risks and indeed there is a rich literature that further, Conger et al. (1992) studied farm families
shows both race differences in BMI and impor- under stress from economic decline. The key
tant life course contingencies in such differences. theme in such work is an emphasis on the fact that
With respect to race and sex, Cossrow and human lives do not re-write themselves at each
Falkner (2004) report that the proportion obese successive stage and instead that the conditions of
among adult women in the US is 30 % for non- the past have formative effects on possible futures.
Hispanic Whites, 49.7 % for African Americans, Indeed, the central organizing feature of research
and 39.7 % for Hispanics. Smaller differences are on social mobility within- and across-generations
found for adult males, 27 %, 28.1 %, and 28.9 %. is the relationship between status origins, typically
Still, there appear to be important life course con- family socioeconomic status, and status destina-
tingencies at work in that race differences among tions, typically achieved occupational status
teenage and childhood males are stark (12.8 %, (Hauser and Warren 1997; see also, chapter Three
20.7 % and 27.5 % for Non-Hispanic whites, Generation Studies: Methodological Challenges
African Americans, and Hispanics male children, and Problems by Thornberry, this volume).
respectively and 12 %, 17 %, and 27.3 % for Studies of the precursors to obesity in adult-
Non-Hispanic whites, African Americans, and hood have focused considerable attention on
Hispanics male teens, respectively). Even larger childhood risk factors. One particularly well-
race differences among teenagers and children developed area of research shows that body
are found among females. As such information is mass in childhood and adolescence is strong
based on repeated cross-sectional analysis, we predictor of risk of being overweight in mid
can further explore such issues and illustrate the adulthood (e.g., Guo and Chumlea 1999). In
techniques and value of growth curve analysis addition, important elements of social structure
with specic identication to variation in are implicated in weight gain in later life.
between- and within-person change. Parsons and colleagues (1999) for example
show that socioeconomic status of ones family
of origin is consistently correlated with adult
3.2 Longitudinal Associations obesity, but does not have a strong association
with excess weight in childhood (Parsons et al.
A second key principle of a life course perspective 1999). Such sleeper effects are increasingly
is the idea that events or experiences, even dispa- key aspects of theoretical understanding (Cook
rate ones, are linked over the life span and that et al. 1979). Still other research focuses more
linked experiences often have important long-term on behavioral aspects of childhood and their
outcomes (McLeod and Almazan 2003). Here, consequences. Reilly and colleagues (2005),
much work focuses on links that both cross life for example, show that excessive hours spent
stages and that represent either causal or structural watching television in early childhood is asso-
relationships. Not surprisingly much of this work ciated with increased risk of childhood obesity.
focuses on links between childhood and subse- Importantly however, there is not a lot of lon-
quent life stages. In one particularly inuential ger-term longitudinal work that links childhood
body of work, Sampson and Laub (1995), Laub behaviors and circumstances to obesity risk in
and Sampson (2009) used reconstructed data from adolescence and adulthood.
The Logic and Practice of Growth Curve Analysis: Modeling Strategies for Life Course Dynamics 545

3.3 Trajectories, Transitions, been increasing with successive age cohorts in


and Turning Points the U.S.
Trajectories and transitions can combine at
Whether and how much past circumstances and particular moments in time to create turning
behaviors shapes future patterns of behavior (such points in the life course when former paths of
as obesity) is relevant to another set of concepts in experience or behavior are abruptly truncated and
a life course perspective: trajectories, transitions, new paths are developed (Laub and Sampson
and turning points. In the earliest formulations, 1993). A very important line of life course
the concept of trajectories referred to temporally research focuses on how trajectories that are redi-
dened patterns of behavior, particularly social rected by non-normative events or, in some cases,
roles (Elder 1985). Examples abound in social unanticipated transitions Such effects are seen for
demography such as trajectories of schooling, desistance from crime (Laub et al. 1998; Nagin
work, partnership and marriage, and childbearing et al. 1995), dynamics of health (Bauldry et al.
where theoretical and empirical attention has 2012; Ferraro et al. 1997) and mental health
focused on continuities and discontinuities in (Mirowsky and Ross 2007; Wheaton and Gotlib
both the acquisition of roles and the sequences of 1997; McLeod and Shanahan 1996), and the
roles over time (Hogan 1978; Marini 1984). probability of a range of non-normative events
Importantly, trajectories often have a particular (Furstenberg 2003; McLeod and Fettes 2007;
shape or form that indicates a particular pattern of Rindfuss et al. 1987). At an individual level, a
human development over time (Elder 1998). sudden illness, an encounter with the criminal
Transitions are fundamentally connected to justice system, or a religious conversion may
trajectories in that they indicate movement into suddenly redirect a life trajectory in another
or out of social roles when linked in a sequence direction, for better or worse.
over time. Finishing schooling, getting married, To return to our example of obesity, there is not
getting divorced, starting work, stopping work a lot of research that formally or even informally
are all key transitions that given meaning to tra- examines trajectories of BMI or weight over the
jectories by shaping both the timing of transitions life course and even less that explores how trajec-
and directly inuencing the content of human tories, once set, are reversed. Still, existing work
experience and behavior. Importantly, transitions is provocative. On the general issue of temporal
can and typically are bundled together into patterns over the life course, Baltrus and col-
what has been called role congurations. leagues (2005) examined trajectories of weight
Congurations of roles often represent prescribed gain in a multi-cohort sample. Highlighting the
packages reecting social approved trajectories importance of social structures, the research
of success (or failure). In this respect, role con- shows that race and sex interacted in important
gurations imply a whole that is greater than the ways creating different trajectories for race and
sum of its parts (Macmillan and Eliason 2003). gender sub-populations. While Black males had
The idea that role congurations and trajectories lower initial weights, they had trajectories of
assume meaning and social signicance is nicely growth, around 0.02 kg per year, that were similar
illustrated by a disturbing trend in the U.S. to the sample average, Black females had both
referred to as NEETs, young adults not in educa- greater weight at baseline assessment and greater
tion, not employed, and not in training (Monaghan growth in weight, around 0.10 kg per year. This
2014). NEETS can be seen as a highly problem- race-sex variation in trajectories of weight gain
atic life course trajectory where growing num- was largely accounted for by the dynamics of
bers of young adults, especially males are failing socioeconomic position over the life course (see
to make the transition to adulthood, at least from also Guo et al. 2000). More generally, there is
an economic and social perspective. And, it is evidence of both dened trajectories in body
likely that these problematic trajectories have mass, as well as suggestive evidence of individual
546 R. Macmillan and F. Furstenberg

and group heterogeneity in such trajectories. Still, ment of familial responsibilities. How all this
efforts to understand the dynamics of body mass occurs however is obviously complex, and more
could benet from further analysis of trajectories robust treatments of the multilevel processes in
of weight gain, particularly with reference to the play emphasize interconnections of social struc-
complex interactions of social structural and ture, opportunities and risks, resources and
social and behavioral risk in early life. human agency, temporality and trajectories of
social roles, and role-related attainments (Ferraro
and Shippee 2009). Such work emphasizes the
3.4 Accumulation, Accentuation ways in which institutions structure opportunities
and Leveling and how processes involving selection and allo-
cation of resources are directed to those who have
The features of a life course perspective outlined more to offer.
so far ultimately provide the backdrop for a core A complementary idea emphasizes the notion
theoretical principle/debate over the implications of accumulation (Caspi et al. 1987). Initial differ-
of aging for social advantage or disadvantage. ences across people increase the likelihood of
Some researchers hypothesize that aging is associ- continuity of state or position by virtue of selec-
ated with processes that accumulate advantages tion or allocation into like circumstances in the
taking the form of path-dependent trajectories. future. For example, early exposure to poverty
Often traced to the seminal thesis of Robert Merton may result in diminished opportunities for educa-
(1968), scholars in a variety of elds have offered tion by virtue of geographic isolation and segre-
the thesis that initial (dis)advantages appear to gation that then translate into limited attainment,
grow exponentially over time, whether one is ref- poorer employment options, and increased risk of
erencing aging, the progression of careers, the further poverty in later life. From another per-
maturation of states, or personal or social develop- spective, the continuity of state may imply accu-
ment (see chapter Life Course Lens on Aging and mulation of advantage or disadvantage in that
Health by Ferraro this volume). people who are consistently non-poor may have
Importantly, scholars have paid considerable quite different experiences when compared to the
attention to the mechanisms that amplify initial sporadically poor or being chronically poor.
differences. In a number of inuential articles, An alternative hypothesis about accumulation
Dannefer (1987, 2003) offered a thesis on cumu- of (dis)advantage discussed above is the proposi-
lative advantage/disadvantage that emphasized tion that individual differences become muted
the role of social structures and social dynamics over time. Typically associated with studies of
in producing divergence in given resources (e.g., health and aging, the leveling hypothesis pro-
money, health, status) over time. In what is often poses that biological, psychological, and social
termed accentuation, prior advantage offers differences across people, particularly those
access to preferred pathways into more advanta- associated with socioeconomic stratication,
geous positions. Think, for example, of the diminish over time (see discussions in Dupre
enduring power of elite universities both to select 2007 and Herd 2006). With respect to physiol-
and reinforce the socially advantaged. In a com- ogy, it is not unreasonable to assume that differ-
plimentary way, ORand (1996) emphasizes the ences in robustness may diminish as entire
role of institutions and life course trajectories of cohorts become progressively frailer with
schooling, of work, and of family in producing advancing age (Lynch 2008). Similarly, one
complex patterns of cumulative advantage and could imagine that racial, educational and eco-
disadvantage. One can imagine that people are nomic gradients in health would shrink as physi-
endowed with different levels of cognitive ability ological aging and senescence become more
and how these would be magnied over time pronounced. To date, the literature is mixed, but
through different school opportunities, particular several studies provide evidence that aging is
forms of employment and the strategic manage- indeed a leveler (e.g., Dowd and Bengtson 1978).
The Logic and Practice of Growth Curve Analysis: Modeling Strategies for Life Course Dynamics 547

Still, the existing literature largely encompasses 2011). Similar dynamics are suggested by other
comparisons of different birth cohorts to proxy cross-sectional analyses that show socioeco-
individual aging and hence does not speak nomic and race differences in BMI and obesity
directly to life course concerns about trajectories risk present in adulthood (Ogden et al. 2006).
of within-person change. Still, without appropriate longitudinal data and
Applying these perspectives on change over analyses of trajectories of BMI with aging, we
the life course to questions of BMI and excess lack key information about both trends and het-
weight, there are a number of provocative theses. erogeneity in trends of BMI over the life course,
First, there might be cumulative advantage or dis- about potential time-varying dynamics of risk
advantage in weight gain over time. As will be factors, and how these congurations of condi-
discussed in more detail below, psychometric tions cohere to explain the dynamic patterns of
principles around ceiling and oor effects in cumulative advantage or disadvantage and the
measures direct attention to the correlation potential of leveling of life course inequalities
between the initial level of an outcome of interest with advancing age.
such as BMI and subsequent change over time. If In the remainder of this chapter, we turn to
range is bounded (and there is some central ten- the logic and mechanics of a growth curve ana-
dency in direction of change), one might expect a lytic strategy that allows researchers to capture
negative correlation. Those who are initially high such life course dynamics and we turn to data
on some outcome would be expected to show from the National Longitudinal Survey of Youth
relatively smaller gain over time given the pre- -1997 and the issue of BMI to provide illustra-
existing ceiling. Those who are low on some out- tions of how the key life course analytic con-
come will have less room to drop. In the instance cepts can be operationalized within a growth
of BMI, there is no necessary limit or the limits curve framework. Table 1 provides an overview
are determined by functionality and extreme, and of the relationship between the conceptual
hence one could imagine a range of potential issues discussed above and the various analyses,
growth curves. High BMI in early life may be tables, and gures describing the empirical
associated with steeper trajectories of weight aspects of our work.
gain in subsequent years. This may reect the
self-reinforcing character of weight and caloric
intake. Alternatively, it is possible that a leveling 4 Growth Curve Analysis
process occurs whereby weight gain that is age- and Trajectories of BMI: Data
driven may mute initial differences. Structure and Statistical
A second and equally interesting implication Foundations
is that social gradients in BMI may respond
directly to the process of aging. There is good The basis of growth curve analysis of any sort
evidence that socioeconomic and racial differ- involves some form of panel data and typically a
ences in body mass are apparent in early life, random-effects statistical framework. In terms of
including the preschool years. Whitaker and data structure, units are observed at multiple
Orzol (2006) show data indicating that Hispanics points in time and data is collected repeatedly
have almost twice the rates of obesity as non- (e.g., annually) on specic outcomes (e.g., height
Hispanic whites at 3 years of age, while gradi- and weight). Table 2 shows a typical at or
ents of obesity risk are more or less linear with wide le structure for a ctitious sample of
respect to maternal education and household children studied from age 10 to age 14. In these
poverty. Likewise, studies of short-term panel data, each case represents a single individual
data such as the Early Childhood Longitudinal with time-invariable or stable measures com-
Study-Birth Cohort (ELCS-B) suggest continu- bined with time-varying measures that have some
ity of risk, but do not allow for further elabora- particular variable label that denotes time of mea-
tion to other life stages (e.g., Moss and Yeaton surement (e.g., HEIGHT1).
548 R. Macmillan and F. Furstenberg

Most statistical programs for growth curve 4.1 The Variance-Components


analysis require that the data be in a person- Model
period format. Here the data are simply recon-
gured from the wide format so that each For the following illustrations, we estimate mod-
specic period is embedded within a person. els using the XTMIXED routine in Stata 13.1.
The reconguration of the same data from Given this, the notation that we adopt parallels
Table 2 with the addition of calculated BMI is that of Rabe-Hesketh and Skrondal (2008) who
shown in Table 3. In this example, each respon- have produced an excellent text on multilevel and
dent contributes ve periods where the time- longitudinal modeling using Stata. In general,
invariant factors can vary across respondents efforts to understand how outcomes for person j
but are constant within respondents and the at period i are determined can usefully turn to a
time-varying factors can be variable across both standard regression model:
respondents and periods.
BMI ij = b + V j + e ij (1)
An initial starting point for any analysis would
be consideration of the individual growth trajec- where the body mass index score for person j at
tories. Practically, such trajectories show both the time i is a function of some sample mean , a
initial level of BMI as well as the within-person residual component j that is specic to each indi-
change over time. The illustrative data from vidual and constant across periods and a compo-
Tables 1 and 2 are portrayed in Fig. 1. nent ij that is specic to each subject at each time
This information is clearly valuable. In this period. Here the component , often termed the
case, there is clear variation in the starting random intercept, has a mean of zero and a vari-
points of observation and clear variation in ance and indicates the composite deviation of
growth with age such that knowing where the persons mean from the overall mean . The
someone was at the start of the observation component ij is the within-person residual that
period is of limited value for knowing where indicates the period specic deviations off of the
they ended up. At the same time, a more formal person-specic mean. It too has an assumed
accounting of the various quantities, incorpora- mean of zero and a variance of . A particularly
tion of statistical uncertainty through sampling important quantity from such models is the inter-
error, and covariate adjustment to better under- class correlation which assess the amount of vari-
stand the nature of the sample heterogeneity ance that is between subjects and the amount of
requires a statistical framework that both variance that is within-subjects over time:
accounts for its unique character and yields the
y
quantities of interest. Here, we can turn to the r= . (2)
family of random effects models. y +q
As can be seen, the statistic itself is a ratio of
the between cluster variance, , to the total vari-
Table 1 Overview ance and, as growth curve models involve time
Theory Formulae Tables Figures ordered observations clustered within individu-
Social structure 3, 4 4, 5, 6, 7 3, 5, 6, 7 als, the statistic indicates how much of the vari-
and development ance exists across individuals or within
Longitudinal 6 5, 6, 7 5, 6, 7 individuals over time. For growth curve analysis,
associations this statistic serves two purposes. First, it tells us
Trajectories, 5 3, 4, 5, 6 1, 3, 5, 6, 7 whether there is sufcient variance within sub-
transitions, and
turning points jects to even begin growth curve analysis. Second,
Accumulation, 6, 7 5, 6, 7 it provides a quick and informative metric on
accentuation, and how much explanatory power exists at each of
leveling the different levels.
Table 2 Data in wide le format
ID Race Sex PARED HEIGHT1 HEIGHT2 HEIGHT3 HEIGHT4 HEIGHT5 WEIGHT1 WEIGHT2 WEIGHT3 WEIGHT4 WEIGHT5
1 NHW F LTHS 50 52,2 54,4 56,6 58,8 78 91 104 117 130
2 NHW M 4YR 57 58 59 60 61 90 102 114 126 138
3 AA M HS 50 51,8 53,6 55,4 57,2 75 84 93 102 111
4 AA F SC 54 55,1 56,2 57,3 58,4 65 76 87 98 109
5 H F HS 58 60 62 64 66 80 91 102 113 124
6 NHW M HS 52 53,1 54,2 55,3 56,4 80 86 92 98 104

K H M 2YR 54 54,9 55,8 56,7 57,6 80 95 110 125 140
The Logic and Practice of Growth Curve Analysis: Modeling Strategies for Life Course Dynamics
549
550 R. Macmillan and F. Furstenberg

Table 3 Data in person-period le format


ID Period Race Sex Pared Height Weight BMI
1 1 NHW F LTHS 50 78 21,9
1 2 NHW F LTHS 52,2 91 23,5
1 3 NHW F LTHS 54,4 104 24,7
1 4 NHW F LTHS 56,6 117 25,7
1 5 NHW F LTHS 58,8 130 26,4
2 1 NHW M 4YR 57 90 19,5
2 2 NHW M 4YR 58 102 21,3
2 3 NHW M 4YR 59 114 23,0
2 4 NHW M 4YR 60 126 24,6
2 5 NHW M 4YR 61 138 26,1
3 1 AA M HS 50 75 21,1
3 2 AA M HS 51,8 84 22,0
3 3 AA M HS 53,6 93 22,8
3 4 AA M HS 55,4 102 23,4
3 5 AA M HS 57,2 111 23,8
4 1 AA F SC 54 65 15,7
4 2 AA F SC 55,1 76 17,6
4 3 AA F SC 56,2 87 19,4
4 4 AA F SC 57,3 98 21,0
4 5 AA F SC 58,4 109 22,5
5 1 H F HS 58 80 16,7
5 2 H F HS 60 91 17,8
5 3 H F HS 62 102 18,7
5 4 H F HS 64 113 19,4
5 5 H F HS 66 124 20.0
6 1 NHW M HS 52 80 20,8
6 2 NHW M HS 53,1 86 21,4
6 3 NHW M HS 54,2 92 22,0
6 4 NHW M HS 55,3 98 22,5
6 5 NHW M HS 56,4 104 23,0

K 1 H M 2YR 54 80 19,3
K 2 H M 2YR 54,9 95 22,2
K 3 H M 2YR 55,8 110 24,8
K 4 H M 2YR 56,7 125 27,3
K 5 H M 2YR 57,6 140 29,7

To incorporate covariates into the model, the where the 1 coefcient indicates the degree to
standard model is simply elaborated. For exam- which variation in educational attainment shifts,
ple, if we are interested in knowing whether there up or down, the predicted value of BMIij from the
are substantive differences by educational attain- population mean 0. A graphic representation of
ment on BMI, the resulting model is this is shown in Fig. 2. This graphic shows how a
given individuals growth rate is dened by some
BMI ij = b0 + b1education1 j + V j + e ij (3)
average distance between the respondents time
The Logic and Practice of Growth Curve Analysis: Modeling Strategies for Life Course Dynamics 551

31

29

27
Case 1
25 Case 2
BMI Score

Case 3
23 Case 4

21 Case 5
Case 6
19 Case 7

17

15
10 11 12 13 14

Fig. 1 Hypothetical trajectories of BMI over time

BMI

30.0

1j
28.0

26.0

1j
24.0

22.0

1
18.0
0

LTHS HSGRAD SOMECOLL 2YRDEG 4YRDEG

Fig. 2 Illustration of random-intercept model

ordered observations, how the individual specic coefcients. Random coefcients or random
growth rate varies off of a population average slopes allow the effects of covariates to vary
growth rate by the extent 1j and how the time- across units. One very simple representation of
specic observations vary off of the unobserved this is to think of being able to run separate
trajectory by amount . regression analyses for each of the j units in the
The key adaptation that moves one from the sample. Each of these regressions would indicate
standard random-effect framework to a growth variation in effects across units by virtue of a
curve analysis is the introduction of random person-specic intercept and person-specic
552 R. Macmillan and F. Furstenberg

BMI

30.0 1j

28.0
2j
26.0
1

24.0

22.0
1j
18.0 1
0

LTHS HSGRAD SOMECOLL 2YRDEG 4YRDEG

Fig. 3 Random-coefcient representation

slope. Yet to identify the extent or magnitude of


variation and statistical signicance, it is neces- indexes the population slope of the change in BMI
sary to formalize the relations within a linear for unit changes in educational attainment, while
model. Again, consider the example of educa- 2j represents the unique deviation of the refer-
tional differences in BMI that we encountered enced individuals slope off the population slope.
earlier. In the previous specication, we simply To produce a growth curve model within this
examined how much person specic BMI shifted framework, one simply incorporates some dimen-
depending upon level of educational attainment. sion of time as the key random coefcient. Here,
In a random coefcient model, the education one could think of any markers of aging or devel-
effect itself is variable and parameterized as such: opment, including age, standard time (e.g., days,
weeks, months, years), developmental stages
BMI ij = 0 + 1education1 j + 1 j (e.g., school grade), or even life stage (e.g.,
+ 2 j education + ij infancy, childhood, the teenage years, early adult-
(4)
hood, mid adulthood, old age). Importantly, the
where the core parameters are dened as before model will accommodate any temporal dimen-
while 1j represents the deviation of person j from sion suitable with the qualication that the
the mean intercept 0 and 2j represents the devia- research needs to specify appropriate scaling that
tion of person js slope from the mean slope 1. lends itself to interpretation. So, for a growth
Figure 3 represents this graphically. Again, the s curve analysis within a random coefcient frame-
dene the distances between the best tting slope work, one would estimate
across levels of education and the actual data
BMI ij = a + b1 Ageij + V1 j + V 2 j Ageij + e ij (5)
points for a given individual. 0 captures the devi-
ation of the population intercept off of zero, while where BMI for person j at time i is a function of
1j captures the deviation of the referenced indi- a population mean, denoted , a regression
viduals intercept off of the population intercept. 1 parameter 1 that shows the association between
The Logic and Practice of Growth Curve Analysis: Modeling Strategies for Life Course Dynamics 553

BMI and age for person j at time i, 1j is a random best interpretation is how the rate of growth
intercept, and an idiosyncratic error term for changes by level or value of the covariate. For
person j at time i. The 1 parameter in this model variables that index states (e.g., married), particu-
captures the degree of growth over time as it larly dummy variables, the interpretation is best
indexes change in BMI associated with (a one- seen as how the growth rate changes when one is
unit) change in age. Moreover, the second 1j in a particular state as opposed to not being in a
parameter species a random slope of age that particular state. Importantly, this has consider-
allows for variation in their overall rate of growth able afnity with ideas of linked trajectories and
over time. transitions and the potential for turning points in
From this initial model, the operationalization the life course. In the case of less discrete mea-
of the core life course concepts and principles sure, the best interpretation is how growth is
guides which covariates are added and how the accelerated or stied when combined with some
model is structured. Covariates in a growth curve of level of some local life circumstance. Horney
framework are usefully differentiated into those et al. (1995) for example show how trajectories of
that are time invariant and those that are time- offending are inuenced by the extent to which
varying. In the former case, values of the covari- one is using drugs.
ate are xed over the entire range of the
observation window (i.e., do not change). In the
latter case, covariate values can and do change 5 Trajectories of BMI
over time. In either case, covariates can simply in the National Longitudinal
be added to the part of the model predicting ini- Survey of Youth: 1997
tial level or intercept and/or to the part of the and Their Implications
model predicting change over time, although for Life Course Theory
interpretation changes depending upon covariate and Research
type and place in the model. For example, if one
was interested in knowing whether socioeco- To illustrate the logic and mechanics of this
nomic status in ones family of origin inuences growth curve approach, we examine trajectories
trajectories of BMI, and again following the style of BMI in the NLSY-97 data. A full description
of Rabe-Hesketh and Skrondal (2008), one would of the data and measures is found in the accom-
estimate parameters for the following equation: panying appendix.
An initial model rst evaluates a random-
BMI ij = a + b1 Ageij + b2 FSES j + V 1 j intercept versus a random-coefcient specica-
+ V 2 j Ageij + b3 FSES j * AGEij + e ij (6) tion of unconditional growth (i.e., no other
covariates beyond time). In theoretical terms, this
The term 1 captures the rate of change over model is simply assessing trajectories of BMI
age which is a random coefcient (due to 2j), 2 across subjects and adjudicating whether it is rea-
captures the effect of time-invariant family socio- sonable to assume that trajectories have a single
economic status (FSESj) on the initial level of form or whether they are multifaceted.
BMI, and 3 captures the degree to which growth Empirically, this model including a single predic-
rates vary according to family socioeconomic tor indexing aging and the relevant parameters
status. For time-varying covariates, the model set are shown in Table 4. The table, and all subse-
up is the same but interpretation is more nuanced. quent tables, is partitioned in two: the xed part
In the case of effects on the intercept, the inter- shows the relationship between the outcome BMI
pretation is strongest as the average level of the and the covariates in the model, in the rst case
time-varying covariate over the observation win- aging, while the random part shows the relevant
dow and its association with the intercept value parameters for the variance components.
as requirements of temporal ordering are vio- The initial models provide a wealth of infor-
lated. In the case of effects on the growth rate, the mation about the dynamics of BMI from
554 R. Macmillan and F. Furstenberg

adolescence to adulthood. First, the intra-class Table 4 Unconditional growth model with and without
random-slopes
correlations () show how much of the variance
in BMI is across individuals or within-individuals Model 1 Model 2
over time. For both models, has a value of 0.77, Random intercept Random coefcient
indicating that 23 % of the variance in BMI is (SE) (SE)
associated with change over time. The intercept Fixed part
value indicates that the average BMI between the Aging 0.405 0.002 0.409 0.004
ages of 12 and 16 is approximately 22.7 Intercept 22.303 0.058 22.277 0.051
(22.303 + 0.405 = 22.708), almost dead center of Random part
4.818 0.040 4.215 0.038
the normal range (i.e., 2024.9). Rate of growth y 11
is shown by the coefcients and indicates that
0.327 0.003
on average BMI increases by 0.405 per year, a y 22
growth rate that is highly statistically signicant. 21 0.015 0.014
The 21 parameter incorporates a correlation 2.561 0.006 2.132 0.005
q
between the random-intercept and the random-
slope terms in model 2 and indicates no signi- Log 233,226 223,271
likelihood
cant correlation. Finally, we perform a
likelihood-ratio test comparing the random-
intercept (model 1) and random-coefcient Table 5 Conditional growth model with random-slopes
(model 2) models to see which model ts the Model 3 Model 4
data better. With two degrees of freedom for the Random Random
addition of the random-coefcient and for the coefcient coefcient
correlation between random components, the LR (SE) (SE)
test statistic (which is distributed chi-square) has Fixed part
a value of 22,285 that is highly statistically sig- Aging 0.409 0.004 0.393 0.008
nicant and indicates that the random-slope spec- White females 0.946 0.131 0.958 0.132
ication provides a much better t to the data. As Black males 0.646 0.174 0.690 0.176
an overall summary, the initial models provide Black females 1.400 0.174 1.254 0.176
strong evidence that the dynamics of BMI have Hispanic 0.987 0.179 0.921 0.180
males
strong temporal components that are highly vari-
Hispanic 0.057 0.187 0.020 0.188
able across individuals and hence is usefully females
understood through a growth curve approach. Interactions with aging
Importantly, not all data have this quality, in White females 0.008 0.011
which case a trajectory approach may not be Black males 0.029 0.015
warranted. Black females 0.093 0.015
With evidence of signicant heterogeneity in Hispanic 0.043 0.015
both intercepts for and growth of BMI, we next males
examine whether there is social structural varia- Hispanic 0.025 0.016
tion in initial levels and growth rates with specic females
attention to race, sex and social class. In model 3 Intercept 22.204 0.091 22.229 0.091
of Table 5, we being by simply including dummy Random part
4.146 0.034 4.144 0.037
variables identifying white females, Black males y 11
and females, and Hispanic males and females
0.327 0.003 0.325 0.003
with white males being the reference category. y 22
Here, white females have lower initial values of 21 0.025 0.014 0.023 0.014
BMI ( 22.204 .946+ 0.409 = 21.667 ), while 2.132 0.005 2.132 0.005
q
black males (22.204 + .646 + 0.409 = 23.259 )
and females ( 22.204 + 1.400 + 0.409 = 24.013 ) Log likelihood 223,150 223,120
The Logic and Practice of Growth Curve Analysis: Modeling Strategies for Life Course Dynamics 555

and Hispanic males ( 22.204 + .987 + 0.409 = tuated over time. At the same time, there is no
23.600 ) have higher BMIs in mid adolescence. clear evidence of attenuation or leveling.
As would be expected, the addition of race-sex Extending the analysis, we incorporate a mea-
indicators reduces the random-intercept standard sure of social class: fathers educational attain-
deviation from 4.818 to 4.146. This reduction ment. Social class is a further indicator of position
indicates that different race-sex groups have dif- in social structural position and a key variable
ferent starting points in adolescence for BMI that should shape trajectories of experience and
and illuminates a particular inuence of social state over time. Relevant parameter estimates are
structural position on life course developmental shown in Model 5 in Table 6. When fathers edu-
pathways of BMI. cation is included in the model as an inuence on
Perhaps more interesting, model 4 includes the random-intercept, two interesting ndings
the interactions between race and sex and the emerge. The rst is a small, negative effect on
growth parameter. This model provides a better initial BMI. Each additional year of a fathers
t to the data as the 2 log likelihood value is education decreases initial BMI by 0.096 points.
reduced by 30 with 5 degrees of freedom Considering that initial BMI has a standard devi-
(p < 0.001). Here, the growth in BMI over time is ation of around 4, a unit change in educational
substantively similar for white males (0.393), attainment of fathers is equivalent to 1/20th of a
white females (0.401 = 0.393 + 0.008), Black standard deviation. Second, the inclusion of
males (0.364 = 0.393+ 0.029), and Hispanic fathers educational attainment reduces race-sex
females (0.418 = 0.393 + 0.025) as none of the differences, particularly for Black males and
product terms are statistically signicant. In con- females and for Hispanic males. In comparison to
trast, there are statistically signicant, larger the coefcient indexing race-sex differences in
growth rates for Black females model 4, the effects shown in model 5 are
(0.486 = 0.393 + 0.093) and Hispanic males decreased by 19 %, 9 %, and 34 %, respectively.
(0.436 = 0.393 + 0.43). At the same time, there are strong conditional
Combining the intercept estimates with the effects on trajectories of growth. Here, unit
growth parameters creates a unique lens on the increases in fathers education are associated
dynamics of BMI across race and sex (Fig. 4). with a 0.009 point decrease in annual change in
For example, while the rate of growth is the same, BMI, independent of the race-sex effects on
white females have lower initial BMIs in mid growth. We also investigated whether effects of
adolescence compared to white males while fathers education on growth rate was conditional
black males have higher initial BMIs. Black on race and sex and found no convincing evi-
females show a different pattern in that their ini- dence. As a result, the estimated model assumes
tial level of BMI is substantially (and statisti- that the effect of fathers education is consistent
cally) higher than all other race-sex groups and across race-sex groups.
their rate of growth is the largest. This could be Given the lack of further moderation, we need
interpreted as a pattern of earlier onset of excess to consider race-sex subgroups and fathers educa-
weight given that the initial level of BMI is almost tion effects simultaneously to identify processes
at the threshold for being over weight (24). that accentuate or attenuate the trajectory of
Hispanic males also show a similar pattern of ear- BMI. In one respect, conditional race-sex differ-
lier onset of high BMI coupled with larger than ences in growth of BMI are both accentuated and
average growth. In contrast, Hispanic females are attenuated with the inclusion of fathers education.
within the error-bounds of average for initial The latter is seen with respect to Black females
BMI but show a slightly accelerated growth rate. (0.0930.082) and Hispanic males (0.0430.014).
Taken as a whole, the accelerated growth of BMI The former is seen for Black males (0.029 to
trajectories for Black females and Hispanic males 0.042). In another respect, the negative effect of
may be seen as an interesting case of cumulative fathers educational attainment combines with
disadvantage as initial differences become accen- race-sex effects to produce cumulative patterns of
556 R. Macmillan and F. Furstenberg

Table 6 Conditional growth model with random-slopes


Model 5 Model 6
Random coefcient Random coefcient
(SE) (SE)
Fixed part
Aging 0.516 0.021 0.497 0.021
White females 0.965 0.132 1.008 0.131
Black males 0.559 0.177 0.580 0.176
Black females 1.146 0.176 1.100 0.176
Hispanic males 0.612 0.188 0.617 0.187
Hispanic females 0.274 0.195 0.325 0.194
Fathers education 0.096 0.017 0.081 0.017
Role transitions
Marriage 1.178 0.096
Divorce/separation 0.273 0.246
Enrolled in school 0.752 0.044
Employed 0.230 0.037
Has child 0.810 0.075
Interactions with age
White females 0.008 0.011 0.006 011
Black males 0.042 0.015 0.046 0.015
Black females 0.082 0.015 0.080 0.015
Hispanic males 0.014 0.016 0.012 0.016
Hispanic females 0.003 0.016 006 0.016
Fathers education 0.009 0.001 0.010 0.001
Marriage 0.075 0.009
Divorce/separation 0.004 0.021
Enrolled in school 0.093 0.005
Employed 0.031 0.005
Has child 0.053 0.008
Intercept 23.532 0.247 23.648 0.008
Random part
4.134 0.037 4.115 0.037
y 11
0.324 0.003 0.323 0.003
y 22
21 0.029 0.014 0.024 0.014
2.132 0.014 2.114 0.014
q
Log likelihood 223,078 222,386

growth. For example, a one standard deviation thirds (0.042 + 0.027 = 0.069) with similar
(3) increase in fathers education would decrease advantages in family background. For the other
growth in BMI by 0.027 points. This implies that groups, non-signicant differences are simply
approximately one-third of the growth disadvan- modied in line with linear increases in fathers
tage of Black females (0.082 + 0.027 = 0.55) educational attainment.
would be offset if fathers had higher educational The previous analyses have investigated links
attainment. In contrast, the growth advantage of between social structure and the development of
Black males increases by approximately two- weight gain. But, we can elaborate life course
The Logic and Practice of Growth Curve Analysis: Modeling Strategies for Life Course Dynamics 557

Fig. 4 Trajectories of BMI by race and sex, NLSY-97

theory further by considering life course dynam- predicting BMI. Unlike the previous models,
ics in response to particular role transitions. As these are time-varying variables that index
discussed earlier, role transitions are pivotal com- whether someone is or is not in a given state in
ponents of trajectories with important transfor- any given year. Referencing widely accepted
mative potential (Shanahan 2000). In extreme markers of the transition to adulthood (Shanahan
cases, transitions can radically alter or wholesale 2000), we focus attention on school enrollment,
change trajectories and produce turning points in full-time employment, marriage, marital dissolu-
life course trajectories (Sampson and Laub 1995). tion, and parenthood. Consistent with our earlier
This notion that a single event can have a radical descriptions, the model includes effects on both
impact on redirecting the life course may be chal- the intercept and the growth parameter yet orga-
lenging to demonstrate empirically, but turning nizes discussion of interpretation largely on the
points that dramatically shift the life course has latter.
long been recognized. A sudden illness or acci- As discussed earlier the effects of role transi-
dent, an unexpected job opportunity, or a chance tions on the random-intercept are not particu-
encounter are but a few examples of events that larly interpretable. One cannot say much other
constitute turning points, reasonably exoge- than to note that there are large effects on aver-
nous occurrences that suddenly alter an expected age levels of BMI. Being married, for example,
trajectory of behavior. We consider this next by is associated with 1.178 higher BMI. A slightly
examining the idea that role shedding and acqui- smaller positive association is seen for parent-
sition in the transition to adulthood may expose hood (0.810), while employment also shows a
people to new environments, networks, and expe- positive association but one that is much smaller
riences that may alter their current states of being. (0.230). In contrast, being enrolled in school is
We examine this issue in Model 6 of Table 6 associated with 0.752 lower BMI points, while
by including specic role transitions in models being divorced or separated has no signicant
558 R. Macmillan and F. Furstenberg

association with average BMI. Effects on growth or average level of BMI and growth in BMI over
are much more interpretable and in most case time. In the former case, each change in educa-
substantial. For example, when respondents are tional attainment is associated with a 0.311
married, their growth rate for BMI decreases by higher average BMI. As before, this parameter is
0.075 points per year. Growth in BMI is also particularly difcult to interpret as educational
muted when one is employed or a parent, by attainment, both in the data and in general, is
0.031 and 0.053 points per year, respectively. sequential and age-graded. As such there are no
For purposes of illustration, rather than sys- cases of high educational attainment on the initial
tematic investigation, we briey consider the intercept. At the same time, the conditional effect
issue of role congurations. Here, we do so by of fathers educational attainment is 0.103.
simply estimating whether the effects of mar- In contrast, effects on growth yield a number
riage and parenthood are conditional on the pres- of interesting ndings that have much clearer
ence of the other state. As with family social interpretation. It is immediately apparent that an
class, there are good reasons to think that the individuals educational attainment has large,
meaning of particular role is shaped by the pres- negative effects on BMI growth ( = 0.043). In
ence or absence of other roles (Macmillan and addition to being highly statistically signicant,
Copher 2005). The relevant parameters are again the inclusion of respondents education substan-
product terms but this time capturing the co- tially increases t to the data ( 2LL = 192, 2
occurrence of time-varying marriage and parent- df). This parameter indicates that growth rates
hood. Shown in model 8 of Table 7, neither decelerate or trajectories are atter at higher lev-
parameter is statistically signicant, indicating els of education. Equally important, the inclusion
that both marriage and parenthood matter for of respondents educational attainment reduces
BMI and trajectories of BMI but operate inde- the effect on growth of fathers educational
pendent of one another. attainment by about 50 % to 0.005 (compare with
As a nal set of analyses, we try to knit Model 6 in Table 6). Interestingly, this is the only
together the various themes by emphasizing life variable that is substantially mediated: both the
course longitudinal associations as further con- race-sex and role transition indicators are largely
necting ideas about social structure, transitions unchanged from that seen in earlier models.
and trajectories, and accumulation, accentuation, The relative independence of these factors
and leveling over the life course. We do so by leads one to further consider accentuation and
including a measure of respondents educational attenuation associated with combinations of
attainment, as both a direct and conditional effect. characteristics. As one example, the growth rate
An individuals educational attainment is a key for BMI for white males who are still in school
indicator of achieved socioeconomic status and have made none of the role transitions char-
(Hauser and Warren 1997), and an essential ele- acteristic of the transition to adulthood is 0.667
ment in models of socioeconomic mobility with (0.573 + 0.094 = 0.667) indicating that BMI
fundamental associations with educational attain- would increase approximately 2 points every 3
ment and social class position of parents (Blau years. In contrast, a white male who has left
and Duncan 1967). Educational status, of course, school, is employed, married, and has a child,
has a wide range of consequences for life chances, and has a 4 year college degree would have a
including health dynamics (Ross and Wu 1995). growth rate of only 0.122 indicating that BMI
Model 9 in Table 8 includes respondents edu- could be expected to increase by only 1 point
cational attainment as a time-varying covariate in every eight years. In this example, we can see
the model with effects on both the random- cumulative advantages that accrue from norma-
intercept and the random-coefcient indexing tive role transitions accompanied by high educa-
heterogeneity in growth over time. There are sev- tional attainment and how educational attainment
eral important ndings. First, educational attain- can produce increased variance in trajectories of
ment has countervailing effects on the intercept BMI over time.
Table 7 Conditional growth model with random-slopes
Model 7 Model 8
Random coefcient Random coefcient
(SE) (SE)
Fixed part
Aging 0.523 0.032 0.499 0.022
White females 1.007 0.131 1.009 0.131
Black males 0.579 0.176 0.577 0.176
Black females 1.110 0.176 1.092 0.176
Hispanic males 0.620 0.187 0.616 0.187
Hispanic females 0.327 0.194 0.327 0.176
Fathers education 0.060 0.022 0.081 0.169
Role transitions
Marriage 1.547 0.374 1.299 0.132
Divorce/separation 0.683 1.148 0.300 0.247
Enrolled in school 0.505 0.183 0.744 0.044
Employed 0.272 0.153 0.229 0.037
Has child 1.058 0.313 0.859 0.082
Married with child 0.225 0.182
Interactions with fathers education
Marriage 0.031 0.030
Divorced/separated 0.034 0.092
Enrolled in school 0.020 0.014
Employed 0.003 0.012
Has child 0.020 0.026
Interactions with age
White females 0.006 0.011 0.006 0.011
Black males 0.046 0.015 0.045 0.015
Black females 0.080 0.015 0.082 0.015
Hispanic males 0.012 0.016 0.012 0.016
Hispanic females 0.005 0.016 0.005 0.016
Fathers education 0.012 0.002 0.010 0.001
Marriage 0.038 0.020 0.087 0.012
Divorce/separation 0.028 0.094 0.001 0.021
Enrolled in school 0.073 0.022 0.092 0.005
Employed 0.038 0.020 0.031 0.005
Has child 0.085 0.032 0.060 0.009
Married with child 0.024 0.017
Interactions with fathers education
Married 0.002 0.003
Divorced/separated 0.003 0.008
Enrolled in school 0.002 0.002
Employed 0.001 0.002
Has child 0.003 0.003
Intercept 23.394 0.309
Random part
4.115 0.037 4.115 0.037
y 11
0.323 0.003 0.323 0.003
y 22
21 0.024 0.014 0.024 0.014
2.114 0.005 2.114 0.005
q
Log likelihood 222,383 222,385
560 R. Macmillan and F. Furstenberg

Before concluding with some summary graph- and the relevant coefcient is highly statistically
ics, we consider a clear example of accentuation signicant. There are different ways of interpret-
as whether advantages of in ones family of ori- ing the effects given the nature of the variables,
gin are amplied by ones own status achieve- but they are generally complementary. In this
ments. Such an effect is seen via product terms illustration, we focus on the degree to which
indexing higher levels of fathers education by fathers educational attainment conditions the
higher levels of respondents education that are contemporaneous effects of respondents educa-
included in model 10 in Table 8. Again, we con- tional attainment on trajectories of BMI. This
centrate on the effects on the growth parameter interpretation is conceptually more tractable: the

Table 8 Conditional growth models with random slopes


Model 9 Model 10
Random coefcient Random coefcient
(SE) (SE)
Fixed part
Aging 0.573 0.032 0.424 0.043
White females 1.043 0.131 1.040 0.131
Black males 0.605 0.176 0.610 0.176
Black females 1.037 0.176 1.105 0.176
Hispanic males 0.615 0.187 0.601 0.187
Hispanic females 0.382 0.194 0.383 0.194
Fathers education 0.103 0.023 0.158 0.027
Rs education 0.311 0.019 0.079 0.080
Role transitions
Marriage 1.499 0.374 1.173 0.378
Divorce/separation 0.821 1.146 0.531 1.147
Enrolled in school 0.694 0.183 0.868 0.188
Employed 0.200 0.152 0.326 0.154
Has child 0.991 0.312 0.928 0.313
Interactions with fathers education
Marriage 0.048 0.030 0.067 0.030
Divorced/separated 0.024 0.092 0.001 0.092
Enrolled in school 0.014 0.015 0.029 0.015
Employed 0.008 0.012 0.018 0.012
Has child 0.017 0.026 0.011 0.026
Rs education 0.018 0.006
Interactions with age
White females 0.013 0.011 0.013 0.011
Black males 0.052 0.015 0.054 0.015
Black females 0.091 0.015 0.088 0.015
Hispanic males 0.011 0.016 0.015 0.016
Hispanic females 0.004 0.016 0.005 0.016
Fathers education 0.005 0.002 0.006 0.003
Marriage 0.087 0.035 0.119 0.036
Divorce/separation 0.041 0.093 0.009 0.094
Enrolled in school 0.094 0.022 0.128 0.023
Employed 0.032 0.020 0.047 0.020
(continued)
The Logic and Practice of Growth Curve Analysis: Modeling Strategies for Life Course Dynamics 561

Table 8 (continued)
Model 9 Model 10
Random coefcient Random coefcient
(SE) (SE)
Has child 0.074 0.032 0.062 0.032
Rs education 0.043 0.002 0.001 0.009
Also by fathers education
Married 0.004 0.003 0.006 0.003
Divorced/separated 0.001 0.008 0.001 0.008
Enrolled in school 0.003 0.002 0.006 0.002
Employed 0.001 0.002 0.002 0.002
Has child 0.002 0.003 0.001 0.003
Rs education 0.003 0.001
Intercept 23.055 0.309 23.758 0.006
Random part
4.118 0.033 4.120 0.033
y 11
0.321 0.003 0.321 0.004
y 22
21 0.026 0.014 0.028 0.014
2.110 0.005 2.111 0.014
q
Log likelihood 222,191 222,177

most straightforward interpretation of effects on vidual whose father had high educational attain-
growth with time-varying variables is the degree ment would see virtually no change in BMI over
to which growth rates increase or decrease when time (0.424 + [0.059*7 = 0.011).
one is in a given state or at a given level of the We conclude by returning to graphic represen-
moderating factor. From this perspective, a unit tations to show how the various estimates com-
change in educational attainment is associated bine to produce distinctly different trajectories of
with a 0.019 decrease in growth of BMI when BMI illustrating the different types of life course
fathers educational attainment is low (two stan- effects we have described. Here, we highlight
dard deviations below the mean). When fathers ve dimensions of stratication, race, sex, family
educational attainment is average (12 years), a SES, achieved education, and role acquisition in
unit change in educational attainment is associ- the transition to adulthood, and examine their
ated with 0.040 decrease in growth of BMI. For cumulative effects on BMI. We do so by estimat-
those whose fathers who have high levels of edu- ing expected values for each year for subgroups
cational attainment (two standard deviations identied through the cross-classication of the
above the mean), effect of educational attainment measures of stratication. For purposes of simple
on growth is 0.059. As a simple lens into the representation, we rst contrast two extreme
implications of such differences, the growth rate cases. The rst comprises black females whose
for BMI for someone (e.g., white male who is fathers had low levels of education (12 years or
exited school but made no other role transitions) less), who themselves have low levels of educa-
with high educational attainment (i.e., graduate tional attainment (high school graduate or less)
school) but whose fathers had a low level of edu- and who move into their 20s without employ-
cational attainment would be 0.291 ment, marriage, or parenthood. The second com-
(0.424 + [0.019*7]) or just over 1 BMI point prises white females whose fathers had high
increase every 4 years. In contrast, a similar indi- educational attainment (16 years or greater), who
562 R. Macmillan and F. Furstenberg

Fig. 5 Predicted values of BMI by age, stratied by race, fathers educational attainment, respondents educational
attainment, and role acquisition, NLSY97

themselves had high educational attainment lating from the data, the average white female
(some college or more) and who were employed, would fall into the obese range sometime in their
married, and had children in their transition to late 40s or early 50s, assuming that the trajectory
adulthood. This leads to a second set of analyses maintains itself. In contrast, the average Black
where we show the effects of fathers education female in that subgroup passes into the obese
(low versus high), respondents education (low range at age 25. Because the effects of obesity on
versus high), and role acquisition (empty ver- a wide range of health liabilities is cumulative
sus full) for each of the race-sex groups. In the (Ferraro and Kelley-Moore 2003), the patterns
latter cases, we refer to the rst group as disad- manifest stark inequalities in health over the life
vantaged and the second group as course.
advantaged. Figure 6 shows the trajectories of BMI for
Figure 5 shows the simulated trajectories females that are stratied by race and level of
based on estimated values of BMI from model advantage or disadvantage. As a reference point,
10. There are three interesting elements. First, black females have the highest BMI in mid ado-
there is a clear pattern of accentuation in that the lescence and show almost perfect linear growth
BMI differential across groups was only about 2 through to the late 20s. Here, BMI increases from
BMI points in late adolescence, yet grows to over 23.5 to 31.5, an 8 point increase. For more advan-
5 points by the late 20s. Second, the trajectories taged black females, the trajectory of BMI begins
of BMI growth are dramatically different. For the at age 19 (when there is variation on accumulated
subgroup of black females, BMI is growing life course role transitions (i.e., employment and
approximately 2 points every 5 years. In contrast, marriage and parenthood) and grows approxi-
the growth rate for the subgroup of white females mately 1 BMI point through age 28. In compara-
is only 1 point every 5 years. Third, and extrapo- tive terms, BMI for disadvantaged Black females
The Logic and Practice of Growth Curve Analysis: Modeling Strategies for Life Course Dynamics 563

Fig. 6 Predicted values of BMI by age: females stratied by race and (dis)advantage, NLSY97

grew 4 points from age 18 to age 28 while BMI enthood. At the same time, there is signicant
for advantaged Black females grew only 1 point. evidence of accentuation is also present among
Given the structure of the model, the same trend males. For white males, as indicated, initial dif-
gap applies to Hispanic and white females with ferences in BMI actually favor those disadvan-
the important qualication that initial levels dif- taged (25 versus 23). By the late 20s, the
fer by approximately 1 BMI point. pattern of advantage has ipped with advantaged
Among males we nd similar patterns of tra- white males now having BMIs that are 2 points
jectories of BMI (see Fig. 7). For disadvantaged lower than those disadvantaged. For Hispanic
males, BMI in mid adolescence is greater for and Black males, there are no real differences in
Black males, although both intercepts fall within BMI in the early 20s (26.5). However, differ-
1 BMI point of one another and are not signi- ences are notable by the late 20s.
cantly different from one another. Substantively,
both show initial levels that sit at the center of the
distribution for the normal range (i.e., 2024.9). 6 Further Issues
Growth however is quite steep with an increase
of approximately 0.5 a BMI point each year. There are numerous further issues that we have
Trajectories for more advantaged groups could not covered in depth but warrant some discus-
hardly be more different. White males, for exam- sion. First, growth curve models are analytic
ple, have higher BMI in mid adolescence (25.5 at strategies that can be implemented in a variety of
age 16) but have a relatively at growth rate. ways with varying statistical elements. In addi-
Between age 16 and age 28, BMI only increases tion to the random-effects equation approach that
by about 1.5 points. Trajectories for advantaged we use, there are also structural equation model
Black males are similar but start later, due to later (SEM) and hierarchical linear model (HLM)
movement into employment, marriage, and par- approaches. In the former case, the growth rate is
564 R. Macmillan and F. Furstenberg

Fig. 7 Predicted values of BMI by age: males stratied by race and (dis)advantage, NLSY97

treated as a latent construct derived from a par- come in panel data and then these are used as out-
ticular scaling of factor loadings for the time comes in a second stage with covariates tted to
variable. For example, simple linear growth can each to predict shifts, up or down, in intercept
be estimated by xing time variables to have and slope values. In the former case, covariates in
equal differences (e.g., time1 = 1, time2 = 2, the intercept equation determine whether differ-
time3 = 3). Accelerated growth simply involves ent sub-samples within a population have differ-
alternative scaling (e.g., time1 = 1, time2 = 4, ent starting points, while covariates in the slope
time3 = 9). The model is then estimated as a full- equation determine whether different sub-
information structural equation model with the samples within a population have different
outcome at time t being a function of a latent growth rates. Importantly, method of estimation
intercept with a xed value and a latent slope or parameterization is entirely a matter of choice
with a start value, the outcome at time t + 1 esti- and different approaches will produce estimates
mated as a function of the latent intercept with a that are substantively similar but not necessarily
xed value and a latent slope with a scaled up identical estimates given the elements in the
value, and so on for all the relevant time period. models and peculiarities of the estimator.
As with all structural equation models, variances We have also not placed much emphasis on
are estimated and covariances, including those effect sizes and their relationship to statistical
among errors, can be estimated if appropriate. signicance. In this area, growth models are no
A second alternative strategy is the hierarchi- different than any other generalized linear model
cal linear model (HLM) approach. Conceptually, in that the slope parameters of any sort indicate
this is a two level model, two-stage model (even the magnitude of the association between the
although all estimation is simultaneous) where an covariate and model parameter (i.e., intercept or
initial stage estimates random-variable slope). When interpreting such things, it is impor-
intercept(s) and slope(s) for some particular out- tant to consider the scaling of independent
The Logic and Practice of Growth Curve Analysis: Modeling Strategies for Life Course Dynamics 565

variables and to consider (a) how much change in the intersection of historical time and biographi-
intercept or slope is associated with change in the cal time and the interlocked nature of develop-
independent variable; and (b) how much residual ment within historical context. In the most
variance is reduced when particular parameters general sense, it emphasizes temporal variation
are added to the model. There are no hard and fast in the occurrence and meaning of social events
rules but one can think about the issue as a gen- and social, psychological, and physiological
eral example of how much variation in the out- states. Against this backdrop, the dynamics of
come is associated with a particular covariate and social structure, particularly social position,
how large is the reduction in the variance of any exerts a powerful inuence on variation in life
given component when a particular parameter is course trajectories. Trajectories involve the inter-
added. On the issue of statistical signicance, the connection of roles and states over time and ulti-
main question is how one should incorporate the mately generate life course patterns of accentuated
non-independence of units associated with the or attenuated advantage. As a set of statistical
clustering of observations within individuals. tools, random-effect, growth analysis are particu-
Typically, one would simply adjust standard larly well suited to modeling heterogeneity in tra-
errors for non-independence using either a jectories of states, social roles, or social positions
Huber-White sandwich estimator or a clustered by focusing explicitly on within-person change
sandwich estimator. Under most conditions, over time and its connection to time-invariant and
these increases the size of the standard error and time-varying factors. The life course perspective
hence decrease the likelihood of type-II errors. includes a exible set of tools that can accom-
An equally sensible approach is to adjust upward modate a wide range of relevant concepts and
the critical value on the hypothesis test of a given provide rigor of thought and method in the speci-
parameter and hence require a more signicant cation of testable hypotheses and the generation
effect before deeming it to be non-zero in the of empirical parameters.
population. The specications and examples provided in
this chapter are meant to orient researchers to the
mechanics of growth curve analysis, the logical
7 Conclusion stages of implementation, the key statistical
quantities that can be generated, and best prac-
The purposes of this chapter are to articulate tices for interpretation. Our example of BMI
some key concepts of a life course perspective, to illustrates the various elements and was moti-
translate them into a statistical structure, and vated by the specic health contexts of contem-
show how a random-effects, growth model can porary cohorts in OECD type countries that are
articulate such concepts to speak to key questions rapidly witnessing obesity epidemics and the
in life course social science. We use the example developmental nature of body mass. There are a
of excess weight in the transition to adulthood as wide range of applications beyond body mass
a means of describing both how such models and any quasi-continuous outcome is fair game,
work and how the relevant parameters can be as is any dichotomous outcome that can reason-
interpreted to address key theoretical questions. ably be t with respect to linear probability. The
Specically, our analyses highlight the complex eld of growth curve analyses is quite dynamic
interactions of social structure, life course link- and increasingly broad in scope. We have not
ages, trajectories and turning points, and accen- covered all the potential applications nor have we
tuation and attenuation in life course dynamics, elaborated all elements in its statistical founda-
both a general concepts and as relevant for stud- tions. Still, as conceptualization and theorization
ies of excess weight and their consequences. expand and generate new and exciting questions,
A life course approach is organized around the so to will the utility of random-effect approaches
concept of time. It is explicitly concerned with to provide illuminating and provocative answers.
566 R. Macmillan and F. Furstenberg

8 Appendix: Data when included in a linear model t to panel data


and Measures captures the nature of change over time in the
independent variable. This measure is the key
The data that we use in this research come from parameter in that we can elaborate its effects in a
the National Longitudinal Survey of Youth number of ways to study the ways in which vari-
1997 (hereafter NLSY97). The NLSY97 consists ous life course dynamics are implicated in stabil-
of an initial sample of 8,984 youths who were ity and change in BMI over time.
between the ages of 12 and 16 in 1997. When We capture various aspects of social structure
possible the respondents were re-interviewed by measuring race -sex group that differentiates
annually and data were collected on a range of white males, white females, Black males, Black
topics on the transition to adulthood. As of 2014, females, Hispanic males, and Hispanic females.
there are 15 waves of data that cover an age range We also include the highest level of fathers edu-
of 12 to 31. In addition to non-Hispanic whites, cational attainment based on the highest attain-
the NLSY97 oversampled blacks and Hispanics ment of either the residential father in the
such that there are relatively large samples of six household or the biological father if the former is
race-sex groups. Compared to other national sur- missing. We capture an alternative measure of
veys, panel retention is excellent with 83 % of the social structure and stratication through the
sample retained at wave 15. respondents educational attainment. Although
For a study of the dynamics of BMI, we capi- there are a number of conceptualizations, we
talize on the record structure of the NLSY97 data treat attainment as a set of dummy variables
and its position in the history of population health indexing high school/GED, some college, a
in America. For the former, the multi-panel two-year degree, or a four-year college degree
record structure provides annual, repeated mea- or greater with the reference category being less
sures of self-reported height and weight, coupled than a high school degree. This allows us to cap-
a rich set of time-stable and time varying mea- ture a range of meaningful contrasts in education
sures. In the latter case, the obesity epidemic in as they relate to health and allows us to assess
the United States has had profound effects on the linearity or consider nonlinearities if apparent.
age structure of health liabilities. As Harris Finally, we capture key life course transitions
(2010) notes, numerous data, including studies and their effects on BMI. These include school
such as the National Longitudinal Study of enrolment, independent residence, employment,
Adolescent Health show unequivocally that obe- marriage, marital disruption, and parenthood.
sity is harbinger of both short-term and longer- Each of these is measured as a dichotomous time-
term chronic health problems and that a range of varying variable.
serious health problems (e.g., type II diabetes,
hypertension) are increasingly visible through
the early adult years. References
The key outcome of interest is annual mea-
sures of respondents body mass index score or Baltrus, P., Lynch, J., Everson-Rose, S., Raghunathan, T.,
& Kaplan, G. (2005). Race/ethnicity, life-course
BMI constructed from self-reported height and socioeconomic position, and body weight trajectories
weight. There are clear anomalies and apparent over 34 years: The Alameda County Study. American
coding errors in a small subset of person-periods Journal of Public Health, 95, 15951601.
and these introduce a number of extreme and Bauldry, S., Shanahan, M. J., Boardman, J. D., Miech, R. A.,
& Macmillan, R. (2012). A life course model of self-
unrealistic values for BMI. We deal with these by rated health through adolescence and young adult-
setting all values less than 12 and all values hood. Social Science & Medicine, 75(7), 13111320.
greater than 50 to be missing. Blau, P. M., & Duncan, O. D. (1967). The American occupa-
The central parameter of interest from which tional structure. New York: The Free Press.
Caspi, A., Elder, G. H., & Bem, D. J. (1987). Moving
we extrapolate interpretation is aging. This is a away from the world: Life-course patterns of shy chil-
measure that indexes the passage of time and dren. Developmental Psychology, 23, 308313.
The Logic and Practice of Growth Curve Analysis: Modeling Strategies for Life Course Dynamics 567

Conger, R. D., Conger, K. J., Elder, G. H., Lorenz, F. O., Ferraro, K. F., & Shippee, T. P. (2009). Aging and cumula-
Simons, R. L., & Whitbeck, L. B. (1992). A family pro- tive inequality: How does inequality get under the
cess model of economic hardship and adjustment of skin? The Gerontologist, 49, 333343.
early adolescent boys. Child Development, 63(3), Ferraro, K., Farmer, M., & Wybraniec, J. (1997). Health
526541. trajectories: Long-term dynamics among black and
Coale, A., & Demeny, P. (1966). Regional model life white adults. Journal of Health and Social Behavior,
tables and stable populations. Princeton: Princeton 38, 3854.
University Press. Furstenberg, F. (2003). Reections on the future of the life
Cook, T., Gruder, C., Hennigan, K., & Flay, B. (1979). course. In J. Mortimer & M. J. Shanahan (Eds.),
History of the sleeper effect: Some logical pitfalls in Handbook of the life course (pp. 661670). New York:
accepting the null hypothesis. Psychological Bulletin, Kluwer.
86, 662679. Guo, S., & Chumlea, W. (1999). Tracking of body mass
Cossrow, N., & Falkner, B. (2004). Race/ethnic issues in index in children in relation to overweight in adult-
obesity and obesity-related comorbidities. Journal of hood. The American Journal of Clinical Nutrition, 70,
Clinical Endocrinology and Metabolism, 89, 145s148s.
25902594. Guo, S., Huang, C., Maynard, L., Demerath, E., Towne,
Cummins, S., & Macintyre, S. (1999). The location of B., Chumlea, W., & Siervogel, R. (2000). Body mass
food stores in urban areas: A case study in Glasgow. index during childhood, adolescence and young adult-
British Food Journal, 101, 545553. hood in relation to adult overweight and adiposity:
Dannefer, D. (1987). Aging as intracohort differentiation: The Fels Longitudinal Study. International Journal of
Accentuation, the Matthew effect, and the life course. Obesity and Related Metabolic Disorders, 24,
Sociological Forum, 2, 211236. 16281635.
Dannefer, D. (2003). Cumulative advantage/disadvantage Harris, K. (2010). An integrative approach to health.
and the life course: Cross-fertilizing age and social Demography, 47, 122.
science theory. The Journals of Gerontology Series B: Hauser, R. M., & Warren, J. R. (1997). Socioeconomic
Psychological Sciences and Social Sciences, 58, indexes for occupations: A review, update, and cri-
S327S337. tique. Sociological Methodology, 27, 177298.
Dong, C., Sanchez, & Price, R. (2004). Relationship of Herd, P. (2006). Do functional health inequalities decrease
obesity to depression: A family-based study. in old age? Educational status and functional decline
International Journal of Obesity, 28, 790795. among the 19311941 birth cohort. Research on
Dowd, J., & Bengtson, V. (1978). Aging in minority popu- Aging, 28, 375392.
lations an examination of the double jeopardy hypoth- Hogan, D. P. (1978). The variable order of events in the life
esis. Journal of Gerontology, 33, 427436. course. American Sociological Review, 43, 573586.
Dupre, M. E. (2007). Educational differences in age- Horney, J., Osgood, D. W., & Marshall, I. H. (1995).
related patterns of disease: Reconsidering the cumula- Criminal careers in the short-term: Intra-individual
tive disadvantage and age-as-leveler hypotheses. variability in crime and its relation to local life circum-
Journal of Health and Social Behavior, 48, 115. stances. American Sociological Review, 60, 655673.
Elder, G. H., & Rockwell, R. C. (1979). The life-course and Khaodhiar, L., McCowen, K., & Blackburn, G. (1999).
human development: An ecological perspective. Obesity and its comorbid conditions. Clinical
International Journal of Behavioral Development, 2(1), Cornerstone, 2, 1731.
121. Laub, J. L., & Sampson, R. J. (1993). Turing points in the
Elder, G. H., Jr. (1985). Life course dynamics: Trajectories life course: Why change matters to the study of crime.
and transitions 19681980. Ithaca: Cornell University Criminology, 31, 301325.
Press. Laub, J., Nagin, D., & Sampson, R. (1998). Trajectories of
Elder, G. H., Jr. (1994). Time, human agency, and social change in criminal offending: Good marriages and the
change: Perspectives on the life course. Social desistance process. American Sociological Review, 60,
Psychology Quarterly, 57, 415. 225238.
Elder, G. H., Jr. (1998). The life course as developmental Laub, J. H., & Sampson, R. J. (2009). Shared beginnings,
theory. Child Development, 69, 112. divergent lives: Delinquent boys to age 70. Cambridge:
Elder, G. H., Jr., Johnson, M. K., & Crosnoe, R. (2003). Harvard University Press.
The emergence and development of life course theory. Ltkepohl, H. (2007). New introduction to multiple time
In J. Mortimer & M. J. Shanahan (Eds.), Handbook of series analysis. New York: Springer.
the life course (pp. 314). New York: Kluwer. Springer, Lynch, S. (2008). Race, socioeconomic status, and health
New York. in life-course perspective introduction to the special
Elder, G. H., Jr. (1999). Children of the great depression. issue. Research on Aging, 30, 127136.
Boulder: Westview Press. Macmillan, R., & Copher, R. (2005). Families in the life
Ferraro, K., & Kelley-Moore, J. (2003). Cumulative dis- course: Interdependency of roles, role congurations,
advantage and health: Long-term consequences of and pathways. Journal of Marriage and Family, 67,
obesity? American Sociological Review, 68, 707729. 858879.
568 R. Macmillan and F. Furstenberg

Macmillan, R., & Eliason, S. (2003). Characterizing the ORand, A. (1996). The precious and the precocious:
life course as role congurations and pathways: A Understanding cumulative disadvantage and cumula-
latent structure approach. In J. Mortimer & M. J. tive advantage over the life course. The Gerontologist,
Shanahan (Eds.), Handbook of the life course 36, 230238.
(pp. 529554). New York: Kluwer. Ogden, C. L., Carroll, M. D., Curtin, L. R., McDowell,
Marini, M. M. (1984). The order of events in the transition M. A., Tabak, C. J., & Flegal, K. M. (2006). Prevalence
to adulthood. Sociology of Education, 57, 6384. of overweight and obesity in the United States, 1999
Massey, D., & Denton, N. (1993). American apartheid: 2004. JAMA, 295, 15491555.
Segregation and the making of the underclass. Ogden, C. L., Carroll, M. D., Kit, B. K., & Flegal, K. M.
Cambridge, MA: Harvard University Press. (2014). Prevalence of childhood and adult obesity in
Masters, R., Reither, E., Powers, D., Yang, Y., Burger, A., the United States, 20112012. JAMA, 311, 806814.
& Link, B. (2013). The impact of obesity on US mor- Parsons, T. J., Power, C., Logan, S., & Summerbelt, C. D.
tality levels: The importance of age and cohort factors (1999). Childhood predictors of adult obesity: A
in population estimates. American Journal of Public systematic review. International Journal of Obesity,
Health, 103, 18951901. 23, S1S107.
McLeod, J. D., & Almazan, E. P. (2003). Connections Petersen, T. (1993). Recent advances in longitudinal
between childhood and adulthood. In J. Mortimer & methodology. Annual Review of Sociology, 19,
M. J. Shanahan (Eds.), Handbook of the life course 425454.
(pp. 391411). New York: Kluwer. Rabe-Hesketh, S., & Skrondal, A. (2008). Multilevel and
McLeod, J. D., & Shanahan, M. J. (1996). Trajectories of longitudinal modeling using Stata. College Station:
poverty and childrens mental health. Journal of Stata Press.
Health and Social Behavior, 37, 207220. Reilly, J. J., Armstrong, J., Dorosty, A. R., Emmett, P. M.,
McLeod, J. D., & Fettes, D. L. (2007). Trajectories of fail- Ness, A., Rogers, I., Steer, C., & Sherriff, A. (2005).
ure: The educational careers of children with mental Early life risk factors for obesity in childhood: Cohort
health problems. American Journal of Sociology, 113(3), study. BMJ Open, 330, 17.
653. Rindfuss, R. R., Swicegood, C. G., & Rosenfeld, R. A.
Meredith, W., & Tisak, J. (1990). Latent curve analysis. (1987). Disorder in the life course: How common and
Psychometrika, 55, 107122. does it matter? American Sociological Review, 52,
Merton, R. (1968). The Matthew effect in science. 785801.
Science, 159, 5663. Ross, C. E., & Wu, C. L. (1995). The links between educa-
Mirowsky, J., & Ross, C. E. (2007). Life course trajectories tion and health. American Sociological Review, 60,
of perceived control and their relationship to educa- 719745.
tion1. American Journal of Sociology, 112, 13391382. Sampson, R. J., & Laub, J. H. (1995). Crime in the making:
Modell, J., Furstenberg, F. F., Jr., & Hersberg, T. (1976). Pathways and turning points through life. Cambridge:
Social change and life course development in historical Harvard University Press.
perspective. Journal of Family History, 1, 18601975. Sewell, W. H., Jr. (1992). A theory of structure: Duality,
Monaghan, A. (2014, August 21). Number of young peo- agency, and transformation. American Journal of
ple classed as Neets falls as labour market improves. Sociology, 98, 129.
The Guardian. Shanahan, M. J. (2000). Pathways to adulthood in chang-
Moss, B. G., & Yeaton, W. H. (2011). Young childrens ing societies: Variability and mechanisms in life course
weight trajectories and associated risk factors: Results perspective. Annual Review of Sociology, 26, 667692.
from the Early Childhood Longitudinal Study-Birth Shanahan, M. J., & Macmillan, R. (2008). Biography and
Cohort. American Journal of Health Promotion, 25, the sociological imagination: Contexts and contingen-
190198. cies. New York: WW Norton.
Muthn, B., & Muthn, L. K. (2000). Integrating person- Singer, J. D., & Willett, J. B. (2003). Applied longitudinal
centered and variable-centered analyses: Growth mixture data analysis: Modeling change and event occurrence.
modeling with latent trajectory classes. Alcoholism: Oxford: Oxford University Press.
Clinical and Experimental Research, 24, 882891. Stearns, P. N. (2002). Fat history: Bodies and beauty in the
Nagin, D. S. (2009). Group-based modeling of develop- modern west. New York: NYU Press.
ment. Cambridge, MA: Harvard University Press. Swinburn, B. A., Sacks, G., Hall, K. D., McPherson, K.,
Nagin, D. S., Farrington, D. P., & Moftt, T. E. (1995). Finegood, D. T., Moodie, M. L., & Gortmaker, S. L.
Life-course trajectories of different types of offenders. (2011). The global obesity pandemic: Shaped by
Criminology, 33, 111139. global drivers and local environments. The Lancet,
Ng, M. M., et al. (2014). Global, regional, and national 378, 804814.
prevalence of overweight and obesity in children and U.S. Department of Health and Human Services. (2011).
adults during 19802013: A systematic analysis for Healthy weight, overweight, and obesity among U.S.
the Global Burden of Disease Study. The Lancet, 384, adults. Washington, DC: Centers for Disease Control
766781. and Prevention. National Center for Health Statistics.
The Logic and Practice of Growth Curve Analysis: Modeling Strategies for Life Course Dynamics 569

Wheaton, B., & Gotlib, I. H. (1997). Trajectories and turn- Whitaker, R., & Orzol, S. (2006). Obesity among US
ing points over the life course: Concepts and themes. urban preschool children: Relationships to race, eth-
Stress and adversity over the life course: Trajectories nicity, and socioeconomic status. Archives of
and turning points. In I. H. Gotlib & B. Wheaton Pediatrics & Adolescent Medicine, 160, 578584.
(Eds.), Stress and adversity over the life course: Wilson, W. J. (1987). The truly disadvantaged: The inner
Trajectories and turning points (pp. 125). Cambridge: city, the underclass, and public policy. Chicago:
Cambridge University Press. University of Chicago Press.
Three Generation Studies:
Methodological Challenges
and Promise

Terence P. Thornberry

1 Introduction have profound effects on their children. It is true


too that the behavior of a child can affect the par-
Human development takes place in the context of ents behavior. Bidirectional relationships
intertwined social relationships and the shape of between parent and child have been observed as
ones life course is inuenced by the shape of the early as toddlerhood and extend throughout the
life courses of others. Actors do not behave or life course. For example, child temperament and
decide as atoms outside a social context Their the manner in which children react to a parent
attempts at purposive action are instead embed- inuence parental behavior (Patterson et al. 1992;
ded in concrete, ongoing systems of social rela- Scaramella and Conger 2003). Clearly the lives
tions (Granovetter 1985:487). One of the most of parents and children are inextricably linked.
intimate and inuential of these relationships is Although parents and children have mutual,
that between parent and child. As Elder noted: reciprocal inuences on one another, for the
Each generation is bound to fateful decisions design of three generation studies we are espe-
and events in the others life course (1985:40). cially interested in the manner in which a parent
A parents catastrophic illness, sudden wealth or inuences the growth and development of their
sudden unemployment, divorce or remarriage, children.1 Indeed, there are two central questions
imprisonment, and so forth not only affect that addressed by three generation studies. First, what
individual, but ripple across to affect members of are the levels of continuity and discontinuity in
other generations. They often have profound psy- the behavior of interest across the generations?
chological and material effects on their children Second, what are the mediating and moderating
and, in some cases, their grandchildren. Similarly, inuences that help explain the manner in which
such events can also impact prior generations of the lives of parent and child are linked?
the familyparents and grandparents. The rst question concerns the extent to which
Generations are linked not only by fateful a parents involvement in a particular behavior
events but by mundane, run-of-the-mill events as inuences the likelihood that their child will also
well. A parents style of parenting, happiness or be involved in that same behavior at the same
depression, and partner relationships can also developmental stage. The typical hypothesis
proffered is one of continuity. That is, we expect
that children follow in the footsteps of their
T.P. Thornberry (*)
Department of Criminology and Criminal Justice,
1
University of Maryland, College Park, MD, USA There are exceptions to this, for example, intergenera-
e-mail: thornbet@umd.edu tional studies in gerontology, which are discussed below.

Springer International Publishing Switzerland 2016 571


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_25
572 T.P. Thornberry

parents such that parental behavior increases the ity and discontinuity? If parents and children are
odds that the child will also be involved in that similar with respect to a particular behavior, the
same behavior. But, as discussed below, there are central question is: Why? If the behavior of inter-
substantial levels of intergenerational discontinu- est is antisocialfor example, drug usewhat
ity, an understudied area, that also require are the mediating processes by which the parents
explanation. involvement in drug use increases the childs risk
The examination of intergenerational continu- of also being involved in drug use? There may be
ity and discontinuity in three generation studies direct inuences such as modeling the behavior,
(hereinafter 3G studies) is appropriate for and indirect inuences such as the negative conse-
has been used to investigate a wide range of quences of the parents earlier drug use on their
behaviors. They include conventional or proso- later life course development and on their ability
cial activities such as educational attainment, to effectively socialize their children, or most
work and career, marriage and intimate partner likely both direct and indirect effects. If the
relations, fertility, positive parenting styles, reli- behavior of interest is prosocialfor example,
gious afliation, political orientation, and so high educational attainmentthe parallel ques-
forth. They also include the investigation of prob- tion emerges. What are the processes by which
lem behaviors such as smoking, drug use, vio- the parents success is translated into the childs
lence, delinquency, health problems such as success? Correctly answering these questions
obesity, mental health problems such as depres- informs our theoretical understanding of the
sion, maladaptive forms of parenting such as long-term consequences of the parents behav-
child maltreatment, and others. Intergenerational iordrug use and educational attainment in these
continuities are not limited to individual-level examplesas well as enhancing our understand-
behaviors. For example, as Sharkey (2008) dem- ing of the early origins of these behaviors from
onstrates, there are substantial levels of continu- the perspective of the childs involvement.
ity in neighborhood context across multiple Understanding these processes provides vital
generations. information for the development of programs to
Intergenerational studies can address both enhance prosocial and to decrease antisocial
homotypic and heterotypic continuity in behav- outcomes.
ior. With homotypic continuity parent involve- In contrast, if the parent and child are dissimi-
ment in a particular behaviorfor example, drug lar with respect to the behavior of interestthat
usesignicantly increases the risk that the is, if there is intergenerational discontinuitythe
child will also be involved in that same behavior. central question is: What moderating inuences
With heterotypic continuity, parent involvement account for the discrepancy? There are two forms
in drug use signicantly increases the risk that of discontinuity as illustrated by the example of
the child will be involved in other problem drug use. In some cases, the child will be signi-
behaviors such as depression, violence, etc. The cantly less involved in drug use than the parent
same issues apply to prosocial behaviorsfor and the task is to explain intergenerational resil-
example, high educational attainmentwhere ience. That is, what are the protective factors that
the investigation can focus on either homotypic reduce the transfer of risk from the parents drug
or heterotypic continuity. Whichever type of con- use to their childs drug use? In other cases, the
tinuity is the focus, in a sense this rst core ques- child will use drugs when the parent did not and
tion of intergenerational study is descriptive: the task then is to account for de novo drug use.
How similar (or dissimilar) are parents and chil- Intergenerational studies, therefore, confront an
dren with respect to a particular behavior? interesting, and often difcult, challengeto
The second core question investigated in 3G account for both continuity and discontinuity in
studies is explanatory. What are the mediating the behavior of interest across the generations.
and moderating inuences that help account for Increasingly, these issues are being addressed
the observed levels of intergenerational continu- in prospective studies based on three adjacent
Three Generation Studies: Methodological Challenges and Promise 573

generations of the same families. These 3G stud- 1991; Farrington 1993; Fuller et al. 2003).
ies provide a powerful lens through which to Obviously, there are many aws with this type of
view the extent to which the lives of parent and design. First, the individual is reporting with
child are linked and to identify the life course imperfect knowledge on the behavior of their
processes by which those linkages are likely to parents and their children. Second, these studies
emerge. The remainder of this chapter focuses on have to rely on retrospective data for at least one
the methodological components of 3G studies. In generation, and often on retrospective data with a
doing so it rst describes the core design charac- very long recall period. Third, the reporters cur-
teristics of 3G studies characteristics that, when rent behavior can inuence their views of parent
combined, distinguish them from traditional lon- and child behavior; parents who are highly
gitudinal studies. Second, it discusses a set of aggressive or who have serious alcohol problems,
methodological challenges that are unique, or at for example, may well project that behavior onto
least peculiar, to 3G studies, and some possible their perceptions of behavior in the other genera-
solutions to them. Third, in keeping with one of tions. If intergenerational studies are going to
the central themes of the Handbook, I offer sev- contribute to our understanding of the origins and
eral suggestions for future research. Throughout, course of behavior, they need to have greater con-
I will illustrate some of these points by referring ceptual and methodological integrity.
to the Rochester Intergenerational Study, a long- With that in mind, I would like to offer a de-
term, 3G study of antisocial behavior that began nition, or at least a description, of what a scien-
in 1999 (Thornberry 2009). A brief description of tically credible intergenerational study would
the design of this study is presented in Appendix entail in order to validly answer the two core
A, both to describe what I think is a reasonably questions of intergenerational study. In doing so,
typical 3G study and to provide the methodologi- I use the notational system of Generation 1 (G1)
cal detail needed for later discussions. Also, to refer to the grandparent generation, Generation
much of the presentation in this chapter is inu- 2 (G2) to refer to the parent generation, and
enced by my own interest in the study of adoles- Generation 3 (G3) to refer to the child genera-
cent problem behaviors but I believe the core tion. Typically, the G3 child is the focal partici-
argument extends to other patterns of behavior pant in an intergenerational study and it is that
such as educational attainment, parenting, and so persons behavior that we are attempting to
forth. explain. In my view, an intergenerational study
should, at a minimum, meet four central design
criteria.
2 Three Generation Studies First, a 3G study should have prospective data
of Behavior on a G2 parents and on a G3 childs involvement
in the behavior of interest. Prospective data on a
Although discussed in the literature for quite G1 grandparents behavior is also valuable (see
some time, scientically acceptable 3G studies below), but it is not necessary to investigate inter-
are relatively new to the social sciences and, as a generational continuity and discontinuity
result, there is a lack of agreement about what between G2 and G3, the core issue of intergen-
constitutes an intergenerational study of a partic- erational study. There are, of course, variations
ular behavior. A number of early intergenera- on this aspect of an intergenerational design. For
tional studies used cross-sectional designs and example, it can be extended to additional genera-
asked a single respondent to report on their own tions as in the remarkable Trans-5 Study, a pro-
behavior as well as the behavior of the preceding spective study of ve generations of the same
and following generations; for example, The families using ofcial records in the Netherlands
intergenerational transmission of aggression (Bijleveld and Wijkman 2009). Also, some
across three generations (Doumas et al. 1994) studies combine prospective data on G2 with
used this approach (see also Alexander et al. cross-sectional, or largely cross-sectional, data
574 T.P. Thornberry

on G3 to address intergenerational themes In combination, these four design criteria pro-


(Kaplan and Tolle 2006; Knight et al. 2014; vided a template for considering what constitutes
Smith and Farrington 2004). Nevertheless, as dis- a 3G study, one with the potential to provide
cussed below there are distinct advantages to information that extends beyond what is typically
having prospective data on both G2 and G3 in learned in a longitudinal design. I recognize that
order to fully address the two core questions of these methodological criteria are not unique to
intergenerational study: identifying and account- intergenerational studies; for example, in any
ing for continuity and discontinuity. longitudinal study prospective data are generally
Second, the measures on each generations viewed as psychometrically preferable to retro-
involvement in the behavior of interest should be spective data. It is the combination of these four
as independent as possible and, whenever possi- criteria that provide the dening qualities of 3G
ble, based on different reporters. For example, studies. That is, when all four are present a study
each could self-report on their own behavior. If can adequately address the core questions of
the same reporter provides information on both intergenerational inquirydescribing the level
generations, the independence of the independent of continuity and discontinuity as well as identify
and dependent variables is compromised. Also, mediating and moderating inuences to help
using different reporters minimizes some of the explain continuity and discontinuity. When any
methodological problems mentioned earlier, such one of these criteria is absent, the validity of the
as reliance on retrospective data for at least one study in addressing these questions is compro-
of the generations. mised. For example, if the study does not meet
The third design criterion is the key dening the fourth criterioni.e., it does not have data on
element of an intergenerational study and the G2 life course developmentit would be able to
one that most clearly separates intergenerational determine the level of continuity in a particular
studies from traditional longitudinal studies. behavior but it would be severely challenged to
There is a growing consensus (Capaldi et al. explain how that similarity occurs, that is, it
2012; Kaplan and Tolle 2006; Knight et al. would be difcult to identify mediating pro-
2014) that intergenerational studies should have cesses. This challenge is faced by many studies
comparable measures of G2 and G3 behavior that rely entirely on ofcial records (e.g., van de
covering the same ages or the same develop- Rakt et al. 2008; Frisell et al. 2011). In contrast, a
mental stages. I return to a discussion of this study that does not meet the second criterioni.e.,
issue and its centrality to intergenerational it relies on a single reporter for both genera-
study, when I compare intergenerational and tionsmay have extensive information on medi-
longitudinal studies. ating processes but its ability to validly measure
Fourth, intergenerational studies should have the behavior of interest in both generations, and
detailed prospective data on G2 life-course devel- therefore estimate the level of continuity is com-
opment. This facilitates the identication of promised as we noted earlier (Doumas et al.
mediating and moderating processes that account 1994). When all four of these criteria are met,
both for intergenerational continuity and discon- however, a 3G study is in strong position both to
tinuity in the behavior of interest as well as the identify levels of continuity and to explore medi-
investigation of gene-by-environment interac- ating and moderating processes.
tions if genetic data are also available. Essentially, The saliency of these criteria in dening a 3G
information on G2s movement along major life- study may vary somewhat across behaviors of
course trajectories (e.g., work, family formation, interest and, as noted above, my own view is no
etc.) is required. Absent such developmental doubt shaped by my current research agenda. For
data, levels of similarity in the behavior could be example, if the behavior of interest is relatively
established but there would be little ability to staticsuch as educational attainment or age at
identify the mediating processes that help explain rst marriagethe use of prospective designs may
why that similarity occurs. not be as essential. For behaviors that are more
Three Generation Studies: Methodological Challenges and Promise 575

changeable, howeversuch as school engage- behavior, the behavior of interest in that study
ment or the quality of intimate relationships (Thornberry et al. 2003). G1 parents varied con-
they are quite salient. An important topic for siderably in age, from their late 20s to their late
future research is to continue this denitional dis- 50s at the beginning of the study. The Rochester
cussion to arrive at a broader consensus as to Intergenerational Study then sampled the oldest
what constitutes an intergenerational study and biological child age 2 or older of each G2 partici-
what distinguishes them from other types of pant, starting in 1999. Obviously, the G2 partici-
study designs, for example, traditional longitudi- pants became parents over a very wide age span
nal studies (see below). (currently ages 1439) and G3 participants cur-
rently range in age from 2 to 28.

2.1 Origins of Intergenerational


Studies 2.2 Intergenerational Versus
Longitudinal Designs
Three generation studies typically begin as exten-
sionsas add-onsto an existing longitudinal Although many 3G studies are themselves pro-
study that followed an original G2 participant spective extensions of an earlier study, the com-
over time. In some cases the studies also followed bination of the four core design characteristics
a G1 participant. Thus, intergenerational studies described earlier distinguish them from tradi-
are rarely planned from the outset, largely tional longitudinal studies. I will use a behavior
because of the length of time that would have to that occurs during adolescence to illustrate these
elapse between the time the study began and the relationships. The behavior could be prosocial
time the G3 child exhibits an outcome of (e.g., attitudes toward school engagement and
interest. success) or antisocial (e.g., drug use). Figure 1
This observation about typical origins has two presents an overview of the intergenerational
important methodological consequences. First, relationships in behavior that can be examined
since the original longitudinal study was designed based on three types of research designs. The rst
to study something other than intergenerational (Fig. 1a) focuses on traditional longitudinal
linkages, 3G studies inherit a set of design deci- designs that examine the G1 and G2 generations.
sions about sampling, measurement, and so forth. The particular focus of longitudinal studies, at
While those design decisions may have been least when using prospective data, is on the con-
ideal for the initial purpose, they may not be ideal current relationship between the parents adult
for the investigation of intergenerational issues. behavior and the childs adolescent behavior.
In a sense, 3G studies are akin to secondary data This is indicated by the path labeled A.
analysis projects; they have to build on and adapt Intergenerational designs shift the focus of the
to an existing foundation of already collected research in two important ways as represented in
data. Fig. 1b. First, attention shifts from G1 and G2 to
Second, and more importantly, the origins of G2 and G3. The focal participants are the chil-
3G studies create a very uneven age distribution dren in the third generation and it is their behav-
across the three generations. Researchers only ior that we are attempting to explain. Similar to
have control over the selection of the initial G2 traditional longitudinal studies, 3G studies can
participants and, therefore, over their ages (or examine the concurrent impact of the G2 parents
birth cohorts). The original researchers dene the adult behavior on the G3 childs adolescent
population of interest and decide which G2 par- behavior (Path A). In addition, however, 3G stud-
ticipants to sample. For example, the original ies can uniquely estimate Path B, the impact of a
Rochester Youth Development Study selected parents adolescent behavior on their childs
adolescents in the seventh and eighth grade adolescent behavior. That is, they can look at con-
because of the typical age of onset of delinquent tinuity (or discontinuity) between the generations
576 T.P. Thornberry

a Longitudinal Design
Adult Behavior
G1
A

Adolescent Behavior
G2

b Intergenerational Design
Adolescent Behavior Adult Behavior
G2
B A

G3 Adolescent Behavior

c Intergenerational Extension
Adolescent Behavior Adult Behavior
G2
C B A

Childhood Behavior Adolescent Behavior


G3

Fig. 1 Intergenerational relationships for behaviors based on longitudinal and intergenerational designs

for the same behavior at the same developmental turn, have strong and proximal impacts on the
stage and, importantly, do so using prospective child (Capaldi et al. 2003; Conger et al. 2003).
data. In addition, as indicated in Fig. 1c, intergen- Although all of those concurrent relationships are
erational studies can be extended to other devel- important, results about Path A do not inform us
opmental stages. For example, they can examine about the direction, signicance, or magnitude of
the impact of G2 adolescent behavior on G3 either Path B or Path C. For example, with respect
childhood behavior (Path C). More generally, to antisocial behavior, if we only have data on
intergenerational studies can look at relationships Path A we may well underestimate the true inter-
between G2 and G3 at various, but clearly generational effect since we would be assessing
dened, developmental stages. the parents antisocial behavior well into their
This does not imply that results from longitu- adult years when its prevalence and frequency
dinal studies, as illustrated in Fig. 1a, are unim- drop precipitously and when there is less vari-
portant in explaining these behaviors. Whatever ability in the behavior. As a result, it is quite pos-
the behavior of interest is, the parents adult sible that the Path A relationship would be weaker
involvement in that behavior is likely to have than the total intergenerational effect. In fact,
important direct and indirect effects on their ndings from several intergenerational studies of
childs involvement in that behavior. For exam- antisocial behavior are consistent with this con-
ple, the parents behavior can inuences the over- tention (Huesmann et al. 1984; Thornberry et al.
all family environment in which the child is 2006; Wu and Kandel 1995). Future research
raised, it can directly model the behavior of inter- should systematically investigate the extent to
est for the child, and it can impact, either posi- which estimates of continuity differ when based
tively or negatively, parenting styles which, in on Path A versus Path B and whether or not
Three Generation Studies: Methodological Challenges and Promise 577

different mediators are identied in these two For example, there are a range of counterfactual
approaches to research. Also, these questions approaches that help in this regard (Pearl 2009;
should be investigated for different types of Rosenbaum and Rubin 1983) and they have been
behaviors. used in intergenerational investigations (e.g.,
Moreover, focusing on Path B draws our atten- Thornberry and Henry 2013). Nevertheless, they
tion to new and different pathways that link par- are not the equivalent of experimental designs
ents and children with respect to behavior. The and their ability to assess causality has been
mediators of the Path B linkage, often separated debated. It is beyond the scope of this chapter to
by 2025 years, are likely to be quite different resolve this issue but it is an important one for
from the mediators associated with concurrent future research to address, not only for 3G
similarity in behavior across the generations designs but, more generally, for all non-
(Path A). Intergenerational mediators focus atten- experimental studies.
tion on long-term developmental processes in
addition to the more immediate family context
(Thornberry 2005, 2009). Identifying these medi- 2.3 Grandparental Inuences
ators can offer new and different insights into the
origins of behavior that complement those of tra- To this point I have focused entirely on the G2 and
ditional longitudinal studies. The results of these G3 generations. G1 grandparents, however, also
studies will also lead to new, and earlier, contribute substantially to our understanding of
approaches to the development of programs. intergenerational processes. Figure 2 presents a
Indeed, they have the potential of identifying very simplied view of grandparental inuences.
intervention targets to increase prosocial out- The right hand, vertical portion represents the
comes and reduce antisocial outcomes that exist types of relationships that can be investigated in
even before the child is born. longitudinal studies. For example, we can examine
Overall, in my view there are four dening how G2 parenting inuences G3 behavior, in this
design features that constitute a 3G study. They case school engagement, and we can investigate
include: prospective data on G2 and G3 behavior; how concurrent factors, represented here by G2
independent measures in each generation; mea- depressive symptomatology, inuence those par-
sures of the behavior of interest at comparable enting behaviors. There is a robust literature dem-
ages or developmental stages; and data on long- onstrating that experiencing depressive symptoms
term mediating processes. Studies with all of reduces effective parenting (Conger et al. 2010).
these features are better able to address the cen- That type of analysis, of course, can be extended to
tral questions of intergenerational study. In this other concurrent inuences on G2 parenting
regard they differ from traditional longitudinal behaviors, both positive (e.g., partner satisfaction
studies and, therefore, their results enable us to or stable employment) and negative (e.g., ongoing
extend our understanding of the origins and substance use or marital conict).
course of antisocial behavior. While we know that parenting styles have a
In assessing these core questions, it would major inuence on a childs behavior, we know
also be ideal to identify causal relationships. But, relatively little about the origins of parenting
as is true of all non-experimental designs, the styles from longitudinal studies (Belsky 1993;
issue of identifying causal relationships is dif- Conger et al. 2010). By incorporating informa-
cult in the context of intergenerational studies. tion on G1 behavior patterns, however, we can
Typically, we simply observe or measure the greatly extend our investigation of this issue. For
independent variable (the parents and perhaps example, we can examine how parenting styles
grandparents behavior), the dependent variable are reproduced across the generations (see
(the childs behavior) as well as the mediators. As Fig. 2). That is, we can examine the extent to
a result, 3G designs are no more or less able to which G2s parenting style exhibited towards G3
assess causality than other observational studies. reects the parenting styles to which G2 parents
578 T.P. Thornberry

G2 Depressive
Symptomatology
_
_

+
G1 Effective Parenting G2 Effective Parenting
+

+ + +

G2 School Engagement + G3 School Engagement


+

Fig. 2 Grandparental inuences in intergenerational designs

were exposed when they themselves were chil- G1 on G3. For example, Thornberry et al. (2006)
dren. Conger and colleagues (Conger et al. 2003; found that G1s earlier drug use directly increased
Neppl et al. 2009; Scaramella and Conger 2003) the likelihood of G3 drug use (independent of the
have shown that between G1 and G2 there is effect of G2 drug use) as long as G1 had ongoing
homotypic continuity for harsh parenting and for contact with G3. Indirect effects are illustrated in
positive parenting; that is, exposure to harsh par- Fig. 2. First, G1 parenting style is hypothesized to
enting leads to signicantly higher levels of harsh indirectly inuence G3 school engagement via its
parenting and exposure to positive parenting impact on G2 adolescent school engagement
leads to signicantly higher levels of positive which itself has both direct and indirect (via G2
parenting. At the same time, they did not nd het- parenting) effects on the G3 outcome. Another
erotypic continuity from one type of parenting to indirect pathway of G1 parenting operates through
the other; that is, exposure to harsh parenting did G2 depressive symptomatology. All of these rela-
not signicantly reduce positive parenting nor tionships enhance our understanding of the G3
did exposure to positive parenting signicantly childs behavior by eshing out our understanding
reduce harsh parenting. Furthermore, Neppl et al. of the long-term origins of what we know to be
(2009) also show that the mediators of these powerful, proximal causes of that behavior.
intergenerational linkages are different for posi- Although understanding the contributions that
tive parenting versus harsh parenting. Thornberry grandparents make to an understanding of inter-
(2009) reports similar ndings that vary by G2 generational inuences is an important contribu-
gender, with harsh parenting being more impor- tion of 3G studies, in the interest space I will focus
tant for G2 fathers and positive parenting being primarily on relationships between G2 and G3 in
more important for G2 mothers. If we can exam- the remainder of this chapter.
ine the reproduction of parenting styles across
the G1 and G2 generations, we can also examine
other factors that can have important indirect 2.4 Variations in Three
inuences on G3 behavior. They include the Generation Studies
reproduction of poverty and structural adversity,
neighborhood context, school performance, fer- Thus far I have focused on prospective 3G stud-
tility patterns, psychopathology, marital conict, ies that begin from an ongoing longitudinal study
and many others. of G2 in discussing core design elements. This
Intergenerational studies with this design type of design enables one to address what I con-
can also examine both direct and indirect effects of sider to be the basic issues of intergenerational
Three Generation Studies: Methodological Challenges and Promise 579

studyexamining and explaining continuity and tremendous amount of natural attrition in the
discontinuity in a particular behavior at the same grandparent generation, from 516 participants to
developmental stage. There are related designs just 61. As with other 3G studies the LSOG,
that can also address this core issue. with a fully elaborated generation-sequential
In a traditional longitudinal study, if the G2 design, allows comparisons of sets of aging par-
participant is followed far enough into the life ents and children at the same stage of life during
course, for example into the adult years, they different historical periods (Bengston 1971 ff;
would overlap in age with G1 when they were emphasis added).
included in the original study as an adult. In this Clearly, there are variations of the basic 3G
case Path B in Fig. 1 would represent continuity design with which to address the two core inter-
in the behavior during the adult years but the age- generational issues of describing and explaining
specic rates for G1 and for G2 would still be continuity and discontinuity in behavior. In this
separated by about 2030 years. That is, it would chapter, I focus on the 3G design described ear-
not reect a concurrent relationship but an inter- lier which emanates from an ongoing longitudi-
generational one. In general, intergenerational nal study of G2 participants; however, I will
continuity in adult behaviors can be studied in comment on variations in this approach at appro-
designs of this nature, for example, studies of priate points.
intergenerational continuity in criminal behavior
(e.g., Besemer and Farrington 2012; van de Rakt
et al. 2008). Conger and colleagues have exam- 3 Methodological Issues
ined continuity and discontinuity in parenting Related to Three Generation
styles by comparing G1 parenting towards G2 Studies
when they were children and adolescents with G2
parenting towards G3 when they were children Given the typical origins of 3G studies as exten-
and adolescents (Conger et al. 2003; Neppl et al. sions of existing longitudinal studies, researchers
2009; Scaramella and Conger 2003). Here also, face a number of design decisions and issues that
the focus of investigation is on an intergenera- create particular challenges for the investigation
tional relationship, not a concurrent one. of intergenerational themes. The remainder of
Another variation in 3G study designs occurs this chapter identies and discusses various ways
in the eld of gerontology with its focus on the of responding to them. These design issues are
investigation of family relations with respect to themselves interrelated and the decisions made to
aging. These projects are, in a sense, the mirror resolve one has implications for, and often con-
image of the 3G studies described above which strains, decisions for the others. I will illustrate
focus on childhood and adolescent behaviors. some of these issues using the Rochester
Bengstons Longitudinal Study of Generations Intergenerational Study (Thornberry 2009).
(LSOG) is a classic example of this approach
(Bengston et al. 2002). At its initiation, the focal
participants were a sample of G1 grandparents 3.1 Additional Design Issues
who were eligible for the study if they were mar-
ried, had children, and had at least one grandchild Earlier I presented a discussion of what I consider
between the ages of 16 and 26. Thus, the LSOG to be the core dening characteristics of a 3G
acquired three generations of the same families study. In addition to those dening characteristics,
by starting with grandparents and then selecting there are other design elements that, when present,
their children and their grandchildren, following greatly extend the power and scope of 3G studies
all three generations across time, from 1971 to to address issues of continuity and discontinuity.
1997, a span of 26 years. Of course, there was a In this section, I discuss two of them.
580 T.P. Thornberry

3.1.1 Inclusion of Both Mothers multiple mediating pathways that help account
and Fathers for IG continuity for fathers. Adjustment to work
It is not essential for a 3G study to have a G2 and stable employment appear to be particularly
sample that includes both G2 mothers and G2 important (Conger et al. 1993; Moftt et al.
fathers but having both greatly extends the reach 2002). Also, ongoing contact and involvement
of 3G investigation. To be clear, we are not refer- between the father and child is likely to play a
ring to mother-father pairs in this discussion (that strong role in moderating IG continuity and pro-
issue is dealt with later in the chapter). Here we cesses for fathers, but not for mothers. Father
refer to studies that have an initial adolescent absence is generally detrimental to child develop-
sample of males (who become G2 fathers) and ment (Carlson 2006; Magnuson and Berger
females (who become G2 mothers). In other 2009), although the absence of highly antisocial
words, the 3G study includes a sample of G2 fathers may be benecial (Jaffee et al. 2003).
fathers, their G3 children (and perhaps the childs Father absence also has a variety of indirect
mother) as well as G2 mothers, their G3 children effects on child development via economic stress
(and perhaps the childs father). When both G2 (Martin et al. 2010), the lack of a co-parent
fathers and G2 mothers are included a number of (Hetherington and Kelly 2002), and marital con-
important issues can be addressed. ict (Kelly 2000).
First, what we know about father effects, rela- The overall point, of course, is that a sample
tive to what we know about mother effects, in that includes both G2 mothers and G2 fathers
explaining a childs behavior is relatively sparse allows us to examine these issues empirically.
(Phares et al. 2005). Indeed, most of what we That is, we can examine the extent to which inter-
know about parental inuences on child devel- generational inuences are similar or different
opment is, in fact, an understanding of maternal for mothers and fathers. It may be, for example,
inuences. G2 samples that include both mothers that hypothesized mediators, like parental depres-
and fathers can therefore contribute to our inves- sion or work related stress, are equally damaging
tigation of how fathers affect their children. in the families of G2 fathers and G2 mothers
Second, and perhaps more importantly, since the (Thornberry et al. 2009a) but that can only be
G2 fathers and mothers in a 3G study are drawn determined by the direct comparison of mothers
from the same population and followed over time and fathers. Direct comparisons of maternal and
in identical ways, 3G studies that include both paternal inuences in these 3G designs should be
G2 mothers and G2 fathers can uniquely enable a high priority for future research.
direct comparisons of the similarities and differ-
ences in the ways in which fathers and mothers 3.1.2 Other Caregivers
impact their children. Intergenerational studies often have a curious
For example, Thornberry (2009) hypothesizes quality of operating as if each child only had one
that the strength of some relationships will vary parentthe G2 participant in the initial longitu-
by G2 gender. In American culture, mothers are dinal study. It is also quite important to include
cast as the primary parent and norms are the G3 childs other parent or other primary care-
stricter on the centrality and endurance of the giver and to collect prospective data from that
mother-child dyad (Doherty et al. 1998). As a person. It is particularly important in the case of
result, family factors could be more important G2 fathers, as the other caregiver is overwhelm-
mediators of IG continuity for mothers than for ingly G3s biological mother who is usually the
fathers (Martin et al. 2010). In contrast, the childs primary caregiver. In the case of G2 moth-
impact of fathers appears to be more diverse. ers, the other caregiver is often G3s biological
Since fathers play numerous roles within the father but in many cases it is a grandmother,
family, as spouses, protectors, models, moral stepparent, or another relative (see Appendix A),
guides, teachers, breadwinners (Lamb and and the likelihood of these relationships occur-
Tamis-LeMonda 2004:10), there are likely to be ring varies considerably by the composition of
Three Generation Studies: Methodological Challenges and Promise 581

the sample in terms of such factors as SES, race/ inuences. Collecting genotypic data from the
ethnicity, urbanicity, and so forth. full triothe child and both biological parents
Incorporating the other primary caregiver into enhances our ability to identify genetic risk and
a 3G study helps intergenerational analyses in protective factors as well as to investigate gene-
several ways. First, it is useful to be able to iden- environment interactions in accounting for vari-
tify major concurrent inuences on the childs ous behaviors of interest. As described in detail
behavior, especially by the primary caregiver, in elsewhere (Laird and Lange 2006), family-based
order to better interpret how intergenerational designs that include both biological parents are
inuences operate. For example, the G2 parents robust to population stratication in which small
inuence on the childs behavior may be primar- allele frequency differences across socially
ily indirect, operating through the other caregiver, dened racial and ethnic groups lead to false pos-
effects that could be missed if only the G2 parent itives in genetic epidemiological research.
is included in the design. Specically, these models enable researchers to
Second, the role of the other caregiver may be evaluate the association between a risk allele and
particularly important in accounting for intergen- some behavioral trait but condition the associa-
erational resilience. One reason why a G2 parent tion on the probability that a particular individual
with a substantial history of problem behaviors will have inherited the risk allele given their par-
may have a G3 child who avoids those behaviors ents genotypes. This eliminates the risk of spuri-
could be the presence of another caregiver who is ous false positives. Equally important, the
generally more prosocial than G2. For example, availability of family trios also allows more pre-
more prosocial other caregivers are more likely to cise statistical estimates because within and
make orderly transitions to adulthood, experience between family information can be used to reduce
lower levels of stressors and structural adversity, the burden of multiple testing.
and have effective parenting styles, all of which
may reduce G3 problem behavior. They can also
be a source of control and social capital to 3.2 Sampling Issues
increase G2 prosocial opportunities and parent-
ing skills (Conger et al. 2010). There is some evi- 3.2.1 Sample Size
dence of protective effects from prosocial other and Statistical Power
caregivers in accounting for intergenerational Three generation studies that develop from an
resilience (Dong and Krohn 2014; Lovegrove original longitudinal study of G2 participants
2010). typically have a relatively small number of fami-
Third, the inclusion of data from the other lies and G3 participants. Although there are some
caregiver reduces the likelihood of having to rely exceptions that start with a large G2 sample such
on a common reporter in assessing intergenera- as the AddHealth study (Harris et al. 2009), most
tional relationships. If parental reports of G3 longitudinal studies have samples that range from
behavior are collected from the other caregiver, several hundred families (Besemer and Farrington
then the other caregiver report can be used to 2012; Capaldi et al. 2012) to a few thousand
assess the relationship between G2 behavior and (Knight et al. 2014). Following the G2 partici-
G3 behavior, rather than relying on the G2 report pants over time until they become parents inevi-
for both. This can also help in the measurement tably involves some degree of attrition and only
of mediating and moderating variables. For those who remain in the study form the pool for
example, in the RIGS, in addition to self reports, the new intergenerational study. Of those who
we collect G2s assessment of the other caregiv- remain, only some become parents; estimates
ers parenting style and the other caregivers indicate that approximately 16 % of women and
assessment of G2s parenting style. 24 % males will remain childless (Martinez et al.
Finally, including the G3 childs other biologi- 2012). Finally, of those who become parents it is
cal parent enhances the investigation of genetic inevitable that not all the G3 children and their
582 T.P. Thornberry

G2 caretakers agree to participate in the 3G study. 3.2.2 G3 Age Heterogeneity


This is particularly the case for G2 fathers. The issue of sample size is further complicated
Fathers have varying degrees of involvement by G3 age heterogeneity. As noted earlier there is
with their children, ranging from being the pri- no control over when G3 participants are born
mary caregiver to quite literally not knowing and they typically span a considerable age range.
where the child is. When fathers have little con- The G3 children in the Rochester study currently
tact with the child, researchers need to locate and range from toddlers to young adults. Thus, the
elicit cooperation from the childs primary care- total sample size, which is often already small, is
giver, typically the biological mother, to enroll also divided across many birth cohorts. The most
G3 in the study. That is often difcult, especially feasible way for 3G studies to gain adequate sta-
if there is conict between the parents. This is tistical power to test theoretical models is to clus-
typically not the case with G2 mothers as they are ter G3 participants of the same ages, drawn from
almost universally the childs primary caregiver. different birth cohorts, into synthetic cohorts of
Thus, there is an inevitable, and often substan- adequate size that represent meaningful segments
tial, erosion of families as one moves to the third of the overall life course, for example, late ado-
generation. The Rochester study began with 1,000 lescence. In other words, 3G studies are almost
G2 participants, but it is likely to have a maximum inevitably a special case of cohort sequential or
of approximately 530 G3 participants despite hav- accelerated longitudinal designs (Prinzie and
ing high cooperation and retention rates. As a Onghena 2005).
result, statistical power is always an issue that Combining different cohorts, however, is not
needs to be carefully assessed in 3G studies. While necessarily straightforward. First, G3 partici-
the researcher had control over the initial G2 sam- pants of similar ages will be drawn from different
ple size and presumably insured adequate power birth cohorts that may be separated by 20 or more
to estimate theoretical models of interest, there is years. Therefore, cohort and period effects need
no such guarantee for the G3 sample. Nevertheless, to be considered. Second, some G3 participants
the purpose of the 3G study is to estimate similar, will be the children of teenage parents and others
and often more complicated, models but inevita- will be born to parents who become parents for
bly with a smaller sample. the rst time in their early 40s. There is a robust
Relatedly, if the sample of G2 participants is literature indicating that the timing of rst births,
properly selected, that is, if it is a probability especially teen parenthood versus a more age-
sample from a well-dened population, it can be normative onset of parenthood, affects child
considered to be representative of that popula- development (Hofferth and Goldscheider 2010;
tion. It is less clear that either the G1 or the G3 Nagin et al. 1997). Finally, at the selected ages,
participants would constitute a representative some G3 participants will be interviewed during
sample. For example, G1 parents are not neces- the earlier waves of the 3G study while others
sarily representative of all parents in their genera- will be interviewed toward the end of the study.
tion. Similarly, the G3 children do not necessarily Panel bias or testing effects may be embedded in
represent the children of their generation nor the longitudinal data (Thornberry 1989; Torche
even the children of the full G2 cohort given the et al. 2012), so that too will need to be
inevitable loss of families through attrition and considered.
noncooperation. The central point is that it is The most common response to the necessity
likely that only the generation initially sampled at of combining multiple birth cohorts in one
the beginning of the study (G2 in the case of the analysis is to control either for G3 age (e.g.,
Rochester study and G1 in the case of the Bailey et al. 2013) or for G2 age at the birth of G3
Bengston study) can be considered a representa- (e.g., Capaldi et al. 2012; Thornberry et al. 2009a,
tive sample. The implications of this issue are not b). A more formal way of testing the appropriate-
well discussed in the literature and should be ness of combining multiple cohorts is to estimate
more fully investigated. a multilevel model to examine the extent to which
Three Generation Studies: Methodological Challenges and Promise 583

constructs of interest vary as a function of birth changes with the addition of pertinent control
cohort.2 The variance in the construct at each variables that we consistently include in our
level may be estimated and then used to calculate models: G2 gender and race/ethnicity, G2 birth
the intraclass correlation (ICC), which quanties year, G1 education, the arrest rate and poverty
the proportion of variance that exists at each of rate of the neighborhood that G2 grew up in, G3
the upper levels. If we can safely combine cohorts age, G3 gender, and family poverty at each year
in an analysis, we would expect to see very small the outcome was measured. In this model, the
ICCs associated with birth cohort indicating that ICC at Level 3 was reduced to 1.4 %, indicating
little variance is accounted for by differences that the birth cohort differences can be readily
across cohorts. explained by a carefully thought out set of
To illustrate this approach we used data from covariates.
the RIGS to examine three important outcomes in In this illustration there is compelling evi-
studies of adolescent problem behaviors. In par- dence for the appropriateness of combining mul-
ticular, we examine G3 involvement in external- tiple birth cohorts for analysis. It would be wise
izing problems, internalizing problems, and for 3G researchers generally to assess ICCs as
delinquent behavior as reported by the parent they start new analyses that combine participants
using the Child Behavior Checklist, a well- from different birth cohorts in order to determine
known instrument for assessing childhood prob- the appropriateness of doing so and to identify
lem behaviors (Achenbach 1991). We selected pertinent control variables to include in the base
these measures because in the RIGS they have model to minimize the conditional ICCs.
been measured a sufcient number of times
from ages 4 to 17and for a sufcient number of 3.2.3 Selection of G3 Participants
birth cohortsbirth years range from 1986 to G2 participants obviously can have multiple chil-
2006to constitute a rigorous test of whether or dren. A fundamental design question, therefore,
not there are likely to be cohort differences. is which G3 children to include in the intergen-
We started with an unconditional model with erational sample. Ultimately, the answer to this
three levels: repeated measures of G3 behavior question depends upon the behavior of interest
(Level 1), nested in G3 (Level 2), nested in birth and the specic theoretical questions that are
cohort (Level 3). The Level 3 ICC is of particular being posed about that behavior. (It also depends
interest when determining the extent to which G3 on such practical considerations as the available
are non-independent as a function of when they budget and the interests of funding agencies.)
were born (and given the RIGS design, the age of Nevertheless, there are a number of general
G2 when G3 was born). The unconditional mod- dimensions that should be taken into account in
els produced the following ICCs: Delinquent making this decision and existing 3G studies dif-
Behavior (Level 1 = 49.6 %, Level 2 = 44.7 %, fer on the choices that have been made.
Level 3 = 5.7 %), Externalizing Behavior (Level Some 3G studies only include biological chil-
1 = 37.0 %, Level 2 = 61.7 %, Level 3 = 1.3 %), dren (e.g., Bailey et al. 2006) while others include
Internalizing Behavior (Level 1 = 39.0 %, Level a mix of biological, step, adopted, and foster chil-
2 = 60.0 %, Level 3 = 0.9). Very little of the vari- dren (e.g., Kaplan and Tolle 2006). Some studies
ance is due to G3 birth year differences, particu- only include G3 children if they live with or are
larly for externalizing and internalizing behavior. being reared by the G2 parent (at least at the
The Level 3 ICC for delinquency is a bit higher beginning of the study), while others include all
(close to 6 %). We therefore examined the extent G3 children who meet the main inclusion crite-
to which the unconditional ICC for delinquency rion whether or not there is ongoing contact
between the G3 child and the G2 parent. Some
2 studies include only one G3 child per family,
I would like to thank my colleague Kimberly Henry of
Colorado State University for contributing this analysis to while others include multiple G3 children.
the chapter. Regardless of the decisions made about the above
584 T.P. Thornberry

dimensions, specic decisions on which G3 child esized to be core mediators of intergenerational


or children to sample must also be made. If a continuity (Thornberry 2009) and these families
single child per family is included, should it be therefore drop out of these analyses, reducing
the oldest child, youngest child, or a random sample size and statistical power. This is a clear
selection? If multiple children are included, deci- disadvantage. Similarly, studies that include bio-
sions must be made about the number per family logical children and stepchildren can examine
and, again, which ones. whether there are different levels of continuity
Obviously, by combining these (and other and discontinuity in these different types of rela-
dimensions) a variety of specic 3G study tionships and whether or not the mediating and
designs are available and can be illustrated by moderating processes are the same or different.
several ongoing 3G studies of adolescent prob- RIGS cannot address this issue at all.
lem behaviors. The Intergenerational Project The basic point is that there are a number of
(TIP), the 3G extension of the Seattle Social core dimensions on which to make decisions
Development Project, sampled the oldest biologi- about the selection of G3 participants. There is
cal child who had face-to-face contact with the no right answer as each of these decisions entails
G2 parent on at least a monthly basis (Bailey a set of advantages and disadvantages. They need
et al. 2006). The Oregon Youth Study (OYS) to be carefully assessed with respect to the theo-
sampled all children and cohabiting stepchildren retical questions being addressed to maximize
of the G2 fathers in the initial OYS. Later, each projects validity. They also need to be coor-
because of budgetary limitations, they sampled dinated with other aspects of the projects design
only the rst two biological children per pairing in order to ensure a coherent study. There are no
of an OYS man with a woman (i.e., OYS men methodological studies of the consequences of
who fathered children with more than one woman these different design decisions. That too should
could have more than two children followed) be investigated in future research.
(Kerr et al. 2012:892). In the RIGS, we sampled
the oldest biological child of each G2 regardless 3.2.4 G3 Age at Study Initiation
of whether or not G2 had contact with the child. Another important decision concerns the age of
These three studies address the same basic the G3 participants at the beginning of the 3G
intergenerational questions about continuity and study. This is largely dependent on the age of
discontinuity of adolescent problem behaviors onset and course of the behavior of interest, and
and yet have somewhat different designs. All that can vary widely. For example, aggression
these choices have advantages and disadvan- and general antisocial behavior often begin dur-
tages. For example, by including G3 children ing toddlerhood (Tremblay and Nagin 2005)
whether or not they have ongoing contact with while drug use does not typically begin until mid-
the G2 parent allows the RIGS to investigate adolescence and peaks during late adolescence
whether or not contact moderates the level of and early adulthood (Knight et al. 2014). Clearly
intergenerational continuity in problem behav- these behaviors could lead to different decisions
iors.3 It turns out that it is quite a powerful mod- about when to start following G3 participants.
erator (Thornberry et al. 2009a, b). This is a clear Other behaviors, for example occupational attain-
advantage. At the same time, the inclusion of G2 ment (Mare 2011) or parenting styles (Conger
fathers with little or no contact with the G3 child et al. 2012), are adult behaviors and would lead to
is somewhat wasteful. It is pointless to ask them yet different decisions. In the latter cases as noted
about some parenting behaviors such as monitor- earlier, the design itself might change to include
ing, supervision, and discipline which are hypoth- a focus on continuity (or discontinuity) between
G1 and G2, assuming appropriate data are avail-
3 able for the G1 generation.
This can only be investigated for G2 fathers since virtu-
ally all G2 mothers, about 95 %, are the childs primary Although the natural course of the behavior of
caregiver. interest will inuence decisions about the age at
Three Generation Studies: Methodological Challenges and Promise 585

study initiation, some general guidelines seem sion. Several current studies collect data from
appropriate. First, it is ideal to begin the assess- and about the G3 participants on an annual basis
ments of G3 prior to the typical age of onset of (e.g., Capaldi et al. 2012; Conger et al. 2012;
the behavior. This allows the project to better Thornberry 2009). Other studies, however, add
capture age of onset and provides important what is primarily cross-sectional data about G3
information about the precursors of the behavior. to the ongoing prospective investigation of G2.
Too many longitudinal studies enter the life For example, Kaplan and Tolle (2006) followed
course midstream and one of the contributions an initial panel of seventh graders from 1971 to
that 3G studies can make, since they are often the late 1990s. At the last interview they also
extensions of existing studies, is to begin earlier. interviewed all of the children who were 12 and
Second, many behaviors of interest co-occur and older, a total of 7,519 children from the 5,467
are sequentially related over the life course. For original participants who remained in the study.
example, earlier aggression and delinquency Similarly, at Waves 11 and 12 of the National
often precede serious substance use; educational Youth Survey, G3 children ages 1224 were
attainment precedes and inuences occupational interviewed twice, in 2003 and 2004 (Knight
attainment, and so forth. Even though the ulti- et al. 2014). Finally, in the Cambridge Study in
mate behavior of interest may onset later in life Delinquent Development Farrington and col-
course, beginning assessments earlier has the leagues collected parent reports about the con-
advantage of facilitating the investigation of duct problems of the G3 children at the age 32
these behavioral chains. interviews of the G2 participants (Smith and
Regardless of when it is desirable to begin Farrington 2004) and also interviewed 550 of the
assessing the G3 participants, the natural course offspring of the original male participants. The
of human fertility will inuence those decisions. G3 children averaged 25.4 years of age at the
For example, in most samples a relatively small time of the interview (Auty et al. 2014).
number of G3 children will be born to young The more cross-sectional 3G studies often
teenage parents and for a while there will be a have considerably larger samples of G3 partici-
very small number of G3 participants available pants. For example, there are 1,227 G3 partici-
for study. It is often difcult to secure funding to pants in the National Youth Survey (Knight et al.
launch a study under those conditions. Waiting 2014) and 7,519 in the Houston survey by Kaplan
for a large enough number of G3 children to be and Tolle (2006). They also collected data about
born in order to demonstrate that the sample size multiple G3 children per family. In contrast, the
is adequate for the proposed project means that prospective 3G studies have smaller sample sizes,
children in the right-hand tail of the age distribu- typically between 300 and 500, and usually col-
tion may enter the study at somewhat older ages lect data on one or two G3 children per family.
than desired. For example, at Year 1 of the RIGS, Each strategy has strengths and limitations.
half of the G3 participants were already of school The former studies obviously have more statisti-
age even though it would have been desirable to cal power and may be able to provide more pre-
begin following all G3 participants during the cise estimates of intergenerational similarities. It
preschool years. There is little that can be done to should be noted though that the nesting of chil-
avoid this problem entirely but confronting it dren within families needs to be taken into
directly in the initial proposal may minimize its account in analytic models and statistical power
consequences. may be more related to the number of indepen-
dent families than to the total number of G3
3.2.5 Length and Frequency children. Nevertheless, the larger sample sizes
of Assessment enable more rened analysis, for example, inter-
Finally, the number and frequency of assess- generational models that take both G2 gender and
ments in the 3G study has to be established. G3 gender into account. On the other hand, prospec-
Existing studies vary considerably on this dimen- tive 3G studies, while based on smaller samples,
586 T.P. Thornberry

collect data on risk and protective factors over development is hypothesized to inuence the
time allowing for more rened estimation of quality of the family environment G2 provides to
mediating and moderating processes. Both of G3 including family structure, family climate,
these data collection strategies have important, and parenting styles (Patterson et al. 1992;
albeit somewhat distinct, advantages. Scaramella et al. 2008), all of which are related to
IG continuity in adolescent problem behaviors.
Obviously, to test intergenerational models like
3.3 Measurement Issues this one requires comprehensive prospective data
on movement along major life-course trajectories
In order to more accurately address the central such as education, work, and family formation.
questions of intergenerational study, it is valuable Only by having this detailed information on G2s
to have as broad a measurement space as possible life course can we empirically assess cascading
for each of the generations. Doing so helps to models that attempt to explain both continuity
identify more accurately the level of intergenera- and discontinuity across the generations. It is also
tional continuity by addressing potential sources important to have similar information on G3s
of spuriousness and by enabling the estimation of life-course trajectories to incorporate intragen-
models that more closely approximate causal erational inuences into the story.
analysis, such as propensity score models. A
broad approach to measurement also provides the 3.3.2 Genotypic Data
indicators necessary to examine mediating and It is also important to collect genotypic data as
moderating inuences. In this section I discuss part of 3G studies as several ongoing intergenera-
only three of the types of measures that are help- tional studies are doing (e.g., Bailey et al. 2013;
ful for these purposes. Capaldi et al. 2012; Conger et al. 2003;
Thornberry et al. 2009a). There is considerable
3.3.1 G2 Developmental Data evidence of heritability for most behaviors
There are likely to be many, time-varying mediat- (Plomin 1989); for example, heritability esti-
ing pathways that link a parents adolescent mates for antisocial behavior average about
behavior with their childs involvement in that 4050 % (Miles and Carey 1997; Rhee and
same behavior 2025 years later. In other words, Waldman 2002). Collecting DNA from intergen-
there are multiple developmental pathways that erational study participants, especially the trio of
could account for Path B in Fig. 1. Elsewhere I G3 and both biological parents, allows us to
have presented a life-course theory of intergen- investigate genetic main effects as well as gene-
erational continuity in antisocial behavior environment correlations and gene-by-
(Thornberry 2005, 2009). In a nutshell, it argues environment interactions (e.g., Caspi et al. 2002).
that one way to understand how a parents antiso- The centrality of genetics to the primary
cial behavior creates risk for their children is research questions has implications for other
through the cascading consequences that serious design features and decisions. For example, if it
and prolonged involvement in adolescent delin- is important to conduct a genome wide associa-
quency, drug use, and related problem behaviors tion study (GWAS), then that may lead to the use
have on later development. Among other conse- of an intergenerational project that relies primar-
quences, those behaviors increase the risk of dis- ily on cross-sectional data about G3 since those
orderly transitions from adolescence to adulthood studies typically have a larger sample size and
(Rindfuss et al. 1987; Wickrama et al. 2010), therefore the greater statistical power that is
later stressors (e.g., depressive symptomatology) needed for a GWAS approach. However, those
(McLoyd 1990), structural adversity (e.g., pov- studies have more limited assessments of the G3
erty, marital and family instability) (Conger et al. phenotype, based on either point estimates or ret-
2010), and continued involvement in antisocial rospective data, and may also have less detailed
behavior (Piquero et al. 2003). This course of prospective data for the investigation of
Three Generation Studies: Methodological Challenges and Promise 587

gene- environment interactions. In contrast, stud- individual behaviors and levels of continuity and
ies that have prospective G3 data over longer por- discontinuity. That is an important avenue for
tions of the life course typically have smaller future 3G projects to investigate.
sample sizes and may have to use a candidate It would also be advantageous to have detailed
gene approach. While having richer prospective measures on the broader family context, includ-
data on the phenotype and on environmental fac- ing the structure of the household, the presence
tors, the range of genetic risk and protective fac- of extended family members either in the house-
tors that can be examined will be more limited. hold or nearby, the number of siblings, and so
Both approaches have advantages for enhancing forth. All of these broader familial characteristics
our understanding of the role of genetics in can inuence parent-child interactions, the level
accounting for intergenerational continuity and of intergenerational continuity, and mediating
discontinuity. Combining the extensive life- processes that account for it. For example, in
course information on G2 and G3 development their investigation of intergenerational continuity
mentioned above with genetic risk and protective in cigarette use, Vuolo and Staff (2013) found
factors is an important topic for future research that one of the most powerful mediators was the
that will ultimately enable a very comprehensive presence of an older sibling who also smoked.
assessment of intergenerational linkages.
3.3.4 Similarity of Measures
3.3.3 Contextual Data Regardless of the specic measures that are
It is also helpful to have data on the broader con- included, a 3G study also needs to consider
text in which the linkage between G2 and G3 whether the same measures are used to assess G2
occurs. For example, the neighborhood of resi- and G3 at comparable ages. Recall that G3
dence inuences individual behaviors as well as assessments are likely to occur 2025 years later.
the structure and interaction patterns of families If identical or virtually identical measures are
(Anderson 1990; Burton 1996; Wilson 1987). used, there is obviously a high degree of compa-
Moreover, Sharkey (2008) has demonstrated a rability in the assessment of key concepts across
substantial degree of intergenerational continuity the generations. That facilitates the evaluation of
in neighborhood context. That is, the socioeco- whether there is change or stability in the behav-
nomic status of the parents neighborhood has a iors of interest and in mediating and moderating
pronounced impact on the likely socioeconomic inuences. However, it is likely that new and
status of the childs neighborhood once they grow improved measures for at least some of the con-
to adulthood, independent of individual and fam- cepts have been developed in the interim and
ily characteristics. Exposure to neighborhood large improvements in psychometric properties
poverty and disadvantage is often reproduced, may mandate the use of new measures. If differ-
especially for African-American families. ent measures are used, then, by denition, there
Measuring these processes is valuable in its own is less comparability in the way key concepts are
right. But, in addition, understanding the neigh- assessed. If there is change in the prevalence or
borhood context in each generation and, perhaps frequency of the behavior of interest, for exam-
more importantly, its continuity across genera- ple, it is difcult to conclude whether that reects
tions, has the potential to add new insights into an actual change in the behaviorperhaps a
identifying the mechanisms by which individual period or birth cohort effectversus a change in
level behaviors, either prosocial or antisocial, are the measurement approach.
transmitted across generations. There is obviously no correct answer to this
There is relatively little research on the repro- issue. My own view is that it is not necessary to
duction of social contexts, including neighbor- use identical measures across the generations. An
hood context, or the mechanisms that bring about underlying theory of intergenerational continuity
(Sharkey 2008). There is also a little research on posits that adjacent generations will be similar on
how the reproduction of social context inuences some concept of interest. As long as the same
588 T.P. Thornberry

concept is measured with acceptable validity and cated analytic techniques such as group-based
reliability in both generations, the specic mea- trajectory models (Nagin 2009) and latent class
sure used should not matter. Indeed, if we only models (McCutcheon 1987) that divide the sam-
observe intergenerational continuity when identi- ple into meaningful subgroups. Also, looking at
cal measures are used, it raises the possibility that the biological relatedness of pairs of participants,
the similarities are generated not by the behavior as done by Frisell et al. (2011), can contribute to
of interest but by a particular approach to mea- our understanding of genetic and environmental
surement. Also, the extent to which intergenera- inuences. At the same time reliance entirely on
tional relationships are replicated across different archival data comes at a cost. These studies are
studies that are likely to use different measure- almost entirely descriptive and are focused only
ment approaches is important in assessing the on the rst core intergenerational questionthe
overall validity of the intergenerational theory. level of continuity and discontinuity. They have
few if any explanatory variables and therefore
3.3.5 Source of Measures very little ability to address the second core ques-
Some behaviors of interest can only be measured tionwhich mediating and moderating inu-
by direct contact with the G2 and G3 participants, ences help to explain continuity and
for example, by interviews or videotaped obser- discontinuity.
vations of parent-child interactions (Scaramella Studies with direct contact measures suffer
et al. 1999). For this issue I view the use of infor- from the opposite set of problems. Samples tend
mantsfor example, parent reports about the to be small, especially if complicated measure-
behavior of young children or about the parenting ment strategies such as videotaped observations
of the parents partneras a special case of direct are used (e.g., Capaldi et al. 2012; Conger et al.
contact data. Other behaviors can be measured 2012). Thus, they often encounter issues of statis-
either by direct contact or by ofcial or archival tical power which, as noted above, may limit the
data. For example, with the exception of child application of some analytic techniques.
maltreatment, the study of intergenerational pat- Nevertheless, the richness of data that can only
terns of parenting can only be investigated by be collected via direct contact with the families
direct contact. In contrast, the study of delinquent enables these studies to investigate mediating and
behavior can be measured either by self-reports, moderating inuences in greater detail.
archival data such as arrest records, or both.
Both strategies have been used in the study of
antisocial behavior and, again, there are advan- 4 Contributions of Three
tages and disadvantages to these approaches. Generation Studies
Studies based on archival records typically have
the largest sample sizes since data collection is Intergenerational studies have the potential to
relatively inexpensive. Van de Rakt et al. (2008) advance our understanding of a number of themes
used the Dutch registry data to study intergenera- that are central to the life-course perspective. In
tional continuity in criminal convictions of the remainder of this chapter I comment on some
fathers and children with a sample of 8,085 chil- of them and, in doing so, address the potential
dren. Frisell et al. (2011) examined whether vio- that 3G studies have and suggest some additional
lent crime runs in families based on a total themes for future research.
population study of 12.5 million individuals in
Sweden. In addition, it is easier for archival stud-
ies to be extended beyond three generations as in 4.1 Consequences of Behavior
Bijlevelds Dutch study of ve generations of the
same families (Bijleveld and Wijkman 2009). A central theme of the life-course approach is to
These very large samples obviously increase sta- describe and understand the long-term conse-
tistical power and facilitate the use of sophisti- quences of movement along major developmen-
Three Generation Studies: Methodological Challenges and Promise 589

tal trajectories such as family, education, work, tional effects (Mare 2011). That is an important
involvement in antisocial behavior, and so forth. gap for 3G studies to investigate in the future.
For example, there is a persistent theme related to Nevertheless, 3G studies provide a more com-
the timing of transitions with a major assumption plete understanding of the consequences that a
that off-time and disorderly transitions will be persons fateful decisions and events can have
problematic both in the short-term and long-term by extending our view past the consequences for
(Elder 1985). Accordingly, many longitudinal only the individual to consequences for other
studies have examined this issue. Focusing on generations. These investigations are quite con-
adolescence as I have throughout this chapter, sistent with developmental psychologys concern
there is ample evidence that successful adapta- with multinality, the notion that a single behav-
tion along prosocial trajectories (for example, ior or characteristic can have causal linkages to
school engagement and satisfactory perfor- many different, and seemingly disparate,
mance) has many long-term benets for the indi- outcomes.
vidual while failure in prosocial arenas (for Identifying more fully the range of conse-
example, school disengagement, dropping out, quences that a particular behavior can generate
and involvement in problem behaviors) has many has important policy implications for the devel-
long-term costs. opment of programs to enhance prosocial behav-
Most longitudinal studies that have addressed iors (e.g., school engagement) and to reduce
this issue have been concerned with the conse- problem behaviors (e.g., drug use). Assume for
quences of a persons earlier development on that the moment that these behaviors have multigen-
same individuals later development. 3G studies erational consequencesconsequences for the
have the potential to increase our understanding individual, his or her children, and grandchildren.
of the scope of this issue by examining the extent Efforts that successfully increase prosocial
to which those consequences, both positive and behaviors and reduce problem behaviors have the
negative, also extend across the generations to potential to benet children and grandchildren,
affect the persons children. These studies can even though they may not yet be born. Those
examine whether the effects are direct or indirect benets, while quite likely of a very modest size,
and, if indirect, the mediating pathways through improve the cost-benet ratio associated with the
which they operate. There is growing evidence program and provide additional evidence of value
from 3G studies that these intergenerational (Aos et al. 2004). Also, by providing information
effects can be quite strong and persistent. about the mediating pathways associated with
Moreover, 3G studies have the potential to intergenerational effects, 3G studies can provide
extend this line of inquiry by also investigating crucial evidence about the proper content of these
grandparental effects to see whether behavioral programs.
consequences extend across multiple generations
of family lineage. Evidence for independent
grandparental effects is mixed and may depend 4.2 Origins of Behavior
on the particular behavior being investigated.
Cherlin and Furstenberg (1992) found few inde- A second important theme of the life-course
pendent grandparental effects on child develop- approach is to identify the origins of behavior.
ment. In contrast, Thornberry et al. (2006) found There is little difference in whether the behavior
that grandmothers drug use had an effect on the is prosocialsuch as warm/nurturing parent-
grandchilds drug use if she had ongoing contact ingor antisocialsuch as child maltreatment.
with the grandchild and Sharkey (2008) found A central question is to identify precursors, and
that continuity in the neighborhood context per- ideally causal inuences for that behavior. Given
sisted across several generations. There are, the complexity of human behavior there are
unfortunately, too few investigations of grandpa- likely to be multiple pathways that lead to a
rental effects and, more generally, multigenera-
590 T.P. Thornberry

particular behavior, what developmental psy- cantly less likely to engage in problem behaviors
chology refers to as equinality. as compared to parents, also adds novel informa-
Recently, the social sciences have relied on tion to our understanding of the causes of prob-
longitudinal studies to investigate this theme, in lem behavior. Similarly, for prosocial behaviors,
large part because of the advantages those studies understanding discontinuityfor example why
offer with respect to temporal ordering. But lon- children are more successful than their parents in
gitudinal studies primarily focus on intragenera- the educational arenaalso enhances our under-
tional inuences such as the persons earlier standing of the origins of educational success.
development and the concurrent relationships Findings from 3G studies about discontinuity
between parents and children in explaining a par- have important policy and programmatic impli-
ticular behavior. As Mare has noted: social sci- cations. Identifying protective factors that create
entists tend to maintain a two-generation view of intergenerational resilience for problem behav-
the world (2011:8). 3G studies have the poten- iors or promotive factors that enhance better out-
tial to move beyond this orientation and greatly comes for prosocial behaviors provides specic
extend our understanding of the origins of any information for the content of programmatic
particular behavior by addressing the two major efforts to improve human development. That is,
questions of intergenerational study. by better understanding the processes that lead
First, they provide a more complete under- some children not to follow in the footsteps of
standing of the origins of a particular behavior by their parents, we will have identied areas of
examining not only how a G2 parent concurrently development that programmatic efforts should
inuences G3 behavior but by examining how a attempt to enhance.
parents earlier developmental history also inu- Unfortunately, there are very few studies that
ences the behavior. In part, this is done by exam- focus specically on discontinuity (e.g., Conger
ining both homotypic and heterotypic continuity et al. 2012; Dong and Krohn 2014; Haller and
in behavior as illustrated earlier by Path B in Fig. Chassin 2010; Lovegrove 2010) and one specic
1. area for future research is increased investigation
Second, 3G studies provide a more complete of this topic. The study of intergenerational dis-
understanding by examining the mediating path- continuity raises a new set of conceptual and
ways that account for observed levels of intergen- methodological issues that needed to be carefully
erational continuity (or discontinuity) in behavior. addressed. For example, if the G3 participants are
These mediators are likely to be quite different not followed far enough into their life courses it
from those that account for concurrent similarity. is easy to misclassify them as experiencing inter-
Many of them exist before the child is born and generational discontinuity when, with additional
refer to long-term movement along the parents follow-up, they would exhibit the same behavior
life course trajectories. They are therefore likely as their parents, demonstrating continuity.
to provide novel information compared to that Hopefully, 3G studies will begin to address issues
typically reported in a longitudinal study. of discontinuity more fully as they extend their
Third, 3G studies invariably uncover a sub- follow-up of the G3 samples.
stantial level of intergenerational discontinuity in
which children do not follow in the footsteps of
their parents. Indeed, for the behaviors we have 4.3 Linked Lives
investigated in the RIGS, while continuity coef-
cients are statistically signicant, a majority of Another life-course theme at the heart of 3G
the children do not exhibit the same problem studies is the one with which we beganlinked
behaviors as their parents (e.g., Thornberry et al. lives. Clearly, the life courses of individuals
2006, 2009a; Thornberry and Henry 2013). become intertwined, mutually inuencing one
Understanding discontinuity, especially intergen- another over time. That is true of all intimate rela-
erational resilience where children are signi- tionships and, aside from spouses, no relationship
Three Generation Studies: Methodological Challenges and Promise 591

may be as intertwined as that between parent and approach 3G studies from the perspective of a
child. Intergenerational or 3G studies typically social scientist and not a geneticist or a behav-
focus on how parental development and behav- ioral geneticist. All of that no doubt limits my
iors inuence child development and behaviors. understanding of 3G studies, the topics they can
As just noted, doing so has greatly enhanced our address, the methodological issues they confront,
understanding of the multiple consequences of and how those issues can be resolved. If this
behavior as well as the multiple origins of behav- chapter had been written by a scholar focusing on
ior. But in both cases the focus is typically on the gerontology or demography (e.g., Bengston 1996
G2 parent acting on the G3 child. In part, this is or Mare 2011), for example, I am quite sure that
quite sensible given the temporal, indeed, genera- it would have been different in tone and content.
tional, ordering of these behaviors. For so many Nevertheless, I think the major themes identi-
behaviors investigated in 3G studieseduca- ed here are central to most intergenerational
tional attainment, fertility, adolescent problem studies and provide an integrity to them that sep-
behaviors, occupations, etc.the ow of causal- arates them from traditional longitudinal studies
ity is overwhelmingly from parent to child. and therefore raises the potential for them to
Nevertheless there is little focus within the extend our understanding of human behavior. At
context of 3G studies on examining the impact of their heart intergenerational studies address two
how the behavior of the G3 child also inuences fundamental issues. The rst is to identify the
the behavior of the G2 parent. For example, I level of intergenerational continuity and disconti-
cannot think of any study that examines how G3 nuity in a particular behavior at the same devel-
drug use inuences the likelihood of the initiation opmental stage for each generation. The second
of G2 drug use or, more generally, how a childs is to identify the mechanisms that help us to
antisocial behavior inuences a parents initia- understand how adjacent generations are linked
tion or escalation of antisocial behavior. Clearly, and, therefore, to inform programmatic efforts to
it is plausible that the child can introduce a parent improve prosocial outcomes and ward off antiso-
to these and other forms of behavior but there is cial outcomes.
little empirical information on this issue. It obvi- In designing 3G studies to address these
ously can be addressed in the context of a 3G themes there are, as I hope this chapter has dem-
study and that too should be a topic for future onstrated, a wide range of specic designs that
research. Doing so emphasizes the bidirectional can be implemented. As discussed above there
quality of the concept of linked liveshow indi- are many sampling, research design, measure-
viduals mutually inuence one another. That ment, and data analytic decisions to be made in
interactive quality should be a more prominent creating a particular study. By mixing and match-
feature of 3G studies and intergenerational ing across those various decision points, 3G stud-
analyses. ies offer researchers tremendous exibility in
designing a study to address the particular behav-
ior of interest and the specic theoretical ques-
5 Closing Comments tions being posed. Indeed, even studies that
address the same basic theoretical questions
The focus of this chapter has been on method- about the same behavior often end up with some-
ological issues that 3G studies of human behavior what different designs. We should welcome that
need to confront. I recognize that there are many exibility, while at the same time maintaining a
limitations to the approach presented here. My keen focus on the core issues of intergenerational
own scholarly focus is on the study of adolescent study, as we continue to pursue themes central to
and early adult development with a particular the life-course perspective. As more and more
emphasis on understanding the causes and conse- prospective 3G studies become available, they
quences of delinquency, drug use, gang member- have the potential to greatly expand our under-
ship, and related problem behaviors. Also, I standing of the causes and consequences of
592 T.P. Thornberry

human behavior and how behaviorsboth proso- delinquency. Because of that, we oversampled
cial and antisocialbecome reproduced across males, 3 to 1, given their greater involvement in
the generations. serious violent offending. We also oversampled
youth who resided in neighborhoods with a high
Acknowledgement Support for the Rochester Youth resident arrest rate. The sample can be weighted
Development Study has been provided by the National to account for this stratication. The nal sample
Institute on Drug Abuse (R01DA020195). Work on this
of G2 participants averaged 14 years of age; 73 %
project was also aided by grants to the Center for Social
and Demographic Analysis at the University at Albany were males and 27 % were females. The full G2
from NICHD (P30HD32041) and NSF (SBR-9512290). panel was followed forward with 12 interviews
Points of view or opinions in this document are those of from age 14 to 23, at which point the retention
the author and do not necessarily represent the ofcial
rate was 85 %. In addition, we also interviewed
position or policies of the funding agencies. I would like
to thank Jason Boardman, Kimberly Henry, Marvin one of their G1 parents 11 times; 85 % of the G1
Krohn, Thomas Loughran, and Pamela Porter for com- respondents were the G2 childs biological
menting on and helping with this paper. mother and another 10 % were stepmothers. In
addition to the interviews that were conducted at
regular intervals, we also collected information
Appendix: The Rochester from schools, police, and child protective ser-
Intergenerational Study vices. Details of the Rochester study design are
presented in Thornberry et al. (2003).
The Rochester Intergenerational Study started as The Rochester Intergenerational Study
an add-on to an existing longitudinal study, the (RIGS), described in the right-hand portion of
Rochester Youth Development Study, which Fig. 3, began in 1999. The focal participants are
began in 1988. Its core design is summarized in the oldest biological child of each of the initial
the left-hand portion of Fig. 3. We began with a G2 participants. In Year 1 (1999), 370 children
sample of 1,000 seventh and eighth graders from (ages 213, average age 6) and their families
the Rochester, New York public schools. The ini- enrolled in the study. In each subsequent year we
tial purpose of the project was to study the causes identied additional rst-born children and
and consequences of serious, chronic, and violent enrolled them in the study as they turned 2. A

1988 1997 1999 2010

G1 = Grandparent

G2 = Parent
(ages) (14) (23) (25) (41)
n 1000 846

OCG = Other Caregiver

G3 = Oldest Biological Child


(ages) (2-13) (2-29)
n Mean=6 Mean=19
370 530
Official Data:
(e.g., Police, CPS)

Fig. 3 Research design for the Rochester Intergenerational Study


Three Generation Studies: Methodological Challenges and Promise 593

total of 530 G3 children have enrolled in the Bailey, J. A., Hill, K. G., Guttmannova, K., Oesterle, S.,
Hawkins, J. D., Catalano, R. F., & McMahon, R. J.
study, currently ranging in age from 2 to 29 with
(2013). The association between parent early adult
an average age of 19. Importantly, therefore, drug use disorder and later observed parenting prac-
many now overlap in age with the age of the G2 tices and child behavior problems: Testing alternate
parents when they were rst assessed as part of models. Developmental Psychology, 49, 887899.
Belsky, J. (1993). Etiology of child maltreatment: A
the original Rochester study.
developmental-ecological analysis. Psychological
For each family in the intergenerational study Bulletin, 114, 413434.
we annually interview three participants. We Bengtson, V. L. (1971, 1985, 1988, 1991, 1994, 1997,
assess the G2 parents (the original adolescent 2000). Longitudinal study of generations, [California].
ICPSR22100-v2. Ann Arbor: Inter-university
participants in the Rochester Study), the G3 child
Consortium for Political and Social Research [distrib-
when they are 8 years and older, and another pri- utor], 2009-05-12. http://doi.org/10.3886/
mary caregiver (OCG) for each G3 child. In the ICPSR22100.v2.
case of G2 fathers, the OCG is almost invariably Bengtson, V. L. (1996). Continuities and discontinuities in
intergenerational relationships over time. In V. L.
(93 %) the childs biological mother. In the case
Bengtson (Ed.), Adulthood and aging: Research on
of G2 mothers, however, the OCGs are grand- continuities and discontinuities (pp. 271303).
mothers (47 %), biological fathers (31 %), stepfa- New York: Springer.
thers (6 %), aunts (7 %), and others (9 %). At Bengtson, V. L., Biblarz, T. J., & Roberts, R. E. (2002).
How families still matter: A longitudinal study of
younger ages when G3 could not be interviewed,
youth in two generations. New York: Cambridge
we conducted videotaped observations of G3 University Press.
interacting with G2 and with OCG. We also col- Besemer, S., & Farrington, D. P. (2012). Intergenerational
lected DNA from a subsample of the families. transmission of criminal behaviour: Conviction trajec-
tories of fathers and their children. European Journal
Finally, we collect data from ofcial agencies
of Criminology, 9, 120141.
such as police and social services. More detail Bijleveld, C. J. H., & Wijkman, M. (2009).
about the RIGS design can be found in Thornberry Intergenerational continuity in convictions: A ve-
(2009). generation study. Criminal Behaviour and Mental
Health, 19, 142155. doi:10.1002/cbm.714.
Burton, L. M. (1996). Age norms, the timing of family
role transitions, and intergenerational caregiving
References among aging African American women. The
Gerontologist, 36(2), 199208.
Achenbach, T. M. (1991). Manual for the child behavior Capaldi, D. M., Pears, K. C., Patterson, G. R., & Owen,
checklist/4-18 and 1991 prole. Burlington: University L. D. (2003). Continuity of parenting practices across
of Vermont, Department of Psychology. generations in an at-risk sample: A prospective com-
Alexander, P. C., Moore, S., & Alexander, E. R., III. parison of direct and mediated associations. Journal of
(1991). What is transmitted in the intergenerational Abnormal Child Psychology, 31, 127142.
transmission of violence? Journal of Marriage and the Capaldi, D. M., Pears, K. C., Kerr, D. C. R. (2012). The
Family, 53, 657668. Oregon Youth study three-generational study: Theory,
Anderson, E. (1990). Streetwise: Race, class, and change design, and ndings. The International Society for the
in an urban community. Chicago: The University of Study of Behavioural Development Bulletin, Serial
Chicago Press. No. 62 (pp. 2933).
Aos, S., Lieb, R., Mayeld, J., Miller, M., & Pennucci, A. Carlson, M. J. (2006). Family structure, father involve-
(2004). Benets and costs of prevention and early ment, and adolescent behavioral outcomes. Journal of
intervention programs for youth. Olympia: Washington Marriage and Family, 68, 137154.
State Institute for Public Policy. Caspi, A., McClay, J., Moftt, T. E., Mill, J., Martin, J.,
Auty, K. M., Farrington, D. P., & Coid, J. W. (2014). Craig, I. W., . . . Poulton, R. (2002). Role of genotype
Intergenerational transmission of psychopathy and in the cycle of violence in maltreated children. Science,
mediation via psychosocial risk factors. British 297, 851854.
Journal of Psychiatry, 206, 2631. Cherlin, A. J., & Furstenberg, F. F., Jr. (1992). The new
Bailey, J. A., Hill, K. G., Oesterle, S., & Hawkins, J. D. American grandparent: A place in the family, a life
(2006). Linking substance use and problem behavior apart. Cambridge, MA: Harvard University Press.
across three generations. Journal of Abnormal Conger, R. D., Lorenz, F. O., Elder, G. H., Jr., Simons,
Psychology, 34, 263282. R. L., & Ge, X. (1993). Husband and wife differences
594 T.P. Thornberry

in response to undesirable life events. Journal of Research Design [WWW document]. URL: http://
Health and Social Behavior, 34, 7188. www.cpc.unc.edu/projects/addhealth/design.
Conger, R. D., Neppl, T. K., Kim, K. J., & Scaramella, Hetherington, E. M., & Kelly, J. (2002). For better or for
L. V. (2003). Angry and aggressive behavior across worse. New York: Norton.
three generations: A prospective, longitudinal study of Hofferth, S. L., & Goldscheider, F. (2010). Family struc-
parents and children. Journal of Abnormal Child ture and the transition to early parenthood.
Psychology, 31, 143160. Demography, 47, 415437.
Conger, R. D., Conger, K. J., & Martin, M. J. (2010). Huesmann, L. R., Eron, L. D., Lefkowitz, M. M., &
Socioeconomic status, family processes, and individ- Walder, L. O. (1984). Stability of aggression over time
ual development. Journal of Marriage and Family, 72, and generations. Developmental Psychology, 20,
685704. 11201134.
Conger, R. D., Schoeld, T. J., & Neppl, T. K. (2012). Jaffee, S. R., Moftt, T. E., Caspi, A., & Taylor, A.
Intergenerational continuity and discontinuity in harsh (2003). Life with (or without) father: The benets of
parenting. Parenting: Science and Practice, 12, living with two biological parents depend on the
222231. fathers antisocial behavior. Child Development, 74,
Doherty, W. J., Kouneski, E. F., & Erickson, M. F. (1998). 109126.
Responsible fathering: An overview and conceptual Kaplan, H. B., & Tolle, G. C., Jr. (2006). The cycle of devi-
framework. Journal of Marriage and the Family, 60, ant behavior: Investigating intergenerational parallel-
277292. ism. New York: Springer.
Dong, B., & Krohn, M. D. (2014). Exploring intergenera- Kelly, J. B. (2000). Childrens adjustment in conicted
tional discontinuity in problem behavior: Bad parents marriage and divorce: A decade review of research.
with good children. Youth Violence and Juvenile Journal of the American Academy of Child and
Justice. doi:10.1177/1541204014527119. Adolescent Psychiatry, 39, 963973.
Doumas, D., Margolin, G., & John, R. S. (1994). The Kerr, D. C. R., Capaldi, D. M., Pears, K. C., & Owen,
intergenerational transmission of aggression across L. D. (2012). Intergenerational inuences on early
three generations. Journal of Family Violence, 9, alcohol use: Independence from the problem behavior
157175. pathway. Development and Psychopathology, 24,
Elder, G. H., Jr. (1985). Perspectives on the life course. In 889906.
G. H. Elder Jr. (Ed.), Life course dynamics (pp. 2349). Knight, K. E., Menard, S., & Simmons, S. B. (2014).
Ithaca: Cornell University Press. Intergenerational continuity of substance use.
Farrington, D. P. (1993). Have any individual, family or Substance Use & Misuse, 49, 221233.
neighbourhood inuences on offending been demon- Laird, N. M., & Lange, C. (2006). Family-based designs in
strated conclusively? In D. P. Farrington, R. J. the age of large-scale gene-association studies. Nature
Sampson, & P.-O. H. Wikstrm (Eds.), Integrating Reviews Genetics, 7, 385394. doi:10.1038/nrg1839.
individual and ecological aspects of crime (pp. 737). Lamb, M. E., & Tamis-Lemonda, C. S. (2004). The role of
Stockholm: Allmaana Forlaget. the father: An introduction. In M. E. Lamb (Ed.), The
Frisell, T., Lichtenstein, P., & Langstrom, N. (2011). role of the father in child development (pp. 131).
Violent crime runs in families: A total population New York: Wiley.
study of 12.5 million individuals. Psychological Lovegrove, P. J. (2010). Explaining discontinuous cross-
Medicine, 41, 97105. generational patterns of delinquency. Unpublished
Fuller, B. E., Chermack, S. T., Cruise, K. A., Kirsch, E., doctoral dissertation, University of Colorado.
Fitzgerald, H. E., & Zucker, R. A. (2003). Predictors Magnuson, K., & Berger, L. M. (2009). Family structure
of aggression across three generations among sons of states and transitions: Associations with childrens
alcoholics: Relationships involving grandparental and well-being during middle childhood. Journal of
parental alcoholism, child aggression, marital aggres- Marriage and Family, 71, 575591.
sion and parenting practices. Journal of Studies on Mare, R. D. (2011). A multigenerational view of inequal-
Alcohol, 64, 472483. ity. Demography, 48, 123.
Granovetter, M. (1985). Economic action and social struc- Martin, M. J., Conger, R. D., Schoeld, T. J., Dogan, S. J.,
ture: The problem of embeddedness. American Widaman, K. F., Donnellan, M. B., & Neppl, T. K.
Journal of Sociology, 91, 481510. (2010). Evaluation of the interactionist model of
Haller, M. M., & Chassin, L. (2010). The reciprocal inu- socioeconomic status and problem behavior: A devel-
ences of perceived risk for alcoholism and alcohol use opmental cascade across generations. Development
over time: Evidence for aversive transmission of and Psychopathology, 22, 697715.
parental alcoholism. Journal of Studies on Alcohol Martinez, G. M., Daniels, K., & Chandra, A. (2012).
and Drugs, 71, 588596. Fertility of men and women aged 1544 years in the
Harris, K. M., Halpern, C. T., Whitsel, E., Hussey, J., United States: National survey of family growth,
Tabor, J., Entzel, P., Udry, J. R. (2009). The National 20062010 National Health Statistics report.
Longitudinal Study of Adolescent to Adult Health: Hyattsville: National Center for Health Statistics.
Three Generation Studies: Methodological Challenges and Promise 595

McCutcheon, A. L. (1987). Latent class analysis increases in adolescent internalizing and externalizing
(Quantitative Applications in the Social Sciences problems. Journal of Research on Adolescence, 9,
Series No. 64). Thousand Oaks: Sage. 111141.
McLoyd, V. C. (1990). The impact of economic hardship Scaramella, L. V., Neppl, T. K., Ontai, L. L., & Conger,
on black families and children: Psychological distress, R. D. (2008). Consequences of socioeconomic disad-
parenting, and socioemotional development. Child vantage across three generations: Parenting behavior
Development, 61, 311346. and child externalizing problems. Journal of Family
Miles, D. R., & Carey, G. (1997). Genetic environment Psychology, 22, 725733.
architecture of human aggression. Journal of Sharkey, P. (2008). The intergenerational transmission of
Personality and Social Psychology, 72, 207217. context. American Journal of Sociology, 113(4),
Moftt, T. E., Caspi, A., Harrington, H., & Milne, B. J. 931969.
(2002). Males on the life-course-persistent and Smith, C. A., & Farrington, D. P. (2004). Continuities in
adolescence-limited antisocial pathways: Follow-up at antisocial behavior and parenting across three genera-
age 26 years. Development and Psychopathology, 14, tions. Journal of Child Psychology & Psychiatry &
179207. Allied Disciplines, 45, 230248.
Nagin, D. S. (2009). Group-based modeling of develop- Thornberry, T. P. (1989). Panel effects and the use of self-
ment. Cambridge, MA: Harvard University Press. reported measures of delinquency in longitudinal stud-
Nagin, D. S., Pogarsky, G., & Farrington, D. P. (1997). ies. In M. W. Klein (Ed.), Cross-national research in
Adolescent mothers and the criminal behavior of their self-reported crime and delinquency (pp. 347369).
children. Law and Society Review, 31, 137162. Dordrecht: Kluwer Academic Publishers.
Neppl, T. K., Conger, R. D., Scaramella, L. V., & Ontai, Thornberry, T. P. (2005). Explaining multiple patterns of
L. L. (2009). Intergenerational continuity in parenting offending across the life course and across genera-
behavior: Mediating pathways and child effects. tions. The Annals of the American Academy of Political
Developmental Psychology, 45, 12411256. and Social Science, 602, 156195.
Patterson, G. R., Reid, J. B., & Dishion, T. J. (1992). Thornberry, T. P. (2009). The apple doesnt fall far from
Antisocial boys. Eugene: Castalia Publishing the tree (or does it?): Intergenerational patterns of anti-
Company. social behavior. Criminology, 47, 297325.
Pearl, J. (2009). Causality: Models, reasoning, and infer- Thornberry, T. P., & Henry, K. L. (2013). Intergenerational
ence. Cambridge: Cambridge University Press. continuity in maltreatment. Journal of Abnormal
Phares, V., Fields, S., Kamboukos, D., & Lopez, E. (2005). Child Psychology, 41, 555569.
Still looking for poppa. American Psychologist, 60, Thornberry, T. P., Krohn, M. D., Lizotte, A. J., Smith,
735736. C. A., & Tobin, K. (2003). Gangs and delinquency in
Piquero, A. R., Farrington, D. P., & Blumstein, A. (2003). developmental perspective. New York: Cambridge
The criminal career paradigm. In M. Tonry (Ed.), University Press.
Crime and justice: A review of research (Vol. 30, Thornberry, T. P., Krohn, M., & Freeman-Gallant, A.
pp. 359506). Chicago: University of Chicago Press. (2006). Intergenerational roots of early onset sub-
Plomin, R. (1989). Behavioral genetics. Journal of stance use. Journal of Drug Issues, 36, 128.
Nervous and Mental Disease, 177, 645. Thornberry, T. P., Freeman-Gallant, A., & Lovegrove, P. J.
Prinzie, P., & Onghena, P. (2005). Cohort sequential (2009a). The impact of parental stressors on the inter-
design. In B. S. Everitt & D. C. Howell (Eds.), generational transmission of antisocial behavior.
Encyclopedia of statistics in behavioral science (Vol. Journal of Youth and Adolescence, 38, 312322.
1, pp. 319322). Chichester: Wiley. Thornberry, T. P., Freeman-Gallant, A., & Lovegrove, P. J.
Rhee, S. H., & Waldman, I. D. (2002). Genetic and envi- (2009b). Intergeneration linkages in antisocial behavior.
ronmental inuences on antisocial behavior: A meta- Criminal Behaviour and Mental Health, 19, 8093.
analysis of twin and adoption studies. Psychological Torche, F., Warren, J. R., Halpern-Manners, A., &
Bulletin, 128, 490529. Valenzuela, E. (2012). Panel conditioning in a longitu-
Rindfuss, R. R., Swicegood, C. G., & Rosenfeld, R. dinal study of adolescents substance use: Evidence
(1987). Disorder in the life course: How common and from an experiment. Social Forces, 90, 891918.
does it matter? American Sociological Review, 52, Tremblay, R. E., & Nagin, D. S. (2005). The developmen-
785801. tal origins of physical aggression in humans. In R. E.
Rosenbaum, P. R., & Rubin, D. B. (1983). The central role Tremblay, W. W. Hartup, & J. Archer (Eds.),
of the propensity score in observational studies for Developmental origins of aggression (pp. 83106).
causal effects. Biometrika, 70, 4155. New York: The Guilford Press.
Scaramella, L. V., & Conger, R. D. (2003). Intergenerational van de Rakt, M., Nieuwbeerta, P., & de Graaf, N. D.
continuity of hostile parenting and its consequences: (2008). Like father, like son: The relationships
The moderating inuence of childrens negative emo- between conviction trajectories of fathers and their
tional reactivity. Social Development, 12, 420439. sons and daughters. British Journal of Criminology,
Scaramella, L. V., Conger, R. D., & Simons, R. L. (1999). 48, 538556.
Parental protective inuences and gender-specic
596 T.P. Thornberry

Vuolo, M., & Staff, J. (2013). Parent and child cigarette Wilson, W. J. (1987). The truly disadvantaged: The inner
use: A longitudinal, multigenerational study. city, the underclass, and public policy. Chicago: The
Pediatrics, 132, e568e577. University of Chicago Press.
Wickrama, T., Wickrama, K. A. S., & Baltimore, D. L. Wu, P., & Kandel, D. B. (1995). The roles of mothers and
(2010). Adolescent precocious development and fathers in intergenerational behavioral transmission:
young adult health outcomes. Advances in Life Course The case of smoking and delinquency. In H. B. Kaplan
Research, 15, 121131. (Ed.), Drugs, crime, and other deviant adaptations:
Longitudinal studies (pp. 4981). New York: Plenum
Press.
Neighborhood, Place, and the Life
Course

Christopher R. Browning, Kathleen A. Cagney,


and Bethany Boettner

1 Introduction ing research and the need for novel strategies to


address ongoing challenges (Harding et al. 2011).
The health and developmental consequences of In this chapter we take stock of contemporary
residential and other spatial environments across advances in neighborhood research and their
the life course have been a focus of social science implications for understanding life course pro-
disciplines for decades. The last 25 years, in par- cesses. We highlight the fundamental challenges
ticular, have seen a marked increase in the inves- the eld faces and novel directions that hold the
tigation of neighborhood effects (particularly promise to usher in a new era of neighborhood
in urban settings) on collective and individual and broader place effects research. At the core
wellbeing. These efforts have yielded compelling of our argument is the claim that neighborhood
evidence of neighborhood inuences on a range research has neglected the role of actual mobility
of outcomes. Largely focused on variation in as captured by day-to-day activity patterns and
socioeconomic factors, extant research has exposures. Incorporating individual level activity
offered robust evidence of neighborhood effects spaces i.e., the set of places individuals come
on mental and physical health, delinquency and into contact with as a result of their routine activi-
crime, educational achievement, sexual and ties into theory, data, and measurement will
fertility-related behavior, among other outcomes substantially advance research on neighborhoods
(Sampson 2012). Despite the dramatic upsurge in and extend the eld beyond a focus on arbitrarily
interest and the cumulative body of ndings sig- dened neighborhood boundaries (Golledge and
naling the relevance of neighborhoods, the eld Stimson 1987; Inagami et al. 2007; Newsome
has nevertheless been the subject of extensive et al. 1998; Schnfelder and Axhausen 2002,
criticism. A number of scholars have called for 2003). A shift toward activity space thinking in
recognition of the signicant limitations of exist- neighborhood research is already evident in an
emerging, interdisciplinary theoretical and
empirical literature (Browning and Soller 2014;
C.R. Browning (*) B. Boettner Cagney et al. 2013; Graif et al. 2014; Jones and
Department of Sociology and Institute for Population Pebley 2014; Krivo et al. 2013; Kwan 2013;
Research, Ohio State University,
Matthews 2011; Matthews and Yang 2013;
Columbus, OH, USA
e-mail: browning.90@osu.edu Mennis and Mason 2011; Perchoux et al. 2013).
K.A. Cagney
Nevertheless, this trend remains inchoate.
Department of Sociology and Population Research Moreover, efforts to understand how activity pat-
Center, University of Chicago, Chicago, IL, USA terns concatenate in urban settings to form what

Springer International Publishing Switzerland 2016 597


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_26
598 C.R. Browning et al.

we term ecological networks with implications places (more generally) in shaping life
for the collective functioning of neighborhood outcomes.
and activity space environments are only just The life course approach provides an impor-
emerging (Browning and Soller 2014; Graif et al. tant organizing frame with which to understand
2014; Sampson 2012). variability in neighborhood and other place-
We begin with a brief overview of key princi- based exposures and inuences. Fundamentally,
ples of life course research and their application the concepts of trajectory sequences of roles
to understanding neighborhood and place effects. and experiences and transition changes in
We focus specically on the role of agency in the roles or state (Elder et al. 2003) implicitly incor-
construction of the sociospatial contexts of every- porate the contribution of place to life course pro-
day experience, the historical and spatial speci- cesses. Residential moves are critical transitions
city of life course trajectories and transitions, in the life course, potentially reshaping trajecto-
and the linkage of lives in space and time. We ries across a number of life domains. Educational
then turn to a short history of research on neigh- or work transitions imply signicant changes of
borhood effects, from the origins of the neighbor- settings involving a high level of exposure. Even
hood approach in the work of the Chicago School without explicit moves, residential, education
of Sociology, to the mid-twentieth century ebb in and work contexts can change around us, imply-
interest, to the current revitalization of the eld ing transitions of another kind and acknowledg-
across a number of disciplines. The core of the ing the dynamic interplay of individual experience
chapter reviews a range of challenges facing con- and context. Thus, place is heavily implicated in
temporary neighborhood research and the poten- understanding the consequences of roles and
tial for new directions rooted in more rigorous experiences across the life course.
emphasis on exposure processes. We focus pri- Elder and colleagues outline a series of prin-
marily on issues in theorizing and estimating ciples to organize the key contributions of life
multilevel inuence processes, but also discuss course research (Elder et al. 2003). A number of
concerns regarding the appropriate unit of analy- these principles are relevant to understanding the
sis, selection and causation, data resources, and connection of life course with neighborhood and
the tendency for contextual research to be siloed. place approaches. First, the life course perspec-
tive emphasizes the role of agency in the con-
struction of life course trajectories. Decisions
2 Principles of Life Course regarding residential moves, everyday activity
Research: The Role of Place patterns, and organizational involvements pro-
duce a complex constellation of person-
With Gieryn (2000), we understand places as environment interactions that is not well captured
not only geographic locations, but the material by conventional neighborhood effects research
form, social practices, and cultural representa- designs. For instance, the assumption that indi-
tions of these locations. Neighborhoods are viduals are nested within some neighborhood
residential places, but they are also composed, to unit that constitutes an equivalent and encom-
some degree, of smaller scale places such as passing space for residents (e.g., persons embed-
schools, stores, parks, and so on. Research on ded in census tracts) implies a kind of residential
neighborhoods as units of exposure is thus, determinism that belies individual agency in the
implicitly or explicitly, also research on the clus- production of daily exposures.
ter of places that give the neighborhood its mate- Second, the life course perspective explicitly
rial form and shape its social and cultural recognizes the embeddedness of developmental
organization. From the standpoint of individuals, experiences in history and place (Dannefer 1984;
residential neighborhoods are places to which Elder 1974). Neighborhood approaches add
people are regularly (but variably) exposed and richness to research focused exclusively on the
worthy of emphasis in understanding the role of micro level contexts of development, but too
Neighborhood, Place, and the Life Course 599

often such research is limited to incorporation of ciated with more prevalent social pathology?
abstract structural conditions (e.g., the neighbor- Are there features of urban social life that encour-
hood poverty level) without attention to actual age such outcomes, above and beyond the char-
exposure settings. The principle of time and acteristics of the individuals that settle in urban
place highlights the historical and temporal spec- environments? How and why are areas within cit-
icity of place-based exposures, drawing atten- ies differentiated with respect to collective and
tion to experiences in actual routine activity individual well-being? The insights that emerged
settings and their cumulative consequences. from the Chicago School integrated Durkheimian
Third, the emphasis on linked lives acknowl- views on the potentially disruptive aspects of
edges the interdependence of life course trajecto- social change with human ecological models of
ries, an insight that is particularly relevant in urban growth (Park and Burgess 1925). Thomas
understanding the social processes responsible and Znaniecki (1919) introduced the concept of
for place effects on well-being. Linked lives may social disorganization to describe the process
be understood both temporally (e.g., intergenera- by which normative structures of immigrant cul-
tional linkages) and spatially (e.g., concurrent tures became attenuated when imported to new
inuences of proximate others). Research has contexts, with implications for family function-
focused heavily on understanding the inuence ing and immigrant child outcomes. Shaw and
of social network ties on life course outcomes McKay (1942), in turn, applied the social disor-
(e.g. Cornwell et al. 2008). A more sophisticated ganization concept to the neighborhood level to
emphasis on place, however, extends the notion explain area level variation in crime rates.
of linked lives to incorporate shared routine Focusing on the implications of poverty, residen-
activities and the potentially critical role of tial instability, and race/ethnic heterogeneity for
encounters in space that would not be considered the ability of communities to realize shared val-
conventional social network ties (e.g., friendship, ues and maintain effective social controls, these
kinship, occupational, etc.). Despite the central- authors pioneered the investigation of
ity of this form of ecologically-based interaction neighborhood-level inuences on delinquency
to a number of prominent theories (e.g. Wilson and crime. A key contribution of their work was
1987), few efforts to formalize the notion of the explicit distinction between the criminogenic
linked lives in space have been undertaken. We structural features of neighborhood contexts and
consider these principles of life course research the individual residents of urban neighborhoods.
and their application to an improved understand- Subsequent studies of social disorganization
ing of neighborhood and place effects throughout extended the investigation of neighborhood inu-
the chapter. ences to include mental health outcomes (Faris
and Dunham 1939).
The framework pioneered by Chicago School
3 Neighborhood Research: scholars spawned a body of research in the mid-
A Brief History twentieth century documenting substantial spa-
tial inequality across a number of dimensions.
Research on neighborhood can be traced to the This spate of research, however, was largely
origins of the sociological discipline, most nota- characterized by macro-level analytic approaches
bly in the work of the Chicago School of that limited the capacity of researchers to disen-
Sociology. Reacting to a historical context tangle individual and contextual effects (i.e.,
marked by rapid urbanization, large-scale immi- through multilevel designs). Moreover, research
gration, and increasing salience of social prob- in the social disorganization tradition became the
lems such as urban poverty, crime, and mental subject of criticism as theoretical models of the
illness, Chicago School scholars articulated a set mechanisms translating neighborhood structural
of research questions that continue to preoccupy disadvantage into diminished wellbeing were
the social sciences. Why are urban contexts asso- increasingly recognized to be underdeveloped
600 C.R. Browning et al.

(Bursik 1988). In combination with the emerging outcomes. The empirical research that followed
dominance of the survey research model, the cri- largely corroborated the expectation that residence
tique led to relatively slow growth in macro-level in disadvantaged neighborhoods inuenced a wide
research on neighborhoods in US-based urban range of outcomes, including child and adolescent
sociology and related subelds. Population development (Brooks-Gunn et al. 1993; Elliott
researchers, however, continued to make substan- et al. 1996), sex, fertility and family formation
tive and methodological advances in this area (Baumer and South 2001; Brewster 1994), educa-
(Casterline 1985; Entwisle 2007). tional achievement (Ainsworth 2002; Harding
The 1987 publication of William Julius 2003), labor market opportunities (Massey and
Wilsons seminal text, The Truly Disadvantaged, Shibuya 1995), exposure to crime (Krivo and
marked a critical turning point in research on Peterson 1996), mental and physical health
neighborhood contexts. Wilsons core thesis (Aneshensel and Sucoff 1996; Ross 2000), and
shifted disciplinary attention away from later life wellbeing (Krause 1996).
individual-level factors as principal mechanisms Consistent support for neighborhood struc-
explaining the highly unequal outcomes of urban tural inuences on life course outcomes drew
populations by race, ethnicity, and socioeco- attention to the need for more extensive data col-
nomic status, to the sociospatial conditions to lection to assess neighborhood inuences and
which the most disadvantaged urban residents careful empirical assessment of potential mediat-
were exposed. Wilson documented dramatic ing mechanisms. Consequently, the mid- to late-
inequalities in structural aspects of neighborhood 1990s saw unprecedented, large-scale data
environments, articulating a model of both the collection efforts explicitly designed to capture
origin of concentrated poverty neighborhoods multilevel inuences on individual outcomes.
and their consequences for life prospects. Wilson The Project on Human Development in Chicago
presented a number of hypotheses regarding the Neighborhoods (PHDCN) (Earls et al. 1995), ini-
actual mechanisms accounting for neighborhood tiated in 1994, was a massive, multi-faceted data
poverty effects on resident individuals, with a collection undertaking designed to test a range of
signicant emphasis on social isolation from con- hypotheses regarding the potential social mecha-
ventional institutions, the lack of salient main- nisms that linked neighborhood structural disad-
stream role models, and compromised informal vantage to the developmental outcomes of urban
social control at the neighborhood level. The youth. A second major data collection develop-
Truly Disadvantaged ultimately set the stage for ment was the Moving to Opportunity
a veritable explosion in research on neighbor- Demonstration (MTO) an experimental study
hood contexts (Sampson et al. 2002). of the impact of moving public housing residents
Following the publication and broad impact of from high- to low-poverty environments (Goering
The Truly Disadvantaged, researchers renewed and Feins 2003). The PHDCN and MTO data
focus on the potential inuence of residential collection projects are among the more visible
places, particularly for child development. Jencks indicators of the substantial excitement that
and Mayers (1990) review of the literature high- attended research on neighborhoods in the latter
lighted the role of collective socialization practices, half of the 1990s and the promise the perspective
peer inuences, and local institutional resources as held for shedding light on the origins of dispari-
factors linking disadvantaged neighborhood envi- ties in outcomes across the life course.
ronments to child outcomes. The Massey and Research emerging in the wake of these and
Denton (1993) publication of Massey and Dentons other studies offered additional evidence in sup-
American Apartheid further energized the focus on port of the already widely held contention that
neighborhood inuences and joined the debate neighborhood structural disadvantage negatively
regarding the foundational role of residential racial inuenced a wide variety of life course outcomes
segregation in exacerbating the impact of concen- (Sampson et al. 2008; Sanbonmatsu et al. 2012).
trated poverty on urban African American Beyond the inuence of neighborhood structure,
Neighborhood, Place, and the Life Course 601

however, the PHDCN, in particular, offered the 4 Neighborhood Research:


opportunity to assess aspects of neighborhood Challenges and New
social climate and physical conditions that might Directions
independently inuence individual-level wellbe-
ing and mediate the effects of neighborhood We focus primarily on challenges concerning
structural factors. Sampson and colleagues theory and empirical investigation of multilevel
(1997) emphasized the role of collective ef- inuence processes, but also consider concerns
cacy the combination of mutual trust and regarding the appropriate unit of analysis (e.g.,
expectations regarding the collective capacity to neighborhood boundary), selection and causal
act on behalf of shared goals as a critical inference, limitations in available data, and the
community-level ingredient accounting for the siloed nature of contextual research. We draw
association between neighborhood disadvantage attention to potential new directions for neigh-
and crime, among other outcomes. Other borhood and place research and their relevance
approaches emphasized social capital in the form for understanding life course processes. We
of network ties (Bellair 1997; Bursik and emphasize, in particular, more sophisticated
Grasmick 1993; Haynie et al. 2006), social con- attention to exposure processes through the con-
tagion (Crane 1991), social and physical disorder cepts of activity space and ecological networks.
(visible cues of community decline) (Browning
et al. 2013; Cohen et al. 2000; Taylor 2001), and
local institutional resources (Small and 4.1 Multilevel Inuence Processes
McDermott 2006), among other processes.
The availability of high quality data on multi- The social process turn (Sampson et al. 2002)
ple aspects of community environments beyond placed heavy emphasis on identifying and empir-
structural background initiated a period of inten- ically investigating the mechanisms through
sive effort to adjudicate among possible mecha- which neighborhood effects are channeled. This
nisms channeling neighborhood inuence. Other research has yielded complex and equivocal nd-
large-scale studies emerged to advance this effort, ings. We have chosen to label this section multi-
including the Los Angeles Family and level inuence processes in recognition of the
Neighborhood Study (L.A.FANS) and the need to address not only what substantive mecha-
Chicago Community Adult Health Study nisms operate to channel neighborhood inu-
(CCAHS) (Morenoff et al. 2007; Sastry et al. ence, but also two other critical issues: the
2006). The theoretical and empirical emphasis on potential for differential exposure to neighbor-
neighborhood social processes constituted a sig- hood factors among residents and differential
nicant turning point in neighborhood research reception of neighborhood exposures (i.e., the
(Sampson et al. 2002). Indeed, the eld remains potential for individuals to experience the same
heavily focused on identifying the neighborhood environments differently, with implications for
processes and conditions that are most relevant outcomes).
for individual level wellbeing (Galster 2012),
which neighborhood processes are relevant for 4.1.1 Neighborhood Mechanisms
distinct aspects of wellbeing (health, behavior, The move toward more rigorous investigation of
educational outcomes, etc.; Leventhal et al. mechanisms, largely occurring during the late
2009), and the extent to which relevant processes 1990s and 2000s, has spawned a substantial lit-
actually mediate neighborhood structural inu- erature, adequate review of which is beyond the
ences on key outcomes. Though much has been scope of the current chapter. Signicant progress
accomplished since researchers began investigat- has been made with respect to the measurement
ing neighborhood social processes in earnest, the and testing of neighborhood mechanisms.
eld nevertheless faces a range of ongoing chal- Nevertheless, a sampling of research examining
lenges, to which we now turn. effects of theoretically-informed neighborhood
602 C.R. Browning et al.

mechanisms on health (mental and physical) and from victimization may also be linked with social
risk behavior (e.g., violence, crime, substance psychological outcomes and improved mental
use, and risky sexual activity)1 corroborates the health. A number of studies have found collective
assertion that ndings on the impact of neighbor- efcacy and cohesion effects on depression and
hood social processes remain equivocal. We other mental health related outcomes at various
focus on the role of collective efcacy, social net- stages of the life course (Ahern and Galea 2011;
work and contagion processes, organizational Echeverra et al. 2008; Maimon et al. 2010; Xue
resources, social/physical disorder, and routine et al. 2005). Collective efcacy has also been
activities.2 associated with physical health outcomes such as
self-rated health (Fan and Chen 2012), mortality
Collective Efcacy Collective efcacy has been (Cohen et al. 2003; Wen and Christakis 2005),
widely emphasized in the work of Sampson and physical activity (Fisher et al. 2004; Kimbro et al.
colleagues as capturing expectations for action 2011), asthma (Cagney et al. 2007), and obesity
relevant for the group-level capacity to realize (Cohen et al. 2006).
shared values. Rooted in Kornhausers (1978) Although evidence supporting collective ef-
classic reformulation of the social disorganiza- cacy as a benecial neighborhood-level mecha-
tion perspective, collective efcacy is theoreti- nism has been substantial, extant research is by
cally linked with a broad range of benets to no means conclusive. Indeed, prominent studies
neighborhood environments. For instance, norms have found no association between collective
and expectations for pro-social action may con- efcacy and key behavioral outcomes. For
tribute to regulation of the misuse of public space, instance, a study of adolescent participation in
with direct implications for the prevalence of violence using data from the PHDCN found no
problem behavior within urban communities. evidence of an association between collective
Sampson and colleagues have demonstrated the efcacy and survey-reported participation in vio-
substantial effects of collective efcacy on the lent behavior (Sampson et al. 2005). Fagan and
prevalence of violence within urban neighbor- Wright (2012) nd no association between col-
hoods (Morenoff et al. 2001; Sampson et al. lective efcacy and self-reported offending
1997, 1999) and additional studies have found among males, and a positive relationship, oppo-
regulatory effects of collective efcacy on aflia- site of the theorized direction, among females. A
tion with deviant peers (Brody et al. 2001) and number of other studies nd no evidence of a
delinquency among adolescents (Molnar et al. relationship between collective efcacy and
2008; Simons et al. 2005). Collective efcacy is physical and mental health outcomes, including
also hypothesized to promote socialization pro- self-rated health (Cagney et al. 2005; Franzini
cesses with indirect implications for a wide range et al. 2005), obesity (Burdette et al. 2006), physi-
of developmental outcomes, including the preva- cal activity (Pabayo et al. 2014), and adolescent
lence of risky sexual behavior (Browning et al. internalizing behaviors (Dupr et al. 2012).
2008) and drug use (Fagan et al. 2014). The prev- Although not denitive, ndings on the breadth
alence of trust among local neighborhood resi- and impact of collective efcacy on individual
dents and norms regarding mutual protection level wellbeing are, on balance, impressive, indi-
cating that continued investigation of this social
1
process is likely to be a productive area of inquiry.
We make only occasional reference to the now extensive
neighborhood research on educational achievement in this
section given the recent publication of high quality Social Networks Theory and empirical research
reviews of this literature (Sastry 2012; Sharkey and Faber on neighborhood social networks have been
2014).
2
longstanding interests of urban social scientists.
Space precludes adequate review of the literature on
Seminal research (Kasarda and Janowitz 1974)
neighborhood cultural mechanisms. See Small et al.
(2010) for a recent review of cultural approaches in urban heightened interest in systemic network-based
sociology. processes by which neighborhoods achieve regu-
Neighborhood, Place, and the Life Course 603

latory capacity (Bursik and Grasmick 1993). which ties may yield positive or negative
Research on the individual-level health benets outcomes.
of social network ties and support (Berkman
et al. 2000; House 2002; Thoits 2011) also Organizational resources Local organizations
prompted interest among social scientists and such as schools, churches, youth services, health
social epidemiologists in the potential inuence services, and local commerce are places that may
of macro-level neighborhood networks through provide a range of resources relevant for both the
social support and inuence processes (Browning individual level wellbeing of urban residents as
and Cagney 2002; Carpiano 2006; Kawachi and well as the effective functioning of communities
Berkman 2000). Indeed, some research has found (Marwell 2009). First, the effects of
positive effects of neighbor-based social interac- neighborhood-based organizations may operate
tion for health. Morenoff (2003), for instance, through the individual exposures to organizations
found protective effects of the extent of neighbor- of varying quality (Cubbin et al. 2012), availabil-
based reciprocal exchange on the prevalence of ity (Curley 2010), diversity (Murphy and Wallace
low-birth weight in Chicago. Network exchange 2010), and affordability. For instance, studies of
has also been linked with lower levels of adult the joint inuence of neighborhoods and schools
depression (Kim 2010; Mair et al. 2010) and are expanding rapidly; these studies offer robust
improved self-rated health (Mohnen et al. 2011). evidence of both the association between neigh-
Yet, in total, research on the role of interper- borhood and school quality and the inuence of
sonal ties in urban communities has not yielded the latter on developmental outcomes (Cook
consistent evidence of benecial effects. With et al. 2002; Kirk 2009; Owens 2010; Teitler and
respect to behavioral outcomes, for instance, Weiss 2000).
studies of neighbor networks have yielded incon- Local organizations may also be studied by
clusive evidence regarding protective effects on examining the consequences of organizations for
crime and violence (Bellair 1997; Browning et al. effective neighborhood-level functioning. Small
2004; Merry 1981; Simcha-Fagan and Schwartz and colleagues (Allard and Small 2013; Small
1986; Warner and Rountree 1997) and risk 2009) argue that organizations are constitutive of
behaviors such as alcohol use (slund and informal social processes that are consequential
Nilsson 2013; de Haan et al. 2010; Ennett et al. for urban social organization. Small (2009) found
2008; although see Fagan et al. 2007; Steen that residents of a high poverty neighborhood
2010). Indeed, some have hypothesized that who frequented a local child care organization
extensive network ties may carry negative conse- beneted not only from the direct services pro-
quences for urban communities if they serve as vided by the organization, but also from exposure
conduits for health compromising or risk-oriented to network-based resources that were, in many
behavior. Wilson (1996) has argued that highly cases, incidental outcomes of organizational par-
disadvantaged, socially isolated communities ticipation. Informal networks even those based
may combine both dense network ties and lack of on relatively weak ties emerging from local
contact with mainstream institutions and organi- organization-based interactions are a key feature
zations. In such contexts, network inuences are of the organizational embeddedness perspec-
less likely to be shaped by embeddedness in set- tive Small proposes, in which the origin of net-
tings characterized by conventional norms and work ties is critical to understanding their
expectations. Thus the consequences of network resource potential. Smalls work on the implica-
ties for collective and individual well-being are tions of organizational participation for the avail-
uncertain and potentially contingent. Extant nd- ability of valuable social resources highlights the
ings suggest the importance of research on the potential for benecial collective features of
types of network ties that are most likely to yield neighborhood life to emerge from viable organi-
benecial inuence and the conditions under zational bases in urban communities.
604 C.R. Browning et al.

Similarly, Sampson points to the relevance of homelessness, prostitution, and people drinking
organization-based resources in contributing to in public (social disorder) has received consid-
benecial neighborhood social processes erable attention in the neighborhood literature as
(Sampson 2012). Sampson measured such a precursor to both behavioral and health prob-
resources through a multi-item index tapping the lems. A number of scholars have argued that such
survey-reported presence of organizations such signs of disorder signal the weakening of local
as community newspapers, neighborhood watch social order, emboldening potential offenders
groups, block group or tenant associations, crime and inducing health-consequential fear and phys-
prevention programs, substance abuse and men- iological stress in neighborhood residents.
tal health treatment programs, family planning Research on the behavioral implications of disor-
clinics, youth centers, and afterschool recre- der or incivilities has offered some evidence
ational programs for youth, among others. that disorder contributes to crime, but notable
Organization-based resources predicted collec- challenges to this model have also appeared.
tive efcacy in longitudinal models from the Sampson and Raudenbush (1999), for instance,
PHDCN, net of a range of controls for neighbor- call into question whether disorder is a causal
hood structure and potentially confounding social factor in neighborhood variation in crime and
processes, including lagged collective efcacy. other behavioral problems or simply a correlated
Thus an emerging strand of research points to outcome determined by the same factors that
organizational features of neighborhood environ- generate spatial variability in crime (see also
ments as a critical backdrop for the development Boggess and Maskaly 2014; Geller 2007; Yang
of other potentially benecial informal social 2010). Research has offered some evidence of
processes. Findings on the consequences of neighborhood disorder effects on outcomes such
neighborhood organizational environments for as sexual risk behavior (Latkin et al. 2007), drug
aspects of wellbeing are consistent with these use (Latkin et al. 2005) and alcohol use among
studies. Research examining the effects of orga- youth (Steen 2010; Yabiku et al. 2007), although
nizational density have found evidence of protec- ndings have been inconsistent with respect to
tive effects on the prevalence of crime (Peterson the latter (Jackson et al. 2014).
et al. 2000; Slocum et al. 2013) and adolescent Arguably more robust are associations
aggression and delinquency (Molnar et al. 2008). between aspects of disorder and psychosocial
Despite signicant theoretical attention to the and health outcomes. A number of studies have
role of organizations in urban community context found links between neighborhood disorder and
and an emerging body of empirical research, mental health (Ford and Rechel 2012; Hill et al.
extant research has not assessed the role of aggre- 2005; Ross and Mirowsky 2009), sexually trans-
gate patterns of shared organizational involve- mitted disease (Cohen et al. 2000), and self-rated
ment in the everyday lives of urban residents. health and chronic conditions (Hale et al. 2010;
Studies to date have focused on specic types of Ross and Mirowsky 2001). These ndings sug-
organizations (Small 2009) or survey-reported gest that disorder cues operate more consistently
and administrative measures of organizational through fear and stress-related responses, with
density and participation. However, research on consequences for mental and physical health
the extent to which local organizations collec- (Ross and Jang 2000). Some behavioral out-
tively bring neighborhood residents together, comes (e.g., drug use and risky sexual behavior),
with implications for the development of infor- in turn, may be an indirect result of efforts to
mal social ties and mobilization capacity, has yet cope with ongoing stressful exposures (Latkin
to emerge. et al. 2007). The expanding literature on expo-
sure to violence (ETV) also offers robust evidence
Social and Physical Disorder Social and physi- of negative health consequences. Often consid-
cal disorder or visible cues of decline such as ered among disorder-related outcomes, ETV has
grafti, litter on the street (physical disorder), been associated with internalizing behavior
Neighborhood, Place, and the Life Course 605

(Browning et al. 2014; Fowler et al. 2009), child on neighborhoods has taken a cross-sectional
physical and self-rated health (Boynton-Jarrett approach with characteristics of the neighbor-
et al. 2012; Sternthal et al. 2010), and adult hood of residence at a given time point hypothe-
hypertension (Ford and Browning 2014; Mujahid sized to capture exposure to environments of
et al. 2010). In contrast to more equivocal results varying levels of advantage. Sharkey and Faber
for crime and delinquency, associations between (2014) rightly point out that understanding cumu-
neighborhood disorder, crime and health out- lative exposure to neighborhood environments
comes appear to be more robust. over the life course or even inter-generationally
(Sharkey 2013; Sharkey and Elwert 2011) is criti-
Routine Activities Finally, routine activity pat- cal to adequately assess their consequences for
terns at the neighborhood level have been empha- life prospects. Sampson and colleagues (2008)
sized in a number of recent literature reviews have attempted to estimate the inuence of cumu-
(Browning and Soller 2014; Harding et al. 2011; lative exposure to concentrated disadvantage on
Sharkey and Faber 2014). Patterns of space use verbal ability through capturing the effects of
may differ considerably across neighborhoods neighborhood of residence across multiple waves
due to variability in the availability and quality of of a longitudinal survey, employing marginal
local organizations and amenities that draw resi- structural models (Robins et al. 2000) to account
dents into the local area. Calls have increased in for time-varying confounders. Other recent work
recent years for incorporation of routine activities has found neighborhood poverty effects consis-
in theoretical models of neighborhood effects tent with cumulative exposure models of health
and data collection efforts designed to capture (Clarke et al. 2014) and fertility (South and
mobility patterns more precisely. Sampson Crowder 2010; Wodtke 2013) outcomes.
(2006), for instance, argues that the intersection Ironically, the spatial dimension of exposure
of adults and children in neighborhood space as a i.e., the actual locations urban residents encoun-
function of routine activity patterns may have ter on a regular basis has only recently received
important implications for the effectiveness of more focused attention. Yet, establishing the
informal social control efforts. However, few actual content of exposures should ideally pre-
empirical investigations of neighborhood rou- cede the effort to determine the duration of expo-
tines have emerged. This mechanism is central to sure. The emphasis on agency in life course
our discussion of future directions for neighbor- research challenges the somewhat deterministic
hood and place-based analyses; hence we reserve approach to spatial exposure taken in conven-
a more extended discussion of routine activities tional neighborhood research. For instance,
until Sect. 4.1.2 below. extant multilevel studies of neighborhood effects
on health and behavioral outcomes have tended
We next turn to a discussion of two additional to assume that residential neighborhoods encom-
issues in considering multilevel inuence pro- pass the relevant set of spatial exposures for sub-
cesses related to neighborhood effects the jects at virtually all life stages (Cagney et al.
potential for differential exposures among resi- 2013). As an operationalization of this approach,
dents of the same local area and differential researchers often assign a neighborhood context,
reception of local area inuences by individual typically a census unit such as a tract or block
level characteristics. group in the US context, to the subject based on
residential address and then attach characteristics
4.1.2 Differential Exposures of the chosen administrative unit such as poverty
A recent review of literature examining neigh- level or, when, available, social process indica-
borhood effects on achievement outcomes high- tors in order to model the effects of these contex-
lights emerging research on the temporal tual characteristics on some outcome. This
dimension of neighborhood exposure (Sharkey approach is problematic from the standpoint
and Faber 2014). The vast majority of research of both theory and measurement. For instance,
606 C.R. Browning et al.

measures of residential neighborhood exposure to assessing actual exposures among a sample of


do not capture variability in non-residential expo- British youth. They also collected independent
sures that may be included within resident activ- data on urban areas operationalized as clusters of
ity spaces. Actual day-to-day routine exposures approximately 125 households. Measures of col-
are unlikely to be neatly contained within the lective efcacy for each areal unit are then linked
boundaries of administrative units employed to with ne-grained data on the travel paths of youth
proxy neighborhoods. Moreover, assumptions over a 4-day period. In a series of studies,
about the equivalence of within-neighborhood Wikstrm and colleagues (2012) found evidence
exposures (i.e., that time spent within the neigh- that activity space exposures to areas character-
borhood boundary results in a common set of ized by varying levels of collective efcacy pre-
experiences) may also be problematic. dicts instances of delinquent and criminal activity
It follows that more precise assessment of the (although only for a subset of youth). The pro-
time and place of everyday exposures a key nounced effect of collective efcacy on youth
principle of life course research might offer behavior when assessed using activity space
more accurate representation of spatial environ- measures suggests that the equivocal ndings for
ments than conventional neighborhood measure- collective efcacy in prior research may be due to
ment. A number of contemporary studies have the use of conventional neighborhood designs.
demonstrated the extent to which traditional With respect to health outcomes, Zenk et al.
administrative units fail to capture the spatial (2011) demonstrate associations between the
extent of non-home exposures (Chaix 2009; Zenk presence of fast food outlets proximate to the
et al. 2011). Even for adolescents a life stage activity spaces of urban adolescents and increased
typically thought to be associated with greater saturated fat intake and reduced whole grain
exposure to the local neighborhood administra- intake. Kestens and colleagues (2012) employ
tively dened neighborhood boundaries appear cell phone based technologies to examine the
to misrepresent actual exposures. For instance, in links between activity space food environment
a recent study tracking the travel paths of a sam- and overweight status. Inagami and colleagues
ple of youth over the course of a day, Basta and (2007) found that non-residential exposures
colleagues found that half of the subjects spent moderated the impact of traditionally measured
91.5 % or more of their outside-the-home time in neighborhood characteristics on self-rated health.
a census tract other than the census tract where Although incipient, research on the health and
their home was located (2010, p. 1947). behavioral consequences of activity space expo-
Individuals will vary, however, in the extent to sures offers evidence that more precise measure-
which their own resources (money, time, social ment of spatial exposures is a fruitful direction
support) will enable choice regarding activity for research on place and context effects (Chaix
locations. Such choices will also be made within et al. 2013).
the spatial constraints presented by the availabil- Finally, the principle of linked lives illumi-
ity of options for routine activities both within nates the potential role of intersecting routine
and beyond the immediate neighborhood activity patterns for understanding a range of
(Furstenberg et al. 1999). social processes shaping the life course. Network
A number of researchers have focused more concepts offer an important resource for advanc-
directly on the role of activity space exposures in ing understanding of the structure of shared
assessing the relevance of context for health and exposures. Sampson, for instance, argues inter-
behavioral outcomes. Literature on delinquency neighborhood networks based on residential
and crime, for instance, has incorporated rela- moves offer insight into the social origins of links
tively sophisticated assessments of activity space. between neighborhoods. Beyond spatial proxim-
Wikstrm and colleagues (2012) developed a ity, Sampson and colleagues nd that
space-time budget approach (Anderson 1971; inter-neighborhood mobility is driven by demo-
Goodchild and Janelle 1984; Hgerstrand 1970) graphic similarity as well as features of social
Neighborhood, Place, and the Life Course 607

climate such as levels of collectively perceived community-based routine activity involving chil-
disorder (Graif and Sampson 2010; Sampson dren and adults who do not have close (e.g., fam-
2012). Building on this insight, Graif et al. (2014) ily) social network ties, but nevertheless interact
argue that identifying the extent, structure, and in meaningful ways.
inuence of non-residential neighborhood expo- More recently, Browning and colleagues
sures is an important next step in contextual (Browning et al. 2015a,b; Browning and Soller
research on crime and delinquency. They suggest 2014) examined the impact of ecological net-
that an expanded focus on the network of neigh- works measured from survey data provided by
borhoods generated through shared residential the Los Angeles Family and Neighborhood
and activity space exposures offers an opportu- Survey (L.A.FANS). Using routine activity loca-
nity to understand inuence channels across tion reports for L.A.FANS sampled households
environments linked through residential moves within 65 census tracts (roughly 40 households
and shared routines (independent of spatial per tract), they estimate the structure of tract-
proximity). based eco-networks, nding that more extensive
Browning and Cagney (Browning and Soller linkage within the network (i.e., more overlap in
2014; Cagney et al. 2013) employ the concept of routine activity locations between households) is
an ecological (or eco-) network to capture the positively associated with neighborhood collec-
intersection between people and places either tive efcacy, intergenerational closure, and net-
through residential mobility or as people engage work interaction among neighbors (Browning
in daily routine activities. More formally, the et al. 2015a). They also nd protective effects of
eco-network can be understood as an instance of more densely tied eco-networks on the preva-
a two-mode network where the nodes are loca- lence of adolescent health risk behavior
tions and actors. Eco-network processes gure (Browning et al. 2015b).
prominently in research on urban social organi- In summary, the spatial dimension of expo-
zation and child development. For instance, sure has been relatively neglected in research on
Wilson (1987, 1996) argues that urban neighbor- neighborhood and place effects on life course
hoods characterized by high levels of unemploy- outcomes. Yet exposure processes are critical for
ment limit the extent to which resident youth understanding the relevance of context. Indeed,
encounter adults engaged in work-related daily we argue that the continued lack of precise infor-
routines. In this view, regular exposure in public mation on exposures in research on place effects
space to adults engaged in employment-related exacerbate a range of additional challenges to the
practices reinforces norms about the role of work eld.
in adult lives and expectations for the life course
including future employment. In an important 4.1.3 Differential Reception
contribution to life course research (Elder et al. of Neighborhood Inuences
2003), Barker and colleagues (Barker and Alongside attention to the exposure processes
Schoggen 1973) argue that changes in the ten- through which neighborhood effects operate is
dency toward age-specialization in behavior set- increasing recognition of the potential for differ-
tings occurring in mid-twentieth century rural ential reception of neighborhood contextual
Kansas resulted in increasing spatial segregation inuences at the individual level. The term
of youth routine activities (primarily through reception is used here to capture the processes
concentrating youth activities within schools). by which people encounter the environment, rec-
These changes led to fewer opportunities for ognizing that individuals are agents in these
involvement in (and apprenticeship through) encounters, not merely passive recipients of
non-school community-based activities led by inuence. A range of studies has investigated
adults, with implications for socialization. Like cross-level interactions between features of
Wilson, Barker emphasizes the socialization con- neighborhoods and individual level characteris-
sequences of intersecting ecologies of tics. Age has been among the most prominent
608 C.R. Browning et al.

individual-level conditioning factors investigated processes are increasingly inuential compared


in such studies (Robert and Li 2001; Vartanian to other developmental periods.
and Buck 2005). Critical period approaches, Neighborhood exposures may also be particu-
for instance, suggest that exposure to features of larly relevant for older adults who experience an
neighborhood contexts during particularly sensi- increasing sense of physical vulnerability. For
tive phases of the life course may amplify effects instance, older adults may be more susceptible to
on wellbeing (Kuh and Ben-Shlomo 1997; Lynch the effects of disorder and mobility-impeding
and Smith 2005). physical decay (e.g., sidewalk cracks, potholes,
No consensus exists on which periods of the and broken curbs) (Balfour and Kaplan 2002;
life course are most likely to enhance the impact Clarke et al. 2008; Mendes de Leon et al. 2009;
of neighborhood exposures. With respect to and for a review see Yen et al. 2009). Adult years
childhood and adolescence, Anderson et al. accrued in the same neighborhood may also have
(2014) review two conceptual models of neigh- an impact in its own right. Personal and nancial
borhood effects across the early life course rele- investments that lead to linked lives reinforce
vant to understanding differential reception neighborhood inuence and may create an iden-
processes. First, the early exposure and carry for- tity informed by residential location (Rubinstein
ward model proposes that neighborhood inu- and Parmelee 1992). As noted in the discussion
ences during early periods are crucial; early of life course variability in exposures, however,
experiences set trajectories of development in generalizations about the differential salience or
motion, with long-term implications (consistent impact of neighborhood and other spatial expo-
with Vartanian and Houser 2010; Wheaton and sures across age are likely to misrepresent the
Clarke 2003). Second, the adolescent exposure complexity of environmental inuence processes
model emphasizes the importance of neighbor- (Cagney et al. 2013).
hood exposures during the teen years, a develop- A number of other characteristics have been
mental period when environments outside the considered as potential factors shaping reception
home become more salient (Sastry and Pebley of neighborhood processes most notably gen-
2010). The results of research investigating the der, but also race and ethnicity. Although too
critical period hypothesis are mixed, however. voluminous to review here, we note that the num-
For instance, Sastry and Pebley (2010) found that ber of studies investigating differential reception
the effects of average neighborhood income on of neighborhood effects have mounted rapidly
test scores were particularly strong for those aged (and are far more common than research on dif-
812 and 1317 compared to children aged 37. ferential exposures). However, three issues com-
Wheaton and Clark (2003) nd that early expo- plicate interpretation of ndings from this body
sure to neighborhood disadvantage at ages 611 of research. First, theoretical justications for
explains the association between neighborhood exploring differential reception hypotheses
and mental health in early adulthood. Anderson remain relatively underdeveloped (Sharkey and
et al. (2014) use longitudinal data to assess these Faber 2014). Although the gendered reception of
models, nding the strongest relationship adverse neighborhood environments, for instance,
between neighborhood effects and achievement has received comparatively extensive theoretical
and internalizing symptoms measured concur- treatment (Leventhal et al. 2009), research on
rently in early childhood, with some modest sup- other potential demographic or socioeconomic
port for the carry-forward model from childhood conditioning factors has been more limited.
into adolescence for reading scores. Despite prior A second concern relates to the potential for
theoretical support for the heightened importance differential reception processes to be confounded
of neighborhoods during adolescence (Phillips with exposures. Hypotheses regarding differen-
and Shonkoff 2000; Steinberg and Morris 2001), tial reception are difcult if not impossible to test
Anderson et al. do not nd support for an adoles- in the absence of detailed data on exposures. For
cent critical period during which neighborhood instance, conclusions regarding the gendered
Neighborhood, Place, and the Life Course 609

reception of neighborhood environments require (Sperling 2012) is difcult to resist, given data
that differences in exposure by gender likely constraints and the force of custom within the
substantial in some urban environments be con- eld. Despite longstanding concerns, few efforts
trolled. Finally, the foregoing discussion high- to develop theoretically motivated aggregate
lights the potential for reciprocal associations exposure units that move beyond census-dened
between differential reception and exposure pro- boundaries have been undertaken (see also Logan
cesses. Parents, for instance, may react differ- 2012, pp. 520521).
ently to comparable neighborhood To the extent that the unit of analysis question
exposures e.g., a neighborhood-based crime motivates effective measurement of exposures,
event (or changes in crime trends) may lead some activity space information introduces the possi-
parents to increase supervision levels or remove bility of corroboration i.e., the ability to address
their children from exposure to neighborhood the question of what unit of analysis best contains
environments while others may be more con- the theoretically relevant exposures. Construction
strained in their capacity to respond to local of eco-networks also allows for alternative speci-
crime patterns. Differences in responses to the cations of the relevant aggregate exposure space
environment then set in motion exposure pro- rooted in people- as opposed to exclusively
cesses with potentially substantial impact on life spaced-based designations (Kwan 2009). By
course trajectories (e.g., choice of a non- specifying the intersection of people and places
neighborhood school). In the absence of informa- as an ecological network, community detection
tion about exposures, hypotheses regarding algorithms may be employed to identify struc-
reciprocal relations between exposure and recep- tures of shared routines (i.e., clusters of people
tion processes cannot be examined, and biased that tend to encounter one another at higher rates
conclusions favoring differential reception or clusters of places that are linked by visits from
hypotheses may be drawn. Thus studies attempt- the same people). As sources of contextual inu-
ing to address the potential for differential recep- ence, these eco-communities may play an inde-
tion must carefully assess underlying exposure pendent (and potentially powerful) role when
assumptions. compared to residential neighborhoods.
These approaches offer potential for improved
measurement of communities in rural and less
4.2 Additional Challenges densely populated areas as well. A long history
of well-developed urban theory, coupled with
The range of issues we label multilevel inu- readily available data sources, has led to a pre-
ence processes remain central to ongoing ponderance of contextual research focused on
debates regarding neighborhood effects. A num- urban communities. Suburban and rural contexts
ber of other concerns, however, have also drawn currently lack sufcient research attention and
considerable attention, including the appropriate data to examine the collective processes that lead
unit of analysis, selection and causation, data to prosocial behavior or individual wellbeing.
concerns, and the siloed nature of contextual Studies that engage these communities, and that
research. provide opportunity for comparative analyses,
First, questions regarding the nature of expo- are needed (York Cornwell and Cagney 2014).
sures raise obvious concerns about the appropri- Second, causality and selection have been par-
ate unit of analysis when conducting ticularly intractable problems in studies of neigh-
neighborhood research (Hipp 2007). The concept borhood effects. Although observational research
of a neighborhood poses substantial problems of has offered relatively consistent effects of neigh-
operationalization an issue that has beset the borhood structural disadvantage on a range of
eld since its inception (Isaacs 1948; McKenzie outcomes (Brooks-Gunn et al. 1997; Pickett and
1921; Mumford 1954). As several researchers Pearl 2001; Robert 1999), these studies cannot
have noted, the tyranny of census geography demonstrate a causal relationship with certainty.
610 C.R. Browning et al.

Observational studies face the inevitable criti- ued, despite direct acknowledgement by a num-
cism that families who select into disadvantaged ber of scholars of the need for richer information
contexts may be at high risk for compromised on spatial exposures. As Harding and colleagues
outcomes. In turn, unobserved family-level risk state (in a review of the literature on neighbor-
factors may explain apparent links between fea- hood effects and educational attainment):
tures of the neighborhood context and youth Neighborhood research desperately needs new and
outcomes. far more nuanced data. In particular, we need data
Although studies such as MTO have attempted that measure how individuals and families of dif-
to approximate an experimental design (Goering ferent types allocate their time between different
places, the extent of exposure to different people
and Feins 2003; Kling et al. 2007; Sanbonmatsu and locations, as well as the consequent inuences
et al. 2012), the capacity of such designs to shed on individual behavior. (2011, p. 276; see also
light on the nature of neighborhood inuence has Sampson 2006; Sharkey and Faber 2014)
been questioned (Sampson 2008). A key concern
is the lack of information regarding the assump- The PHDCN and L.A.FANS have provided
tion that the residential move from a high to a important data resources allowing for a number
low-poverty neighborhood actually resulted in of signicant advances in neighborhood research.
sustained differences in the routine spatial expo- The next generation of neighborhood and contex-
sures of study participants (Clampet-Lundquist tual research more generally, however, will
and Massey 2008). The current status of the require up-to-date and far more detailed data on
debate on causality and selection suggests that the dynamics of urban exposures and the mecha-
estimating unbiased contextual effects is nisms through which environments inuence life
extremely difcult. course trajectories. Several projects are already
Clearly, causal inference on the effects of capitalizing on efcient survey-based methods
neighborhoods on individual outcomes is a for collecting location data, for example, mobile
thorny problem to which activity space technologies and GPS for collecting continuous
approaches provide no easy solution. travel path information, and Ecological
Nevertheless, notions fundamental to causal Momentary Assessment (surveys administered
thinking, including treatment effects and dose- over smartphones collecting information on
response, require information on the timing and exposures in real time) (Browning and Soller
duration of contextual exposures. In the context 2014; Palmer et al. 2013; Trull and Ebner-Priemer
of neighborhood and other place effects research, 2013). Moreover, the capacity to leverage Big
activity space data provide a means of capturing Data resources on the characteristics of locations
exposure to treatments more directly. Moreover, is increasing at breakneck pace with the prolifer-
exogenous shocks in the context of urban envi- ation of administrative data resources, e.g.,
ronments are rare events; however, research on through new urban data transparency initiatives
such occurrences (e.g., the community context of (City of Chicago Data Portal 2010), volunteered
adaptation to natural disasters) is far more useful geographic information (VGI; Goodchild 2007),
when information on the timing and duration of social media data, and sensor data (Kaplan and
exposures to exogenously changing local Stone 2013). Indeed, the exponential increase in
resources or risks is available (Frankenberg et al. the quantity and availability of data on places
2012; Fussell et al. 2010; Kirk 2012). holds the potential to fundamentally transform
Third, the collection of activity space data is the capacity to estimate exposures precisely.
still relatively rare and new technologies for gath- Finally, the increasing interest in neighbor-
ering spatially referenced data (e.g., mobile tech- hood and other place effects on individual out-
nologies) have only recently been integrated into comes has spawned efforts to understand
neighborhood research (Browning and Soller multi-contextual inuence processes. This
2014). The limited attention to such data in extant movement is an important corrective to the
neighborhood-oriented study designs has contin- longstanding problem of contextual research
Neighborhood, Place, and the Life Course 611

silos. For instance, with respect to youth social network interactions and cannot simply be
outcomes, Cook and colleagues (2002) empha- assumed away in understanding the consequences
size the joint inuence of neighborhoods, of such interaction (Feld 1981). Thus combining
schools, peers, and families on youth develop- activity space and network thinking in a more
mental outcomes. In this view, a singular focus comprehensive contextual model is likely to be a
on any given context may neglect the cumula- fruitful research direction.
tive inuence of all contexts, the potential for
features of contexts to interact with one another,
and the possibly mediating role of some con- 5 Conclusion
texts (e.g., peers) in the association between
others (e.g., neighborhood or school) and indi- In recent decades, neighborhood and place-based
vidual outcomes. The multi-contextual approach research has expanded rapidly with respect to
highlights the limitations of siloed contextual productivity, interdisciplinary interest, method-
research (e.g., separate literatures for school and ological advances, and multilevel data resources.
neighborhood effects) and the need for broader Although the eld faces a number of ongoing
recognition of the range of contexts relevant for challenges, we have argued that the emerging
development. Similar challenges might be lev- emphasis on more sophisticated theoretical and
eled at literatures focused on other phases of the empirical assessment of exposures holds the
life course. For instance, the need to recognize potential to substantially advance the eld. A
the joint role of activity spaces, institutional set- fundamental advantage of an emphasis on expo-
tings, and social networks in a more encompass- sures employing activity space and ecological
ing contextual theory of older adult well-being network concepts is the ability to capture actual
has been acknowledged in recent work on aging individual level and aggregate (shared) exposures
(Cagney et al. 2013). both within and beyond conventionally dened
An emphasis on multiple contextual inu- neighborhood environments.
ences on individual outcomes is an important Beyond incorporating exposures into theory
research direction. At the same time, acknowl- and data collection efforts, we see a number of
edgment of the simultaneous and potentially additional areas of potential growth. First, more
interactive effects of people and places on indi- attention to relatively neglected phases of the life
vidual outcomes adds clarity to the multi- course, such as early childhood and mid-life, is
contextual insight. For instance, a focus on warranted. Much of the extant research particu-
neighborhoods, schools, peers, and families as larly that focused on neighborhood inuences
key contexts for youth development combines has focused on child and adolescent development,
two presumably spatially-bounded places (neigh- with a burgeoning literature on the role of neigh-
borhoods and schools) with two aspects of youth borhoods in the lives of older adults. Although
social networks (peers and family). In practice, mid-life has not been ignored in this literature,
face-to-face social network interactions always neither has it been an explicit or central focus.
take place in a shared spatial setting. Settings for Research that includes adults in mid-life has
interaction might include schools, public spaces given insufcient attention to the distinct and
in residential neighborhoods, or homes. Family- potentially variable demands of this life stage.
based network interactions frequently take place Research on early childhood has not been as
in a home setting, but many occur outside the prevalent in the neighborhood literature, again
home. Similarly, face-to-face interactions with possibly due to an assumption that the more lim-
peers may occur virtually anywhere. Interactions ited exposure radii during this period may render
on social media do not typically occur in shared non-home exposures less relevant. In both cases,
spaces, but parties to such interactions always we would argue that place and activity space
occupy some setting at the time of contact. In exposures are likely to contribute signicantly to
short, space is a virtually constant feature of individual level outcomes.
612 C.R. Browning et al.

Second, an area that remains in its infancy is Aneshensel, C. S., & Sucoff, C. A. (1996). The neighbor-
hood context of adolescent mental health. Journal of
the assessment of place trajectories over time.
Health and Social Behavior, 37(4), 293310.
Rapid change in the form of gentrication doi:10.2307/2137258.
(Pattillo 2007), the suburbanization of poverty slund, C., & Nilsson, K. W. (2013). Social capital in
(Holliday and Dwyer 2009; Howell and relation to alcohol consumption, smoking, and illicit
drug use among adolescents: A cross-sectional study
Timberlake 2014), increasing segregation by
in Sweden. International Journal for Equity in Health,
income (Reardon and Bischoff 2011) vs. race and 12(1), 33. doi:10.1186/1475-9276-12-33.
ethnicity, and more general increases in spatial Bachrach, C., Adler, N. E., Chase-Lansdale, L., Hofferth,
inequality suggest the importance of collecting S. L., & Tran, V. C. (2013). The contribution of a
national network of social observatories to improving
ne-grained, theoretically relevant measures of
population health. Social Observatory Coordinating
key place-based attributes across time Network White Paper. http://socialobservatories.org/
(Raudenbush and Sampson 1999). Initiatives to papers/paper_20130703.0001
generate a longitudinal database on places have Balfour, J. L., & Kaplan, G. A. (2002). Neighborhood
environment and loss of physical function in older
emerged in recent years, but have yet to result in
adults: Evidence from the Alameda County Study.
large-scale efforts to collect rich place-level data American Journal of Epidemiology, 155(6), 507515.
that could be incorporated into longitudinal, mul- doi:10.1093/aje/155.6.507.
tilevel analyses of life course processes (Bachrach Barker, R. C., & Schoggen, P. (1973). Qualities of com-
munity life. Oxford: Jossey-Bass.
et al. 2013; Social Observatories Coordinating
Basta, L. A., Richmond, T. S., & Wiebe, D. J. (2010).
Network 2012) Neighborhoods, daily activities, and measuring health
We can only assume that research on the role risks experienced in urban environments. Social
of place in understanding inequality in life pros- Science & Medicine, 71(11), 19431950.
doi:10.1016/j.socscimed.2010.09.008.
pects will continue to grow as contextual
Baumer, E. P., & South, S. J. (2001). Community effects on
approaches to health, behavior, and other domains youth sexual activity. Journal of Marriage and Family,
increasingly attract interest. As the eld expands 63(2), 540554. doi:10.1111/j.1741-3737.2001.00540.x.
in visibility and inuence, research will benet Bellair, P. E. (1997). Social interaction and community
crime: Examining the importance of neighbor net-
from ongoing efforts to link life course processes
works. Criminology, 35(4), 677704.
to the inevitably situated nature of experiences in doi:10.1111/j.1745-9125.1997.tb01235.x.
place and time. Berkman, L. F., Glass, T., Brissette, I., & Seeman, T. E.
(2000). From social integration to health: Durkheim in
the new millennium. Social Science & Medicine, 51(6),
843857. doi:10.1016/S0277-9536(00)00065-4.
References Boggess, L. N., & Maskaly, J. (2014). The spatial context
of the disordercrime relationship in a study of Reno
Ahern, J., & Galea, S. (2011). Collective efcacy and neighborhoods. Social Science Research, 43, 168
major depression in urban neighborhoods. American 183. doi:10.1016/j.ssresearch.2013.10.002.
Journal of Epidemiology, 173(12), 14531462. Boynton-Jarrett, R., Rosenberg, L., Palmer, J. R., Boggs,
doi:10.1093/aje/kwr030. D. A., & Wise, L. A. (2012). Child and adolescent
Ainsworth, J. W. (2002). Why does it take a village? The abuse in relation to obesity in adulthood: The black
mediation of neighborhood effects on educational womens health study. Pediatrics, 130(2), 245253.
achievement. Social Forces, 81(1), 117152. doi:10.1542/peds.2011-1554.
doi:10.1353/sof.2002.0038. Brewster, K. L. (1994). Neighborhood context and the
Allard, S. W., & Small, M. L. (2013). Reconsidering the transition to sexual activity among young black
urban disadvantaged the role of systems, institutions, women. Demography, 31(4), 603614.
and organizations. The ANNALS of the American doi:10.2307/2061794.
Academy of Political and Social Science, 647(1), Brody, G. H., Conger, R., Gibbons, F. X., Ge, X., McBride
620. doi:10.1177/0002716213479317. Murry, V., Gerrard, M., & Simons, R. L. (2001). The
Anderson, J. (1971). Space-time budgets and activity inuence of neighborhood disadvantage, collective
studies in urban geography and planning. Environment socialization, and parenting on African American
and Planning A, 3(4), 353368. doi:10.1068/a030353. childrens afliation with deviant peers. Child
Anderson, S., Leventhal, T., & Dupr, V. (2014). Development, 72(4), 12311246.
Exposure to neighborhood afuence and poverty in doi:10.1111/1467-8624.00344.
childhood and adolescence and academic achievement Brooks-Gunn, J., Duncan, G. J., Klebanov, P. K., &
and behavior. Applied Developmental Science, 18(3), Sealand, N. (1993). Do neighborhoods inuence child
123138. doi:10.1080/10888691.2014.924355.
Neighborhood, Place, and the Life Course 613

and adolescent development? American Journal of What difference does the neighborhood make? The
Sociology, 99(2), 353395. doi:10.1086/230268. Journals of Gerontology Series B: Psychological
Brooks-Gunn, J., Duncan, G. J., & Aber, J. L. (1997). Sciences and Social Sciences, 60(4), S181S190.
Neighborhood poverty: Context and consequences for doi:10.1093/geronb/60.4.S181.
children (Vol. 1). New York: Russell Sage Foundation Cagney, K. A., Browning, C. R., & Wallace, D. M. (2007).
Publications. The Latino paradox in neighborhood context: The
Browning, C. R., & Cagney, K. A. (2002). Neighborhood case of asthma and other respiratory conditions.
structural disadvantage, collective efcacy, and self- American Journal of Public Health, 97(5), 919925.
rated physical health in an urban setting. Journal of doi:10.2105/AJPH.2005.071472.
Health and Social Behavior, 43(4), 383399. Cagney, K. A., Browning, C. R., Jackson, A. L., & Soller,
doi:10.2307/3090233. B. (2013). Networks, neighborhoods, and institutions:
Browning, C. R., & Soller, B. (2014). Moving beyond An integrated activity space approach for research
neighborhood: Activity spaces and ecological net- on aging. In L. J. Waite & T. J. Plewes (Eds.), New
works as contexts for youth development. Cityscape, directions in the sociology of aging (pp. 6080).
16(1), 165196. Washington, DC: National Academies Press.
Browning, C. R., Dietz, R. D., & Feinberg, S. L. (2004). Carpiano, R. M. (2006). Toward a neighborhood resource-
The paradox of social organization: Networks, collec- based theory of social capital for health: Can Bourdieu
tive efcacy, and violent crime in urban neighbor- and sociology help? Social Science & Medicine, 62(1),
hoods. Social Forces, 83(2), 503534. doi:10.1353/ 165175. doi:10.1016/j.socscimed.2005.05.020.
sof.2005.0006. Casterline, J. B. (1985). The collection and analysis of
Browning, C. R., Burrington, L. A., Leventhal, T., & community data: WFS seminar on collection and anal-
Brooks-Gunn, J. (2008). Neighborhood structural ysis of data on community and institutional factors,
inequality, collective efcacy, and sexual risk behavior 2023 June 1983. Voorburg: International Statistical
among urban youth. Journal of Health and Social Institute.
Behavior, 49(3), 269285. Chaix, B. (2009). Geographic life environments and
doi:10.1177/002214650804900303. coronary heart disease: A literature review, theoretical
Browning, C. R., Soller, B., Gardner, M., & Brooks-Gunn, contributions, methodological updates, and a research
J. (2013). Feeling disorder as a comparative and agenda. Annual Review of Public Health, 30(1),
contingent process: Gender, neighborhood conditions, 81105. doi:10.1146/annurev.
and adolescent mental health. Journal of Health and publhealth.031308.100158.
Social Behavior, 54(3), 296314. Chaix, B., Mline, J., Duncan, S., Merrien, C., Karusisi,
doi:10.1177/0022146513498510. N., Perchoux, C., et al. (2013). GPS tracking in neigh-
Browning, C. R., Gardner, M., Maimon, D., & Brooks- borhood and health studies: A step forward for envi-
Gunn, J. (2014). Collective efcacy and the contingent ronmental exposure assessment, a step backward for
consequences of exposure to life-threatening violence. causal inference? Health & Place, 21, 4651.
Developmental Psychology, 50(7), 18781890. doi:10.1016/j.healthplace.2013.01.003.
doi:10.1037/a0036767. City of Chicago Data Portal. (2010). https://data.cityof-
Browning, C. R., Calder, C. A., Soller, B., Jackson, A. L., chicago.org/
& Dirlam., J. (2015a). Ecological networks and neigh- Clampet-Lundquist, S., & Massey, D. S. (2008).
borhood social organization. Presented at the annual Neighborhood effects on economic selfsufciency: A
meetings of the American Sociological Association, reconsideration of the moving to opportunity experi-
Chicago, IL. ment. American Journal of Sociology, 114(1), 107
Browning, C. R., Soller, B., & Jackson, A. L. (2015b). 143. doi:10.1086/588740.
Neighborhoods and adolescent health-risk behavior: Clarke, P., Ailshire, J. A., Bader, M., Morenoff, J. D., &
An ecological network approach. Social Science & House, J. S. (2008). Mobility disability and the urban
Medicine, 125, 163172. doi:10.1016/j. built environment. American Journal of Epidemiology,
socscimed.2014.06.028. 168(5), 506513. doi:10.1093/aje/kwn185.
Burdette, H. L., Wadden, T. A., & Whitaker, R. C. (2006). Clarke, P., Morenoff, J. D., Debbink, M., Golberstein, E.,
Neighborhood safety, collective efcacy, and obesity Elliott, M. R., & Lantz, P. M. (2014). Cumulative expo-
in women with young children. Obesity, 14(3), 518 sure to neighborhood context consequences for health
525. doi:10.1038/oby.2006.67. transitions over the adult life course. Research on Aging,
Bursik, R. J., Jr., & Grasmick, H. G. (1993). Neighborhoods 36(1), 115142. doi:10.1177/0164027512470702.
and crime: The dimensions of effective community Cohen, D. A., Spear, S., Scribner, R., Kissinger, P., Mason,
control. New York: Lexington Books. K., & Wildgen, J. (2000). Broken windows and the
Bursik, R. J., Jr. (1988). Social disorganization and theo- risk of gonorrhea. American Journal of Public Health,
ries of crime and delinquency: Problems and pros- 90(2), 230236. doi:10.2105/AJPH.90.2.230.
pects. Criminology, 26(4), 519552. doi:10.1111/ Cohen, D. A., Farley, T. A., & Mason, K. (2003). Why is
j.1745-9125.1988.tb00854.x. poverty unhealthy? Social and physical mediators.
Cagney, K. A., Browning, C. R., & Wen, M. (2005). Social Science & Medicine, 57(9), 16311641.
Racial disparities in self-rated health at older ages: doi:10.1016/S0277-9536(03)00015-7.
614 C.R. Browning et al.

Cohen, D. A., Finch, B. K., Bower, A., & Sastry, N. In J. T. Mortimer & M. J. Shanahan (Eds.), Handbook
(2006). Collective efcacy and obesity: The potential of the life course (pp. 319). New York: Springer.
inuence of social factors on health. Social Science & doi:10.1007/978-0-306-48247-2_1.
Medicine, 62(3), 769778. doi:10.1016/j. Elliott, D. S., Wilson, W. J., Huizinga, D., Sampson, R. J.,
socscimed.2005.06.033. Elliott, A., & Rankin, B. (1996). The effects of neigh-
Cook, T. D., Herman, M. R., Phillips, M., Settersten, J., & borhood disadvantage on adolescent development.
Richard, A. (2002). Some ways in which neighbor- Journal of Research in Crime and Delinquency, 33(4),
hoods, nuclear families, friendship groups, and 389426. doi:10.1177/0022427896033004002.
schools jointly affect changes in early adolescent Ennett, S. T., Foshee, V. A., Bauman, K. E., Hussong, A.,
development. Child Development, 73(4), 12831309. Cai, L., Luz, H., et al. (2008). The social ecology of
doi:10.1111/1467-8624.00472. adolescent alcohol misuse. Child Development, 79(6),
Cornwell, B., Laumann, E. O., & Schumm, L. P. (2008). 17771791. doi:10.1111/j.1467-8624.2008.01225.x.
The social connectedness of older adults: A national Entwisle, B. (2007). Putting people into place. Demography,
prole. American Sociological Review, 73(2), 185 44(4), 687703. doi:10.1353/dem.2007.0045.
203. doi:10.1177/000312240807300201. Fagan, A. A., & Wright, E. M. (2012). The effects of
Crane, J. (1991). The epidemic theory of ghettos and neighborhood context on youth violence and delin-
neighborhood effects on dropping out and teenage quency: Does gender matter? Youth Violence and
childbearing. American Journal of Sociology, 96(5), Juvenile Justice, 10(1), 6482.
12261259. doi:10.1177/1541204011422086.
Cubbin, C., Jun, J., Margerison-Zilko, C., Welch, N., Fagan, A. A., Van Horn, M. L., Hawkins, J. D., & Arthur,
Sherman, J., McCray, T., & Parmenter, B. (2012). M. (2007). Using community and family risk and pro-
Social inequalities in neighborhood conditions: Spatial tective factors for community-based prevention plan-
relationships between sociodemographic and food ning. Journal of Community Psychology, 35(4),
environments in Alameda County, California. Journal 535555. doi:10.1002/jcop.20163.
of Maps, 8(4), 344348. doi:10.1080/17445647.2012. Fagan, A. A., Wright, E. M., & Pinchevsky, G. M. (2014).
747992. The protective effects of neighborhood collective ef-
Curley, A. M. (2010). Neighborhood institutions, facili- cacy on adolescent substance use and violence follow-
ties, and public space: A missing link for HOPE VI ing exposure to violence. Journal of Youth and
residents development of social capital? Cityscape, Adolescence, 43(9), 115. doi:10.1007/
12(1), 3363. s10964-013-0049-8.
Dannefer, D. (1984). Adult development and social the- Fan, Y., & Chen, Q. (2012). Family functioning as a medi-
ory: A paradigmatic reappraisal. American ator between neighborhood conditions and childrens
Sociological Review, 49(1), 100116. health: Evidence from a national survey in the United
doi:10.2307/2095560. States. Social Science & Medicine, 74(12), 1939
De Haan, L., Boljevac, T., & Schaefer, K. (2010). Rural 1947. doi:10.1016/j.socscimed.2012.01.039.
community characteristics, economic hardship, and Faris, R. E. L., & Dunham, H. W. (1939). Mental disor-
peer and parental inuences in early adolescent alco- ders in urban areas: An ecological study of schizo-
hol use. Journal of Early Adolescence, 30(5), 629 phrenia and other psychoses (Vol. xxxviii). Chicago/
650. doi:10.1177/0272431609341045. London: University of Chicago Press.
Dupr, V., Leventhal, T., & Vitaro, F. (2012). Feld, S. L. (1981). The focused organization of social ties.
Neighborhood processes, self-efcacy, and adolescent American Journal of Sociology, 86(5), 10151035.
mental health. Journal of Health and Social Behavior, doi:10.2307/2778746.
53(2), 183198. doi:10.1177/0022146512442676. Fisher, K. J., Li, F., Michael, Y., & Cleveland, M. (2004).
Earls, F. J., Brooks-Gunn, J., Raudenbush, S. W., & Neighborhood-level inuences on physical activity
Sampson, R. J. (1995). Project on human development among older adults: A multilevel analysis. Journal of
in Chicago neighborhoods 19952002 [Computer Aging & Physical Activity, 12(1), 4563.
le]. ICPSR13714-v1. Ann Arbor: Inter-university Ford, J. L., & Browning, C. R. (2014). Effects of exposure
Consortium for Political and Social Research to violence with a weapon during adolescence on adult
[distributor]. hypertension. Annals of Epidemiology, 24(3), 193
Echeverra, S., Diez-Roux, A. V., Shea, S., Borrell, L. N., 198. doi:10.1016/j.annepidem.2013.12.004.
& Jackson, S. (2008). Associations of neighborhood Ford, J. L., & Rechel, M. (2012). Parental perceptions of
problems and neighborhood social cohesion with the neighborhood context and adolescent depression.
mental health and health behaviors: The multi-ethnic Public Health Nursing, 29(5), 390402.
study of atherosclerosis. Health & Place, 14(4), 853 doi:10.1111/j.1525-1446.2012.01015.x.
865. doi:10.1016/j.healthplace.2008.01.004. Fowler, P. J., Tompsett, C. J., Braciszewski, J. M., Jacques-
Elder, G. H., Jr. (1974). Children of the great depression: Tiura, A. J., & Baltes, B. B. (2009). Community vio-
Social change in life experience. Chicago: University lence: A meta-analysis on the effect of exposure and
of Chicago Press. mental health outcomes of children and adolescents.
Elder, G. H., Jr., Johnson, M. K., & Crosnoe, R. (2003). Development and Psychopathology, 21(01), 227259.
The emergence and development of life course theory. doi:10.1017/S0954579409000145.
Neighborhood, Place, and the Life Course 615

Frankenberg, E., Nobles, J., & Sumantri, C. (2012). Hale, L., Hill, T. D., & Burdette, A. M. (2010). Does sleep
Community destruction and traumatic stress in post- quality mediate the association between neighborhood
tsunami Indonesia. Journal of Health and Social disorder and self-rated physical health? Preventive
Behavior, 53(4), 498514. Medicine, 51(34), 275278. doi:10.1016/j.
doi:10.1177/0022146512456207. ypmed.2010.06.017.
Franzini, L., Caughy, M., Spears, W., & Fernandez Harding, D. J. (2003). Counterfactual models of neighbor-
Esquer, M. E. (2005). Neighborhood economic condi- hood effects: The effect of neighborhood poverty on
tions, social processes, and self-rated health in low- dropping out and teenage pregnancy. American
income neighborhoods in Texas: A multilevel latent Journal of Sociology, 109(3), 676719.
variables model. Social Science & Medicine, 61(6), doi:10.1086/379217.
11351150. doi:10.1016/j.socscimed.2005.02.010. Harding, D. J., Gennetian, L., Winship, C., Sanbonmatsu,
Furstenberg, F. F., Cook, T. D., Eccles, J., & Elder, G. H., L., & Kling, J. R. (2011). Unpacking neighborhood
Jr. (1999). Managing to make it: Urban families and inuences on education outcomes: Setting the stage
adolescent success. Chicago: University of Chicago for future research. In G. J. Duncan & R. J. Murnane
Press. (Eds.), Whither opportunity? Rising inequality,
Fussell, E., Sastry, N., & VanLandingham, M. (2010). schools, and childrens life chances (pp. 277296).
Race, socioeconomic status, and return migration to New York: Russell Sage Foundation.
New Orleans after Hurricane Katrina. Population and Haynie, D., Silver, E., & Teasdale, B. (2006).
Environment, 31(13), 2042. doi:10.1007/ Neighborhood characteristics, peer networks, and ado-
s11111-009-0092-2. lescent violence. Journal of Quantitative Criminology,
Galster, G. C. (2012). The mechanism(s) of neighbour- 22(2), 147169. doi:10.1007/s10940-006-9006-y.
hood effects: Theory, evidence, and policy implica- Hill, T. D., Ross, C. E., & Angel, R. J. (2005).
tions. In M. van Ham, D. Manley, N. Bailey, Neighborhood disorder, psychophysiological distress,
L. Simpson, & D. Maclennan (Eds.), Neighbourhood and health. Journal of Health and Social Behavior,
effects research: New perspectives (pp. 2356). 46(2), 170186. doi:10.1177/002214650504600204.
Dordrecht: Springer. Hipp, J. R. (2007). Block, tract, and levels of aggregation:
doi:10.1007/978[-]94-007-2309-2_2. Neighborhood structure and crime and disorder as a
Geller, A. (2007). Neighborhood disorder and crime: An case in point. American Sociological Review, 72(5),
analysis of broken windows in New York City. 659680. doi:10.1177/000312240707200501.
Columbia University. Retrieved from http://grad- Holliday, A. L., & Dwyer, R. E. (2009). Suburban neigh-
works.umi.com/32/66/3266579.html borhood poverty in U.S. metropolitan areas in 2000.
Gieryn, T. F. (2000). A space for place in sociology. City & Community, 8(2), 155176.
Annual Review of Sociology, 26(1), 463496. doi:10.1111/j.1540-6040.2009.01278.x.
doi:10.1146/annurev.soc.26.1.463. House, J. S. (2002). Understanding social factors and
Goering, J. M., & Feins, J. D. (Eds.). (2003). Choosing a inequalities in health: 20th century progress and 21st
better life? Evaluating the moving to opportunity century prospects. Journal of Health and Social
social experiment. Washington, DC: The Urban Behavior, 43(2), 125142. doi:10.2307/3090192.
Institute. Howell, A. J., & Timberlake, J. M. (2014). Racial and eth-
Golledge, R. G., & Stimson, R. J. (1987). Analytical nic trends in the suburbanization of poverty in U.S.
behavioural geography. London: Croom Helm. metropolitan areas, 19802010. Journal of Urban
Goodchild, M. F. (2007). Citizens as sensors: The world Affairs, 36(1), 7998. doi:10.1111/juaf.12030.
of volunteered geography. GeoJournal, 69(4), 211 Inagami, S., Cohen, D. A., & Finch, B. K. (2007). Non-
221. doi:10.1007/s10708-007-9111-y. residential neighborhood exposures suppress neigh-
Goodchild, M. F., & Janelle, D. G. (1984). The city around borhood effects on self-rated health. Social Science &
the clock: Space-time patterns of urban ecological Medicine, 65(8), 17791791. doi:16/j.
structure. Environment and Planning A, 16(6), 807 socscimed.2007.05.051.
820. doi:10.1068/a160807. Isaacs, R. R. (1948). Are urban neighborhoods possible?
Graif, C., & Sampson, R. J. (2010). Inter-neighborhood Journal of Housing, 6(1), 177180.
networks and the structure of urban residential mobil- Jackson, N., Denny, S., & Ameratunga, S. (2014). Social
ity. Cambridge, MA: Harvard University, Department and socio-demographic neighborhood effects on ado-
of Sociology. lescent alcohol use: A systematic review of multi-level
Graif, C., Gladfelter, A. S., & Matthews, S. A. (2014). studies. Social Science & Medicine, 115, 1020.
Urban poverty and neighborhood effects on crime: doi:10.1016/j.socscimed.2014.06.004.
Incorporating spatial and network perspectives. Jencks, C., & Mayer, S. E. (1990). The social conse-
Sociology Compass, 8(9), 11401155. doi:10.1111/ quences of growing up in a poor neighborhood. In
soc4.12199. L. E. Lynn Jr. & M. G. H. McGeary (Eds.), Inner-city
Hgerstrand, T. (1970). What about people in Regional poverty in the United States (pp. 111185).
Science? Papers of the Regional Science Association, Washington, DC: National Academies Press.
24(1), 621. doi:10.1007/BF01936872.
616 C.R. Browning et al.

Jones, M., & Pebley, A. R. (2014). Redening neighbor- Kuh, D., & Ben-Shlomo, Y. (Eds.). (1997). A life course
hoods using common destinations: Social characteris- approach to chronic disease epidemiology. New York:
tics of activity spaces and home census tracts Oxford University Press.
compared. Demography, 51(3), 727752. doi:10.1007/ Kwan, M.-P. (2009). From place-based to people-based
s13524-014-0283-z. exposure measures. Social Science & Medicine, 69(9),
Kaplan, R. M., & Stone, A. A. (2013). Bringing the labo- 13111313. doi:10.1016/j.socscimed.2009.07.013.
ratory and clinic to the community: Mobile technolo- Kwan, M.-P. (2013). Beyond space (as we knew it):
gies for health promotion and disease prevention. Toward temporally integrated geographies of segrega-
Annual Review of Psychology, 64(1), 471498. tion, health, and accessibility. Annals of the Association
doi:10.1146/annurev-psych-113011-143736. of American Geographers, 103(5), 10781086. doi:10
Kasarda, J. D., & Janowitz, M. (1974). Community .1080/00045608.2013.792177.
attachment in mass society. American Sociological Latkin, C. A., Williams, C. T., Wang, J., & Curry, A. D.
Review, 39(3), 328339. doi:10.2307/2094293. (2005). Neighborhood social disorder as a determinant
Kawachi, I., & Berkman, L. F. (2000). Social cohesion, of drug injection behaviors. Health Psychology, 24(1),
social capital, and health. In L. F. Berkman & 96100. doi:10.1037/0278-6133.24.1.96.
I. Kawachi (Eds.), Social epidemiology (pp. 174190). Latkin, C. A., Curry, A. D., Hua, W., & Davey, M. A.
New York: Oxford University Press. (2007). Direct and indirect associations of neighbor-
Kestens, Y., Lebel, A., Chaix, B., Clary, C., Daniel, M., hood disorder with drug use and high-risk sexual part-
Pampalon, R., et al. (2012). Association between ners. American Journal of Preventive Medicine, 32(6),
activity space exposure to food establishments and S234S241. doi:10.1016/j.amepre.2007.02.023.
individual risk of overweight. PLoS ONE, 7(8), Leventhal, T., Dupr, V., & Brooks-Gunn, J. (2009).
e41418. doi:10.1371/journal.pone.0041418. Neighborhood inuences on adolescent development.
Kim, J. (2010). Neighborhood disadvantage and mental In R. M. Lerner & L. Steinberg (Eds.), Handbook of
health: The role of neighborhood disorder and social adolescent psychology (3rd ed., Vol. 2, pp. 411443).
relationships. Social Science Research, 39(2), 260 New York: Wiley.
271. doi:10.1016/j.ssresearch.2009.08.007. Logan, J. R. (2012). Making a place for space: Spatial
Kimbro, R. T., Brooks-Gunn, J., & McLanahan, S. (2011). thinking in social science. Annual Review of Sociology,
Young children in urban areas: Links among neighbor- 38(1), 507524. doi:10.1146/
hood characteristics, weight status, outdoor play, and annurev-soc-071811-145531.
television watching. Social Science & Medicine, Lynch, J., & Smith, G. D. (2005). A life course approach
72(5), 668676. doi:10.1016/j. to chronic disease epidemiology. Annual Review of
socscimed.2010.12.015. Public Health, 26(1), 135. doi:10.1146/annurev.
Kirk, D. S. (2009). Unraveling the contextual effects on publhealth.26.021304.144505.
student suspension and juvenile arrest: The indepen- Maimon, D., Browning, C. R., & Brooks-Gunn, J. (2010).
dent and interdependent inuences of school, neigh- Collective efcacy, family attachment, and urban ado-
borhood, and family social controls. Criminology, lescent suicide attempts. Journal of Health and Social
47(2), 479520. Behavior, 51(3), 307324.
doi:10.1111/j.1745-9125.2009.00147.x. doi:10.1177/0022146510377878.
Kirk, D. S. (2012). Residential change as a turning point Mair, C., Diez Roux, A. V., & Morenoff, J. D. (2010).
in the life course of crime: Desistance or temporary Neighborhood stressors and social support as
cessation? Criminology, 50(2), 329358. predictors of depressive symptoms in the Chicago
doi:10.1111/j.1745-9125.2011.00262.x. Community Adult Health Study. Health & Place,
Kling, J. R., Liebman, J. B., & Katz, L. F. (2007). 16(5), 811819. doi:10.1016/j.
Experimental analysis of neighborhood effects. healthplace.2010.04.006.
Econometrica, 75(1), 83119. Marwell, N. P. (2009). Bargaining for Brooklyn:
doi:10.1111/j.1468-0262.2007.00733.x. Community organizations in the entrepreneurial city.
Kornhauser, R. (1978). Social sources of delinquency. Chicago: University of Chicago Press.
Chicago: University of Chicago Press. Massey, D. S., & Denton, N. (1993). American apartheid:
Krause, N. (1996). Neighborhood deterioration and self- Segregation and the making of the underclass.
rated health in later life. Psychology and Aging, 11(2), Cambridge, MA: Harvard University Press.
342352. doi:10.1037/0882-7974.11.2.342. Massey, D. S., & Shibuya, K. (1995). Unraveling the tan-
Krivo, L. J., & Peterson, R. D. (1996). Extremely disad- gle of pathology: The effect of spatially concentrated
vantaged neighborhoods and urban crime. Social joblessness on the well-being of African Americans.
Forces, 75(2), 619648. doi:10.1093/sf/75.2.619. Social Science Research, 24(4), 352366. doi:10.1006/
Krivo, L. J., Washington, H. M., Peterson, R. D., ssre.1995.1014.
Browning, C. R., Calder, C. A., & Kwan, M.-P. (2013). Matthews, S. A. (2011). Spatial polygamy and the hetero-
Social isolation of disadvantage and advantage: The geneity of place: Studying people and place via ego-
reproduction of inequality in urban space. Social centric methods. In L. M. Burton, S. A. Matthews,
Forces, 92(1), 141164. doi:10.1093/sf/sot043. M. Leung, S. P. Kemp, & D. T. Takeuchi (Eds.),
Neighborhood, Place, and the Life Course 617

Communities, neighborhoods, and health (pp. 3555). Mumford, L. (1954). The neighborhood and the neighbor-
New York: Springer. hood unit. Town Planning Review, 24(4), 256.
doi:10.1007/978[-]1-4419-7482-2_3. Murphy, A. K., & Wallace, D. (2010). Opportunities for
Matthews, S. A., & Yang, T.-C. (2013). Spatial Polygamy making ends meet and upward mobility: Differences
and Contextual Exposures (SPACEs): Promoting in organizational deprivation across urban and subur-
activity space approaches in research on place and ban poor neighborhoods. Social Science Quarterly,
health. American Behavioral Scientist, 57(8), 1057 91(5), 11641186.
1081. doi:10.1177/0002764213487345. doi:10.1111/j.1540-6237.2010.00726.x.
McKenzie, R. D. (1921). The neighborhood: A study of Newsome, T. H., Walcott, W. A., & Smith, P. D. (1998).
local life in the city of Columbus, Ohio. II. American Urban activity spaces: Illustrations and application of
Journal of Sociology, 27(3), 344363. a conceptual model for integrating the time and space
Mendes de Leon, C. F., Cagney, K. A., Bienias, J. L., dimensions. Transportation, 25(4), 357377. doi:10.1
Barnes, L. L., Skarupski, K. A., Scherr, P. A., & Evans, 023/A:1005082827030.
D. A. (2009). Neighborhood social cohesion and dis- Owens, A. (2010). Neighborhoods and schools as compet-
order in relation to walking in community-dwelling ing and reinforcing contexts for educational attain-
older adults: A multi-level analysis. Journal of Aging ment. Sociology of Education, 83(4), 287311.
and Health, 21(1), 155171. doi:10.1177/0038040710383519.
doi:10.1177/0898264308328650. Pabayo, R., Molnar, B. E., Cradock, A., & Kawachi, I.
Mennis, J., & Mason, M. J. (2011). People, places, and (2014). The relationship between neighborhood socio-
adolescent substance use: Integrating activity space economic characteristics and physical inactivity
and social network data for analyzing health behavior. among adolescents living in Boston, Massachusetts.
Annals of the Association of American Geographers, American Journal of Public Health, 104(11), e142
101(1), 272291. doi:10.1080/00045608.2010.53471 e149. doi:10.2105/AJPH.2014.302109.
2. Palmer, J. R. B., Espenshade, T., Bartumeus, F., Chung,
Merry, S. E. (1981). Urban danger: Life in a neighbor- C., Ozgencil, N., & Li, K. (2013). New approaches to
hood of strangers. Philadelphia: Temple University human mobility: Using mobile phones for demo-
Press. graphic research. Demography, 50(3), 11051128.
Mohnen, S. M., Groenewegen, P. P., Vlker, B., & Flap, doi:10.1007/s13524-012-0175-z.
H. (2011). Neighborhood social capital and individual Park, R. E., & Burgess, E. W. (1925). The city. Chicago:
health. Social Science & Medicine, 72(5), 660667. University of Chicago Press. Reprint 1984.
doi:10.1016/j.socscimed.2010.12.004. Pattillo, M. (2007). Black on the block: The politics of
Molnar, B., Cerda, M., Roberts, A., & Buka, S. (2008). race and class in the city. Chicago: University of
Effects of neighborhood resources on aggressive and Chicago Press.
delinquent behaviors among urban youths. American Perchoux, C., Chaix, B., Cummins, S., & Kestens, Y.
Journal of Public Health, 98(6), 10861093. (2013). Conceptualization and measurement of envi-
doi:10.2105/AJPH.2006.098913. ronmental exposure in epidemiology: Accounting for
Morenoff, J. D. (2003). Neighborhood mechanisms and activity space related to daily mobility. Health &
the spatial dynamics of birth weight. American Place, 21, 8693. doi:10.1016/j.
Journal of Sociology, 108(5), 9761017. healthplace.2013.01.005.
doi:10.1086/374405. Peterson, R. D., Krivo, L. J., & Harris, M. A. (2000).
Morenoff, J. D., Sampson, R. J., & Raudenbush, S. W. Disadvantage and neighborhood violent crime: Do
(2001). Neighborhood inequality, collective efcacy, local institutions matter? Journal of Research in Crime
and the spatial dynamics of urban violence. and Delinquency, 37(1), 3163. doi:10.1177/0022427
Criminology, 39(3), 517558. 800037001002.
doi:10.1111/j.1745-9125.2001.tb00932.x. Phillips, D. A., & Shonkoff, J. P. (Eds.). (2000). From neu-
Morenoff, J. D., House, J. S., Hansen, B. B., Williams, rons to neighborhoods: The science of early childhood
D. R., Kaplan, G. A., & Hunte, H. E. (2007). development. Washington, DC: National Academies
Understanding social disparities in hypertension prev- Press.
alence, awareness, treatment, and control: The role of Pickett, K. E., & Pearl, M. (2001). Multilevel analyses of
neighborhood context. Social Science & Medicine, neighbourhood socioeconomic context and health out-
65(9), 18531866. doi:10.1016/j. comes: A critical review. Journal of Epidemiology and
socscimed.2007.05.038. Community Health, 55(2), 111122. doi:10.1136/
Mujahid, M. S., Roux, A. V. D., Cooper, R. C., Shea, S., & jech.55.2.111.
Williams, D. R. (2010). Neighborhood stressors and Raudenbush, S. W., & Sampson, R. J. (1999). Ecometrics:
race/ethnic differences in hypertension prevalence (the Toward a science of assessing ecological settings, with
Multi-Ethnic Study of Atherosclerosis). American application to the systematic social observation of
Journal of Hypertension, 24(2), 187193. doi:10.1038/ neighborhoods. Sociological Methodology, 29(1),
ajh.2010.200. 141. doi:10.1111/0081-1750.00059.
618 C.R. Browning et al.

Reardon, S. F., & Bischoff, K. (2011). Income inequality Sampson, R. J., Morenoff, J. D., & Gannon-Rowley, T.
and income segregation. American Journal of (2002). Assessing neighborhood effects: Social pro-
Sociology, 116(4), 10921153. doi:10.1086/657114. cesses and new directions in research. Annual Review
Robert, S. A. (1999). Socioeconomic position and health: of Sociology, 28, 443478. doi:10.2307/3069249.
The independent contribution of community socioeco- Sampson, R. J., Morenoff, J. D., & Raudenbush, S. (2005).
nomic context. Annual Review of Sociology, 25, 489 Social anatomy of racial and ethnic disparities in vio-
516. doi:10.2307/223514. lence. American Journal of Public Health, 95(2), 224
Robert, S. A., & Li, L. W. (2001). Age variation in the 232. doi:10.2105/AJPH.2004.037705.
relationship between community socioeconomic sta- Sampson, R. J., Sharkey, P., & Raudenbush, S. W. (2008).
tus and adult health. Research on Aging, 23(2), 234 Durable effects of concentrated disadvantage on ver-
259. doi:10.1177/0164027501232005. bal ability among African-American children.
Robins, J. M., Hernn, M. A., & Brumback, B. (2000). Proceedings of the National Academy of Sciences,
Marginal structural models and causal inference in 105(3), 845852. doi:10.1073/pnas.0710189104.
epidemiology. Epidemiology, 11(5), 550560. Sanbonmatsu, L., Potter, N. A., Adam, E., Duncan, G. J.,
Ross, C. E. (2000). Neighborhood disadvantage and adult Katz, L. F., Kessler, R. C., et al. (2012). The long-term
depression. Journal of Health and Social Behavior, effects of moving to opportunity on adult health and
41(2), 177187. doi:10.2307/2676304. economic self-sufciency. Cityscape, 14(2), 109136.
Ross, C. E., & Jang, S. J. (2000). Neighborhood disorder, Sastry, N. (2012). Neighborhood effects on childrens
fear, and mistrust: The buffering role of social ties achievement: A review of recent research. In R. King
with neighbors. American Journal of Community & V. Maholmes (Eds.), The Oxford handbook of pov-
Psychology, 28(4), 401420. erty and child development (pp. 423447). New York:
doi:10.1023/A:1005137713332. Oxford University Press. doi:10.1093/
Ross, C. E., & Mirowsky, J. (2001). Neighborhood disad- oxfordhb/9780199769100.001.0001/
vantage, disorder, and health. Journal of Health and oxfordhb-9780199769100-e-24.
Social Behavior, 42(3), 258276. doi:10.2307/3090214. Sastry, N., & Pebley, A. R. (2010). Family and neighbor-
Ross, C. E., & Mirowsky, J. (2009). Neighborhood disor- hood sources of socioeconomic inequality in chil-
der, subjective alienation, and distress. Journal of drens achievement. Demography, 47(3), 777800.
Health and Social Behavior, 50(1), 4964. doi:10.1353/dem.0.0114.
doi:10.1177/002214650905000104. Sastry, N., Ghosh-Dastidar, B., Adams, J., & Pebley, A. R.
Rubinstein, R. I., & Parmelee, P. A. (1992). Attachment to (2006). The design of a multilevel survey of children,
place and the representation of the life course by the families, and communities: The Los Angeles Family
elderly. In I. Altman & S. M. Low (Eds.), Place and Neighborhood Survey. Social Science Research,
attachment (pp. 139163). New York: Springer. 35(4), 10001024. doi:10.1016/j.
doi:10.1007/978-1-4684-8753-4_7. ssresearch.2005.08.002.
Sampson, R. J. (2006). How does community context Schnfelder, S., & Axhausen, K. W. (2002). Measuring
matter? Social mechanisms and the explanation of the size and structure of human activity spacesthe
crime rates. In P.-O. H. Wikstrm & R. J. Sampson longitudinal perspective. In Arbeitsbericht Verkehrs-
(Eds.), The explanation of crime: Context, mecha- und Raumplanung 135. Institut fr Verkehrsplanung
nisms, and development (pp. 3160). Cambridge: und Transportsysteme (IVT), ETH Zrich, Zrich.
Cambridge University Press. doi:10.3929/ethz-a-004444846.
Sampson, R. J. (2008). Moving to inequality: Schnfelder, S., & Axhausen, K. W. (2003). Activity
Neighborhood effects and experiments meet social spaces: Measures of social exclusion? Transport
structure. American Journal of Sociology, 114(1), Policy, 10(4), 273286. doi:10.1016/j.
189231. doi:10.1086/589843. tranpol.2003.07.002.
Sampson, R. J. (2012). Great American city: Chicago and Sharkey, P. (2013). Stuck in place: Urban neighborhoods
the enduring neighborhood effect. Chicago: University and the end of progress toward racial equality.
of Chicago Press. Chicago: University of Chicago Press.
Sampson, R. J., & Raudenbush, S. W. (1999). Systematic Sharkey, P., & Elwert, F. (2011). The legacy of disadvan-
social observation of public spaces: A new look at dis- tage: Multigenerational neighborhood effects on cog-
order in urban neighborhoods. American Journal of nitive ability. American Journal of Sociology, 116(6),
Sociology, 105(3), 603651. doi:10.1086/210356. 19341981. doi:10.1086/660009.
Sampson, R. J., Raudenbush, S. W., & Earls, F. J. (1997). Sharkey, P., & Faber, J. W. (2014). Where, when, why, and
Neighborhoods and violent crime: A multilevel study for whom do residential contexts matter? Moving
of collective efcacy. Science, 277(5328), 918924. away from the dichotomous understanding of neigh-
doi:10.1126/science.277.5328.918. borhood effects. Annual Review of Sociology, 40(1),
Sampson, R. J., Morenoff, J. D., & Earls, F. J. (1999). 559579. doi:10.1146/annurev-soc-071913-043350.
Beyond social capital: Spatial dynamics of collective Shaw, C., & McKay, H. D. (1942). Juvenile delinquency
efcacy for children. American Sociological Review, and urban areas. Chicago: University of Chicago
64(5), 633660. doi:10.2307/2657367. Press.
Neighborhood, Place, and the Life Course 619

Simcha-Fagan, O., & Schwartz, J. E. (1986). Thoits, P. A. (2011). Mechanisms linking social ties and
Neighborhood and delinquency: An assessment of support to physical and mental health. Journal of
contextual effects. Criminology, 24(4), 667699. Health and Social Behavior, 52(2), 145161.
doi:10.1111/j.1745-9125.1986.tb01507.x. doi:10.1177/0022146510395592.
Simons, R. L., Simons, L. G., Burt, C. H., Brody, G. H., & Thomas, W. I., & Znaniecki, F. (1919). The Polish peasant
Cutrona, C. (2005). Collective efcacy, authoritative in Europe and America: Monograph of an immigrant
parenting and delinquency: A longitudinal test of a group. Chicago: The University of Chicago Press.
model integrating community-and family-level pro- Trull, T. J., & Ebner-Priemer, U. (2013). Ambulatory
cesses. Criminology, 43(4), 9891029. assessment. Annual Review of Clinical Psychology,
doi:10.1111/j.1745-9125.2005.00031.x. 9(1), 151176. doi:10.1146/
Slocum, L. A., Rengifo, A. F., Choi, T., & Herrmann, annurev-clinpsy-050212-185510.
C. R. (2013). The elusive relationship between com- Vartanian, T. P., & Buck, P. W. (2005). Childhood and
munity organizations and crime: An assessment across adolescent neighborhood effects on adult income:
disadvantaged areas of the South Bronx. Criminology, Using siblings to examine differences in ordinary least
51(1), 167216. doi:10.1111/1745-9125.12001. squares and xedeffect models. Social Service
Small, M. L. (2009). Unanticipated gains: Origins of net- Review, 79(1), 6094. doi:10.1086/426718.
work inequality in everyday life. Oxford: Oxford Vartanian, T. P., & Houser, L. (2010). The effects of child-
University Press. hood neighborhood conditions on self-reports of adult
Small, M. L., & McDermott, M. (2006). The presence of health. Journal of Health and Social Behavior, 51(3),
organizational resources in poor urban neighborhoods: 291306. doi:10.1177/0022146510378241.
An analysis of average and contextual effects. Social Warner, B. D., & Rountree, P. W. (1997). Local social ties
Forces, 84(3), 16971724. doi:10.1353/sof.2006.0067. in a community and crime model: Questioning the sys-
Small, M. L., Harding, D. J., & Lamont, M. (2010). temic nature of informal social control. Social
Reconsidering culture and poverty. The ANNALS of Problems, 44(4), 520536. doi:10.2307/3097221.
the American Academy of Political and Social Science, Wen, M., & Christakis, N. A. (2005). Neighborhood
629(1), 627. doi:10.1177/0002716210362077. effects on posthospitalization mortality: A population-
Social Observatories Coordinating Network. (2012). based cohort study of the elderly in Chicago. Health
Vision statement for a national network of social Services Research, 40(4), 11081127.
observatories. Social Observatory Coordinating doi:10.1111/j.1475-6773.2005.00398.x.
Network White Paper. http://socialobservatories.org/ Wheaton, B., & Clarke, P. (2003). Space meets time:
vision. Integrating temporal and contextual inuences on
South, S. J., & Crowder, K. (2010). Neighborhood pov- mental health in early adulthood. American
erty and nonmarital fertility: Spatial and temporal Sociological Review, 68(5), 680706.
dimensions. Journal of Marriage and the Family, doi:10.2307/1519758.
72(1), 89104. Wikstrm, P.-O. H., Oberwittler, D., Treiber, K., &
doi:10.1111/j.1741-3737.2009.00685.x. Hardie, B. (2012). Breaking rules: The social and situ-
Sperling, J. (2012). The tyranny of census geography: ational dynamics of young peoples urban crime.
Small-area data and neighborhood statistics. Oxford: Oxford University Press.
Cityscape, 14(2), 219223. Wilson, W. J. (1987). The truly disadvantaged: The inner
Steen, J. A. (2010). A multilevel study of the role of envi- city, the underclass, and public policy. Chicago:
ronment in adolescent substance use. Journal of Child University of Chicago Press.
& Adolescent Substance Abuse, 19(5), 359371. doi:1 Wilson, W. J. (1996). When work disappears: The world
0.1080/1067828X.2010.502479. of the new urban poor. New York: Knopf.
Steinberg, L., & Morris, A. S. (2001). Adolescent devel- Wodtke, G. T. (2013). Duration and timing of exposure to
opment. Journal of Cognitive Education and neighborhood poverty and the risk of adolescent
Psychology, 2(1), 5587. parenthood. Demography, 50(5), 17651788.
doi:10.1891/194589501787383444. doi:10.1007/s13524-013-0219-z.
Sternthal, M., Jun, H.-J., Earls, F. J., & Wright, R. J. Xue, Y., Leventhal, T., Brooks-Gunn, J., & Earls, F. J.
(2010). Community violence and urban childhood (2005). Neighborhood residence and mental health
asthma: A multilevel analysis. European Respiratory problems of 5- to 11-year-olds. Archives of General
Journal, 36(6), 14001409. Psychiatry, 62(5), 554563. doi:10.1001/
doi:10.1183/09031936.00003010. archpsyc.62.5.554.
Taylor, R. B. (2001). Breaking away from broken win- Yabiku, S. T., Dixon Rayle, A., Okamoto, S. K., Marsiglia,
dows: Baltimore neighborhoods and the nationwide F. F., & Kulis, S. (2007). The effect of neighborhood
ght against crime, grime, fear, and decline. Boulder: context on the drug use of American Indian youth of
Westview Press. the Southwest. Journal of Ethnicity in Substance
Teitler, J. O., & Weiss, C. C. (2000). Effects of neighbor- Abuse, 6(2), 181204. doi:10.1300/J233v06n02_11.
hood and school environments on transitions to rst Yang, S.-M. (2010). Assessing the spatialtemporal rela-
sexual intercourse. Sociology of Education, 73(2), tionship between disorder and violence. Journal of
112132. doi:10.2307/2673240.
620 C.R. Browning et al.

Quantitative Criminology, 26(1), 139163. study. The Journals of Gerontology Series B:


doi:10.1007/s10940-009-9085-7. Psychological Sciences and Social Sciences, 69(Suppl
Yen, I. H., Michael, Y. L., & Perdue, L. (2009). 2), S51S63. doi:10.1093/geronb/gbu052.
Neighborhood environment in studies of health of Zenk, S. N., Schulz, A. J., Matthews, S. A., Odoms-
older adults: A systematic review. American Journal of Young, A., Wilbur, J., Wegrzyn, L., et al. (2011).
Preventive Medicine, 37(5), 455463. doi:10.1016/j. Activity space environment and dietary and physical
amepre.2009.06.022. activity behaviors: A pilot study. Health & Place,
York Cornwell, E., & Cagney, K. A. (2014). Assessment 17(5), 11501161. doi:10.1016/j.
of neighborhood context in a nationally representative healthplace.2011.05.001.
Part V
The Life Course
and Policy: Building the Nexus
Life Course Research
and the Shaping of Public Policy

John H. Laub

1 The Nexus of Life Course one. Ideas must guide public policy. In this chap-
and Public Policy1 ter, I examine the nexus between one idea the
life course and public policy. I argue that the
Social scientists are often frustrated by the appar- life course paradigm is especially relevant for
ent disconnect between sound empirical research shaping public policy.2 A life course perspective
and public policy initiatives. This occurs across a offers ideas, data, and tools that can readily shape
number of domains including education, health, and enhance public policy. Another attraction is
labor, and the focus of this chaptercrime and the fact that a life course perspective is inherently
criminal justice. Recently, there have been sev- interdisciplinary. Indeed, life course researchers
eral attempts to better connect research evidence are drawn from a number of disciplines and elds
and public policy. However, these attempts have including sociology, psychology, economics, life
focused somewhat narrowly on improving sciences, gerontology, criminology, social work,
research methodology (e.g., using more random- public policy, and law.
ized controlled trials in evaluation studies) or by In this rst section, I focus on the role of life
creating new strategies of dissemination (e.g., course theory and life course methodology. I then
what works clearinghouses and the ubiquitous highlight several life course principles that are
one pager). A focus on new means of dissemina- central for shaping effective public policies per-
tion is especially timely given recent changes in taining to crime and criminal justice.
how individuals access information drawing on
new forms of social media (e.g., Twitter).
While these new strategies may well bear fruit 1.1 The Role of Life Course Theory
I believe the challenge is largely an intellectual and Method

1
The focus of this chapter is on how life course research The life course perspective is a theoretical orien-
can inform public policy. I recognize that the life course tation, a research methodology, and an empirical
can also be altered by public policies. As but one example, eld of study. Many of the key ideas from the life
witness the fourfold increase in the rate incarceration
in the United States over the last 40 years (see Travis et al.
2014). Space limitations preclude me from tackling this 2
In an earlier paper, I argued that a life course perspective
second issue.
could serve as a paradigm for the eld of criminology.
J.H. Laub (*) Drawing on known facts about crime, I made the case that
Department of Criminology and Criminal Justice, a life course perspective provided insight and understand-
University of Maryland, College Park, MD, USA ing of the causes and dynamics of crime across the life
e-mail: jlaub@umd.edu span (see Laub 2006).

Springer International Publishing Switzerland 2016 623


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_27
624 J.H. Laub

course social change; trajectories, transitions, trajectories they can redirect paths. Social
turning points; behavioral continuity and change; institutions and triggering life events that may
cohort and historical period generate important modify trajectories (either in a positive or nega-
insight and understanding of crime and human tive direction) include school, work, the military,
development across the life span. Elder (1999) residential change, marriage, parenthood, and
uses the term life course theory to refer to a the- contact with the criminal justice system.
oretical orientation that establishes a common Given this theoretical orientation it is no sur-
eld of inquiry by dening a framework that prise that life course researchers use longitudinal
guides research in terms of problem identica- data to answer questions about human develop-
tion and formulation, variable selection and ratio- ment over life span. Prospective (forward look-
nales, and strategies of design and analysis ing) longitudinal research designs form the heart
(p. 302). of life-course research (as opposed to retrospec-
A life course perspective focuses on temporal tive life-histories). Data covering the period from
dimensions of individual lives and societal struc- childhood through adulthood is preferred. The
ture. Much of social science research takes a good news is that there are a number of statistical
snapshot look at social reality; namely, draw- techniques that can be used to analyze longitudi-
ing on cross-sectional research that captures one nal data (see Elder and Giele 2009). In addition,
point in time. In contrast, the life course perspec- life course researchers often use both quantitative
tive shifts focus to the dynamics of lives and and qualitative data to link social history and
social change. In other words, a life course per- social structure to the unfolding of human lives
spective could be thought of as a movie that (see Elder and Giele 2009; cross-reference
captures the unfolding of lives over time. The life Hermanowicz, chapter Longitudinal Qualitative
course has been dened as pathways through the Research this volume; Macmillan and
age differentiated life span (Elder 1985: 17), in Furstenberg, chapter The Logic and Practice of
particular the sequence of culturally dened Growth Curve Analysis: Modeling Strategies for
age-graded roles and social transitions that are Life Course Dynamics this volume).
enacted over time (Caspi et al. 1990: 15). Two
central concepts underlie life-course dynamics.
A trajectory is a line of development over the life 1.2 Important Life Course
span such as worklife, parenthood, or criminal Principles
behavior. Trajectories refer to long-term patterns
of behavior and are marked by a sequence of A life course perspective recognizes the impor-
transitions. Transitions are marked by life events tance of putting persons in context individ-
such as rst job or rst marriage that are embed- ual, group, and community/ neighborhood
ded in trajectories and evolve over shorter time characteristics matter. Drawing on the ecological
spans (Elder 1985: 3132). perspective of human development from Urie
The interlocking nature of trajectories and Bronfenbrenner (1979), my view is that individu-
transitions may generate turning points or a als are embedded in a broader social develop-
change in the life course (Elder 1985: 32). mental and structural context and that individual
Adaptation to life events is crucial because the behavior is the product of the interaction between
same event or transition followed by different individual development and social context (e.g.,
adaptations can lead to different trajectories family, school, and neighborhood).
(Elder 1985: 35). The long-term view embodied Elder (1999) identied four paradigmatic
by the life-course focus on trajectories implies a principles for the study of the life course. The
strong connection between childhood events and rst principle is The life course of individuals is
experiences in adulthood. However, the simulta- embedded in and shaped by the historical times
neous shorter-term view also implies that transi- and events they experience over their lifetime
tions or turning points can modify life (Elder 1999: 304). Macro level events such as the
Life Course Research and the Shaping of Public Policy 625

Great Depression in the 1930s and World War II The fourth principle is Individuals construct
during the 1940s have had profound effects on their own life course through the choices and
individual development. More recently, the actions they take within the opportunities and
effects of social unrest during the 1960s, the ter- constraints of history and social circumstances
rorist attacks on 9/11/01, and the Great Recession (Elder 1999: 308). This suggests that in a life
in 2008 were equally profound. With respect to course perspective individuals are not passive
crime and criminal justice, growing up during the vessels, but rather they are integral actors exercis-
War on Crime (1960s and 1970s), the War on ing agency in shaping life course outcomes. The
Drugs (1980s and 1990s), and the War on Terror challenge is to assess how social structures affect
(2001 to the present) shaped life course develop- individual actions and in turn to examine how
ment largely through increased contact with the individual actions change social structures.
criminal justice system.
The second principle is The developmental
impact of a succession of life transitions or events 2 Life-Course Criminology
is contingent on when they occur in a persons
life (Elder 1999: 306). This principle draws Since 1987, my colleague, Robert Sampson, and
attention to the timing of both historical and life I have been engaged in a long-term research proj-
events. As Elders research demonstrated experi- ect examining continuity and change in criminal
encing the Great Depression as a young child had offending from childhood through old age. This
different effects compared to those who experi- project generated two award winning books
enced the Great Depression during adolescence (Crime in the Making: Pathways and Turning
(for details see Elder 1999). In criminology, there Points Through Life (1993) and Shared
is good evidence that experiencing an arrest dur- Beginnings and Divergent Lives: Delinquent
ing adolescence could have severe consequences Boys to Age 70 (2003); an edited volume of the
with respect to school completion (see Kirk and Annals (2005); 28 articles, 18 book chapters, and
Sampson 2013). scores of presentations. In 2011, Robert Sampson
The third paradigmatic principle is Lives and I were awarded the Stockholm Prize in
are lived interdependently, and social-historical Criminology for our research on how and why
inuences are expressed through this network offenders stop offending. We coined the term
of shared relationships (Elder 1999: 307). It is life-course criminology to describe our theo-
thus important to consider not only that lives retical and empirical work. In this section, I dis-
are linked horizontally say husbands and wives, cuss our theoretical framework, the longitudinal
but also how lives are linked vertically and data we collected and analyzed, and our key
across generations (grandparents and parents research ndings.
and parents and children). This idea of linked
lives is especially relevant for the study of
crime and justice. For instance, what is the rela- 2.1 Theoretical Framework
tionship between parental criminality and the
criminality of their offspring? Also, the fact is In our research program, we drew on the life
that the targets of criminal justice policies are course perspective to better understand the
often linked to others, especially other family development and patterning of crime across the
members such as children. For example, life span. We were especially interested in docu-
Wakeeld and Wildeman (2014) have exam- menting the trajectories of criminal offending
ined the deleterious consequences of high rates over time and examining key transitions within
of incarceration for children. In addition, whole trajectories that may be turning points in the life
communities can be affected by criminal justice course. In doing so, we developed a theoretical
policies such as mass incarceration (see framework to explain childhood antisocial
Travis et al. 2014). behavior, adolescent delinquency, and crime in
626 J.H. Laub

early adulthood. The general organizing princi- quency age, general IQ, ethnicity, and neigh-
ple was that crime is more likely to occur when borhood socio-economic status. The Gluecks
an individuals bond to society is attenuated. Our were interested in learning how it was that two
analysis of the causes of delinquency shared boys with the same age, IQ, ethnic make-up, and
much in common with the focus in classical neighborhood SES could have different outcomes
social control theory (Hirschi 1969) on adoles- with respect to delinquency. To ensure an ade-
cence, but the reality of later life-course mile- quate comparison, boys in reform school were
stones required us to develop a modied selected for their serious and/or persistent
theoretical perspective. After all, the transition to involvement in crime while boys in public
young adulthood brings with it new social con- schools were selected for their absence of serious
trol institutions and potential turning points that and/or persistent delinquency based on ofcial
go well beyond adolescence. record checks and interviews with parents, teach-
We thus developed an age-graded theory ers, and the boys themselves.
emphasizing informal social controls that are A large amount of biological, psychological,
manifested in shifting and possibly transforma- and sociological data were collected for each boy
tive ways as individuals age. For example, we and their families at wave one (average age 14)
focused on parenting styles (supervision, warmth, (see Glueck and Glueck 1950 for details). But of
consistent discipline) and emotional attachment even greater interest from a life course perspec-
to parents in childhood, school attachment and tive was that the same boys were followed up by
peers in adolescence, and marital stability, mili- the Gluecks at age 25 and again at age 32 (see
tary service, and employment in adulthood. Glueck and Glueck 1968 for details). One of the
Although these are manifestly distinct domains remarkable features of the Gluecks study was
that are age graded, we argued that there are the large number of data sources that were used.
higher-order commonalities with respect to the Before the era of Big Data and Integration
concept of social connectivity through time (see Across Administrative Records, the Gluecks
Sampson and Laub 1993; Laub and Sampson research team conducted interviews with the
2003 for details). boys themselves at all three waves as well as
interviews with parents, teachers, employers,
spouses, friends, community members, social
2.2 The Glueck Data workers, probation ofcers, parole ofcers, and
psychiatrists. In addition, criminal histories were
We used a unique longitudinal data set to test our examined across various state depositaries and
theory of crime. In the rst part of our research from the FBI. Finally, administrative records
project, we reconstructed what is known in crimi- were surveyed across a number of social service
nology as the Unraveling Juvenile Delinquency agencies and from the military. My colleague,
study conducted by Sheldon and Eleanor Glueck George Vaillant referred to the Glueck data
(1950, 1968). This classic study began in 1940 archive as the Hubble Telescope for social sci-
and the initial data collection lasted 10 years. The entists (see also Elder and Giele 2009: 2).
two samples of boys were studied. The rst con- In the rst phase of our project we recon-
sisted of 500 delinquent boys who were incarcer- structed and reanalyzed the Glueck data for our
ated in two reform schools in Massachusetts. The book, Crime in the Making: Pathways and Turning
second consisted of 500 nondelinquent boys who Points Through Life (see Sampson and Laub 1993
were enrolled in the Boston public schools. All of for details). Shortly after the publication of Crime
the boys were born in Boston during the late in the Making, we decided to launch a follow-up
1920s or early 1930s. An unusual feature of the of the delinquent men from the Gluecks study.
Gluecks UJD design was that each delinquent We knew that 475 of the original 500 delinquents
and nondelinquent boy were matched on four had survived to age 32, but the last point of con-
characteristics believed to be related to delin- tact was circa 1960, more than 30 years ago.
Life Course Research and the Shaping of Public Policy 627

Nevertheless, we launched our follow-up study school processes appeared to be most important
focusing on three components. The rst was a in the causal chain (for details see Sampson and
criminal record search both at the state level as Laub 1993: chapters 4 and 5).
well as using FBI rap sheets. The second was a Second, we found continuity in crime and
search of death records both at the state level as other problematic behaviors across the life span.
well as using the National Death Index. The third More specically, in our analyses of the Glueck
and most challenging was to locate and interview data, we found that independent of age, IQ,
a subset of the men as they approached age 70. neighborhood SES, and ethnicity, the original
Using information gleaned from the criminal delinquents and nondelinquents displayed behav-
record search, we selected a stratied sample of ioral consistency both homotypic and hetero-
men ranging from those men with no arrests since typic3 well into adulthood. Indeed, delinquency
adolescence to men who were arrested in each and other forms of antisocial conduct in child-
decade of their life. We ended up conducting life hood were strongly related to troublesome adult
history interviews with 52 men. We merged the behavior including crime, incarceration, eco-
quantitative and qualitative data in our follow-up nomic dependency, unemployment, marital dis-
study from age 32 to roughly age 70 with the data cord, and divorce (for details see Sampson and
collected by Gluecks data from childhood to age Laub 1993: chapter 6).
32 to have a complete set of data on criminal Third, using the data on crime over the long
offending over the full life span. These combined haul, we probed the relationship between age and
data may well be the longest longitudinal study of crime. Our analyses of the Glueck data showed
criminal offenders in the world. that, on the one hand, the aggregate age-crime
curve was not the same as individual age-crime
trajectories, lending apparent support to one of
2.3 Review of Key Research the major claims of the criminal career model.
Findings More precisely, we found enormous variability in
the peak ages of offending and the age at desis-
What did we learn over the course of our long- tance varied markedly across the Glueck men
term project? A number of key research ndings (see Laub and Sampson 2003: chapter 5). On the
emerged from our reanalysis of the reconstructed other hand, we found that crime declines with
Glueck data archive as well as from our own fol- age even for active offenders and that trajectories
low-up study. of desistance cannot be prospectively identied
First, consistent with our theory of age-graded based on typological accounts rooted in child-
informal social control, we found that informal hood and individual differences (see, for exam-
social control processes are crucial in under- ple, Moftt 1993). That is, drawing on extant
standing the onset of delinquency. More speci- theory and a multitude of childhood and adoles-
cally, in our analyses of the Glueck data, we cent risk factors, offenses eventually decline for
found that the strongest and most consistent all groups of offenders. Simply put, desistance
effects on both ofcial and unofcial delinquency processes are at work even for the highest-risk
ow from the social processes of family, school, and predicted life-course persistent offenders.
and peers. Low levels of parental supervision, While prognoses from childhood factors such
erratic, threatening, and harsh discipline, and low IQ, aggressive temperament, or early onset
weak parental attachment were strongly related of antisocial behavior were modestly accurate in
to delinquency. School attachment had large neg-
ative effects on delinquency independent of fam- 3
Homotypic continuity is continuity in similar behaviors
ily processes. Attachment to delinquent peers had across the life course such as the linkage between juvenile
delinquency and adult crime. Heterotypic continuity is
a signicant positive effect on delinquency
continuity in diverse behaviors across the life course such
regardless of family and school process. Further as the linkage between crime and other problem
analyses did reveal, however, that family and behaviors.
628 J.H. Laub

predicting stable differences in adolescent Enactment and the reinforcement of new


offending, they did not yield distinct groupings routines
that were valid prospectively for troubled kids Cognitive identity shifts and new life scripts.
over the long haul. Not only was prediction poor What we referred to as moving from a hell
at the individual level, our data raised questions raiser to a family man.
regarding the categorically distinct groupings
that dominate theoretical and policy discussions
(e.g., life-course-persistent offenders, and/or
super-predators). We found that these group- 3 Translational Criminology:
ings tended to wither when placed under the Using Life Course Research
microscope of long-term observation (Laub and on Crime to Shape Public
Sampson 2003: chapter 5; Sampson and Laub Policy
2003). Our take-away is that trajectories of
crime are inuenced but not determined by prior In order to fully discuss using life course research
childhood differences. on crime to shape public policies regarding crime
Fourth, we found that adult turning points mat- and justice, I must provide some context and per-
ter for desistance from crime. More specically, in sonal history. After 30 years in academe, I had the
our analyses of the Glueck data, we found that job good fortune of being nominated by President
stability and marital attachment in adulthood were Barack Obama to serve as the Director of the
signicantly related to changes in adult crime the National Institute of Justice in the Ofce of Justice
stronger the adult ties to work and family, the less Programs in the Department of Justice. The posi-
crime and deviance. Despite differences in early tion of Director is a presidential appointment with
childhood experiences, adult social bonds to work conrmation by the United States Senate.
and family thus had similar consequences for the Created more than 40 years ago, the National
life-course trajectories of the 500 delinquents and Institute of Justice (NIJ) is the research, develop-
500 nondelinquent controls. These results were ment and evaluation agency of the
consistent for a wide variety of crime outcome U.S. Department of Justice. NIJ is dedicated to
measures, control variables (e.g., childhood anti- improving knowledge and understanding of
social behavior and individual-difference con- crime and justice issues through science. NIJ pro-
structs) and analytical techniques ranging from vides objective and independent knowledge and
methods that accounted for persistent unobserved tools to reduce crime and promote justice, par-
heterogeneity in criminal propensity such as ran- ticularly at the state and local levels. In my view,
dom-effects panel models to analyses of qualita- NIJ has a unique mission as a science agency
tive data (see details Sampson and Laub 1993: focused on policy and practice. Given this posi-
chapters 7 thru 9 and Laub and Sampson 2003: tion, NIJ faces a twofold strategic challenge
Chapters 7 thru 9). We concluded from these anal- generating knowledge that is scientically
yses that the adult life course matters. rigorous and disseminating knowledge that is rel-
What was of particular importance were the evant to policymakers and practitioners.
common features of the turning points we uncov- On July 22, 2010, I was sworn in as the
ered. Regardless of the type of turning point such as Director of the National Institute of Justice.
marriage or serving in the military, there were four Never before in the history of the Institute has the
underlying mechanisms at work. The mechanisms position of Director been lled by someone with
are all consistent with the general idea of informal a Ph.D. in criminology and criminal justice and
social control. Namely, what appears to be impor- with extensive research experience. Some have
tant about institutional or structural turning points called this a turning point for the eld and it is a
are that they all involve, to varying degrees: clear indication that science is and will be an
important part of the mission of NIJ, the Ofce of
A kning off from ones delinquent past. Justice Programs, and the Department of Justice
Monitoring coupled with social support as a whole.
Life Course Research and the Shaping of Public Policy 629

There are other important features of NIJ and nizes research ndings, sets priorities for future
my role as the director that need to be pointed research, and provides inuential guides to pol-
out. First, despite the fact that NIJ is the primary icy and practice. Despite efforts by many to
federal funding source for research on crime and divide theory and research from policy, the fact is
justice, it is woefully underfunded with a base theory, research, and policy are deeply inter-
budget of about $40 million a year. Second, NIJ twined and central to the lives of everyone
resides in the Ofce of Justice Programs in the involved in explaining crime and advancing jus-
Department of Justice. In other words, NIJ sits tice and public safety.
amidst a sea of lawyers which represents a chal- In 2003 when I was President of the American
lenge to NIJ as a science agency. There is simply Society of Criminology,4 I organized the annual
not a strong culture of science in the Justice meeting program around the following theme
Department. Third, and perhaps most important, The Challenge of Practice, The Benets of
in July 2010, the same month that I started my Theory. I did this because I believe that such dis-
stint as Director, the National Research Council tinctions are unnecessary and ultimately counter-
released a report entitled, Strengthening the productive. Furthermore, the distinction is
National Institute of Justice. inconsistent with much of the history of crimi-
This was the rst large scale evaluation of NIJ nology. We have a strong tendency to favor
by an independent entity in more than 30 years dichotomies, all or nothing propositions, and
(see NRC 1977). The 2010 NRC Report offered subsequently we are forced to choose, theory or
ve broad recommendations that focused on the policy. As I wrote in my presidential address, I
need for independence and self-governance at believe that in order to enhance policy and prac-
NIJ; the critical elements essential for a science tice one needs not only sound research, but strong
agency that NIJ purports to be; the need for NIJ theory (see Laub 2004).
to bolster the research infrastructure internally One can legitimately ask What is the role of
and externally; the need for NIJ to embrace sci- the federal government in criminology and crimi-
entic integrity and transparency in all of its nal justice? In the broadest sense, it is to support
activities; and the need for NIJ to embrace a cul- research and data collection, analysis, and dis-
ture of self-assessment. According to the NRC semination. Following James Q. Wilson (1996),
Report, NIJ has lacked the essential tools com- the federal government can be and should be the
mensurate with a science agency: (a) a strong research and development arm of the criminal
management structure; (b) a scientic staff; (c) a justice system. For instance, a key role for NIJ is
budget to support both short- and long-term designing and testing crime prevention and crime
goals; and (d) protections from political shifts. control strategies. In order to accomplish this
Moreover, because of budget constraints and mission, I believe you need to focus on three
directives from Congress starting in 2004, NIJ major areas the nature of crime, the causes of
has shifted attention and resources away from crime, and the response to crime. In my view, this
both basic and applied social science research is the nucleus of a comprehensive research plan
toward capacity building and training, especially in the area of crime and justice and can be
in the area of forensic sciences (see NRC 2010). informed by the three scientic ofces housed at
NIJ social sciences, physical sciences, and
forensic sciences.
3.1 Providing a Conceptual In thinking about the creation of a research
Framework agenda at NIJ, I drew on my experiences as a life

So what should be the priorities of NIJ? In my


view, NIJ lacked a conceptual and intellectual
4
framework. Too often the wrong questions were The American Society of Criminology is one of the larg-
est international organizations in the world whose mem-
posed, the wrong language was used, and false bers are devoted exclusively to the study of and the
choices were presented. Coherent theory orga- prevention and control of crime.
630 J.H. Laub

course researcher. The conceptual framework I Health and Human Services, Labor, Housing
used included the following: and Urban Development, Education, and
Veterans Affairs.
The need for a long-term view. One of my Crime is both a dependent variable and an
favorite sayings at NIJ was one data point independent variable. The example I used
does not make a trend. repeatedly at NIJ was how many of the
There are multiple causes of crime. One major Glueck men who were involved in serious
development in criminology over the last 25 criminal activity as young adults died early
years is the recognition that there is no single and often violently (see Laub and Vaillant
cause or risk factor for crime and violence. 2000). Along the same lines, Lee Robins
Indeed, there are various pathways to crime concluded that antisocial behavior [in child-
and violence. Chronic offenders, in particular, hood] predicts class status more efciently
have multiple risk factors in their background than class status predicts antisocial behavior
including individual factors such as hyperac- (Robins 1966: 305).
tivity, impulsivity, and attention decit; family There is considerable heterogeneity in crimi-
characteristics, especially poor family func- nal offending over the life course. The stabil-
tioning and childrearing practices, school fac- ity of criminal behavior patterns throughout
tors like poor school achievement and low the life course is one of the most consistently
commitment to school, and peer factors, espe- documented patterns found in longitudinal
cially associating with delinquent peers and research on crime. However, not all children
gang membership. In addition, community and youth who are involved in crime grow up
inuences such as poverty and inequality, race to be criminal or antisocial as adults.
and family composition, housing and density, Somewhat paradoxically then, while studies
and neighborhood disorder and change are show that antisocial behavior in children is
important risk factors of criminal offending, one of the best predictors of antisocial behav-
especially violence. These factors tend to be ior in adults, most antisocial children do not
cumulative and interact with one another over become antisocial as adults. Some have called
time. this the Robins paradox (see Robins 1978).
There is a strong interconnection of problem For example, in our study using the Gluecks
behaviors. Based on our research and others, it prospective longitudinal data, we found that
appears that youth problems delinquency, despite continuity in antisocial behavior, there
substance abuse, violence, dropping out of was considerable change in criminal behavior
school, teen pregnancy often share common across the life span. Therefore, there are conti-
risk characteristics (Sampson and Laub 1993; nuities and discontinuities in criminal offend-
Jessor and Jessor 1977). Furthermore, these ing over the life course and this underscores
packages of problems often extend into the need to look at both stability and change in
adulthood (Sampson and Laub 1993; Cairns life course trajectories (see Sampson and Laub
and Cairns 1994). In addition, men and women 1993, 2005; Laub and Sampson 2003).
face numerous collateral consequences when
exiting prison in the mass incarceration era
with respect to employment, education, hous-
ing, and civic life (see Travis et al. 2014 for a 3.2 Moving Criminal Justice Policy
review). Moreover, the reach of the criminal Forward
justice system extends to the courts and police
as well with potentially disturbing conse- One of the ideas that I emphasized at NIJ was
quences (see Goffman 2014). So it makes per- what I call Translational Criminology. I believe
fect sense for the Department of Justice to be translational criminology acknowledges NIJs
interested in working with the Departments of unique mission to facilitate rigorous research that
Life Course Research and the Shaping of Public Policy 631

is relevant to the practice and policy. The idea of implement research evidence in real world prac-
translational criminology is simple, yet powerful. tice settings and to nd out why a program works.
If we want to prevent, reduce, and manage crime, Unpacking the underlying mechanisms of suc-
scientic discoveries must be translated into pol- cessful policies and programs is essential in
icy and practice. Translational criminology aims translational research.
to break down barriers between basic and applied Finally, translational criminology focuses on
research by creating a dynamic interface between dissemination of existing research results as much
research and practice. This process is a two-way as generating new knowledge. What is needed is
street scientists discover new tools and ideas for more systematic study of the process of knowl-
use in the eld and evaluate their impact. In turn, edge dissemination with the recognition that suc-
practitioners offer novel observations from the cessful dissemination of research ndings may
eld setting that stimulate basic investigation. well require multiple strategies including using
This is the knowledge creation process and both new forms of social media. We spend so much
researchers and practitioners play key roles here. time and energy on the front end of the research
In translational medicine, this is referred to as process, but not nearly enough time on making
T1 taking research from the bench (basic sure that critical research ndings make their way
research) to the patients bedside (clinical/ into the eld in a meaningful way. Without robust
applied research) (see http://www.michr.umich. dissemination efforts, research evidence will not
edu/about/clinicaltranslationalresearch). be used the way it was intended to inform crimi-
A unique aspect of translational criminology nal justice policy and practice.
then is the dynamic interface between research Recently, Sampson and colleagues (2013)
and practice and vice versa. To have this kind of articulated a strategy and a set of principles for
exchange assumes a great deal of trust. It should translating causal claims into public policy. The
be recognized that there is skepticism amongst key question they raise is how does policy work
practitioners about researchers, and researchers within a larger social context? They contend the
do not often trust the observations of practitio- answer to this question goes beyond estimating
ners as meaningful and important. Thus, transla- causal effects. According to Sampson and col-
tional criminology requires something that leagues, there are three domains which must be
heretofore has not occurred with much regular- part of the translational process. The rst is the
ity the research community and the practitioner identication of mechanisms and causal path-
community working together as equal partners. ways. The second is an assessment of effect het-
Another goal of translational criminology is to erogeneity. And the third is contextualization.
address the gaps between scientic discovery, The key point they make which is consistent with
program delivery, and effective crime policy. the main theme of this chapter is that theory is
This is the knowledge application process. This essential to understand the policy implications of
is referred to as T2 in translational medicine any research evidence (see also Laub 2004).
enhancing access to and the adoption of So how can we move criminal justice policy
evidence-based strategies in clinical and commu- forward? There are several policies that emerge
nity practice (http://obssr.od.nih.gov/scientic_ from life course research on crime that can shape
areas/translation/index.aspx). and inform public policies.
Translational criminology thus calls for more The rst is the importance of focusing on crime
data on the implementation process. In particular, prevention, especially early on in the life course.
we need to know whether the research evidence Our life course research clearly documented the
is being implemented with delity. Over the last crucial role that that families play in the causation
decade or so, we have spent a considerable effort of crime, especially relating to parent training in
to nd out what programs work and what pro- monitoring, recognizing, and disciplining the mis-
grams dont work. However, this is not enough. behavior of children and efforts should be devoted
What is needed is an understanding of how to to strengthening families in order to prevent crime
632 J.H. Laub

and other problem behaviors (see also Gottfredson Given stability and change in crime and deviance
and Hirschi 1990). Moreover, intervention strate- over the life span, policies and programs are
gies should consider a broad array of antisocial, needed that seek to develop and strengthen social
criminal, and deviant behaviors, and not just limit ties across an array of social institutions.
the prevention focus to one subgroup or crime Moreover, this notion applies to each and every
type. Comprehensive strategies that focus on a stage of an individuals life course. Pathways and
wider range of concurrent problem behaviors are turning points serve as useful metaphors in the
needed. As David Farrington has argued Because development of such policy. The concept of path-
of the link between crime and numerous other ways suggests that some individuals are set on a
social problems, any measure that succeeds in stable track toward delinquency and adult crime
reducing crime will probably have benets that go through the combined negative inuence of poor
far beyond this. Early prevention efforts that parenting, weak school attachment, and cumula-
reduce crime will probably also reduce alcohol tive disadvantage from criminal justice and juve-
abuse, drunk driving, drug abuse, sexual promis- nile justice sanctions (Sampson and Laub 1993).
cuity, and family violence, and probably also This calls for policies that prevent crime. At the
school failure, unemployment, marital dishar- same time, the notion of turning points suggests
mony, and divorce (1990:110). that pathways can be deected by positive devel-
The second is that one needs to be cognizant of opments that strengthen ties to key institutions in
the unintended criminogenic effects of severe sanc- society (Laub and Sampson 1993, 2003). This
tions. In our research using the Glueck data, we calls for policies that reduce recidivism. The
examined the role of criminal behavior and reac- challenge is for the criminal justice system to
tions to it by the criminal justice system and found assist in the facilitation of turning points for those
that delinquent behavior has a systematic attenuat- involved in crime. Recall that one of our ndings
ing effect on the social and institutional bonds link- from the Glueck project was that severe sanctions
ing adults to society (e.g., labor force attachment, can inhibit turning points. Moreover, men and
marital cohesion). More specically, we found that women leaving prison today face numerous col-
social bonds to employment were directly inu- lateral consequences with respect to employ-
enced by criminal sanctions incarceration as a ment, education, housing, and civic life (see
juvenile and as a young adult had a negative effect Travis et al. 2014 for a review). At the very least,
on later job stability, which in turn was negatively we should ensure that the criminal justice system
related to continued involvement in crime over the response writ large does not block potential turn-
life course (see Laub and Sampson 1995). ing points from occurring.
The third is cumulative disadvantage over the Fifth, a key question is how can we create a
life course is a concern. There are spillover choice architecture to nudge offenders
effects from criminal activity into other life toward decisions that will improve their lives? In
domains that may well forestall desistance from their book, Nudge, Thaler and Sunstein (2008)
crime. The idea of cumulative disadvantage pos- draw on recent research on decision-making,
its that delinquency incrementally mortgages the especially from behavioral economics, and they
future by generating negative consequences for make a strong case that people can be nudged to
the life chances of stigmatized and institutional- make better decisions that will improve their
ized youth. For example, arrest and incarceration lives. One way to do this is to organize the con-
may spark failure in school, unemployment, and text in which people make decisions. I think there
weak community bonds, in turn increasing adult is an enormous opportunity to apply the nudge
crime. Serious delinquency in particular leads to idea to change offender behavior. But equally
the kning off of future opportunities such that compelling to consider is how the nudge idea
participants have fewer options for a conven- can also be applied to change the behavior of
tional life (see Sampson and Laub 1997). criminal justice ofcials. As states look to reduce
Fourth, turning points can be conceptualized both the rate and length of incarceration, atten-
and supported by criminal justice institutions. tion needs to be directed to probation and parole
Life Course Research and the Shaping of Public Policy 633

supervision and services. The question is what knowledge used in policy argument beliefs,
can be done to make probation and parole prac- experience, trial and error, reasoning by analogy,
tice more desistance-focused? Through super- and personal or political values are also used in
vision of offenders in the community, probation policy argument (page 8). At NIJ, I argued
and parole ofcers can deter offenders from com- forcefully for science to be at the policy and prac-
mitting new crimes. At the same time, probation tice table. However, I soon recognized that politi-
and parole ofcers can serve as resources for cal considerations and value preferences are also
offenders regarding employment, housing, sub- present in all policy decisions (see also Dornan,
stance abuse treatment and the like. In this man- chapter Longitudinal Studies and Policy for
ner, probation and parole ofcers can act as Children in Low- and Middle-Income Countries
monitors and mentors and provide both supervi- this volume). What were some specic con-
sion and social support, and as a result, behav- straints I faced at NIJ?
ioral change away from crime may well occur.
It is important to note that because I am draw-
ing on the extant research on crime and the life 4.1 Competing Priorities
course most of the discussion and examples
offered here focus on individuals. Of course, the One constraint is competing priorities. The
idea of translational criminology applies to President, the Vice President, the Attorney
macro-level variables as well and could readily General, and the Ofce of Management and
incorporate an examination of how crime rates Budget (OMB) have their priorities and interests.
vary across social aggregates. Moreover, one At the same time, Congress, especially those on
could imagine a rigorous research program with the Appropriations staff, have their priorities and
an eye toward shaping public policies at the state interests as well. The result of all this was crave
level by investigating differences across states in outs by Congress, OMB, the Justice Department
crime rates, laws regarding rearms, laws regard- and/or the White House to direct money to spe-
ing the availability and cost of alcohol, and the cic research activities. For example, in FY 2014,
legalization of marijuana in Colorado and the House Appropriations panel directed NIJ to
Washington to highlight a few examples. In par- spent $4 million dollars to support research on
ticular, we would want to learn how these state domestic radicalization the process by individ-
policies affect the onset of, persistence in, and uals in the United States become violent extrem-
desistance from crime across the life course. ists and seek to engage in terrorist acts here and
Similar questions can be asked regarding the abroad. Similarly, in FY 2014, the Senate
level of resources devoted to family support and Appropriations panel directed NIJ to spent $2
schooling, especially early on in the life course, million dollars on gun safety technology. These
at the state, county, and city levels and their effect directed requests for funding occur virtually
on life course development. Finally, much of our every year.
research on the life course of crime could also be
enhanced by taking a cross-national focus, which
would in turn have important implications for 4.2 Time Pressure
shaping public policies about crime in the U.S.
Another constraint is time pressure. Science
needs time, but policymakers cannot wait. They
4 The Challenge of Bringing must act now. After the tragedy in Newtown,
Research to Public Policy Connecticut, NIJ responded to multiple requests
from Congress, Department of Justice, and the
In 2012, the National Research Council of the White House about gun violence. I was asked to
National Academies issued a report, Using review a joint proposal from the Department of
Science as Evidence in Public Policy. The report Homeland Security and the Ofce for Community
stated that Science is not the only source of Oriented Policing Services calling for scores of
634 J.H. Laub

school resources ofcers to be placed in schools demand side for research. Let me briey touch on
around the country. I asked if there was any three possibilities here.
empirical evidence that school resource ofcers
reduced violence in school. I was told, John,
good question, but what we need now are 5.1 Research on the Use
answers, not questions. of Research

In a masterful review of policing research since


4.3 Fragility of the Evidence 1975, Larry Sherman writes, The examples
offered . . . support a key conclusion: there has
The third constraint is the fragility of the evi- been a massive growth of policing knowledge
dence. In evaluation research in criminal justice over the levels in 1975. Equally important is a
there are few studies that use randomized con- second conclusion: there has been less progress
trolled trials and as a result selection issues loom in using knowledge than in generating it
large. Moreover, single site studies are the norm (2013: 415).
and once positive results are found there is enor- The William T. Grant Foundation has a whole
mous pressure to scale-up quickly. When the portfolio of research on the use of research cover-
results came in from a NIJ funded study of Judge ing a wide range of topical areas affecting youth
Alms HOPE program (Hawaii Opportunity ages 825 (see http://www.wtgrantfoundation.
Probation with Enforcement) showing large org/). The goal of this research program is to
reductions in recidivism NIJ was under pressure build stronger theory and empirical evidence on
from Congress to implement HOPE in all 50 how, when, and under what conditions research
states! Finally, criminal justice research is will be used in policy and practice. Focusing on
plagued by weak data. Ofcial administrative the demand for research, some of the questions
records are often not suited for research purposes being investigated are: Who are the users of
and missing data are a major concern.5 research? How is research dened, acquired,
interpreted, and used? What is the social ecology
of research use? I believe that there are a number
5 Promising Strategies of lessons from this research program that we can
use to enhance our understanding the nexus
One question that is often asked is how can between research and policy. For instance,
social scientists think more effectively about emerging research has demonstrated the impor-
designing research so that it can have an impact. tance of social networks in acquiring research
My response would be the usual pick a compel- evidence (Tseng 2012a). The implication of this
ling topic and use the most rigorous research is that we need to better understand how it is that
methods possible. That said, I do not think that criminal justice practitioners and policymakers
will get us very far. I think we need to focus our hear about life course research. More attention is
attention on other domains, especially the also needed on uncovering the underlying condi-
tions that facilitate the use of research evidence
in policy making and the conditions that inhibit
5
or block the use of research evidence.
One challenge facing life course researchers is convinc-
ing policymakers that longitudinal data collected over
the long term are relevant to the current policy issues
today. We faced this with the Glueck data which was 5.2 Infusing Evidence into
drawn from a sample of white, ethnic delinquents who the Grant Making Process
grew up during the Great Depression and came of age dur-
ing an era when drugs like crack cocaine were nonexistent
and guns were far less frequently used in violent crime On May 18, 2012, a memo from the Ofce of
compared to today. Management and Budget (OMB) was circulated
Life Course Research and the Shaping of Public Policy 635

to all agency heads that contained the following: ence the eld through concerted efforts to
Agencies should demonstrate the use of evi- transform practice and policy (see http://www.hks.
dence throughout their Fiscal Year (FY) 2014 harvard.edu/programs/criminaljustice/research-
budget submissions. This is an example of an publications/executive-sessions/executive-session-
imposed use of evidence government budget- on-policing-and-public-safety-2008-2014).
ing will be based on whether agencies have
adopted programs backed by evidence. Carol
Weiss and her colleagues (2008) have called this 6 Concluding Remarks
The Fairy Godmother approach to research
inuence. The former administrator of the National Oceanic
Of course, many questions remain: What is and Atmospheric Administration, Janet
good evidence? What is a fair and accurate repre- Lubchenco said, Operating in D.C is much
sentation of evidence? While promising, one harder than it needs to be. Its exhausting, its
would not want to see this OMB requirement frustrating, and at times depressing. That said, it
become yet another bureaucratic requirement is possible to get things done (Powell 2013).
without any real meaning. As the Director of NIJ I had a once in a life-
time opportunity to shape criminal justice
research and practice for years to come. I strongly
5.3 Researcher-Practitioner believed that NIJ needed to develop a cutting
Partnerships edge research agenda that addressed the major
topics of interest in the eld. Consider the follow-
I believe we can change the very nature of the ing. In the last 40 years in the eld of criminal
research enterprise by engaging practitioners justice, two of the most important developments
throughout the research process (see also Tseng were the run-up in the rate of incarceration,
2012b). We need to shift attention away from sometimes referred to as mass incarceration, and
pushing research out to the eld and create a the crime rate increases during the 1980s,
stronger research-practice connection from the followed by an equally large crime rate declines
beginning. Insights from practitioners can be during the 1990s and continuing into the new
invaluable in the research process. century. Yet NIJ had no active, systematic
An illustration of a successful researcher- research portfolio on either topic. In FY 2011,
practitioner partnership was NIJs 4-year proj- NIJ co-funded, with the MacArthur Foundation,
ect with Harvards Kennedy School of the National Academy of Sciences to create a
Government on the Executive Session on panel to study the causes and consequences of
Policing and Public Safety. In my view, the high rates of incarceration (see Travis et al.
Executive Sessions are an exemplar of transla- 2014). In FY 2012, NIJ funded the National
tional criminology. Here the leading police Academy of Sciences to host a series of roundta-
executives and researchers came together on a bles to better understanding crime trends. In
regular basis to tackle the major issues facing addition, in FY 2011 and FY 2012, NIJ funded
the eld. The Executive Sessions recognize in a research in a number of new areas including:
direct way that practitioners are partners in the California realignment (the shifting of low-level
research enterprise. It is not about pushing offenders from state prisons to county jails and
research results out to the eld. Rather it is probation); race, crime, and victimization;
focusing on practitioners and nding out what victim-offender overlap; desistance from crime;
research they need to do their jobs better. There police legitimacy; and criminal sanctions. NIJ
are several papers published from these also continued funding work in their signature
Executive Sessions (jointly written by police research programs such as violence against
chiefs and researchers). There is also consider- women, teen dating violence, sexual assault, and
able attention being paid to how best to inu- police ofcer safety.
636 J.H. Laub

In this paper, I argue that a life-course per- Glueck, S., & Glueck, E. (1950). Unraveling juvenile
delinquency. New York: Commonwealth Fund.
spective offers a unique conceptual framework
Glueck, S., & Glueck, E. (1968). Delinquents and nonde-
for thinking about public policies regarding crime linquents in perspective. Cambridge: Harvard
and criminal justice. Embracing what I call University Press.
translational criminology, I drew on key nd- Goffman, A. (2014). On the run: Fugitive life in an
American city. Chicago: University of Chicago Press.
ings from my long-term research project examin-
Gottfredson, M. R., & Hirschi, T. (1990). A general theory
ing the life course of crime from childhood to old of crime. Stanford: Stanford University Press.
age to illustrate how life course research can Hirschi, T. (1969). Causes of delinquency. Berkeley:
shape policies on crime and criminal justice mov- University of California Press.
Jessor, R., & Jessor, S. L. (1977). Problem behavior and
ing forward. Of course, this is not meant to ignore
psychosocial development: A longitudinal study of
the challenges of bringing research to bear on youth. New York: Academic.
public policy. There are real obstacles and at Kirk, D. S., & Sampson, R. J. (2013). Juvenile arrest and
times these are difcult to overcome. Moreover, collateral educational damage in the transition to
adulthood. Sociology of Education, 86, 3662.
recently we have witnessed serious attacks on the
Laub, J. H. (2004). The life course of criminology in the
social sciences in the federal government (see United States: The American Society of Criminology
Prewitt 2013a, b). Regardless, the fact is we need 2003 presidential address. Criminology, 42, 126.
more social science researchers in government. Laub, J. H. (2006). Edwin H. Sutherland and the Michael-
Adler report: Searching for the soul of criminology
Our voices are essential and the academy needs
seventy years later. The American Society of
to recognize this, nurture and encourage it, and Criminology 2006 Sutherland address. Criminology,
ultimately, reward it. My main takeaway message 44, 235258.
is Get Involved! Laub, J. H., & Sampson, R. J. (1993). Turning points in
the life course: Why change matters to the study of
crime. Criminology, 31, 301325.
Acknowledgements I thank Mike Shanahan, Monica Laub, J. H., & Sampson, R. J. (1995). The long-term
Johnson, and Jeylan Mortimer for their comments on an effect of punitive discipline. In J. McCord (Ed.),
earlier draft. I also thank Nicole Frisch and Shradha Coercion and punishment in long-term perspectives
Sahani for their research assistance. (pp. 247258). Cambridge: Cambridge University
Press.
Laub, J. H., & Sampson, R. J. (2003). Shared beginnings,
divergent lives: Delinquent boys to age 70. Cambridge:
References Harvard University Press.
Laub, J. H., & Vaillant, G. E. (2000). Delinquency and
Bronfenbrenner, U. (1979). The ecology of human devel- mortality: A fty year follow-up study of 1,000 delin-
opment. Cambridge: Harvard University Press. quent and nondelinquent boys. American Journal of
Cairns, R. B., & Cairns, B. D. (1994). Lifelines and risks: Psychiatry, 157, 96102.
Pathways of youth in our time. New York: Cambridge Moftt, T. E. (1993). Adolescent-limited and life-course-
University Press. persistent antisocial behavior: A developmental tax-
Caspi, A., Elder, G. H., Jr., & Herbener, E. S. (1990). onomy. Psychological Review, 100, 674701.
Childhood personality and the prediction of life- National Research Council. (1977). Understanding crime:
course patterns. In L. Robins & M. Rutter (Eds.), An evaluation of the National Institute of Law
Straight and devious pathways from childhood to Enforcement and Criminal Justice. Washington, DC:
adulthood (pp. 1335). Cambridge: Cambridge The National Academies Press.
University Press. National Research Council. (2010). Strengthening the
Elder, G. H., Jr. (1985). Perspectives on the life course. In National Institute of Justice. Washington, DC: The
G. H. Elder Jr. (Ed.), Life course dynamics (pp. 2349). National Academies Press.
Ithaca: Cornell University Press. National Research Council. (2012). Using science as evi-
Elder, G. H., Jr. (1999). Children of the great depression: dence in public policy. Washington, DC: The National
Social change in life experience, 25th Anniversary. Academies Press.
Boulder: Westview Press. Powell, A. (2013, May 17). Speaking up for science.
Elder, G. H., Jr., & Giele, J. Z. (Eds.). (2009). The craft of Harvard Gazette.
life course research. New York: The Guilford Press. Prewitt, K. (2013a, May 24). The congressional war on
Farrington, D. P. (1990). Implications of criminal career the social sciences. Pacic Standard.
research for the prevention of offending. Journal of Prewitt, K. (2013b). Is any science safe? Science, 340,
Adolescence, 13, 93113. 525.
Life Course Research and the Shaping of Public Policy 637

Robins, L. N. (1966). Deviant children grown up. M. Tonry (Ed.), Crime and justice in America:
Baltimore: Williams and Wilkins. 19752025 (pp. 377451). Chicago: University of
Robins, L. N. (1978). Sturdy childhood predictors of adult Chicago Press.
antisocial behavior: Replications from longitudinal Thaler, R. H., & Sunstein, C. R. (2008). Nudge: Improving
studies. Psychological Medicine, 8, 611622. decisions about health, wealth, and happiness. New
Sampson, R. J., & Laub, J. H. (1993). Crime in the mak- Haven: Yale University Press.
ing: Pathways and turning points through life. Travis, J., Western, B., & Redburn, S. (Eds.). (2014). The
Cambridge: Harvard University Press. growth of incarceration in the United States: Exploring
Sampson, R. J., & Laub, J. H. (1997). A life-course theory causes and consequences. Washington, DC: The
of cumulative disadvantage and the stability of delin- National Academies Press.
quency. In T. P. Thornberry (Ed.), Developmental the- Tseng, V. (2012a). The use of research in policy and prac-
ories of crime and delinquency: Advances in tice. Social Policy Report, 26, 116.
criminological theory (Vol. 7, pp. 133161). New Tseng, V. (2012b). Partnerships: Shifting the dynamics
Brunswick: Transaction Publishers. between research and practice. New York: William
Sampson, R. J., & Laub, J. H. (2003). Life-course desist- T. Grant Foundation.
ers? Trajectories of crime among delinquent boys fol- Wakeeld, S., & Wildeman, C. (2014). Children of the
lowed to age 70. Criminology, 41, 301339. prison boom: Mass incarceration and the future of
Sampson, R. J., & Laub, J. H. (special issue editors). American inequality. New York: Oxford University
(2005). Developmental criminology and its discon- Press.
tents: Trajectories of crime from childhood to old Weiss, C. H., Murphy-Graham, E., Petrosino, A., &
Age. The Annals of the American Academy of Political Gandhi, A. G. (2008). The fairy godmother and her
and Social Science (Vol. 602). Sage Publications, warts: Making the dream of evidence-based policy
Thousand Oaks, CA. come true. American Journal of Evaluation, 29,
Sampson, R. J., Winship, C., & Knight, C. (2013). 2947.
Translating causal claims: Principles and strategies for Wilson, J. Q. (1996). What, if anything, can the federal
policy-relevant criminology. Criminology & Public government do about crime? (Perspectives on crime
Policy, 12, 587616. and justice: 19961997 lecture series, Vol. I).
Sherman, L. W. (2013). The rise of evidence-based Washington, DC: NIJ Perspectives on Crime and
policing: Targeting, testing, and tracking. In Justice Seminars.
Epidemiological Perspectives
on the Life Course

Michael E.J. Wadsworth and Diana Kuh

1 Introduction cognition and mental health (Rutter 2013). In


contrast, life course epidemiology was only much
The task of epidemiology is to understand the more recently developed in relation to physical
causes of disease and health in populations. health, and that is the concern of this chapter. The
When long-term chronic disease became the pre- evidence base developed rst from questions
dominant health problem in urbanised countries, posed by clinicians and social scientists about the
once infectious disease had been controlled, the persistence and inequitable socio-economic dis-
need for life course epidemiology emerged. tribution of risk of death and illness in infancy
Evidence accumulated that the natural history or (Titmuss 1938; Royal Commission on Population
development of many long-term (chronic) physi- 1949; Douglas 1951; Butler and Bonham 1963;
cal illnesses was considerably longer than origi- Wadsworth 2014).
nally expected (Buttereld 1968), possibly Two principal types of data provide the
beginning in early life (Reid 1969). Clinicians resource for an epidemiological perspective on
began to ask whether the risks of chronic illness, the life course. The rst type is the long-term pro-
which tends rst to become evident in mid-life, spective follow-up study of the same individuals
were in fact to be found in early life (Reid 1969; over many years of life. Some of these studies
Morris 1967). Now, in addition to concerns about begin in early life, that is, at or before birth
the development of illness, life course epidemiol- (e.g., Power and Elliott 2006; Wadsworth 2010;
ogy has begun also to study the processes of age- Kuh et al. 2011; Halfon et al. 2013), and others in
ing which are thought to begin in early life (Kuh mid-life (e.g. Reid et al. 1974; DAgostino et al.
et al. 2011). 1989; Walker et al. 2004; Marmot and Brunner
The idea that environment and experience 2005). The second type of data is derived from
early in life have long-term effects discernible studies which identify populations known to have
throughout adulthood has been well established experienced such early life health risk as poor
in the sciences concerned with behaviour, nutrition before birth or poor growth, and then
measure their health in adulthood in comparison
with others not, or partially, exposed to the early
life risk (e.g. Barker 1991; Lumey et al. 2007).
M.E.J. Wadsworth D. Kuh (*) Great ingenuity has been used to nd in adult-
MRC National Survey of Health and Development
hood individuals who had been exposed to risk in
and MRC Unit for Lifelong Health and Ageing
at UCL, UCL, London, UK early life, for example, to famine because of war
e-mail: d.kuh@ucl.ac.uk (Stein et al. 1975; Lumey et al. 2007) or to poor

Springer International Publishing Switzerland 2016 639


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_28
640 M.E.J. Wadsworth and D. Kuh

socio-economic environments (Barker 1991). data resources for life course epidemiology is
Such catch-up studies are mirrored by follow- also outlined, including the continuation of the
back studies, which begin prospective follow-up original and now maturing longitudinal birth
in adulthood and use record sources and recollec- cohort studies that focus on health, and the
tion to nd out about earlier life exposure to development of their modern equivalents. We
hypothesised risk (e.g. Marmot and Brunner show how the changing history of health con-
2005; DAgostino et al. 2013). cepts inuenced the development and the nature
Life course data collected in prospective stud- of data used in life course perspectives in psycho-
ies has both an inherent problem of research logical, social and biomedical studies, and where
design and an inherent value; each of which novel and imaginative ideas from epidemiolo-
increases as time passes. Scientic concepts and gists acted as catalysts. Within that wider con-
measures develop and expand, but study samples text, our second aim is to show how life course
are rooted in the scientic time of their origins. epidemiology became a recognised eld of
Consequently, for example, an early life DNA epidemiological research at the beginning of the
resource or measure of respiratory function can- twenty-rst century, developing a conceptual
not be available for samples rst studied many framework and set of life course models which
years before it was technically possible to mea- are being applied now to a growing range of
sure those parameters in large population sam- health outcomes, across an increasing number of
ples. It has therefore been necessary to begin new cohort studies with life course data. Finally, we
life course studies as science develops, in order to highlight some current and future opportunities
provide measures and sample characteristics and challenges faced by life course epidemiology
appropriate for contemporary science (Halfon and life course studies.
et al. 2013). Studies begun some time ago are still
inherently valuable because they provide pro-
spective data on early life and later outcomes, as 2 The Development of the Life
well as information on the sequence of exposure Course Concept in Health
to risk and protective elements. Furthermore,
they can show trajectories of change with age in Hamlin (1992) observed that changes in the con-
biological function (e.g. blood pressure, respira- cept of disease and its causes in the nineteenth
tory function) and body state (e.g. body shape century were associated with the development of
and size). epidemiological ideas and practice. We extend
The term life course epidemiology was rst that observation from the nineteenth century to
used in 1997 (Kuh and Ben-Shlomo 1997) even the present, and identify seven eras of health in
though research applying a life course perspec- order to trace ideas that culminated in life course
tive to population health developed during the epidemiology. This section deals with the rst
1970s. Growing empirical evidence that early four eras of the development of ideas about health
life matters for adult health and disease, set and what has become a legacy of essential data
within conceptual life course models and evolu- resources for life course epidemiology in each
tionary frameworks, has strengthened the life historical period. These eras are summarized in
course perspective as a general paradigm for the Table 1.
study of development and ageing, health and During the rst health era (18001840) the
disease. prevalent problem was infectious disease.
The rst aim of this chapter is to trace the his- Concepts of cause ranged widely, but with little
tory of the predominant health problems in understanding, although individual resilience and
Western societies, and how that inuenced the vulnerability were thought to be implicated.
development of life course perspectives in epide- Jenners experiments with vaccination, which
miology and public health. The development of began in 1796, marked the beginning of
Table 1 Summary of development of resources for life course epidemiology 18002014
Health era Predominant health Development of resources for Concepts of disease and
and period problem life course epidemiology prevention
Era 1 18001840 Infectious disease New ideas about the Perception of environmental
individual constitution and causes
personal vulnerability and
resilience
Era 2 18401900 Infectious disease National records of vital Identication of
statistics begin (birth, death, microbiological causes and
the census) transmission
Mandated notication of Preventive measures through
some infections environmental change (e.g.
sewers) and behavioural advice
Era 3 19001940 Infectious disease begins New small longitudinal New understanding of
to give way to chronic studies of child growth and attainment of immunity in
disease development. Large populations as well as
cross-section studies of individuals
nutrition and mental health Recognition that environmental
and genetic effects implicated
in some infectious disease
Setting up of community Effects recognised of poverty
studies of health and on child health, growth and
morbidity (population nutrition
laboratories)
Health insurance studies Overwhelming need for
environmental improvements
Era 4 19401970 Infectious disease New prospective studies Increasing emphasis on chronic
increasingly gives way to begun in early life to study disease, and screening for
chronic disease child and adolescent health, proximal risk factors and
illness, growth, cognition and lifestyle in adulthood, and
behaviour education programmes for
New prospective studies of prevention
specic disease begun in
midlife
New measures of health
function and state show the
population distribution (e.g.
of blood pressure), and
permit studies of pre-
symptomatic illness
Era 5 19701990 Chronic disease, New prospective studies Socioeconomic, psychosocial
disability and ageing begun in early life and developmental models of
Catch-up of samples studied chronic disease causation
in era 3 challenge the dominant adult
lifestyle model
New catch-up studies with Signs that infections are still
samples based on birth important causes of diseases,
records from 1920s even in developed countries
Studies in pregnancy and (the AIDS epidemic is
before are established to identied rst in 1981)
explore Barkers hypothesis
Birth cohort studies begun in
era 4 prepare to study ageing
New and continuing midlife
prospective studies continue
into later life, and new later
life studies begun
Census linked longitudinal
study begins in Britain
(continued)
642 M.E.J. Wadsworth and D. Kuh

Table 1 (continued)
Health era Predominant health Development of resources for Concepts of disease and
and period problem life course epidemiology prevention
Era 6 19902000 Chronic disease, Early life study samples from Chronic diseases of afuence,
disability, ageing, AIDS era 4 approach and enter later AIDS, childhood allergies all
and childhood allergies life, becoming unique life seen as current problems, and
course data resources for ageing as an imminent problem
lifetime studies of ageing
More prospective studies of
ageing established from
midlife
First synthesis of life course
ideas in relation to chronic
disease
Era 7 2000-the Chronic disease, Simplication of methods of Health risks now thought to
present disability and ageing, collecting biological samples include super-hygienic
AIDS and childhood and making DNA resources environment in early life
allergies Life course epidemiology Healthy ageing is increasingly
becomes an established seen as an essential goal and its
discipline denition hotly debated
Emerging omics elds of The concept of health
large-scale data-rich biology broadening to include the
needs life course studies and ability to adapt physiologically,
methods psychologically and socially
Life course resources used to
study epigenetic effects, and
reversibility of environmental
effects
Existing birth cohort studies
show how adult experience
modies the relationship of
early life growth with adverse
risk in adulthood
New, increasingly larger
sample prospective studies
begin in early life, and new
studies of ageing start from
midlife
New biobank studies of very
large adult samples begin,
based mostly on individual
health records and biological
samples
The international rich
resource of life course data is
widely available for testing
life course hypotheses

prevention of infectious and viral disease, but it for future national defense, the work force and
was not until the twentieth century that the major the economy, prompted thinking about age cohort
infectious diseases were brought under control. differences in health (Glass 1973) but did not
Concerns (e.g. of Malthus) about population size, translate effectively into health policy or research
fertility and replacement, and their implications ideas.
Epidemiological Perspectives on the Life Course 643

In the second health era (18401910) mental development (Scottish Council for
scientic discoveries of causes and means of Research in Education 1993) were established
transmission of infectious disease showed the for similar reasons.
role of the environment. That translated in health The two world wars accelerated the develop-
policy terms to environmental intervention ment of curative and preventive health care, from
(e.g. installation of sewers and drains), legislation surgery to nutrition, and reliably effective pre-
about health care (e.g. work environment, food ventive and curative medication against infectious
quality), and advice about personal hygiene and disease began to become available. Improvements
child care. In Britain the surveys of Booth (1891) in environmental health continued, including
and Rowntree (1901) brought awareness of the new emphases on occupational health, such as
extent of poverty and its association with adverse the limitation of effects of working in dusty
health. The poor health, in particular the short atmospheres, for example, in mining (Morris
stature and poor physique of recruits for the war 1967). Epidemiologists showed how population
in South Africa (Great Britain Parliamentary immunity from infectious diseases was to be
Papers 1904), showed the need for improvements achieved through immunisation (Stocks 1975). In
in child health and nutrition in order to have suc- a study of the aetiology of juvenile rheumatism,
cessive generations t enough for military pur- undertaken during this era, Morris and Titmuss
poses and the predominantly manual occupations (1942) noted the long history of risk development
in industry and agriculture. Consequently, sys- and the range of presumptive causes, concluding
tematic national recording was established in that this condition was so sensitive to social
most Western urbanised countries to monitor inequality that emphasis on hereditary inuences
population trends through a regular census, to is premature. Kermack et al. (1934) found cohort
provide mandatory registration of births and differences in patterns of mortality in the UK and
deaths, and mandatory notication of specic Sweden that indicated the importance of child-
diseases known to be infectious. These data hood conditions for adult health (Kuh and Davey
sources provided essential data for the study of Smith 1993). In his study of socio-economic and
epidemics, and revealed falling trends in fertility regional differences in mortality, Titmuss (1938)
and in mortality rates of children and young commented that districts that had high rates of
adults. infant mortality and infant serious illness would
During the third health era (19101940) there give rise to many premature deaths in later life
was increasing concern with the impact of (p. 85), recognising that such change was going
poverty on health, prompted in Europe and the to have age cohort specic effects. This strikingly
US by the economic depressions and mass anticipated early life course epidemiological
unemployment during this period. For Britain, work on this topic published 53 years later
the war in South Africa and contemporary social (Barker 1991).
and economic conditions at home stimulated the The resources of national data established in
political will to begin a national insurance pro- the previous era were continued, and augmented
gramme to reduce the effects of poverty particu- by community studies of population health (e.g.
larly on health, the inauguration of school Collins 1949), by data from clinical follow-up of
medical examinations (1918), and a government patients (e.g. White 1931) and by insurance com-
department responsible for health (1919). In the panies collection of records about age at and
US, research was developed to examine the cause of death in the insured (e.g., Metropolitan
impact of the economic depression on childrens Life Insurance Co 1946). Analyses of insurance
health, growth and development in the Fels study data showed that important changes were taking
(Roche 1992), the Berkeley Growth study (Jones place in Western urbanised countries. As fertility
and Bayley 1941), and the Oakland Growth study continued to fall, and average age at death contin-
(Jones 1939). In Britain a family dietary study ued to rise in those countries, the predominant
(Boyd 1937) and the Scottish study of childrens pattern of ill health changed. Those who now
644 M.E.J. Wadsworth and D. Kuh

escaped premature death from infectious causes adolescent health known as the National Survey
lived on into midlife, the time when symptoms of Health and Development (Wadsworth 2014); it
begin of serious, long-term diseases such as heart later continued as a study of adult health and age-
disease, stroke, cancer and bronchitis. These ing (Kuh et al. 2011). That study was replicated
changes became known in population health as in 1958 and 1970 by two new national studies of
the epidemiological transition. Its cause is much mortality in the rst 7 days of life, intended to
debated, and current thinking considers whether compare early life health and survival since the
it may have been principally brought about by beginning of the National Health Service (in
improvements in environmental health, or nutri- 1948) with data collected in the 1946 study.
tion, or health care (Caldwell 2001). These two studies were begun by clinicians and
During the fourth health era (19401970) in continued as longitudinal studies by epidemiolo-
countries that had completed the epidemiological gists, developmental psychologists and social
transition, there was less concern with poverty as scientists: the National Child Development Study
a potential source of disease risk, and greater and the 1970 Birth Cohort Study (Power and
emphasis on developing hypotheses about the Elliott 2006; Elliott and Shepherd 2006). A com-
natural history of chronic disease. New biologi- parable long-term prospective study was begun
cal measures were developed of the current state in Finland in 1966 (Rantakallio 1988). Initially
of health that could also indicate risk in the lon- these studies showed how poverty was a marked
ger term (e.g. measures of sugar in urine and later disadvantage for many aspects of birth and for
also in blood). Portable and reliable equipment growth in early life. Explanations for these nd-
for measuring physiological function (e.g. blood ings were sought not only within the study sam-
pressure, lung function) was designed for use in ples but also by thinking on a longer time-scale.
prospective studies of large populations in non- Illsley and Kincaid (1963) speculated that risks to
clinical settings. Those measures were the basis, the baby associated with short maternal stature
when repeated over time, for tracing the trajec- may have originated in poor diets of mothers who
tory of change in the individuals risk of chronic in their study were born during the interwar eco-
diseases before symptoms or the full extent of the nomic depression.
disease appeared. That prompted new thinking The challenge of increasing prevalence of
about the natural history of disease, which asked, chronic physical disease in adulthood in the mid-
for example, where in the population distribution twentieth century stimulated new kinds of pro-
of measurements, such as of blood pressure, risk spective studies of health that began in mid-life,
could reliably be said to begin. With the new when the rst signs of many such illnesses appear.
knowledge derived from such work, public health These midlife studies were particularly con-
policy set up screening programmes for the pre- cerned with proximal and behavioural risk fac-
symptomatic identication of disease (e.g. mass tors, and some also focused on the effect of social
miniature x-ray screening for pulmonary tuber- and work environment. They used the new tech-
culosis, and blood pressure screening to detect nology of measurement developed for screening
risk of cardiovascular disease), and programmes purposes. One of the earliest and most well-
to promote personal responsibility for reducing known of the midlife prospective studies is the
the risk of illness through smoking cessation, US Framingham study, begun in 1948, a prospec-
exercise and dietary improvement. tive study of cardiovascular health, disease and
At the beginning of this period, concerns in disease risk (DAgostino et al. 1989). The
Britain among demographers, public health phy- Whitehall I prospective study of cardiovascular
sicians and pediatricians about falling fertility health began in 1967 (Reid et al. 1974). These
and persistently high infant mortality stimulated studies used their longitudinal data to examine
the inauguration of a rst national study of births the dynamics of risk, including interactions of
in 1946 (Wadsworth 2010). Within 2 years that risk factors and accumulation of environmental
became a longitudinal study of child and later risk exposures (for example from stress, nutrition,
Epidemiological Perspectives on the Life Course 645

smoking and diet) over many years of adult life. vulnerability established in childhood, found that
While most midlife cohort studies focused on respiratory disease among young adults in the
specic chronic diseases, the rst ageing cohorts rst British birth cohort study was associated
were established (for example, the Baltimore with lower respiratory tract illness before age 2
Study of Aging began in 1958) (Ferrucci 2008). years, and with indicators of poor socio-economic
Recognition that risk or the actual beginnings conditions at that stage. In cardiovascular disease,
of chronic illness shown in midlife cohort studies Miall and Chinn (1974) showed that blood pres-
might in fact originate in early life began in this sure levels and their change in young adulthood
era. In chronic bronchitis epidemiology it was had predictive value for future illness, and should
suggested that since adult environmental risk did be treated. Where this kind of research could not
not entirely explain the incidence, some part nd a source of data collected sequentially over
might be explained by vulnerability established the course of a samples lifetime, ingenious
as a result of developmental damage caused by methods were devised to approximate that ideal.
childhood or adolescent respiratory disease (Reid Forsdahl (1977) proposed that growing up in
1969). There were other signs that research of poor socio-economic circumstances and then
this kind would be of value. Dubos et al. (1966) living an afuent adult life style increased risk of
had already reported (a) that in animals maternal premature death from cardiovascular disease. He
nutritional state during lactation affected both the conrmed that hypothesis using Norwegian
rate of infant development and the nal adult stat- national statistics to show that districts with high
ure; and (b) that in humans infectious disease rates of infant mortality had high rates of prema-
retarded growth, and nutrition in early life deter- ture adult death from arteriosclerotic heart dis-
mined the extent of damage caused by infection. ease in the same age cohort. He concluded that
They concluded that many problems of adult- the weaker of the cohort die in infancy, [and] the
hood and old age will be found to be the distant more t survive and carry with them a life-long
manifestations of environmental factors that were vulnerability because of the poor living condi-
inuential during the formative years (Dubos tions in early years.
et al. 1966: 799). Perhaps these indications of the importance of
health in early life helped with the establishment
in 1972 of the US Bogalusa study of cardiovascu-
3 How Life Course lar risk in early life (Berenson 2001); it is certain
Epidemiology Came of Age that ndings from the Framingham study were
inuential. In addition, catch-up of participants
At the beginning of the fth health era (1970 in the US Hagerstown morbidity survey in the
1990; see Table 1) life course research proved 1930s showed that those who had been relatively
difcult to fund in the US and Britain. Despite underweight as children, and then overweight in
the promising research in the 3rd and 4th eras adult life, were most at risk of hypertensive vas-
suggesting that the origins of adult health might cular disease (Abraham et al. 1971).
lie in early life, the value of longitudinal studies Concerns about possible early life origins of
was subject to much criticism (Wall and Williams adult ill health certainly encouraged continuation
1970). Nevertheless, unconnected research scat- (in 1979) of the 1946 British birth cohort, and its
tered across the scientic literature on the role of refocussing to include measures of functional
early life in later health gradually built up during change (e.g. blood pressure, respiratory function
this period, and would later be integrated through at the 36 year follow-up in 1982, and additionally
developmental and life course perspectives on cognitive function at the 43 year follow-up in
health and disease. 1989), in order to study inuences on health
For example, Colley and colleagues (1973), change with age (Wadsworth 2014). By the end
building on Reids suggestion that adult onset of this era, this study was contributing to the
of chronic bronchitis might be associated with stronger evidence for the relevance of early life
646 M.E.J. Wadsworth and D. Kuh

for adult health by showing the association of between birth weight and chronic disease were
blood pressure at age 36 years with weight at not usually confounded by lifetime socioeco-
birth (Wadsworth et al. 1985). The rst of nomic circumstances; that metabolic diseases,
Barkers ecological studies (Barker and Osmond such as diabetes, were commonly associated with
1986) replicated Forsdahls (1977) nding that either low or both high and low weight at birth, as
risk of premature death from cardiovascular were interactions with later risk factors,
disease was highest in areas of highest infant particularly adult body mass index; and that adult
mortality. In a catch-up study from birth records classical risk factors did not appear to explain the
Barker then identied the association of low birth relationships. A second objective was the synthe-
weight with premature adult death from cardio- sis of three apparently competing models of the
vascular disease (Barker et al. 1989). causes of adult chronic diseases, which were not
Because of this kind of work on physical mutually exclusive. Life course epidemiology
health, and the contemporary life course work in encouraged the adult lifestyle model to consider
mental health (e.g. by Robins and Rutter 1990; the childhood origins of adult lifestyle; the social
and Stein and Susser, see Stein et al. 1975) and causation model to consider the long-term impact
the social sciences (e.g. by Elder 1974), a life of the early socioeconomic environment; and the
course approach to studying the processes of age- fetal origins model to consider other potential
ing, before later life began, was considered likely sensitive phases in the postnatal period that could
to be fruitful (Riley 1987). leave long-term imprints on the body that might
During the sixth health era (19902000; see increase susceptibility to subsequent chronic dis-
Table 1) life course epidemiology became eases. Whether the impact of the physical and
established. The immediate catalyst for this was social environment on risks to later health pri-
Barkers hypothesis that low birth weight marily occurred during critical periods in early
indicated poor nutrition of the unborn child life when growth and development were rapid, or
during the unique time of development for some whether the risks from environmental impacts
vital organs (e.g. brain, kidneys) and systems accumulated throughout life, was increasingly
(e.g. cardiovascular, respiratory); he consolidated debated in the UK and in North America (Keating
these ideas about the fetal origins of adult disease and Hertzman 1999).
in the biological programming hypothesis The development of life course epidemiology
(Barker 1995). highlighted the scientic contributions from the
A group of mainly UK epidemiologists maturing British birth cohort studies and mobil-
actively involved in the maturing British birth ised future life course research using these and
cohort studies gave wider consideration to these other existing studies and the plethora of new
ideas about the implications of early life risk for cohort studies that were established from the
adult chronic disease (Kuh and Ben-Shlomo 1990s. These studies began to collect biological
1997), and proposed the term life course epide- samples for DNA extraction and for an increas-
miology as the study of the biological, behav- ingly wide range of disease biomarkers. One of
ioural, and psychosocial pathways that link the initial purposes was to investigate whether the
physical and social exposures during gestation, epidemiological associations between early
childhood, adolescence and adult life to chronic growth and later life physical health reected
disease risk. Their rst objective was to critically genetic rather than environmental risk (Hattersley
review all the available evidence for pre-adult and Tooke 1999). New cohort studies were estab-
risk factors on cardiovascular, respiratory and lished to collect greater detail about physical
metabolic disease, seeking replications in other development in early life. In 1991 the British
datasets, and considering risk factors in later Avon Longitudinal Study of Pregnancy and
childhood and adolescence as well as fetal life. Childhood began in pregnancy in order to mea-
Their reviews showed that the relationships sure such prenatal effects as maternal smoking,
Epidemiological Perspectives on the Life Course 647

exercise and nutrition (Golding et al. 2001). The focus on middle and lower income countries
Southampton Womens Survey began in 1998, (e.g. Raghupathy et al. 2010). Findings on the
before rst pregnancy, so that the effect of pre- developmental origins of adult disease increas-
pregnancy health on fetal development and ingly are seen within an evolutionary perspective
outcome could be studied (Inskip et al. 2006). offering a broad unifying and interpretative
At the same time the already long-running cohort framework (Gluckman and Hanson 2004;
studies increased their range of health outcome Gluckman et al. 2009). It is argued, for example,
measures. The British 1946 birth cohort study that the processes of adaptation to the environ-
expanded its midlife measures to include muscle ment in early life may help to ensure later repro-
strength, physical performance and womens ductive success, and have consequences for later
health during the menopause years, in order to life health. Developmental biologists are enhanc-
continue research into the processes of ageing ing understanding of these processes and their
and the modication of early life effects by adult long-term effects (Kuh and Ben Shlomo 2014).
inuences (Wadsworth et al. 2006). The rst bio- This evolutionary perspective on development is
medical follow-up of the 1958 British birth mirrored by evolutionary arguments which
cohort study measured aspects of cardiovascular explain ageing as decline in the force of natural
and respiratory risk. The prospective longitudinal selection which, in the post-reproductive phase of
studies begun in midlife to investigate disease life, gradually reduces the ability to repair the
risk, such as the Framingham and Whitehall II accumulating molecular and cellular damage
studies of cardiovascular disease, continued into (Kirkwood and Austad 2000; Austad 2008). This
the later life of their subjects. In the US, more of perspective provides the wider context for an inte-
the cohort studies developed a focus on ageing, grated life course perspective on ageing (see Sect.
encouraged by the establishment of the National 4.1 below). These developments in epidemiology
Institute on Aging in 1974, whereas in the UK the and biology, coupled with growing life course
rst wave of the English Longitudinal Study of concerns in psychology, sociology, and demogra-
Ageing did not take place until 1998 (Steptoe phy, show how the life course perspective is
et al. 2013). Social science longitudinal studies becoming a common conceptual framework for
primarily concerned with socio-economic dis- interdisciplinary research on development, health
parities in education, occupation and opportunity and ageing (Alwin 2012; Settersten Jr 2009).
began also to include measurements of health, in At the beginning of this seventh era, life
order to study inequity in health and the impact course epidemiology continued to evaluate the
of social diversity on health (www.iser.essex.ac. hypothesis about developmental origins of adult
uk/bhps). disease, and now includes ndings of systematic
During the 7th health era (20002014; see reviews and meta-analyses, for example on birth
Table 1) life course epidemiology became estab- size and adult disease risk (e.g. Huxley et al.
lished as a research eld (Ben-Shlomo and Kuh 2007; Whincup et al. 2008; Kuh and Ben-Shlomo
2002; Kuh et al. 2003; Kuh and Ben-Shlomo 2014). Life course epidemiology also gave more
2004), and is now recognised as part of main- attention to whether early life inuences
stream epidemiology (Ben-Shlomo et al. 2014) explained time trends (Kuh et al. 2002; dos
and public health (Kuh et al. 2014a). Santos Silva 2004), and socio-economic inequal-
Multidisciplinary interest in the fetal origins ities in disease (Davey Smith and Lynch 2004;
of adult disease had, by 2003, broadened into a Leon and Walt 2000). It linked with other disci-
eld of study with a learned society called the plines, and widened its gaze beyond cardio-
Developmental Origins of Health and Disease metabolic and respiratory disease to a range of
(www.mrc-leu.soton.ac.uk/dohad/). It was con- health outcomes and pathways to them, such as
cerned with the long-term effects of postnatal as obesity, cancers, neuropsychiatric outcomes,
well as prenatal development, with a growing reproductive health and menopause, musculo-
648 M.E.J. Wadsworth and D. Kuh

skeletal ageing, depression and body image, scientic questions addressed by life course
infectious diseases, endocrine function, lifetime epidemiology (Fig. 2).
social relationships, and health behaviours. Life The trajectories shown in Fig. 2 reect many
course models were proposed that reected the aspects of biological function (e.g. lung function,
underlying pathways between early life factors muscle function, cognitive function) which dis-
and adult function and disease (Ben-Shlomo and play rapid growth and development in the early
Kuh 2002; Kuh et al. 2003); and innovative ways stages of life until maturity when a peak or pla-
of testing these models were presented (Mishra teau is reached, sometimes described as func-
et al. 2009). tional reserve capacity. This is followed by a
Two groups of models of the development of gradual decline in function with age. In Fig. 2,
risks to future health can be distinguished, trajectory A illustrates normal development and
namely the critical or sensitive period models, decline; trajectory B illustrates the inuence of
and the risk accumulation models. The rst group exposures or genetic factors that result in
of models propose that an exposure acting during sub-optimal development resulting in a reduced
a period of rapid development has effects on the functional reserve at maturity; trajectory C
structure or function of organs, tissues or body illustrates the inuence of exposures or genetic
systems which are unique (critical) or stronger factors acting post maturity that result in acceler-
(sensitive) than effects outside that time window; ated age-related decline; and trajectory D is a
factors acting later in life may or may not modify combination of trajectories B and C. Early life
these effects. By contrast, the accumulation of factors can potentially affect the peak or struc-
risk models propose that the effects of exposures tural reserve and/or the rate of decline but factors
gradually accumulate throughout life, which acting in adult life can only affect the rate of
cause increasing damage to body systems. Figure decline. There is growing knowledge about the
1 summarises four ways in which risks may accu- early life factors that are associated with level of
mulate across life. In model (a) the three effects function at the peak, or at various ages in later
act independently. In model (b) a fourth effect life. But as yet little is known about the early life
(such as social class) inuences each of the other determinants that drive functional change. The
three independent effects. In model (c) the three fundamental tasks of life course epidemiology
effects not only act independently but also inu- are therefore (1) to understand the natural history
ence each other; and in the fourth model (d) the of physiological trajectories, such as lung, mus-
three effects act only through their inuence on cle and cognitive function, to identify periods of
one another. life when important functional change occurs,
During this seventh health era life course epi- and to study how such change inuences subse-
demiologists increasingly focused on measures quent disease risk; (2) to investigate the early life
of biological functional outcomes and how these factors and developmental trajectories associated
changed across life (Hardy et al. 2000; Wills with peak adult function and reserve, and whether
et al. 2011; Lawlor and Hardy 2014) (Johnson these factors also inuence the rate of functional
et al. 2015), how they related to growth and decline; and (3) to show how these earlier life
developmental trajectories (Wills et al. 2012), factors independently, additively, or interactively
and the consequent implications for later disease operate with adult risk factors that inuence the
risk (Ghosh et al. 2014). An adapted version of rate of functional decline. The life course
Strachans original diagram for lung function tra- approach uniquely identies risky trajectories
jectories across life (Strachan 1997) has become and subgroups before disease or disability is
the key, simple gure to show how function manifest, opening up opportunities for preventive
changes across life and to illustrate the main health intervention.
Epidemiological Perspectives on the Life Course 649

Fig. 1 Life course risk accumulation models (Reproduced from Kuh et al. 2003)
650 M.E.J. Wadsworth and D. Kuh

Adult risk factors


Functional/structural
50
reserve

40

30
Early life risk factors
Level of function

20 A
Level below which
limitations may occur B
10
C

D
0
0 10 20 30 40 50 60 70 80 90 100

Years of life

Fig. 2 Life course functional trajectories (Adapted from Kuh and Ben-Shlomo 1997)

4 Three Central Aspects 4.1 A Life Course Epidemiological


of Current Life Course Perspective on Ageing
Epidemiology
Over the last decade, life course epidemiology has
The focus of life course epidemiology is cur- increasingly been applied to the study of ageing.
rently, as it always has been, driven by the This was a natural extension of the longitudinal
requirements of the medical sciences, policy study of change in functional outcomes with age
thinkers, and the social sciences. For the medical throughout life, since an accepted denition of
sciences the strongest demand is associated with biological ageing is the progressive deterioration
the development of new knowledge in relation to of functional capacity. The focus has been on inte-
genetics and the processes of ageing. Policy grating scientic concepts from life course epide-
thinkers are concerned with health management miology with those from ageing research,
in an ageing population and how most effectively conceptually and methodologically, and making
to use new medical knowledge in population best use of the empirical evidence from longitudi-
health care. Social scientists are concerned with nal studies, particularly the maturing birth cohort
social and economic implications of an ageing studies (Kuh et al. 2014). Comparative research
population and gaining greater understanding using data from existing cohort studies is begin-
about the sources and impact of psychological ning to capitalize on their differences in measure-
stress, and with the variation of these in different ment of early life exposures that impact
age cohorts and cohorts in different cultures. development (e.g. in nutrition, exposure to and
Three particularly important topics of current immunisation against infection, parental smoking
concern to life course epidemiology are (1) the and education) in order to study cohort differ-
processes of ageing; (2) the impact on physical ences in growth, adult health and functional
health of chronic exposure to psychological change (e.g. Li et al. 2008). The increasing num-
stress; and (3) how best to manage new require- ber of repeat functional measures in maturing
ments for big data for genetic studies. birth cohort studies and in adult cohort studies
makes it possible to study within person change,
and stimulates further development of methodol-
Epidemiological Perspectives on the Life Course 651

ogy for analysing such change. Policy concerns ight response may be responsible for social
about social and economic costs of population differences in neuroendocrine, physiological, and
ageing, together with a growing policy awareness metabolic variables which are the precursors of
of the practical value of the life course perspec- ill health and disease (p. 27).
tive, have led to increased research funding for an Increasingly the need for a life course per-
interdisciplinary, life course approach to healthy spective in this area has been recognised, and fac-
ageing that facilitates cross-cohort comparative tors from early life that may inuence the
research. The Healthy Ageing across the Life physiological systems concerned are investigated
Course (HALCyon) research programme was one (Ben-Shlom et al. 2014). These factors may
of the rst to study lifetime inuences on ageing. include, for example, maternal anxiety and
It uses nine UK cohort studies, systematic reviews, depression during pregnancy, prenatal substance
cross cohort studies and in depth analysis of exposure (alcohol, tobacco, cocaine), anger
cohorts with unique features (see references on expressed in the family, infection load and poor
www.halcyon.ac.uk). Kuh and colleagues sum- early growth, although the evidence is not always
marised ndings and the wider epidemiological consistent. Such risk factors often cluster and are
literature within an integrated life course concep- imprecisely measured in the midlife cohort stud-
tual framework for ageing research (Kuh and Ben ies, making it difcult to identify which aspects
Shlomo 2014; Kuh et al. 2014). Another network of the early environment or the early stress
of ageing studies, known as the Integrative response have particularly adverse long-term
Analysis of Longitudinal Studies (www.ialsa. effects. Equally challenging is the task of disen-
org), brings together data resources from more tangling whether early life effects operate through
than 90 longitudinal studies, including the continuity in the social environment, or behav-
HALCyon cohorts. It aims to enhance the study of ioural pathways, or through processes of biologi-
change with age in health and illness, the early cal embedding (i.e. how they get under the
detection of neurodegenerative disorders, and the skin). What has become known as the develop-
identication of periods in life when intervention mental biology of social adversity (Hertzman
is likely to have the greatest benet for health. and Boyce 2010; Rutter 2012) combines insights
from disciplines such as neuroscience, genetics
and epigenetics, together with data from birth
4.2 The Life Course cohort and life course studies, in order to unravel
Epidemiological Perspective the biological pathways that link early experi-
on Stress ence to later health and health inequalities (Cohen
et al. 2010; Shonkoff 2012). Effects are seen
There has been continuing concern with the across the social gradient, not just among the
effects of stress on health and disease risk. Cohort most disadvantaged children; socio-economic
studies of health and the social environment, such gradients in physical, socio-emotional and lan-
as the Whitehall II study, have for a long time guage/cognitive development in early life differ-
been concerned with the health impact of chronic entiate access to environments that provide
exposure to stress, including stress through occu- adequate attachment, support, nurturance and
pation, home life, and poverty (Marmot and stimulation (p. 115) (Hertzman 2007).
Brunner 2005), that may be manifest, e.g., as
anger and powerlessness (Williams 2002). The
initial focus was on adult stress; psychosocial and 4.3 The Life Course
biological pathways were identied that could Epidemiological Perspective
account for how environmental and behavioural and Big Data
effects gradually became manifest as physical ill-
ness. Brunner and Marmot (2006) concluded that Combining data from multiple cohorts has been a
disturbance of the usual homeostatic equilib- key feature of the past decade in life course
rium by the repeated activation of the ght-or- research on development, ageing and disease
652 M.E.J. Wadsworth and D. Kuh

development. No one cohort has repeat measures respond to this need; for example, the 1958
across the whole of life, so combining cohorts British birth cohort study acted as the normative
allows lifetime functional and behavioural trajec- comparison sample, or control arm, of the
tories to be described, replication of ndings to Wellcome Trust Case Control Consortium to
be tested, and more robust evidence to be pro- study seven common diseases (Wellcome Trust
vided for policy purposes. For example, cohort Case Control Consortium 2007). From a life
data has been combined to investigate lifetime course epidemiological perspective, the most
trajectories of blood pressure (Wills et al. 2011), interesting genetic studies are those that show:
body size (Johnson et al. 2015) and alcohol con- how genetic associations with body composition
sumption (Britton et al. 2015). This type of or other health-related characteristics vary across
research is building up a growing body of experi- life (Ong et al. 2011); how genetic factors act on
ence on harmonisation of existing cohort data, the tempo of growth in childhood and are associ-
and the development of standardised protocols ated with susceptibility to obesity in adulthood
for new data collections; nevertheless, each lon- (Elks et al. 2012); and how genetic studies can
gitudinal study has to have a scientic niche, as contribute to understanding the environmental
well as a core of common variables, to compete determinants of disease through the use of
successfully for renewed funding. Mendelian randomisation (Davey Smith and
Combining cohorts and establishing new Ebrahim 2003).
mega cohorts and biobanks is a feature of epide- New cohort studies established in this era gen-
miology more generally, driven by the dynamic erally have been much larger than previous cohort
and symbiotic relationship between rapid bio- studies, mainly in order to address genetic ques-
logical and technological advances and greater tions. Adult cohorts, such as UK Biobank,
scientic questioning and understanding of the recruited 0.5 million participants or more (Tyrrell
complex nature of human health and disease et al. 2013). The new prospective birth cohort
development across the life course. The very studies include the Danish National Birth Cohort
existence of rich data archives going back many (Olsen et al. 2001), which has a sample size of
years in an increasing number of cohorts over 80,000 children and data from before birth
promotes a life course perspective. For example, (Olsen et al. 2001), and the Norwegian Mother
the UK Dementia Platform, launched in 2014, and Child Study, which is similar in design and
brings together cohort studies (initially 22 popu- size (Magnus et al. 2006). The UK Life Study
lation or familial disease cohorts) and an infor- (www.lifestudy.ac.uk) is just beginning to collect
matics platform to co-ordinate and integrate data on 80,000 pregnancies and newborn chil-
dementia research. Its aim is nd how to detect dren. The US National Childrens Study aimed to
disease early, improve treatment of symptoms, have a sample of 100,000 recruited before birth
and delay the onset and progression of dementias (Landrigan et al. 2006), but funding was stopped
( http://www.mrc.ac.uk/research/facilities/ at the end of 2014 after the pilot phase.
dementias-research-platform-uk/). Some of the Allied to genomics, the emerging eld of sys-
cohorts involved have long-term follow up over tems biology also requires large scale detailed
many years, including the British 1946 cohort data in order to study the properties of cells,
from birth, that will bring a life course perspec- organisms and tissues, and how they interact.
tive to dementia research. Systems biology, including such elds as lipido-
The rapid development of genetic epidemiol- mics (cellular level study of lipids), and pro-
ogy, from the early candidate gene approach teomics (cellular level study of proteins), is
through to genome-wide association studies, concerned with the relationship between the
required large samples to discover small genetic genome of an organism and its characteristics
effects, to investigate rare as well as common dis- (phenotype). These elds of study will increase
eases, and to explore gene-environment interac- opportunities to study biomarkers and processes
tions. Cohort investigators formed consortia to of development, as well as ageing and disease at
Epidemiological Perspectives on the Life Course 653

the molecular and cellular levels. Particularly and minimally invasive devices which can be
notable is epigenomics, concerned with the com- used in large population samples to measure
plete set of epigenetic modications on the epig- physiological function and behaviours in every-
enome (the genetic material of a cell) and how day life; they are measures which do not need to
these may be affected by environmental inu- be undertaken in a clinical setting. They include,
ences (Relton and Davey Smith 2012). for example, monitoring of blood pressure or glu-
Prospective cohort studies are of great value for cose levels, heart rate or physical activity, in
systems biology which aims to gain insight into order to capture intensity and variability in real
biological mechanisms, test their predictive abil- time. Other technologies capture exposure to the
ity for disease risk and/or functional change, and physical environment (such as long-term pollu-
provide suitable targets for drug interventions. tion exposure). These new technologies are also
Pregnancy and birth cohort studies are of particu- part of big data, characterising better the expo-
lar value because they allow epigenetic changes some that has been described as including inter-
in prenatal and early postnatal life to be studied, nal body processes (e.g. metabolism), inuential
when the epigenome may be particularly suscep- external exposures (e.g. diet, lifestyle), and the
tible to inuence (Ng et al. 2012). These epigen- wider social, economic and psychological inu-
etic changes are thought to be potentially key ences (Wild 2012). That kind of big data for
explanations for the developmental origins of studies which have function and change as their
adult chronic diseases (Waterland and Michels outcomes, as compared with research into rela-
2007; Haggarty and Ferguson-Smith 2014); and tively rare diseases, does not necessarily require
a growing number of studies are investigating very large samples of individuals, and in some
this possibility (Heijmans et al. 2008; Relton cases smaller cohorts that can be intensively phe-
et al. 2012; Godfrey et al. 2011). Epigenetic epi- notyped may be an advantage. Currently the nec-
demiology (the study of inuences, such as essary size of these cohorts and the balance
smoking, on gene regulation of health, develop- between big data strategies and classical epide-
mental and ageing trajectories), and molecular miological approaches are being debated (Kuller
epidemiology (the study of such inuence as it et al. 2013).
happens at the molecular level), also benet from
the methodological tools and conceptual models
that have been developed in life course epidemi- 5 Future Directions of Life
ology. These relatively new elds of study are Course Epidemiology
well poised to investigate development and age-
ing trajectories and to identify the most relevant We cannot do justice here to all the types of epi-
life course models, while taking into account the demiological life course research that ourished
lifetime biological, behavioural and social risk in the last decade and/or will be priorities in the
factors that may confound, mediate or moderate decade to come. Important areas outside the
these trajectories (Ng et al. 2012). It is still too scope of this chapter include (1) family-based
early, and we do not yet have the necessary exper- studies, including intergenerational, sibling and
tise, to properly evaluate the additional value of twin studies, which are increasingly used to study
these omics collaborations in terms of insight within family differences and to strengthen
into the biological understanding or prognostic causal inference (Lawlor and Mishra 2009); (2)
value of functional change and disease develop- comparative studies, over time and location, of
ment, and the extent to which they will reveal populations and their variation in life course
changes at the molecular and cellular levels that health risks (Schooling and Leung 2010); and (3)
mediate or modify the effects of earlier environ- life course intervention studies (Hertzman 2007).
mental risk. Ultimately the purpose of epidemiology,
In addition to the omics technologies, there including life course epidemiology, is to improve
are a growing number of portable, cost effective population health. Strengthening causal infer-
654 M.E.J. Wadsworth and D. Kuh

ence for life course ndings from observational experts in the omics technologies applied to
cohort studies is particularly important as ran- population health have also become key collabo-
domised controlled trials are often impractical or rators. Team science should always be the pre-
unethical for testing life course hypotheses. ferred approach of investigators of prospective
There is a growing number of ways of improving cohort studies seeking answers to fundamental
causal inference in cohort and other observa- questions about development and ageing, and the
tional studies (Academy of Medical Sciences determinants and consequences of lifelong health
Working Group (Chair M Rutter) 2007). These and disease.
include: careful study design and appropriate Life course epidemiological interest began
analytical strategies, including sensitivity analy- principally as a concern with health and growth
ses, and the use of instrumental variables (e.g. in early life in a time of persistently high infant
Mendelian randomisation) or propensity scores mortality (Wadsworth 2010), and continues now
(Hardy et al. 2014); evidence synthesis, asking in a period when ageing is one of the most signi-
whether ndings from experimental and non- cant health problems. The concept of health is
experimental, and from animal and human stud- broadening to include the ability to adapt physi-
ies, are consistent; and the use of family based ologically, psychologically and socially (Huber
studies and cohort comparisons where factors et al. 2011). Within life course epidemiology the
that may potentially confound the relationship key questions on which the eld now concen-
between the exposure of interest and the health trates are:
outcome are differentially distributed.
Over a hundred years ago in the US and the How do human systems and function change
UK, public health ofcials acknowledged, and across life, and do they vary by gender, socio-
new welfare policies reected, the signicance of economic circumstances and opportunities,
the early environment for adult health. At that and ethnicity?
time empirical evidence was limited. Now the During which periods of life are important
evidence base has become stronger, and in the changes in function observed?
last decade life course perspectives are increas- What are the lifetime environmental and
ingly favoured by research funders, and by intrinsic biological determinants of level and
researchers and policymakers searching for change of function?
strategies to improve population health, and child Are the inuences of early life factors modi-
health in particular (Shonkoff et al. 2009; Halfon ed by factors in adulthood, and vice versa?
2012). The need for researchers to focus on trans- What is the relationship between earlier life
lational aspects of life course epidemiology for levels and changes in function with subse-
the development of health services and the deliv- quent disease risk?
ery of care has come to the fore. In the words of
the UKs Chief Medical Ofcer in her annual There is probably most opportunity to address
report for 2012: There is a growing knowledge these questions for cardiovascular function, as
of the complex interplay between psychosocial long-established cohort studies have many repeat
events and biological factors, and we now under- measures. However, a recent review of lifetime
stand that events that occur as a fetus develops trajectories of vascular and metabolic function
and in early life play a fundamental part in later showed that as yet surprisingly little use has been
life, and indeed in the lives of future generations made of the repeat data (Lawlor and Hardy 2014).
(Chief Medical Ofcer 2013). Addressing these questions will help to identify
The development of life course epidemiology when in life interventions may have their greatest
and its data resources began as collaborations impact on maintaining health, maximising func-
between epidemiologists, clinicians, social scien- tion and delaying the diseases of later life, and
tists and health service planners. In the last revealing the population subgroups that may ben-
decade, geneticists, epigeneticists, and other et most from interventions. Knowing how best
Epidemiological Perspectives on the Life Course 655

to intervene requires active collaborations with This kind of new work requires collaboration
those working in the implementation sciences, between life course epidemiology and a broad
policy makers, and those whom the interventions range of life and social sciences, and means that
are designed to help. life course epidemiology is involved in research
at cellular, individual and population levels.
For public health, life course epidemiology
6 Conclusions may be expected to deliver new information to
help in developing innovative preventive care,
Life course epidemiology evolved as hypotheses targeted at the most vulnerable individuals, and at
developed that many chronic diseases were prob- the period of their life course when it will have
ably preceded by a long period of risk develop- maximum benet. The now considerable resource
ment. When those hypotheses were developing, of life course data has great potential for investi-
demographers were studying determinants of gating long-term effects of exposures to health
change in fertility and mortality, and health sci- risks, to show age-cohort differences in risk
entists were beginning to be concerned about the experience, and consequently to model hypothet-
challenges of caring for the health of ageing pop- ical scenarios to show age-cohort differences in
ulations. Consequently epidemiologists, clini- early life risk experience, which reveal likely
cians and demographers developed life course future variation in population health and ageing.
data resources and devised ingenious ways to For the social sciences, life course epidemiol-
study long periods of life and to compare popula- ogy will continue to provide information on
tion health as well as the health of individuals health related determinants of population struc-
across generations. tural changes, including those brought about by
Epidemiological research into how health risk health care innovation. It will also continue to
develops throughout the life course investigates study how social adversity impacts the health of
the beginnings of individual vulnerability, and the individual, and the effects of socio-economic
later inuences that increase or protect that vul- change on population health. The international
nerability; it asks whether processes of accumu- development of new large-scale life course data
lation of risk or protective experience signicantly resources will provide opportunities for cross-
change vulnerability. At the same time research cultural comparative research and research across
in the life sciences has developed practical and time within countries.
reliable measures of health, illness and biological
function. It is now possible to measure trajecto-
ries of change over years of life, thereby increas- References
ing understanding of processes that precede
illness and that are associated with functional Abraham, S., Collins, G., & Nordsieck, M. (1971).
Relationship of childhood weight status to morbidity
decline with age. The genetic sciences have
in adults. HSMHA Health Reports, 86, 273284.
added new dimensions to research into health and Academy of Medical Sciences Working Group (Chair M
the processes of ageing. They now also require Rutter). (2007). Identifying the environmental causes
new life course data to study how genetic effects of disease: How should we decide what to believe and
when to take action? London: Academy of Medical
interact with the environment to increase vulner-
Sciences.
ability to health problems or to protect against Alwin, D. F. (2012). Integrating varieties of life course
risk, as well as to investigate environmental concepts. Journals of Gerontology Series B,
impact on gene expression. New data resources Psychological Sciences and Social Sciences, 67,
206220.
include combining existing life course data
Austad, S. (2008). Making sense of biological theories of
sources, and beginning new large-scale cohort aging. In V. Bengtson, M. Silverstein, N. Putney, &
studies, as well as studies of smaller samples D. Gans (Eds.), Handbook of theories of aging (2nd
with much more detailed data. ed., pp. 163178). New York: Springer.
656 M.E.J. Wadsworth and D. Kuh

Barker, D. J. P. (1991). Fetal and infant origins of adult Colley, J. R. T., Douglas, J. W. B., & Reid, D. D. (1973).
disease. London: BMJ Publishing. Respiratory disease in young adults; Inuence of early
Barker, D. J. P. (1995). Fetal origins of coronary heart dis- childhood lower respiratory tract illness, social class,
ease. British Medical Journal, 311, 171174. air pollution, and smoking. British Medical Journal, 2,
Barker, D. J., & Osmond, C. (1986). Infant mortality, 195198.
childhood nutrition, and ischaemic heart disease in Collins, S. D. (1949). Statistical studies of heart disease.
England and Wales. Lancet, 1(8489), 107781. V. Illness from heart and other cardiovascular-renal
Barker, D. J., Winter, P. D., Osmond, C., Margetts, B., & diseases recorded in general morbidity surveys of fam-
Simmonds, S. J. (1989). Weight in infancy and death ilies. Public Health Reports, 64, 14391497.
from ischaemic heart disease. Lancet, 2(8663), DAgostino, R. B., Kannel, W. B., Belanger, A. J., &
57780. Sytkowski, P. A. (1989). Trends in CHD and risk fac-
Ben-Shlomo, Y., & Kuh, D. (2002). A life course approach tors at age 5564 in the Framingham Study.
to chronic disease epidemiology: Conceptual models, International Journal of Epidemiology, 18, S67S72.
empirical challenges and interdisciplinary perspec- DAgostino, R. B., Sr., Pencina, M. J., Massaro, J. M., &
tives. International Journal of Epidemiology, 31, Coady, S. (2013). Cardiovascular disease risk assess-
285293. ment: Insights from Framingham. Global Heart, 8,
Ben-Shlomo, Y., Gardner, M., & Lightman, S. (2014a). A 1123.
life course approach to neuroendocrine systems: The Davey Smith, G., & Ebrahim, S. (2003). Mendelian ran-
example of the HPA axis. In D. Kuh, R. Cooper, domization: Can genetic epidemiology contribute to
R. Hardy, M. Richards, & Y. Ben-Shlomo (Eds.), A life understanding environmental determinants of disease?
course approach to healthy ageing (1st ed., pp. 133 International Journal of Epidemiology, 32, 122.
145). Oxford: Oxford University Press. Davey Smith, G., & Lynch, J. (2004). Life course
Ben-Shlomo, Y., Mishra, G., & Kuh, D. (2014b). Life approaches to socioeconomic differentials in health. In
course epidemiology. In W. Ahrens & I. Pigeot (Eds.), D. Kuh & Y. Ben-Shlomo (Eds.), A life course
Handbook of epidemiology (2nd ed., pp. 15211549). approach to chronic disease epidemiology (2nd ed.,
New York: Springer. pp. 77116). Oxford: Oxford University Press.
Berenson, G. S. (2001). Bogalusa Heart Study: A long- dos Santos Silva, I. (2004). Time trends in cancer inci-
term community study of a rural biracial (black/white) dence and mortality. In D. Kuh & Y. Ben-Shlomo
population. The American Journal of the Medical (Eds.), A life course approach to chronic disease epi-
Sciences, 322, 267274. demiology (2nd ed., pp. 281306). Oxford: Oxford
Board, E. (2009). The life course is coming of age. University Press.
Advances in Life Course Research, 14, 8386. Douglas, J. W. B. (1951). Health and survival of infants in
Booth, C. (1891). Life and labour of the people. London: different social classes. The Lancet, ii, 440446.
Williams and Norgate. Dubos, R., Savage, D., & Schaedler, R. (1966). Biological
Boyd Orr, J. (1937). Food health and income (2nd ed.). Freudianism. Lasting effects of early environmental
London: Macmillan. inuences. Pediatrics, 38, 789800.
Britton, A., Ben-Shlomo, Y., Benzeval, M., Kuh, D., & Elder, G. H. (1974). Children of the great depression.
Bell, S. (2015). Life course trajectories of alcohol con- Chicago: University of Chicago Press.
sumption in the United Kingdom using longitudinal Elks, C. E., Loos, R. J., Hardy, R., Wills, A. K., Wong, A.,
data from nine cohort studies. BMC Medicine, 13, 47. Wareham, N. J., et al. (2012). Adult obesity suscepti-
Brunner, E., & Marmot, M. (2006). Social organisation, bility variants are associated with greater childhood
stress and health. In M. Marmot & R. G. Wilkinson weight gain and a faster tempo of growth: The 1946
(Eds.), Social determinants of health (2nd ed., British Birth Cohort Study. American Journal of
pp. 630). Oxford: Oxford University Press. Clinical Nutrition, 95, 11501156.
Butler, N. R., & Bonham, D. G. (1963). Perinatal mortal- Elliott, J., & Shepherd, P. (2006). Cohort prole: 1970
ity. Edinburgh: Livingstone. British Birth Cohort (BCS70). International Journal
Buttereld, W. J. H. (1968). Priorities in medicine. of Epidemiology, 35, 836843.
London: Nufeld Provincial Hospitals Trust. Ferrucci, L. (2008). The Baltimore Longitudinal Study of
Caldwell, J. C. (2001). Population health in transition. Aging (BLSA): A 50-year-long journey and plans for
Bulletin of the World Health Organization, 79, the future. The Journals of Gerontology Series A:
159160. Biological Sciences and Medical Sciences, 63,
Chief Medical Ofcer. (2013). Chief Medical Ofcers 14161419.
annual report 2012: Our children deserve better: Forsdahl, A. (1977). Living conditions in childhood and
Prevention pays. Department of Health. https://www. adolescence and important risk factor for arterioscle-
gov.uk/government/publications/chief-medical- rotic heart disease? British Journal of Preventive and
officers-annual-report-2012-our-children- deserve- Social Medicine, 31, 9195.
better-prevention-pays Ghosh, A. K., Hardy, R. J., Francis, D. P., Chaturvedi, N.,
Cohen, S., Janicki-Deverts, D., Chen, E., & Matthews, Pellerin, D., Deaneld, J., et al. (2014, December).
K. A. (2010). Childhood socioeconomic status and Midlife blood pressure change and left ventricular
adult health. Annals of New York Academy of Sciences, mass and remodelling in older age in the 1946 British
1186, 3755.
Epidemiological Perspectives on the Life Course 657

birth cohort study. European Heart Journal, 35(46), Hertzman, C. (2007). Human development, life course,
32873295. intervention and health. In: Pickles A, Maughan B,
Glass, D. V. (1973). Numbering the people; The and Wadsworth M (Eds), Epidemiological methods in
eighteenth-century population controversy and the life course research (pp. 111136). Oxford: Oxford
development of census and vital statistics in Britain. University Press.
Farnborough: Saxon House. Hertzman, C., & Boyce, T. (2010). How experience gets
Gluckman, P. D., & Hanson, M. A. (2004). Living with under the skin to create gradients in developmental
the past: Evolution, development, and patterns of dis- health. Annual Review of Public Health, 31, 329347.
ease. Science, 305, 17331736. Huber, M., Knottnerus, J. A., Green, L., van der Horst, H.,
Gluckman, P. D., Hanson, M. A., Bateson, P., Beedle, Jadad, A. R., Kromhout, D., et al. (2011). How should
A. S., Law, C. M., Bhutta, Z. A., et al. (2009). Towards we dene health? British Medical Journal, 343,
a new developmental synthesis: Adaptive develop- d4163.
mental plasticity and human disease. The Lancet, 373, Huxley, R., Owen, C. G., Whincup, P. H., Cook, D. G.,
16541657. Rich-Edwards, J., Smith, G. D., et al. (2007). Is birth
Godfrey, K. M., Sheppard, A., Gluckman, P. D., Lillycrop, weight a risk factor for ischemic heart disease in later
K. A., Burdge, G. C., McLean, C., et al. (2011). life? American Journal of Clinical Nutrition, 85,
Epigenetic gene promoter methylation at birth is asso- 12441250.
ciated with childs later adiposity. Diabetes, 60, Illsley, R., & Kincaid, J. C. (1963). Social correlates of
15281534. perinatal mortality. In N. R. Butler & D. G. Bonham
Golding, J., Pembrey, M., & Jones, R. (2001). ALSPAC (Eds.), Perinatal mortality (pp. 270286). Edinburgh:
The Avon Longitudinal Study of Parents and Children. Churchill Livingstone.
I Study methodology. Paediatric and Perinatal Inskip, H. M., Godfrey, K. M., Robinson, S. M., Law,
Epidemiology, 15, 7487. C. M., Barker, D. J., & Cooper, C. (2006). Cohort pro-
Great Britain Parliamentary Papers. (1904). Report of the le: The Southampton womens survey. International
Interdepartmental Committee on physical deteriora- Journal of Epidemiology, 35, 4248.
tion. London: His Majestys Stationery Ofce. Johnson, W., Li, L., Kuh, D., & Hardy, R. (2015). How
Haggarty, P., & Ferguson-Smith, A. C. (2014). Life course has the age-related process of overweight or obesity
epigenetics and healthy ageing. In D. Kuh, R. Cooper, development changed over time? Co-ordinated analy-
R. Hardy, M. Richards, & Y. Ben-Shlomo (Eds.), A life ses of individual participant data from ve United
course approach to healthy ageing (1st ed., pp. 198 Kingdom birth cohorts. PLoS Medicine, 12(5),
212). Oxford: Oxford University Press. e1001828.
Halfon, N. (2012). Addressing health inequalities in the Jones, H. E. (1939). Resources for the consultant:
US: A life course health development approach. Social Principles and methods of the Adolescent Growth
Science and Medicine, 74, 671673. Study. Journal of Consulting Psychology, 3(5),
Halfon, N., Larson, K., Lu, M., Tullis, E., & Russ, S. 157159.
(2013). Lifecourse health development: Past, present Jones, H. E., & Bayley, N. (1941). The Berkeley growth
and future. Maternal and Child Health Journal, 18(2), study. Child Development, 12, 167180.
344365. Keating, D. P., & Hertzman, C. (Eds.). (1999).
Hamlin, C. (1992). Predisposing causes and public health Developmental health and the wealth of nations:
in early nineteenth-century medical thought. Social Social, biological and educational dynamics. New
History of Medicine, 5, 4370. York: The Guilford Press.
Hardy, R., Wadsworth, M. E. J., & Kuh, D. (2000). The Kermack, W. O., McKendrick, A. G., & McKinlay, P. L.
inuence of childhood weight and socioeconomic sta- (1934). Death rates in Great Britain and Sweden:
tus on change in adult body mass index in a British Some general regularities and their signicance. The
national birth cohort. International Journal of Obesity, Lancet, 226, 698703.
24, 110. Kirkwood, T. B., & Austad, S. N. (2000). Why do we age?
Hardy, R., Muniz-Terrera, G., & Hofer, S. M. (2014). Nature, 408, 233238.
Design of life course studies of healthy ageing. In Kuh, D., & Ben Shlomo, Y. (2014). Early life origins of
D. Kuh, R. Cooper, R. Hardy, M. Richards, & Y. Ben- adult health and ageing. In L. K. George & K. F.
Shlomo (Eds.), A life course approach to healthy age- Ferraro (Eds.), Handbook of aging and the social sci-
ing (1st ed., pp. 6578). Oxford: Oxford University ences. Elsevier Science Publishing Co Inc, San Diego.
Press. Kuh, D., & Ben-Shlomo, Y. (1997). A life course approach
Hattersley, A. T., & Tooke, J. E. (1999). The fetal insulin to chronic disease epidemiology: Tracing the origins
hypothesis: An alternative explanation of the associa- of ill-health from early to adult life (1st ed.). Oxford:
tion of low birthweight with diabetes and vascular dis- Oxford University Press.
ease. The Lancet, 353, 17891792. Kuh, D., & Ben-Shlomo, Y. (2004). A life course approach
Heijmans, B. T., Tobi, E. W., Stein, A. D., Putter, H., to chronic disease epidemiology (2nd ed.). Oxford:
Blauw, G. J., Susser, E. S., et al. (2008). Persistent epi- Oxford University Press.
genetic differences associated with prenatal exposure Kuh, D., & Davey Smith, G. (1993). When is mortality
to famine in humans. Proceedings of the National risk determined? Historical insights into a current
Academy of Sciences, 105, 1704617049. debate. Social History of Medicine, 6, 101123.
658 M.E.J. Wadsworth and D. Kuh

Kuh, D., dos Santos Silva, I., & Barrett-Connor, E. (2002). (MoBa). International Journal of Epidemiology, 35,
Disease trends in women living in established market 11461150.
economies: Evidence of cohort effects during the epi- Marmot, M., & Brunner, E. (2005). Cohort prole: The
demiological transition. In D. Kuh & R. Hardy (Eds.), Whitehall II study. International Journal of
A life course approach to womens health (pp. 347 Epidemiology, 34, 251256.
373). Oxford: Oxford University Press. Metropolitan Life Insurance Co. (1946). Studies in heart
Kuh, D., Ben Shlomo, Y., Lynch, J., Hallqvist, J., & disease. New York: Metropolitan Life Insurance Co.
Power, C. (2003). Life course epidemiology. Journal Miall, W. E., & Chinn, S. (1974). Screening for hyperten-
of Epidemiology and Community Health, 57, sion: Some epidemiological observations. British
778783. Medical Journal, 3, 595600.
Kuh, D., Pierce, M., Adams, J., Deaneld, J., Ekelund, U., Mishra, G., Nitsch, D., Black, S., De, S. B., Kuh, D., &
Friberg, P., et al. (2011). Cohort prole: Updating the Hardy, R. (2009). A structured approach to modelling
cohort prole for the MRC National Survey of Health the effects of binary exposure variables over the life
and Development: A new clinic-based data collection course. International Journal of Epidemiology, 38,
for ageing research. International Journal of 528537.
Epidemiology, 40, e1e9. Morris, J. N. (1967). Uses of epidemiology (2nd ed.).
Kuh, D., Ben-Shlomo, Y., Tilling, K., & Hardy, R. (2014a). London: Churchill Livingstone.
Life course epidemiology and analysis. In R. Detels, Morris, J. N., & Titmus, R. M. (1942). The epidemiology
M. Gulliford, Q. Abdool Karim, & C. Tan (Eds.), The of juvenile rheumatism. The Lancet, 234, 300307.
Oxford textbook of public health (6th ed.). Oxford: Ng, J. W., Barrett, L. M., Wong, A., Kuh, D., Smith, G. D.,
Oxford University Press. & Relton, C. L. (2012). The role of longitudinal cohort
Kuh, D., Cooper, R., Hardy, R., Richards, M., & Ben studies in epigenetic epidemiology: Challenges and
Shlomo, Y. (2014b). A life course approach to healthy opportunities. Genome Biology, 13, 246.
ageing (1st ed.). Oxford: Oxford University Press. Olsen, J., Melbye, M., Olsen, S. F., Sorensen, T. I., Aaby,
Kuller, L. H., Bracken, M. B., Ogino, S., Prentice, R. L., P., Andersen, A. M., et al. (2001). The Danish National
& Tracy, R. P. (2013). The role of epidemiology in the Birth Cohort Its background, structure and aim.
era of molecular epidemiology and genomics: Scandinavian Journal of Public Health, 29, 300307.
Summary of the 2013 AJE-sponsored Society of Ong, K. K., Elks, C. E., Wills, A. K., Wong, A., Wareham,
Epidemiologic Research Symposium. American N. J., Loos, R. J., et al. (2011). Associations between
Journal of Epidemiology, 178, 13501354. the pubertal timing-related variant in LIN28B and
Landrigan, P. J., Trasande, L., Thorpe, L. E., Gwynn, C., BMI vary across the life course. Journal of Clinical
Lioy, P. J., DAlton, M. E., et al. (2006). The National Endocrinology and Metabolism, 96, E125E129.
Childrens Study: A 21-year prospective study of Power, C., & Elliott, J. (2006). Cohort prole: 1958
100,000 American children. Pediatrics, 118, British birth cohort (National Child Development
21732186. Study). International Journal of Epidemiology, 35,
Lawlor, D. A., & Hardy, R. (2014). Vascular and meta- 3441.
bolic function across the life course. In D. Kuh, Raghupathy, P., Antonisamy, B., Geethanjali, F. S.,
R. Cooper, R. Hardy, M. Richards, & Y. Ben-Shlomo Saperia, J., Leary, S. D., Priya, G., et al. (2010).
(Eds.), A life course approach to healthy ageing (1st Glucose tolerance, insulin resistance and insulin
ed., pp. 146161). Oxford: Oxford University Press. secretion in young south Indian adults: Relationships
Lawlor, D., & Mishra, G. (2009). Family matters: to parental size, neonatal size and childhood body
Designing, analysing and understanding family based mass index. Diabetes Research and Clinical Practice,
studies in life course epidemiology (1st ed.). Oxford: 87, 283292.
Oxford University Press. Rantakallio, P. (1988). The longitudinal study of the
Leon, D., & Walt, G. (2000). Poverty, inequality and northern Finland birth cohort of 1966. Paediatric and
health: An international perspective. Oxford: Oxford Perinatal Epidemiology, 2, 5988.
University Press. Reid, D. D. (1969). The beginnings of chronic bronchitis.
Li, L., Hardy, R., Kuh, D., Lo Conte, R., & Power, C. Proceedings of the Royal Society of Medicine, 62, 316.
(2008). Child-to-adult body mass index and height Reid, D. D., Brett, G. Z., Hamilton, P. J., Jarrett, R. J.,
trajectories: a comparison of 2 British birth cohorts. Keen, H., & Rose, G. (1974). Cardiorespiratory dis-
American Journal of Epidemiology, 168(9), ease and diabetes among middle-aged male Civil
100815. Servants. A study of screening and intervention. The
Lumey, L. H., Stein, A. D., Kahn, H. S., van der Pal-de Lancet, 1, 469473.
Bruin, K. M., Blauw, G. J., Zybert, P. A., et al. (2007). Relton, C. L., & Davey, S. G. (2012). Is epidemiology
Cohort prole: The Dutch Hunger Winter families ready for epigenetics? International Journal of
study. International Journal of Epidemiology, 36, Epidemiology, 41, 59.
11961204. Relton, C. L., Groom, A., St Pourcain, B., Sayers, A. E.,
Magnus, P., Irgens, L. M., Haug, K., Nystad, W., Swan, D. C., Embleton, N. D., et al. (2012). DNA
Skjaerven, R., & Stoltenberg, C. (2006). Cohort pro- methylation patterns in cord blood DNA and body size
le: The Norwegian Mother and Child Cohort Study in childhood. PLoS ONE, 7, e31821.
Epidemiological Perspectives on the Life Course 659

Riley, M. W. (1987). Aging, health and social change: An Tyrrell, J. S., Yaghootkar, H., Freathy, R. M., Hattersley,
overview. In M. W. Riley, J. D. Matarazzo, & A. Baum A. T., & Frayling, T. M. (2013). Parental diabetes and
(Eds.), Perspectives in behavioural medicine: The birthweight in 236 030 individuals in the UK Biobank
aging dimension (pp. 114). Hillsdale: Lawrence Study. International Journal of Epidemiology, 42,
Erlbaum. 17141723.
Robins, L., & Rutter, M. (1990). Straight and devious Wadsworth, M. E. J. (2010). The origins and innovatory
pathways from childhood to adulthood. Cambridge: nature of the 1946 British birth cohort study.
Cambridge University Press. Longitudinal and Life Course Studies, 1, 136.
Roche, A. F. (1992). Growth, maturity, and body composi- Wadsworth, M. E. J. (2014). Focussing and funding a
tion: The Fels longitudinal study. Cambridge: birth cohort study over 20 years: The British 1946
Cambridge University Press. national birth cohort study from 16 to 36 years.
Rowntree, B. S. (1901). Poverty: A study of town life. Longitudinal and Life Course Studies, 1, 121136.
London: Macmillan. Wadsworth, M. E. J., Cripps, H. A., Midwinter, R. A., &
Royal Commission on Population. (1949). London: His Colley, J. R. T. (1985). Blood pressure at age 36 years
Majestys Stationery Ofce. and social and familial factors, cigarette smoking and
Rutter, M. (2012). Achievements and challenges in the body mass in a national birth cohort. British Medical
biology of environmental effects. Proceedings of the Journal, 291, 15341538.
National Academy of Science, 109(Suppl 2), Wadsworth, M., Kuh, D., Richards, M., & Hardy, R.
1714917153. (2006). Cohort prole: The 1946 National birth cohort
Rutter, M. (2013). Developmental psychopathology: A (MRC National Survey of Health and Development).
paradigm shift or just a relabeling? Development and International Journal of Epidemiology, 35, 4954.
Psychopathology, 25, 12011213. Walker, M., Whincup, P. H., & Shaper, A. G. (2004). The
Schooling, C. M., & Leung, G. M. (2010). A socio- British Regional Heart Study 19752004. International
biological explanation for social disparities in Journal of Epidemiology, 33, 11851192.
non-communicable chronic diseases: The product of Wall, W. D., & Williams, H. L. (1970). Longitudinal stud-
history? Journal of Epidemiology and Community ies and the social sciences (1st ed.). London:
Health, 64, 941949. Heinemann Educational Books Ltd.
Scottish Council for Research in Education. (1993). The Waterland, R. A., & Michels, K. B. (2007). Epigenetic
intelligence of Scottish children: A national survey of epidemiology of the developmental origins hypothe-
an age-group. London: University of London Press. sis. Annual Review of Nutrition, 27, 363388.
Settersten, R. A., Jr. (2009). It takes two to tango: The (un) Wellcome Trust Case Control Consortium. (2007).
easy dance between life-course sociology and life- Genome-wide association study of 14,000 cases of
span psychology. Advances in Life Course Research, seven common diseases and 3,000 shared controls.
14, 7481. Nature, 447, 661678.
Shonkoff, J. P. (2012). Leveraging the biology of adver- Whincup, P. H., Kaye, S. J., Owen, C. G., Huxley, R.,
sity to address the roots of disparities in health and Cook, D. G., Anazawa, S., et al. (2008). Birth weight
development. Proceedings of the National Academy of and risk of type 2 diabetes: A systematic review. The
Science, 109(Suppl 2), 1730217307. Journal of the American Medical Association, 300,
Shonkoff, J. P., Boyce, W. T., & McEwen, B. S. (2009). 28862897.
Neuroscience, molecular biology, and the childhood White, P. D. (1931). Heart disease. New York: Macmillan
roots of health disparities: Building a new framework Co.
for health promotion and disease prevention. The Wild, C. P. (2012). The exposome: From concept to
Journal of the American Medical Association, 301, utility. International Journal of Epidemiology, 41,
22522259. 2432.
Stein, S., Susser, M., Saenger, G., & Marolla, F. (1975). Williams, R. (2002). Hostility, psychosocial risk factors,
Famine and human development: The Dutch Hunger changes in brain serotonergic function, and heart dis-
Winter of 194445. New York: Oxford University ease. In S. A. Stansfeld & M. Marmot (Eds.), Stress
Press. and the heart: Psychosocial pathways to coronary
Steptoe, A., Breeze, E., Banks, J., & Nazroo, J. (2013). heart disease (pp. 86100). London: BMJ Books.
Cohort prole: The English longitudinal study of age- Wills, A. K., Lawlor, D. A., Matthews, F. E., Sayer, A. A.,
ing. International Journal of Epidemiology, 42, Bakra, E., Ben-Shlomo, Y., et al. (2011). Life course
16401648. trajectories of systolic blood pressure using longitudi-
Stocks, P. (1975). An appreciation. British Journal of nal data from eight UK cohorts. PLoS Medicine, 8,
Preventative and Social Medicine, 29, 6572. e1000440.
Strachan, D. P. (1997). Respiratory and allergic diseases. Wills, A. K., Lawlor, D. A., Muniz-Terrera, G., Matthews,
In D. Kuh & Y. Ben-Shlomo (Eds.), A life course F., Cooper, R., Ghosh, A. K., et al. (2012). Population
approach to chronic disease epidemiology (pp. 101 heterogeneity in trajectories of midlife blood pressure.
120). Oxford: Oxford University Press. Epidemiology, 23, 203211.
Titmuss, R. (1938). Poverty and population. London:
Macmillan.
The Inuence of Social Welfare
Policies on Health Disparities
Across the Life Course

Pamela Herd

1 Introduction large role in determining health and mortality


outcomes. These two literatures, however, one
Social welfare spending consumes around 60 % focused on the health impacts of policies and the
of the federal budgetand just as a large a frac- other focused on the processes shaping the evolu-
tion of state budgets (CBPP 2015). Moreover, the tion of health disparities across the life course,
magnitude of the cumulative effects of social now exist in parallel streams. The existing body
welfare policies on socioeconomic resources of research based on the life course perspective
throughout the life course, is large. Overall, the pays little attention to social welfare policy,
poverty rate for Americans would be 30 %, despite the likelihood that policies targeting early
instead of 16 %, without these social welfare life environments, education, and income, play a
policies (Sherman et al. 2013). While we tend to critical role in shaping a lifetime of dynamic and
focus on the relative lack of generosity of the US cumulative exposures and experiences, all of
welfare state as compared to our European coun- which contribute to evolving health disparities
terparts and our comparative poor health out- across the life course. Social welfare policies, or
comes, this has possibly led us to underestimate the state more broadly, has not been incorporated
the actual size and inuence of the US welfare theoretically or empirically into life course
stateeven if it is weaker in a relative sense. research on health disparities. In turn, policy
These policies do not just impact those with very research has been limited by its inattention to life
limited socioeconomic resources; from the cradle course processes, focusing, for example, on
to the grave, nearly every American benets from immediate health effects of social welfare poli-
the welfare state. cies, but failing to integrate or test the potentially
Evolving from decades of research that exam- life long implications of poverty reduction poli-
ined how socioeconomic position, especially cies targeted at children or the later life health
income and education, inuences health and con- impacts of policies targeted at working age
sequent health disparities across the life course, adults. Moreover, this literature has generally
there is a large and growing body of research neglected the potential inuence of social welfare
demonstrating that social welfare policies play a policies on adult and later life health, instead
emphasizing early life impacts and inuences in
isolation from the rest of the life course and the
P. Herd (*)
University of Wisconsin-Madison,
family in which individuals are situated.
Madison, WI, USA In this chapter I will show how a life course
e-mail: pherd@lafollette.wisc.edu perspective can help further inform the growing

Springer International Publishing Switzerland 2016 661


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_29
662 P. Herd

body of literature focused on the health impacts that take place within life-stage phases over the
of social welfare policies. I will begin by review- life span have consequences on health (p. 10).
ing our current knowledge regarding how health A central tenet of the life course perspective is
disparities emerge across the life courseas well that early life plays a critical role in shaping
as the inattention in this literature to the social health across the life course. The Barker hypoth-
welfare policies that may be shaping some of esis suggests that most sources of later life health
these disparities. In particular, I will demonstrate disparities can be found in fetal and infant life
the magnitude of these programs. For example, (Barker 1992; Barker et al. 1993). Poor maternal
50 % of all children will benet from social wel- health, which itself is often a product of poverty,
fare policy supports at some point before the age led to an in-utero environment that stagnated
of 18and these policies reduce the poverty rate infant development (Barker 1992; Barker et al.
for children from 30 to 18 % (CBPP 2013). I will 1993; for additional discussion of this hypothe-
then review the existing literature exploring the sis, see chapters by Avison Mental Health,
health impacts of social welfare policiesas Hayward and Sheehan Does the Body Forget?
well as its inattention to the life course conse- Adult Health, Life Course Dynamics, and Social
quences of these investments. I will focus on Change, and Ferraro Life Course Lens on
social welfare policies that foster strong early Aging and Health, this volume). For example,
childhood care and education, increase access to there are large differences in health at birth, as
health care, and supplement income. Finally, I measured by low birth weight, by parental socio-
will conclude by pointing to ways that life course economic status (Aizer and Currie 2014).
scholars focused on health disparities and policy Some of the health effects of poor in-utero
scholars focused on the inuence of social wel- environments, however, may not emerge until
fare policies on health can learn from each other later in life. In essence, this unhealthy in-utero
as to how to further develop our understanding of environment leads to biological programming of
the relationship between socioeconomic position ill health across the life course (Petronis 2010).
and health. The strongest evidence for long term health
effects of poor in-utero environments are from
studies of famines after World War II, which gen-
2 Health Disparities Across erally support the premise that deprived fetal
the Life Course environments increase risks for cardiovascular
risk factors (see Almond and Currie 2011 for a
In this section, I will review the contributions of review).
the life course perspective towards understanding There is more general evidence that childhood
how health disparities emerge and develop across experiences, not simply the fetal environment,
the life course, as well as the limitations of this affect health across the life course. Having par-
research in regards to attention to social welfare ents with limited socioeconomic resources, both
policies. The life course approach focuses on education and income, lead to accumulating
how processes of stratication work over the life health disadvantages across the life course (Case
span of individuals, families, groups, and cohorts et al. 2002; Luo and Waite 2005). Children who
to produce health outcomes in later life (Mortimer grow up poor or with parents with limited educa-
and Shanahan 2004). Essentially, a life course tion and occupational attainment have worse
perspective assumes that social status can affect health outcomes throughout childhood and into
health at any point in ones life, but that the tim- adulthood (Guralnik et al. 2006; Lawlor et al.
ing of exposures and experiences linked to ineq- 2005b; Power et al. 2005; Singh-Manoux et al.
uity may matter more at some ages in the life 2004) and old age (Blackwell et al. 2001; Holland
span. Alwin and Wray (2005) highlight that the et al. 2000). In addition to material resources, the
structure, sequence, and dynamics of events, child development literature has convincingly
transitions, and trajectories (social pathways) shown that early enriching environments, the
The Inuence of Social Welfare Policies on Health Disparities Across the Life Course 663

quality and stability of parental relations, and low very large (Backlund et al. 1996). Moreover,
exposure to adverse events, such as abuse, play a Medicaid pays for nearly one-half of all births in
critical role in shaping well-being in childhood the United States (Markus et al. 2013). Access to
and well into adulthood (Bradely et al. 1988, health insurance, both in terms of how that affects
1989; Estrada et al. 1987; Gottfried et al. 1994; access to health care, but also the degree to which
Pianta et al. 1997; Burchinal et al. 2000; Shonkoff it protects income against high medical costs,
and Phillips 2000; Knudsen et al. 2006; Heckman also has implications for childhood health.
2007, 2012; Ferraro Life Course Lens on Aging Finally, programs that affect children more
and Health, this volume). broadly, ranging from preschool to family leave
Further, the evidence from neuropsychology policies, also have received little attention from
indicates that these material and social resources life course scholars interested in health.
in childhood actually affect physical brain devel- The same can be said for research on adult and
opment and, through it, a host of behaviors later life health. Life course researchers have paid
(Danese et al. 2010; NSCDC 2011). A recent little attention to the social welfare policies that
summary of results from research on brain struc- affect socioeconomic resources in adulthood and
ture and emotional development in children notes later life. The inattention to policy is an important
that stated simply, as young children develop, limitation given the magnitude of the cumulative
their early emotional experiences literally effects of social welfare policies on socioeco-
become embedded in the architecture of their nomic resources in adulthood and, especially, in
brains (National Scientic Council on the later life. Among those aged 1864 the combined
Developing Child 2011). These cognitive and impact of social welfare policies reduces the pov-
psychological resources, in turn, inuence health erty rate from 23 to 15 %. Among older adults the
via their impact on behaviors that regulate expo- poverty rate, largely due to Social Security, is
sure to health risks. These factors have long been reduced from 30 to 18 % (CBPP 2013). Moreover,
considered coping resources that mitigate the the existing policy research has not lled this
effects of a range of stressors, including those hole, instead focusing on childhood health
associated with the SES hierarchy, on health impacts of social welfare policies. Yet, existing
(Pearlin et al. 2005). work focused on adult and later life would indi-
Yet despite this robust and growing body of cate that these policies matter. There is evidence
literature examining health disparities across the that not only do adult socioeconomic circum-
life course, strikingly absent has been the degree stances exert direct effects on adult and later life
to which early childhood policies, which exert health, they may also help offset the inuence of
strong impacts on early childhood resources, are early, and negative, childhood experiences on
playing a role in how these disparities develop long term health (Lynch et al. 1994; Pensola and
over the life course. This research generally fails Martikainen 2003; Luo and Waite 2005; Marmot
to recognize the magnitude of these programs. et al. 2001; Wamala et al. 2001). There is an
Income supports that affect children, including ongoing debate, however, regarding the extent to
the Earned Income Tax Credit (EITC), the Child which adult socioeconomic position (SEP) weak-
Tax Credit, Supplementary Nutritional Aid ens the health effects of childhood SEP. Many
Program (SNAP), housing assistance, and studies that have more effectively adjusted for
Supplemental Security Income (SSI), reduce the adult SEP contradict the Barker hypothesis; they
poverty rate for children from 30 to 18 % (CBPP nd that adult SEP is a better predictor of health
2013). Further, at least 50 % of children will outcomes than childhood SEP, to some extent
access at least one of these programs at some offsetting the health effects of disadvantage
point in their childhood. Given that the bulk of childhood experiences (Lynch et al. 1994;
the relationship between income and health is at Pensola and Martikainen 2003; Luo and Waite
the bottom end of the income distribution, the 2005; Marmot et al. 2001; Wamala et al. 2001).
cumulative impacts of these programs is likely However, most studies do nd that the effects of
664 P. Herd

childhood disadvantage persist into later life. The inuences health and well being (see chapter
remaining questions focus on the magnitude of Living Healthier and Longer: A Life Course
the direct effect of childhood experiences, but not Perspective on Education and Health by Johnson
the importance of the independent effect of adult et al., this volume), there is a growing body of
SEP (for a review see Galobardes et al. 2008).1 evidence focused specically on preschool expe-
riences very early in life, especially for children
from low socioeconomic backgrounds. The most
3 Social Welfare Policies well known studies are of the Perry Preschool
and Health Project and the Carolina Abecedarian Project,
which were experimental studies, done in the
In this section, I will review the existing literature 1960s and 1970s, that attempted to understand
examining the health effects of social welfare how high quality preschool experiences could
policies that foster strong early childhood care inuence child outcomes. Both programs were
and education, increase access to health care, and focused on disadvantaged children. They pro-
supplement income. There are two key limita- duced both short and long term impacts on cogni-
tions that are apparent in this review. First, there tion, educational attainment and earnings in
is a lack of attention to how policies may inu- adulthood (Heckman et al. 2012). More recently,
ence not only early childhood health, but also the Gabrielle Conti and James Heckman (2014)
cumulative impact of these policies over the life found evidence that those who received the pre-
course. Second, there is an inattention in this lit- school treatment also engaged in healthier behav-
erature to direct effects of social welfare policies iors later in life. They attribute these improved
on adult and later life health or the degree to outcomes to reductions in externalizing behavior
which these policies may offset a childhood char- or more broadly improvements in non-cognitive
acterized by limited socioeconomic resources. skills (Conti and Heckman 2013). While these
are very small samples, the long-term ndings on
adult health outcomes are striking. Further,
3.1 Social Welfare Policies: Early focusing on the long-term health effects of pre-
Childhood Care Environments school is very unusualand an important contri-
bution. That said, as Gregory Duncan and
Very early in life, the care environment can have colleagues (2014) emphasize, when one exam-
a profound inuence on early childhood develop- ines evidence from a broader array of preschool
ment. Social welfare policies focused on early experiments and studies, on average, positive
childhood education, particularly preschool, and development impacts, at least in childhood, are
parental leave policies in the immediate period relatively small.
following childbirth, have also generated some A nal critical area of social policy that might
meaningful evidence regarding health outcomes. inuence child-well being is family leave policy.
While there is a wealth of evidence demon- Relative to most industrialized countries, mothers
strating how educational attainment more broadly and fathers in the U.S. have limited access to leave
following the birth of a child. The only existing
1
It is important to point out, however, the methodological federal policy, the Family and Medical Leave Act
challenges inherent in this kind of life course research, (FMLA), provides 3 months of unpaid leave.
specically the interplay between health and socioeco-
Approximately half of workers are eligible because
nomic status, which make it difcult to assess the magni-
tude of these relationships. Poor health does affect FMLA generally only covers large employers. But
socioeconomic trajectories across the life course. A grow- the existing research, both in the U.S. and in other
ing body of literature documents how childhood health countries, shows a positive impact on childhood
affects educational attainment and adult income (Case
health. These policies are linked to better vaccina-
et al. 2005; Haas 2006; Palloni 2006; Johnson et al.
Living Healthier and Longer: A Life Course Perspective tion rates, higher rates of breastfeeding, and lower
on Education and Health, this volume). rates of post-neonatal and child mortality (Berger
The Inuence of Social Welfare Policies on Health Disparities Across the Life Course 665

et al. 2005; Ruhm 2000; Tanaka 2005; Rossin evidence that Medicaid coverage improves child
2011; Daku et al. 2012). health outcomes. Janet Currie and Jeffrey
The large bulk of the evidence examining the Grogger (2002) found that expansions of income
effect of family leave focuses on children rather eligibility thresholds for Medicaid increased pre-
than their mothers or fathers. There is some lim- natal care and reduced fetal deaths. Large
ited evidence, however, regarding the inuence Medicaid expansions in the 1980s and early
on mothers. In short, it does appear that women 1990s reduced childhood mortality (Currie and
experience mental health benets from family Gruber 1996). Janet Currie and colleagues (2008)
leave following the birth of a baby. These include also found that access to Medicaid coverage in
lower levels of depressive symptoms and signi- early childhood inuences later childhood health.
cantly lowered risk of severe depression (Chatterji Most of the research on Medicaid, and espe-
and Markowitz 2008; Dagher et al. 2014). The cially its expansions in coverage over time, has
paucity of this research, however, makes it dif- focused on the health of children. As one study
cult to make any strong statements about the notes, a key beneciarythe mother has been
effect of these policies on health in midlife. completely left out of the analysis (Conway and
Kutinova 2006). With large expansions of
Medicaid since the passage of the Affordable
3.2 Social Welfare Policies: Care Act in 2010, however, there was one recent
Medicaid and Medicare randomized experiment that evaluated the inu-
ence of the program on adult health in the state of
While much of the research on health disparities Oregon (Baicker et al. 2013). The second year
across the life course emphasizes the critical follow-up of this experiment found no inuence
nature of social resources, rather than medical of coverage on the prevalence, diagnosis or treat-
care or access to health insurance, there is a ment for hypertension or high levels of choles-
growing body of evidence that having access to terol. They did nd increased diagnoses of
health insurance does inuence health outcomes. diabetes, but no impact on glycated hemoglobin
The two primary sources of public health insur- levelsa measure of long term glucose levels.
ance for Americans of all ages include Medicaid The prevalence of depression, however, declined
and Medicare and cumulatively provide health signicantly and there was a dramatic decline in
insurance for 35 % of all Americansand nearly catastrophic medical care expenditures. In short,
100 % of those aged 65 and older (United States one might argue that the greatest inuence of the
Census Bureau 2014). program may have been on income security,
Medicaid, the primary source of public health which may have, in turn, inuenced mental
insurance for poor and/or disabled Americans, health.
plays an important role in providing that access. Older adults by and large rely on Medicare for
The program is jointly run between states and the their health insurance. And most receive cover-
federal government, so eligibility and benets age for Medicare at age 65, with the exception of
vary from state to state. That said, the program those who qualify based on disability at earlier
covers nearly all individuals below 100 % of the ages. The evidence is somewhat limited with
poverty line and, in many states those below 135 studies nding generally mixed effects. Most
% of the poverty line. Older adults relying on studies estimate the effects by looking at changes
Medicaid use it to either supplement their in health among the previously uninsured after
Medicare coverage, to pay for things like co- enrollment in Medicare (McWilliams et al. 2007,
payments and deductibles, or if they are receiving 2010; Polsky et al. 2009, 2010). Varying estima-
institutional care. The Medicaid program funds tion strategies produce variation in ndings with
half of all births in the United States and at any either no effect or a positive effect on health
given time 36 % of children receive health insur- (McWilliams et al. 2007, 2010; Polsky et al.
ance via Medicaid. Overall, there is meaningful 2009, 2010). There are two central challenges to
666 P. Herd

this estimation strategy. First, those previously in these localities to ensure compliance (Quadagno
uninsured are generally in poorer health than 2000). Ken Almond and colleagues (Forthcoming)
those with health insurance. A simple pre-post focused on the effects of this desegregationand
access to examining the effects of Medicare on consequent increase in access to medical care
health is difcult to do because even if those who on black infant mortality and the racial gap in
had been previously uninsured continue to face infant mortality. They focused on one of the lead-
health declines after receiving Medicare cover- ing causes of infant mortality during this era
age its not clear that health insurance hasnt had specically due to gastroenteritis and pneumonia,
some kind of impact. It may be that Medicare both of which required access to hospital based
slightly reduced the rate of decline even if health treatment for respiratory distress and dehydration.
is still declining. Even the best observable data They found dramatic reductions in infant mortal-
cannot guarantee that the comparison between ity for black infants but not white infantsduring
previously insured and previously uninsured is the period immediately following desegregation.
truly an equivalent comparison in terms of base- For example the black infant mortality rate in
line health. Second, because access to coverage is Mississippi, which had remained virtually
positively correlated with increasing ageone unchanged from 1955 through 1965, dropped
only becomes eligible for Medicare at 65 and had from 325 in 1000 to 130 in 1000 between 1965
been uninsured at younger agesthis would also and 1967, while white infant mortality remained
bias against nding a positive effect associated virtually unchanged. They were able to draw on
with access to Medicare. In short, because very localized variation in the timing of the deseg-
Medicare is a universal program, it is very dif- regation of these hospitals to demonstrate this
cult to nd plausible exogenous variation in effect throughout much of the South. And these
access to the benet that could lend insights into effects were not observed in the North where hos-
the programs inuence on health. As we will pitals had long been desegregated.
discuss later, the same issue exists for a program
like Social Security.
Interestingly, there is evidence of the impact of 3.3 Social Welfare Policies: Income
Medicare on the health of infantsalbeit via an Supports
indirect impact that was a function of the pro-
grams inuence on the desegregation of Southern In early life, the largest income support policy to
hospitals after the programs implementation in inuence economic security is the Earned Income
1966. In order for hospitals to receive Medicare Tax Credit (EITC). The EITC largely benets
funds they had to demonstrate they did not segre- poor and lower middle class families. It is a
gate. At this point, most Southern hospitals were refundable tax credit. In short, this means that
still segregated, with many hospitals still refusing individuals do not simply pay fewer taxes, many
to accept black patients or having a quota on the receive an income subsidy in the form of a tax
number of black patients they would treat. This creditthat is they pay no income taxes and
effectively reduced or eliminated access to medi- receive an income supplement. One must work to
cal care for many African Americans. However, be eligible for the credit. Up to an earnings
within the span of a relatively short periodabout threshold (or phase-out), the credit rises as earn-
a yearSouthern hospitals were desegregated as ings rise, but then gradually declines after that
a function of the Medicare requirement. But there earnings threshold. Maximum benets top out at
was resistance, especially in the deep South, and over $6000 a year. EITC benets lifted 5.3 mil-
in particular, in Mississippi, leading to variation lion children above the poverty line in 2012.
in when hospitals were actually desegregated. The Further, it is estimated that 50 % of all families
Social Security Administration was on the ground will receive the EITC for at least 1 year while
The Inuence of Social Welfare Policies on Health Disparities Across the Life Course 667

their children are below the age of 18 (Marr et al. and the consequent health of infants and children
2014-CBPP). (Cook et al. 2004). A study using state and county
The existing research focused on the EITC variation in the implementation of the food stamp
nds some meaningful health impacts of the pro- program found large reductions in low birth
gram, particularly early in life. Researchers have weight and smaller reductions in neonatal mor-
used two different approaches to capturing exog- tality associated with the implementation of the
enous variation in EITC benets. The rst program (Almond et al. 2011). While concerns
approach relies on state variation over time in have been raised in recent years that SNAP may
EITC benets. Statues can supplement federal encourage obesity via the consumption of high
EITC benets, and many states do. However, the caloric non-nutritional foods, there is little evi-
supplements are relatively small. As a conse- dence of any meaningful effect in this direction
quence, these estimates are likely underestimated (Kaushal 2007). An associated program, Women,
(Larrimore 2011; Strully et al. 2010). Indeed, Infants and Children (WIC), which provides
Strully and colleagues (2010) and Larrimore nutritional supports and education to pregnant
(2011) utilize this state variation and general women and new mothers nds a positive impact
health as an outcome found little impact of the of the program on reducing low birth weight and
EITC. More recent work, however, has utilized also on breastfeeding, though these studies have
an alternative identication strategy that captures not been able to utilize exogenous variation in
larger changes in benets. In short, the largest benet receipt and instead have attempted to rely
increase in benets associated with the EITC are on careful controls for observable measures that
related to the 1993 expansion, which expanded might confound their ndings (Rossin-Slater
benets more signicantly for those with two or 2013; Bitler and Currie 2005).
more children as compared to those with one While there is an increasingly richly detailed
child or no child. Utilizing the variation in bene- exploration of how the EITC, SNAP, and WIC
ts over time associated with family structure, inuence childhood health, the evidence regard-
the strongest evidence regarding the health ing how these programs might inuence adult
impacts of the EITC focus on low-birth weight. health is far more limited. One important excep-
For example, Hillary Hoynes and colleagues tion to this is a study by William Evans and Craig
(2012) nd that an increase in $1000 worth of Gerthwaite (2010), which examined how EITC
income leads to a 1.62.9 % reduction in the low benets impact mothers health. They focused on
birth weight rate, with the largest impacts the large expansion of the program that occurred
reserved for African American mothers. A central in 1993 and then considered changes in health for
mechanism appeared to be a reduction in mater- women with very limited educational attainment
nal smoking; other research also found improve- and two or more children (the group of women
ments in health behaviors (Rehkopf et al. 2014). most likely to be affected by the policy change)
There is also evidence that EITC benets reduced as compared to those with one child. These
child neglect (Berger et al. 2013). women experienced a reduced likelihood of poor
Second only to the EITC, SNAP (or what we mental health and improved measures of general
think of as Food Stamps), has had a large impact health, including measures that captured bio-
on the economic security of poor Americans and marker data, such as C-Reactive protein, a marker
especially poor children. Between the ages of 1 for inammation that is implicated in diseases
and 20 nearly 50 % of all children will receive ranging from cancer to cardiac disease. In terms
food stamp benets. It is estimated that the pro- of SNAP, one study did consider whether the use
gram lifts nearly 2.5 million children above the of the program increased the risk of obesity out-
poverty line. The SNAP program originated in comes for immigrant adults and found little evi-
the mid 1960s as a part of Lyndon Johnsons dence of an effect (Kaushal 2007). There is no
Great Society agenda. And most evidence indi- existing evidence examining how WIC inuences
cates it has had a large impact on food insecurity mothers health outcomes.
668 P. Herd

A key social welfare policy that could inu- fth of those aged 65 and over (SSA 2014).
ence adult health is unemployment insurance. Furthermore, for those that fall below eligibility
There is an enormous literature that documents guidelines for Social Security or those whose
the largely negative effects of unemployment on incomes fall well below the poverty threshold,
adult health, which a remarkably common expe- there is another safety net to offset extreme pov-
rience for most individuals (for a review see erty; Supplemental Security Income (SSI) subsi-
Roelfs et al. 2011; Dooley et al. 1996). Over 90 dizes incomes for about 6 % of elderly Americans.
% of individuals will have at least one unemploy- Indeed, Social Security and SSI have offset the
ment spell, that is a uninterrupted period of 1 or most severe forms of economic deprivation among
more months in which an individual was unem- the elderly. This is critical from a health perspec-
ployed, across their life and 70 % will have three tive because almost all prior evidence shows that
or more spells. Further, it is an experience that the largest reductions in health are associated with
disproportionately impacts those with low levels changes in income among those with the most lim-
of educational attainment. Those without a high ited incomes (e.g., Backlund et al. 1996). Social
school degree will have an average of 8 spells Security has been remarkably effective at reducing
compared to 3 spells for those with a college poverty rates among the elderly. Between 1960
degree. Unemployment duration varies widely, and 2005 the elderly poverty rate dropped from
but has been increasing from a low of 1015 almost 30 to 10 % (Engelhardt and Gruber 2004).
weeks from the 1950 to the 1980s to 32 weeks in Almost all of this decline can be attributed to ris-
2014though this may decline as the effects of ing Social Security benets (Engelhardt and
Great Recession subside. Despite the common- Gruber 2004). SSI helps to further protect the very
ness of the experience and the well-documented poor. At age 65, one would expect that the combi-
health impacts, there is only one existing study nation of eligibility for full Social Security bene-
that looks at the effect of unemployment insur- ts and SSI would further reduce poverty. Indeed,
ance on offsetting the negative health effects of the percentage of those aged 6569 falling below
unemployment. This study nds evidence that 75 % of the poverty level fell to 4.4 % (compared
while unemployment increases the risk for poor to 6.6 % among those aged 61). These changes in
mental health outcomes, access to unemploy- poverty rates do not prove that Social Security and
ment insurance softens these effects. However, SSI reduce severe poverty, but given the substan-
its effect is reduced if there is evidence of long- tial role that these programs play in providing
term unemployment (Tefft 2011). income for the poorest elderly Americans it cer-
While income support policies exert important tainly supports the point.
impacts for children and adults, the largest There is some evidence as to how Social
income support policy, specically Social Security and the Supplemental Security Income
Security, is targeted at older adultsthough they (SSI) may affect health. Indeed, Social Security
do also serve those under age 62 who are blind, is an obvious place to start given the magnitude
disabled, or a survivor of an individual eligible of the program, its effect on elderly peoples
for Social Security. Focusing on the health effects incomes, and its impact on poverty among elderly
of income support policies targeted at the elderly Americans. But it is difcult to estimate whether
is especially sensible given that there are no other Social Security affects health. Simply examining
points in the life course when incomes are so whether those with higher Social Security bene-
extensively supplemented by income support ts have better health will not indicate whether
policies. Indeed, in 2013, the U.S. delivered $814 Social Security benets improve health because
billion in Social Security benets, which was Social Security benets are based on individuals
approximately 25 % of overall federal spending. prior earnings, which may have been negatively
On average, Social Security comprises 40 % of affected by prior health. Thus, lower Social
annual incomes among those aged 65 and over. Security benets may have been determined by
Moreover, it comprises 80 % of incomes for one- prior health status.
The Inuence of Social Welfare Policies on Health Disparities Across the Life Course 669

One strategy is to examine health impacts of The rst study that examined the health effects
an unanticipated change in income. Snyder and of SSI looked at whether the implementation of
Evans (2006) use this type of design to examine the program had any affect on health. Taubman
the impact of varying Social Security benets on and Sickles (1983) used the Retirement History
mortality. Errant Social Security legislation led to Survey to examine how the health of elderly
a notch, with individuals with the exact same recipients changed after they started receiving
work histories born just before January 1, 1917 SSI. Individuals reported how their health com-
receiving higher Social Security benets in old pared to those of similar agebetter, the same or
age than those born just after this date. The worse. They found that SSI had a positive impact
experimental group, despite receiving about 7 % on the health of elderly beneciaries. The health
higher Social Security benets than the control of individuals eligible for SSI previous to imple-
group, had higher mortality rates after age 65 mentation was statistically signicantly worse
than the control group. While highly cited, this than the health of those not eligible. In both 1975
paper had two signicant problems. First, the and 1977after SSI was implementedthe dif-
poorest beneciaries, who are most likely to have ference in heath was no longer signicantly dif-
health effects from income increases, had negli- ferent between these two groups. There are a few
gible income increases. For example, a notch problems with this study, however. First, declin-
beneciary retiring at age 62 without a high ing differences in health may have been due to
school degree had just a 1 % higher benet or a mortality selectionSSI recipients may have
$5 higher monthly benet. Second, and even reected a more robust group of survivors. Finally,
more concerning, is a recent paper by Elizabeth SSI eligibility also guaranteed access to Medicaid
Weber Handwerker (2007) that found mortality as a supplement to Medicare. Thus, improved
differences between the experimental and control health may have been due to Medicaid, not SSI.
group previous to when these cohorts began col- Work by Herd and colleagues (2008) sought
lecting Social Security benets. Thus, it is possi- to build on Taubmans work, given both its prom-
ble that the higher mortality rates of the notch ise and pitfalls, with an alternative empirical
group likely had little or nothing to do with the design to test whether SSI impacts health. Instead
Social Security benets, but rather the fact that of testing whether the implementation of SSI has
they were sicker previous to receiving their Social an effect on health, they tested whether variation
Security benets. in SSI maximum benets over time within states
While Social Security may have had an impor- predicts changes in health. Though there is a fed-
tant impact on health, it is very difcult to design eral maximum SSI benet, the states can supple-
a study that can appropriately estimate its effect ment the federal benet. The federal minimum is
today, if ever, because of its universality and the set at ~75 % of the poverty line. As with Social
way an individuals health affects their choice Security, SSI is adjusted to account for ination.
regarding when to collect Social Security bene- In 2000, the federal monthly income minimum
ts, in addition to the size of those benets as a was $532 for single individuals and $789 for
function of how health may have inuenced earn- married couples. In 1990 and 2000, 25 and 27
ings over the life courses. SSI, though more lim- states, respectively, supplemented the federal
ited in the population affected, and its total effects benet. Thus, are within state changes in maxi-
on income, has some advantages for testing the mum SSI benets over time correlated with
effects of income supports on health. A key within state changes in disability? They found
advantage is that SSI is targeted at the poorest that increases in the maximum state SSI benet
elderly Americans (though the blind and disabled were correlated with reductions in disability rates
under age 65 are also eligible), and past research among the elderly, net of state and year xed
suggests that income supports that raise the effects and other controls.
incomes of the very poorest should have the larg- If we look outside of the U.S., however, there
est health effects (Backlund et al. 1996). are more promising ways to test the causal effects
670 P. Herd

of income transfer policies on the health of the social resources inuence health, this literature
elderly. Indeed, some of the most promising stud- could benet from a life course perspective. How
ies have been in developing countries, though might we improve our understanding of health
admittedly the extent of their applicability to the disparities by conducting social welfare policy
developed world is unclear. Under an income research in the context of a life course
support experiment, titled PROGRESSA, the perspective?
Mexican government has been providing since The rst area of research that remains unde-
1997 about $800 million in aid to almost one- veloped is the extent to which the inuence of
third of all rural families. The program has cer- policies on childhood health lingers across the
tain conditions that families must meet to obtain life course. Do children who are exposed to lower
aid. Families must seek preventative health care, levels of economic distress as a function of policy
children up to age 5 must have their growth mon- supports become adults with fewer health prob-
itored in clinic visits, and mothers must receive lems? As discussed above, the combination of
prenatal care and receive health education coun- the EITC and SNAP exerts signicant and large
seling. Additional income supplements were also impacts on poverty rates among children and
available if school age children attend school. familiesand there is evidence of childhood
Finally, the income was distributed directly to health effects. But we dont know whether these
mothers, an important distinction in a patriarchal effects continue to hold later in life. An analo-
culture (Gertler 2000). gous comparison would be the experimental lit-
The results showed striking improvements in erature regarding how preschools inuence
health for children, adults, and those over age 50. long-term education and earnings outcomes.
Those over age 50, whose only requirement for There is evidence that these effects subside as
participation was a yearly preventative check up, individuals age (Shager et al. 2012).
had signicant reductions in activity limitations More broadly, the debates among life course
due to illness, fewer days bedridden due to sick- theorists regarding the relative inuence of early,
ness, and more generally an increase in energy as compared to mid life, socioeconomic resources
levels as measured by their ability to walk dis- on later life health outcomes raises questions
tances without signicant fatigue. Children and regarding the relative inuence of social policies
adults also showed improved outcomes. But it depending on the point in the life course at which
could not be proven than income had an indepen- these policies targeted. It may be that childhood
dent effect on the childrens health, due to the is a sensitive period and improvements in eco-
medical care requirements linked to the receipt of nomic security associated with social welfare
income benets. Because of its success the pro- policies have a greater cumulative inuence on
gram is now being generalized to urban Mexico health across the life course as compared to social
and adopted by Argentina, Columbia, Honduras, welfare polices targeted at adults, such as Social
and Nicaragua. Security. Alternately, perhaps socioeconomic
resources, improved by social welfare policies in
adulthood, may offset the inuence of early life
4 Future Research Directions deprivation on mid and later life health.
A key tenet of life course theory is the concept
The gap between the literature focused on the of linked lives and social ties. We do not exist in
health effects of social welfare policies and life isolation from each otherpolicies not only
course research focused on health disparities has inuence individuals, they inuence families. For
left some important lingering questions as to how example, policies targeted at children, like SNAP
health disparities develop and evolve over the life and the EITC, may improve both parent and child
course. While examining how social welfare pol- health via relational mechanisms. Reduced stress
icies inuence childhood health has proved to be associated with economic deprivation may lead
important for understanding how economic and to both direct improvements in health by reduc-
The Inuence of Social Welfare Policies on Health Disparities Across the Life Course 671

ing individual stress, but also indirectly by schooling, this inuences long-term health tra-
improving cognitive and psychological develop- jectories via the inuence of educational attain-
ment in children. There is a growing body of evi- ment on health (Herd 2010). In turn, there is
dence that the stress associated with economic considerable evidence as to how educational
insecurity increases negative interactions attainment inuences the entry intoand exits
between parents and children and negatively frommarriage. Patterns of coupling and decou-
impacts brain development (Lupien et al. 2009; plingpathways into and out of marriageall
Teicher et al. 2003). have implications for health (Williams and
Life course theories could also inform existing Umberson 2004; Umberson et al. 2006).
policy research by focusing attention on the role More broadly, if we want to understand how
of broader social ties and how these might inu- health disparities develop and evolve across the
ence whetheror howsocial welfare policies life course, it is critical that we understand the
inuence health. For example, a longstanding role that social policies have played in inuenc-
interest has been whether housing programs that ing this process. Would health disparities be even
provide vouchers, which allow people to move wider if not for the inuence of redistributory
out of bad neighborhoods, can improve out- policies? That is, to what extent are existing nd-
comes ranging from employment to health. ings in the literature, in part, a function of social
Consequently, a experimental program titled welfare policies? These questions remain unan-
Moving to Opportunity, which was sponsored by swered and hold clues both to what shapes health
the U.S. Department of Housing and Urban disparities, but also how to ameliorate health
Development, randomly assigned housing vouch- disparities.
ers that allowed people in the experimental group The life course literature can also learn from
to move to better neighborhoods. While there existing policy research lessons about how to
were generally improvements in adult health identify exogenous factors that inuence health.
the ndings were slightly more mixed among Scholars operating in the life course perspective
children (see also discussion in Browning et al. are interested in how SEP and health affect each
Neighborhood, Place, and the Life Course, this other in complex ways over the life course to pro-
volume). In particular, male adolescents increased duce health disparities across the life course. This
risky behaviors like smoking. Investigators found includes acknowledging that SEP and health
these effects surprising and hypothesized it may likely have reciprocal relationships across the life
be that boys and girls socialize in different ways course, contributing to chains of risk and accu-
that makes boys more susceptible to negative mulation of risk that individuals face as they age
inuences (Ludwig et al. 2013). However, this (Haas 2008; Hayward and Gorman 2004; Robert
doesnt explain why, in some ways, the boys who et al. 2010). The policy literature provides an
remained in poor neighborhoods fared better than alternative tool set for identifying, for example,
the boys who moved to neighborhoods with less exogenous changes in income (via income sup-
poverty. A central question regards how the social port policies) that helps address the underlying
ties of these children were altered with this move. reciprocal nature of these relationships to better
For example, were they actually more isolated as ensure that results identify a casual impact of
poor individuals in better off neighborhoods income, for example, on health, as opposed to the
reducing the degree to which neighborhood par- reverse. The common techniques employed,
ents might help keep negative behaviors in check? many of which are discussed throughout this
Life course theories might also inform think- chapter, all rely on variation in the implementa-
ing as to how social welfare policies affect transi- tion of policies across geographic areas (often
tions, trajectories and pathways across the life states) and time.
course. For example, if policies like the EITC It is important to remember, however, that this
inuence schooling transitions by increasing the approach is not without limitations. There are
propensity for children to complete more years of two key issues that can undermine this strategy.
672 P. Herd

First, other exogenous variables can be correlated our ability to understand the development and
with the policy change that may confound results evolution of health disparities across the life
showing a relationship between a social welfare course.
policy and a health outcome. For example, stud-
ies focused on the inuence of Medicaid expan-
sions on childhood health that failed to address References
the simultaneous expansion of WIC (an income
supplement) for which many individuals were Aizer, A., & Currie, J. (2014). The intergenerational trans-
mission of inequality: Maternal disadvantage and
automatically identied as eligible for one pro-
health at birth. Science, 344(6186), 856861.
gram if they participated in the other, make it Almond, D., & Currie, J. (2011). Killing me softly: The
difcult to assess whether shown health improve- fetal origins hypothesis. The Journal of Economic
ments were a function of access to health insur- Perspectives, 25(3), 153172.
Almond, D., Hoynes, H. W., & Schanzenbach, D. W.
ance or income supports. Second, it is also the
(2011). Inside the war on poverty: The impact of food
case that the treatment needs to be valid. For stamps on birth outcomes. The Review of Economics
example, if one examines the inuence of income and Statistics, 93(2), 387403.
supports on health, policies that exert a small Alwin, D. F., & Wray, L. A. (2005). A life-span develop-
ment perspective on social status and health. The
inuence on the low end of the income distribu-
Journals of Gerontology Series B, 60B, 714.
tion are unlikely to demonstrate large health Backlund, E., Sorlie, P. D., & Johnson, N. J. (1996). The
effects given the non-linear relationship between shape of the relationship between income and mortal-
income and health (Backlund et al. 1996). ity in the United States: Evidence from the National
Longitudinal Mortality Study. Annals of Epidemiology,
6(1), 1220.
Baicker, K., Taubman, S. L., Allen, H. L., Bernstein, M.,
5 Conclusion Gruber, J. H., Newhouse, J. P., Schneider, E. C.,
Wright, B. J., Zaslavsky, A. M, & Finkelstein, A. N.
(2013). The Oregon experimentEffects of Medicaid
More broadly, the research focused on the devel-
on clinical outcomes. New England Journal of
opment of health disparities across the life course Medicine, 368(18), 17131722.
needs to theoretically tackle the role that the state Barker, D. J. P. (Ed.). (1992). Fetal and infant origins of
has in shaping access to, and the distribution of, adult disease. London: British Medical Journal Books.
Barker, D. J., Gluckman, P. D., Godfrey, K. M., Harding,
economic and social resources via social welfare
J. E., Owens, J. A., & Robinson, J. S. (1993). Fetal
policies, including more broadly other redistribu- nutrition and cardiovascular disease in adult life.
tive systems, especially the tax system. Home Lancet, 10, 938941.
mortgage deductions, child care subsidies, health Berger, L. M., Hill, J., & Waldfogel, J. (2005). Maternity
leave, early maternal employment and child health and
insurance tax subsides, among others, constitute
development in the US*. The Economic Journal,
nearly $470 billion dollars a year that is redistrib- 115(501), F29F47.
uted among American taxpayers, which is equiv- Berger, L. M., Font, S. A., Slack, K. S., & Waldfogel, J.
alent to 20 % of federal spending. While we tend (2013). Income and child maltreatment: Evidence
from the earned income tax credit. Paper prepared for
to focus on the relative lack of generosity of the
presentation at the 2013 Annual Meeting of the
US welfare state as compared to our European Association of Public Policy Analysis and
counterparts and our comparative poor health Management, Washington, DC.
outcomes, this has possibly led us to underesti- Bitler, M. P., & Currie, J. (2005). Does WIC work? The
effects of WIC on pregnancy and birth outcomes.
mate the actual size and inuence of the US wel-
Journal of Policy Analysis and Management, 24(1),
fare stateeven if it is weaker in a relative sense. 7391.
Nonetheless, the relative inattention to the state Blackwell, D. L., Hayward, M. D., & Crimmins, E. M.
in the context of the life course perspective (nota- (2001). Does childhood health affect chronic morbid-
ity in later life? Social Science & Medicine, 52,
ble exceptions including ORand and Henretta
12691284.
1999; Mayer and Schoepin 1989; Leisering Bradley, R. H., Caldwell, B. M., & Rock, S. L. (1988).
2003)as compared to social institutions such as Home environment and school performance: A ten-
work, family, and education has undermined year follow-up and examination of three models of
The Inuence of Social Welfare Policies on Health Disparities Across the Life Course 673

environmental action. Child Development, 59(4), Journal of Health Politics, Policy and Law, 39(2),
852867. 369416.
Bradley, R. H., Caldwell, B. M., Rock, S. L., Ramey, Daku, M., Raub, A., & Heymann, J. (2012). Maternal
C. T., Barnard, K. E., Gray, C., Hammon, M. A., & leave policies and vaccination coverage: A global
Johnson, D. L. (1989). HOME environment and cog- analysis. Social Science & Medicine, 74(2), 120124.
nitive development in the rst 3 years of life: A col- Danese, A., Caspi, A., Williams, B., Ambler, A., Sugden,
laborative study involving six sites and three ethnic K., Mika, J., Werts, H., Freeman, J., Pariante, C. M.,
groups in North America. Developmental Psychology, Moftt, T. E. & Arseneault, L. (2010). Biological
25(2), 217. embedding of stress through inammation processes
Burchinal, M. R., Peisner-Feinberg, E., Bryant, D. M., & in childhood. Brain, Behavior, and Immunity, 24, S8.
Clifford, R. (2000). Childrens social and cognitive Dooley, D., Fielding, J., & Levi, L. (1996). Health and
development and child-care quality: Testing for differ- unemployment. Annual Review of Public Health,
ential associations related to poverty, gender, or eth- 17(1), 449465.
nicity. Applied Developmental Science, 4(3), Duncan, G. J., Magnuson, K., & Votruba-Drzal, E. (2014).
149165. Boosting family income to promote child develop-
Case, A., Lubotsky, D., & Paxson, C. (2002). Economic ment. The Future of Children, 24, 99120.
status and health in childhood: The origins of the gra- Engelhardt, G., & Gruber, J. (2004). Social security and
dient. The American Economic Review, 92, the evolution of elderly poverty (NBER Working Paper
Washington, DC, 13081334. No. 10466). Cambridge, MA: NBER.
Case, A., Fertig, A., & Paxson, C. (2005). The lasting Estrada, P., Arsenio, W. F., Hess, R. D., & Holloway, S. D.
impact of childhood health and circumstance. Journal (1987). Affective quality of the motherchild relation-
of Health Economics, 24, 365389. ship: Longitudinal consequences for childrens school-
Center on Budget and Policy Priorities. (2015). Policy relevant cognitive functioning. Developmental
basics: Where do our tax dollars go? http://www.cbpp. Psychology, 23(2), 210.
org/sites/default/files/atoms/files/4-14-08tax.pdf . Evans, W. N., & Garthwaite, C. L. (2010). Giving mom a
Accessed 6 Aug 2015. break: The impact of higher EITC payments on mater-
Chatterji, P., & Markowitz, S. (2008). Family leave after nal health (No. w16296). Boston, MA: National
childbirth and the health of new mothers (No. Bureau of Economic Research.
w14156). National Bureau of Economic Research. Galobardes, B., Lynch, J. W., & Smith, G. D. (2008). Is
MA: Boston the association between childhood socioeconomic cir-
Conti, G., & Heckman, J. J. (2013). The developmental cumstances and cause-specic mortality established?
approach to child and adult health. Pediatrics, Update of a systematic review. Journal of Epidemiology
131(Supplement 2), S133S141. and Community Health, 62(5), 387390.
Conti, G., & Heckman, J. J. (2014). Economics of child Gertler, P. (2000). Final report: The impact of PROGRESA
well-being (pp. 363401). Dordrecht: Springer. on health. Washington, DC: International Food Policy
Conway, K. S., & Kutinova, A. (2006). Maternal health: Research Institute.
Does prenatal care make a difference? Health Gottfried, A. E., Fleming, J. S., & Gottfried, A. W. (1994).
Economics, 15(5), 461488. Role of parental motivational practices in childrens
Cook, J. T., Frank, D. A., Berkowitz, C., Black, M. M., academic intrinsic motivation and achievement.
Casey, P. H., Cutts, D. B., Meyers, A.F., Zaldivar, N., Journal of Educational Psychology, 86(1), 104.
Skalicky, A., Levenson, S., Heeren, T. & Nord, M. Guralnik, J. M., Butterworth, S., Wadsworth, M. E. J., &
(2004). Food insecurity is associated with adverse Kuh, D. (2006). Childhood socioeconomic status pre-
health outcomes among human infants and toddlers. dicts physical functioning a half century later. Journals
The Journal of Nutrition, 134(6), 14321438. of Gerontology Biological and Medical Sciences, 61,
Currie, J., & Grogger, J. (2002). Medicaid expansions and 694701.
welfare contractions: Offsetting effects on prenatal Haas, S. A. (2006). Health selection and the process of
care and infant health? Journal of Health Economics, social stratication: The effect of childhood health on
21(2), 313335. socioeconomic attainment. Journal of Health and
Currie, J., & Gruber, J. (1996). Saving babies: The ef- Social Behavior, 47, 339354.
cacy and cost of recent changes in the medicaid eligi- Haas, S. A. (2008). Trajectories of functional health: The
bility of pregnant women. Journal of Political long arm of childhood health and socioeconomic
Economy, 6, 12631296. factors. Social Science & Medicine, 66, 849861.
Currie, J., Decker, S., & Lin, W. (2008). Has public health Hayward, M. D., & Gorman, B. (2004). The long arm of
insurance for older children reduced disparities in childhood: The inuence of early life conditions on
access to care and health outcomes? Journal of Health mens mortality. Demography, 41, 87108.
Economics, 27(6), 15671581. Heckman, J. J. (2007). The economics, technology, and
Dagher, R. K., McGovern, P. M., & Dowd, B. E. (2014). neuroscience of human capability formation.
Maternity leave duration and postpartum mental and Proceedings of the National Academy of Sciences,
physical health: Implications for leave policies. 104(33), 1325013255.
674 P. Herd

Heckman, J. J. (2012). The developmental origins of Lynch, J. W., Kaplan, G. A., Cohen, R. D., Wilson, T. W.,
health. Health Economics, 21(1), 2429. Smith, N. L., Kauhanen, J., & Salonen, J. T. (1994).
Heckman, J. J., Pinto, R., & Savelyev, P. A. (2012). Childhood and adult socioeconomic status as predic-
Understanding the mechanisms through which an tors of mortality in Finland. The Lancet, 343(8896),
inuential early childhood program boosted adult out- 524527.
comes (No. w18581). Boston, MA: National Bureau Markus, A. R., Andres, E., West, K. D., Garro, N., &
of Economic Research. Pellegrini, C. (2013). Medicaid covered births, 2008
Herd, P., Schoeni, R. F., & House, J. S. (2008). Upstream through 2010, in the context of the implementation of
solutions: Does the supplemental security income pro- health reform. Womens Health Issues, 23(5),
gram reduce disability in the elderly? Milbank e273e280.
Quarterly, 86, 545. Marmot, M., Shipley, M., Brunner, E., & Hemingway, H.
Herd, P. (2010). Education and health in late-life among (2001). Relative contribution of early life and adult
high school graduates cognitive versus psychological socioeconomic factors to adult morbidity in the
aspects of human capital. Journal of Health and Social Whitehall II study. Journal of Epidemiology and
Behavior, 51(4), 478496. Community Health, 55(5), 301307.
Holland, P., Berney, L., Blane, D., Davey Smith, G., Marr, C., Charite, J., & Huang, C. C. (2014). Earned
Gunnell, D. J., & Montgomery, S. M. (2000). Life Income Tax Credit promotes work, encourages chil-
course accumulation of disadvantage: Childhood drens success at school, research nds. Washington,
health and hazard exposure during adulthood. Social DC: Center on Budget and Policy Priorities. http://
Science and Medicine, 50, 12851295. www.cbpp.org/les/6-26-12tax.pdf
Hoynes, H. W., Miller, D. L., & Simon, D. (2012). Income, Mayer, K. U., & Schoepin, U. (1989). The state and the
the earned income tax credit, and infant health (No. life course. Annual Review of Sociology, 15, 187209.
w18206). Boston, MA: National Bureau of Economic McWilliams, J. M., Meara, E., Zaslavsky, A. M., &
Research. Ayanian, J. Z. (2007). Health of previously uninsured
Kaushal, N. (2007). Do food stamps cause obesity?: adults after acquiring Medicare coverage. JAMA,
Evidence from immigrant experience. Journal of 298(24), 28862894.
Health Economics, 26(5), 968991. McWilliams, J. M., Meara, E., Zaslavsky, A. M., &
Knudsen, E. I., Heckman, J. J., Cameron, J. L., & Ayanian, J. Z. (2010). Commentary: Assessing the
Shonkoff, J. P. (2006). Economic, neurobiological, health effects of medicare coverage for previously
and behavioral perspectives on building uninsured adults: A matter of life and death? Health
Americas future workforce. Proceedings of the Services Research, 45(5p1), 14071422.
National Academy of Sciences, 103(27), Mortimer, J. T., & Shanahan, M. J. (Eds.). (2004).
1015510162. Handbook of the life course. New York: Springer.
Larrimore, J. (2011). Does a higher income have positive National Scientic Council on the Developing Child.
health effects? Using the earned income tax credit to (2011). Childrens emotional development is built into
explore the incomehealth gradient. Milbank the architecture of their brains (Working Paper No. 2,
Quarterly, 89(4), 694727. 3rd Printing). Retrieved from http://developingchild.
Lawlor, D. A., Ebrahim, S., & Davey Smith, G. (2005). harvard.edu/index.php?cID=152
Adverse socioeconomic position across the life course ORand, A. M., & Henretta, J. C. (1999). Age and inequal-
increases coronary heart disease risk cumulatively. ity: Diverse pathways through later life. Boulder:
Journal of Epidemiology & Community Health, 59, Westview Press.
785793. Palloni, A. (2006). Reproducing inequalities: Luck, wal-
Leisering, L. (2003). Government and the life course lets, and the enduring effects of childhood health.
(pp. 205225). New York. Demography, 43, 587615.
Luo, Y., & Waite, L. J. (2005). The impact of childhood Pearlin, L. I., Schieman, S., Fazio, E. M., & Meersman,
and adult SES on physical, mental, and cognitive well- S. C. (2005). Stress, health, and the life course: Some
being in later life. The Journals of Gerontology Series conceptual perspectives. Journal of Health and Social
B: Psychological Sciences and Social Sciences, 60(2), Behavior, 46(2), 205219.
S93S101. Pensola, T. H., & Martikainen, P. (2003). Cumulative
Lupien, S. J., McEwen, B. S., Gunnar, M. R., & Heim, C. social class and mortality from various causes of adult
(2009). Effects of stress throughout the lifespan on the men. Journal of Epidemiology Community Health
brain, behaviour and cognition. Nature Reviews Services, 57, 745751.
Neuroscience, 10(6), 434445. Petronis, A. (2010). Epigenetics as a unifying principle in
Ludwig, J., Duncan, G. J., Gennetian, L. A., Katz, L. F., the aetiology of complex traits and diseases. Nature,
Kessler, R. C., Kling, J. R., & Sanbonmatsu, L. (2013). 465(7299), 721727.
Long-term neighborhood effects on low-income fami- Pianta, R. C., Nimetz, S. L., & Bennett, E. (1997). Mother-
lies: Evidence from moving to ppportunity (No. child relationships, teacher-child relationships, and
w18772). Boston: National Bureau of Economic school outcomes in preschool and kindergarten. Early
Research. Childhood Research Quarterly, 12(3), 263280.
The Inuence of Social Welfare Policies on Health Disparities Across the Life Course 675

Polsky, D., Doshi, J. A., Escarce, J., Manning, W., Shonkoff, J. P., & Phillips, D. (2000). From neurons to
Paddock, S. M., Cen, L., & Rogowski, J. (2009). The neighborhoods: The science of early childhood
health effects of medicare for the nearelderly unin- development. Washington, DC: National Academy
sured. Health Services Research, 44(3), 926945. Press.
Polsky, D., Doshi, J. A., Manning, W. G., Paddock, S., Singh-Manoux, A., Ferrie, J., Chandola, T., & Marmot,
Cen, L., Rogowski, J., & Escarce, J. J. (2010). M. (2004). Socioeconomic trajectories across the life
Response to McWilliams commentary: Assessing the course and health outcomes in midlife: Evidence for
health effects of medicare coverage for previously the accumulation hypothesis? International Journal of
uninsured adults: A matter of life and death?. Health Epidemiology, 33, 10721079.
Services Research, 45(5), 14231429. Snyder, S. E., & Evans, W. N. (2006). The effect of income
Power, C., Hypponen, E., & Davey Smith, G. (2005). on mortality: Evidence from the social security notch.
Socioeconomic position in childhood and early adult The Review of Economics and Statistics, 88,
life and risk of mortality: A prospective study of the 482495.
mothers of the 1958 British birth cohort. American Strully, K. W., Rehkopf, D. H., & Xuan, Z. (2010). Effects
Journal of Public Health, 95, 13961402. of prenatal poverty on infant health state earned
Quadagno, J. B. (2000). Promoting civil rights through income tax credits and birth weight. American
the welfare state: How medicare integrated southern Sociological Review, 75(4), 534562.
hospitals. Social Problems, 47, 6889. Social Security Administration. (2014). Fast facts and g-
Rehkopf, D. H., Strully, K. W., & Dow, W. H. (2014). The ures about social security, 2014. Baltimore: Social
short-term impacts of Earned Income Tax Credit dis- Security Administration. http://www.ssa.gov/policy/
bursement on health. International Journal of docs/chartbooks/fast_facts/index.html. Accessed 2
Epidemiology, 43(6), 18841894. Dec 2014.
Robert, S. A., Cagney, K. A., & Weden, M. M. (2010). A Tanaka, S. (2005). Parental leave and child health across
life-course approach to the study of neighborhoods OECD countries*. The Economic Journal, 115(501),
and health. In C. Bird, P. Conrad, A. Fremont, & F7F28.
S. Timmermans (Eds.), The handbook of medical soci- Taubman, P. J., & Sickles, R. C. (1983). Supplemental
ology (pp. 124143). Nashville: Vanderbilt University social insurance and the health of the poor (NBER
Press. Working Paper No. 1062). Cambridge, MA:
Roelfs, D. J., Shor, E., Davidson, K. W., & Schwartz, J. E. NBER.
(2011). Losing life and livelihood: A systematic Tefft, N. (2011). Insights on unemployment, unemploy-
review and meta-analysis of unemployment and all- ment insurance, and mental health. Journal of Health
cause mortality. Social Science & Medicine, 72(6), Economics, 30(2), 258264.
840854. Teicher, M. H., Andersen, S. L., Polcari, A., Anderson,
Rossin, M. (2011). The effects of maternity leave on chil- C. M., Navalta, C. P., & Kim, D. M. (2003). The neu-
drens birth and infant health outcomes in the United robiological consequences of early stress and child-
States. Journal of Health Economics, 30(2), 221239. hood maltreatment. Neuroscience & Biobehavioral
Rossin-Slater, M. (2013). WIC in your neighborhood: Reviews, 27(1), 3344.
New evidence on the impacts of geographic access to U.S. Census Bureau. (2014). Health insurance coverage
clinics. Journal of Public Economics, 102, 5169. in the United States: 2013 (Current population reports,
Ruhm, C. J. (2000). Parental leave and child health. P60250). Washington, D.C.: U.S. Government
Journal of Health Economics, 19(6), 931960. Printing Ofce.
Shager, H. M., Schindler, H. S., Magnuson, K. A., Duncan, Umberson, D., Williams, K., Powers, D. A., Liu, H., &
G. J., Yoshikawa, H., & Hart, C. M. (2012). Can Needham, B. (2006). You make me sick: Marital qual-
research design explain variation in head start research ity and health over the life course. Journal of Health
results? A meta-analysis of cognitive and achievement and Social Behavior, 47(1), 116.
outcomes. Educational Evaluation and Policy Wamala, S. P., Lynch, J., & Kaplan, G. A. (2001).
Analysis, 0162373712462453. Womens exposure to early and later life socioeco-
Shager, H. M., Schindler, H. S., Magnuson, K. A., Duncan, nomic disadvantage and coronary heart disease risk:
G. J., Yoshikawa, H., & Hart, C. D. (2013). Can The Stockholm Female Coronary Risk Study.
research design explain variation in head start research International Journal of Epidemiology, 30(2),
results? A meta-analysis of cognitive and achievement 275284.
outcomes. Educational Evaluation and Policy Weber Handwerker, E. A. (2007). Empirical essays in the
Analysis, 35(1), 7695. economics of aging. Doctoral dissertation, University
Sherman, A., Trisi, D., & Parrot, D. (2013). Various sup- of California, Berkeley, 2007. Proquest Dissertations
ports for low-income families reduce poverty and have and Theses, 3275438.
long-term positive effects on families and children. Williams, K., & Umberson, D. (2004). Marital status,
Washington, D.C.: Center on Budget and Policy marital transitions, and health: A gendered life course
Priorities. http://www.cbpp.org/sites/default/les/ perspective. Journal of Health and Social Behavior,
atoms/les/7-30-13pov.pdf. Accessed 14 Nov 2014. 45(1), 8198.
Life Course Risks and Welfare
States Risk Management

Martin Diewald

1 Introduction also to the fact that such specic policies are


embedded within the welfare state agencies that
For many years, the formative inuence of the develop and implement them and that consist of
welfare state on the life course has been subject relatively stable laws, institutions, and cultures
of life course theory and research (see, e.g., (Kaufmann 2012). If they are also to be used to
Mayer and Mller 1986). Indeed, the complex study the life course, such concepts must give
intersections between the life course and welfare equal consideration not only to the direct and
state were important motivations for the collec- immediate inuences of institutions and policies
tion of extensive longitudinal data on a large on people but also to the longer-term temporal
scale. More such data than ever have become patterns of inuences over and above the imme-
available over the last two decades, allowing for diate intended ones. That is, to the side effects on
in-depth cross-national and historical analyses in other life domains, and possible consequences
this eld. But despite the opportunities such data for the signicant others of the individuals to
provide, there is still considerable uncertainty whom policy interventions are aimed. In other
about how best to conceptualize the inuences of words, researchers must take into account the
the welfare state on the life course and how to fundamental principles of the life course approach
make full use of the concepts that life course to be able to study the causal nexus between
research has to offer. states and events over lifetimes, interdependent
Given that there is no overarching life course developments in different life domains, and
theory and that there are a variety of approaches linked lives (see Elder 1995).
to examine welfare state inuences, the question In this essay I use the concepts life course
is, what concepts might be most suitable for elu- risks and risk management by the welfare state
cidating features of welfare states and their con- because they offer a promising approach to study
sequences for the life course? If these concepts the relationship between welfare states and life
are to be used to study welfare states, they must courses in general and the production of social
give due consideration not only to the diversica- inequalities in particular. The life course risk
tion and historical transformation of political concept is not new; rather, it is based on the work
economies that result from specic policies, but of several scholars whose concepts I will develop
further, with a focus on how the life course risk
M. Diewald (*) concept can contribute to our understanding of
Department of Sociology, University of Bielefeld, the mechanisms behind the major impact the
Bielefeld, Germany state has on stratication. I rst introduce the life
e-mail: martin.diewald@uni-bielefeld.de

Springer International Publishing Switzerland 2016 677


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_30
678 M. Diewald

course risk concept and its distinction from life et al. (2009: 425) specically mention the role of
course adversity (Part 2). I will give particular the individuals social, economic, and psycho-
attention to life course concepts that relate to logical resources and the mobilization of these
longer-term impacts on the distribution of life resources in shaping life trajectories (see also
chances. I then discuss approaches that embed ORand 2001; Schoon and Bynner 2003). All of
the risk concept in a broader framework of mech- these resources (e.g., signicant others, like par-
anistic approaches to stratication and relational ents or friends, and organizational supports) may
inequality theory (Part 3). Following this discus- help to prevent risks from turning into adversity.
sion, I introduce risk management as an impor- The individuals agency, likewise, can prevent
tant part of welfare states agency that has not yet exposure to risk and disrupt the association
been given the attention it deserves in life course between risk exposures and actual adversity.
research (Part 4). The same is true for state agency, the states
purposive actions to shape the lives of its citizens.
According to DiPrete (2002: 268) risky life
2 Life Course Risks events do not necessarily lead to adverse living
and Adversity conditions because the possible adverse conse-
quences of risky life events may be mitigated by
The term risk is often used in a rather vague welfare states to a considerable degree. He dem-
and ambiguous manner or is used to dene onstrates the striking differences between wel-
merely statistical relationships, but it can also fare state regimes in shaping risks in terms of the
help us develop theory in life course research incidence of risky life events and of buffers
(Mayer 2009: 424). The terms risk and adver- against their adverse consequences. For example,
sity are often confused and used interchange- some welfare states, like the German one, dis-
ably. Schoon and Bynner (2003), for example, courage a decision to divorce in order to avoid
note that adversity, or risk, can encompass this risky event, but, if it occurs, they fail to com-
genetic, biological, psychological, or socioeco- pensate for adversities such divorces negative
nomic factors that are associated with increased consequences for income. Other welfare states,
risk of maladjustment (Schoon and Bynner like the Swedish one, do not try to discourage
2003: 22). In contrast, Ferraro et al. (2009: 422) divorce, but, if it occurs, they buffer effectively
explicitly advocate a conceptual distinction against negative nancial consequences. Thus,
between risk and adversity, dening risk as both the person and state need to conceptualized
the probability of a hazard or negative event in terms of preventing risks and, should they
occurring in the future, stating that once a risk occur, preventing adversity.
eventuates in a negative outcome, it becomes a A multidisciplinary life course framework
disadvantage [or adversity] (see also ORand acknowledges that causal factors for life course
2003: 694). Events and states that may not be risks and adversity may be the result of specic
perceived as adverse or disadvantageous can developments at several levels over time. The
increase the risk of later adversity. For example, starting point is the fundamental question of how
parental pampering of children is often perceived society affects the individuals who live within its
as comfortable but might at the same time pre- boundaries: Whom does the state favor? To iden-
vent the development of tenacity and necessary tify possible causal nexuses that might create
adaptative skills. Furthermore, adversity may or risks over individuals life courses, I distinguish,
may not create risk for future adversity, depend- for heuristic purposes, among three levels of
ing on the availability of coping resources and development, each with two possible outcomes
successful coping strategies. (see also Diewald et al. 2015):
These considerations explicitly connect risk
and adversity to the agency of individuals, net- (a) The development of favorable or unfavorable
works, organizations, and institutions. Ferraro physiological or mental characteristics,
Life Course Risks and Welfare States Risk Management 679

often, to some degree, as a result of genetic are mixed and do not support assumptions of
variation; continuity as unambiguously as life course
(b) the manifestation of such embodied charac- researchers might wish (see Mayer 2009: 417
teristics in favorable or detrimental decisions 418 for an overview). Obviously, there is also
and behaviors; room for reversion to and compensation for early
(c) the transformation of behavior into unequal risk and adversity, and it appears that the mecha-
life chancesrisk and exposureas a result nisms that cause perpetuation or reversion are
of these characteristics and behaviors. quite specic depending on different antecedents,
outcomes, and times and places. This point is
The denition of risk starts with heterogeneity illustrated by research on unemployment and the
in the genetic propensity and physiological pro- question of why unemployment leads to future
cesses to exhibit certain embodied characteris- unemployment. Heckman and Borjas (1980)
tics that play a role for advantage and observe that a continuation of unemployment
disadvantage. A denition of risk at the level of risk may be due to already existing and enduring
individual development has to be aware that gen- skill or motivation decits that repeatedly lead to
eral skills may be relevant for a number of out- such adversity. Insofar as it is due to such indi-
comes in several life domains and life phases. For vidual heterogeneity, it is unlikely that adversity
example, early self-control appears to be related will be compensated later on but rather will con-
to many outcomes in later life. And contrary to tinue due to the same reasons of skill or motiva-
the conventional view, this understanding of tion decits, leading repeatedly to unfavorable
risk should not be limited to the emergence of behaviors. The emergence of these characteris-
negative characteristics and behavior (e.g., tics themselvesthe rst of the three level risk
aggression, anxiety). Risk also involves low or no concept displayed abovethen becomes a major
propensity to exhibit favorable and protective task in identifying the risk chain over the life
characteristics (Nisbett et al. 2012). Finally, the course. The same is true for the transmission of
blocking of detrimental characteristics and risk across family generations (see chapter
behaviors and the activation of favorable ones is Three Generation Studies: Methodological
conducive to socio-economic attainment. Challenges and Problems by Thornberry, this
In addition to these levels of development, we volume). Thus, what appears to be risk and adver-
must still identify possible additional factors sity may in fact be endogenous to very early
outside the individual that may mediate or developmental patterns. This is not a statistical
moderate these interrelated developments, point but rather a substantive one: a life course
namely on the levels of families, social networks, approach to risk and adversity must begin with
neighborhoods, and wider social contexts such as the developmental emergence of basic behav-
organizations and national and supranational ioural patterns and decision-making resources
societies. For example, genetic risk for deviant that the person will then bring to subsequent
behavior may be counteracted by effective social situations.
control (Shanahan et al. 2008), perhaps by the Accordingly, risk analysis must also involve
family or state interventions (e.g., mentoring taking into account an individual-development
programs). perspective in addition to examining events and
There may be a tendency of risks to generate trajectories alone (see Diewald and Mayer 2009).
more risks and of advantage to create more However, this chain of risk is also due to how
advantage over the life course. Trajectories may, experiences of unemployment alter skills and
however, be altered by the mobilization of indi- motivation and how they lead to unfavorable
vidual and social resources and by the activation behaviors. These can be considerably mitigated
of individual agency (Ferraro et al. 2009). through state agency, avoiding such a downward
However, the empirical results on the long-term spiral. Gangl (2004) showed that higher unem-
inuence of early-life events on later-life events ployment benets prolong the duration of
680 M. Diewald

unemployment but, at the same time, allow negotiated in relationships between people and
workers to sustain their search for adequate jobs, status groups. Diewald and Faist (2011) distin-
which ultimately avoided longer-run scarring guished four major groups of attributes of nego-
effects. Gangl showed that avoiding such scar- tiations that characterize relational inequalities:
ring effects is a major advantage in Germany in (1) ascriptive attributes such as distinguishing
contrast to the US. Thus, policy regimes play a physiological features, gender, age, nationality,
critical role in shaping the individuals develop- and ethnicity; (2) cultural preferences, ways of
mental and life course patterns. life, lifestyles, attitudes, and orientations; (3)
To summarize thus far: Risk and adversity are competencies, qualications, and characteristics
two interrelated but distinct concepts. Whether a that are regarded by a society as legitimate mech-
risk leads to adversity is dependent on interven- anisms for the allocation of opportunities or that
ing forces at the levels of the individual, signi- are at least discussed as such; and (4) the differ-
cant others, social contexts, and the mitigating entiation of activities in the context of social divi-
role of welfare state policies and institutions. sion of labor.
Over the life course, the emergence of risk starts How does the emergence of inequalities out of
with the development of unfavorable characteris- such heterogeneities t into the risk and adversity
tics reecting early biological differences (e.g., framework developed above? This is illustrated
genetic and intrauterine experiences), as precur- in Fig. 1: The risk pathway outlined in the previ-
sor of unfavorable decisions and behaviors, ous section characterizes processes over the life
which nally may lead to adversity. However, course shaped by social mechanisms that create
risk is not only the probability that unfavorable inequalities between such related groupsin our
characteristics, behaviors, and adversities occur case with respect to the emergence of risk and
but also the probability that, against the odds, adversity. The most prominent and quite univer-
genetic propensities to favourable characteristics sal social mechanism found in the literature is
do not unfold, that favourable characteristics do social closure. Social closure means conscious or
not lead to favourable behaviors and, nally, do unconscious attempts of one group to exclude
not lead to advantage. other groups from access to scarce cultural and
material resources, like occupational positions,
income, respect, citizenship rights, welfare state
3 Risks and Relational benets, or education. Social closure comprises
Theories of Stratication mechanisms that are often discussed with differ-
ent labels, like boundary making, stereotyping,
Although the circumstances by which risk occurs stigmatization, or opportunity hoarding (Tilly
might indeed be complex and specic for every 1998; Reskin 2003; Diewald and Faist 2011;
single risk (divorce, unemployment, serious ill- Tomaskovic-Devey 2014).
ness), there continues to be a remarkable gap The basic idea is that relational group mem-
between such life course risk concepts and theo- bership (e.g., low and high SES groups) moder-
ries that could put the risk and adversity generat- ates risk pathways, as shown in the gure. For
ing steps described above into a coherent example, Guo and Stearns (2002) compared the
theoretical framework of stratication. Here, I levels of realized genetic potential for intellectual
propose taking mechanistic explanations of rela- development across different social groups. They
tional inequalities for such a framework. showed that children from disadvantaged social
Relational inequality theory (Tilly 1998; Avent- backgrounds have lower chances to realize their
Holt and Tomaskovic-Devey 2014) deals with the genetic potential than those from a social back-
question of why some people get more respect, ground with more resources and no ethnic dis-
resources, and rewards than others. Relational crimination. In other words, children from
means that inequalities are not characteristics disadvantaged backgrounds are excluded from
situated in people or status positions but are opportunities to develop their genetic endowments.
Life Course Risks and Welfare States Risk Management 681

Fig. 1 Risk, adversity, and relational inequalities

This example shows in addition that advantage quence children from lower class background
and disadvantage are not only a possible conse- obtain less education than those from higher class
quence of risk but can themselves become a risk backgrounds with the same ability and school
for further developments in a longitudinal performance.
perspective. That is, life-time patterns of risk and adversity
A similar inequality in risk exposure can be are shaped by membership in relational groups.
demonstrated for the path from characteristics to Moreover, we have an important starting point
decisions and behavior. According to Breen and from which state agency can be evoked: equal
Goldthorpe (1997) children with the same ability opportunity at all steps in the development from
and the same school performance have different genetic variation to adversity (see Roemer 1998
chances for higher educational tracks dependent for an extensive discussion about how this could
on the social class of their parents. The reason is be operationalized). If we use a risk framework
a relative risk aversion mechanism: Educational as outlined above, the question is not simply
decisions are ultimately motivated by the desire whether or not one can escape from the class or
to avoid downward class mobility, and this moti- socio-economic status of origin but whether
vation is stronger than the desire to reach upward social closure based on unequal distribution of
mobility. For higher class families this means risk exposure, risk compensation, and risk accu-
that children have to reach higher educational mulation contributes to shaping the reproduction
degrees than it is necessary for lower class fami- of social origin or downward and upward mobil-
lies. In other words, the motivation for higher ity, respectively.
educational degrees is stronger in higher than in Schoon (2006) has shown that socio-economic
lower classes where the costs of pursuing further adversity in the parental home is a factor that per-
education outweigh the utility of acquiring more sistently presents signicant risks for a variety of
education earlier in the school career. In conse- outcomes over at least the early life course, with
682 M. Diewald

the term persistent referring not only to lifetime the introduction of paternity pay raises parental
but also to historical time, meaning that favorable leave for men. This strand of research often falls
economic conditions did not diminish social ori- short in covering longer life course trajectories,
gin inequalities in risk and adversity. instead concentrating on immediate effects. The
The ubiquitous and astonishingly robust other tradition sees the state as an overarching
impact of social origin on risk development has agency that covers the entire life course from
to be explained better than has been done in exist- birth to death. In this view, the governments of
ing research. One possible explanation is that different societies shape different life course pat-
mechanisms considered to be specic are more terns, which correspond to different holistic wel-
general than has been assumed. To give an exam- fare state regimes or political economies. This
ple, Cardona and Diewald (2014) found that par- idea assumes that governmental institutions and
ents obviously followed similar considerations agency are shaped by an identiable overall logic
not only in school decisions but also in support- that shapes the entire life courses. However, as
ing their childrens skill development: children Mayer (2005: 33), a major proponent of such
from lower classes are promoted only if they dis- holistic ideas himself notes, such abstract types
play high skills, both in the case of skill produc- are ideal types that may be too reductionist to
tion at home and in pursuing higher educational encompass the more differentiated actual regula-
tracks. Parents support for children from higher tions in different subsystems and life domains.
classes is in both cases signicantly less depen- The constructs of both the life course and the
dent on childrens skills and school performance. welfare state are too complex for a uniform cor-
In other words, the same mechanism leads repeat- respondence to be likely (Leisering 2003: 215).
edly to similar inequalities. Even more elaborated typologies (Mayer
To summarize: Integrating relational inequal- 2005: 33) cannot take into consideration the
ity theory in research on risk pathways over the important and inuential differences between
life course makes evident that these risk path- nations and policies within such country clusters
ways are not the same for different groups in the and the changing policies that do not follow a
society, like women and men and people from simple path dependence of established institu-
different social backgrounds. Relational inequal- tions in the face of new paradigms and new risks.
ity theory provides a rich and growing body of The liberal Agenda 2010 launched by the Social
research on social mechanisms that create risk Democrat government in Germany may serve as
pathways for different groups. Especially vari- a prominent example. Although status mainte-
ants of social closure are important to link life nance is a core principle of conservative welfare
course research on risk and adversity to social state regimes like Germany, the Agenda 2010
stratication, social justice considerations, and explicitly restricted status maintenance as a guid-
welfare state intervention. ing principle, for example, in the case of unem-
ployment benets, which proved to be a major
distinction between Germany and the US in
4 Welfare States Risk Gangls study (discussed above) before the
Management Agenda 2010 came into being. Nevertheless,
the idea of overlapping guiding principles should
Existing research in the eld of the state and the not be hastily abandoned.
life course reects two different streams. Social In historical perspective, the welfare state can
policy analysis concentrates on specicities of be seen as an answer to emerging risks of the
the institutional design or single, specic mea- industrial society. In different countries different
sures. For the most part, this research centers on aspects of these risks were identied as the
incentives and disincentives that are set by central theme, and differences in the setting of
policies, and the direct consequences they bear priorities also led to different institutional
for citizens, such as whether and to what degree designs. Whereas in Germany the industrial
Life Course Risks and Welfare States Risk Management 683

working class was at the center of social political ers. Making reference to Fig. 1, risk management
debates, in Great Britain poverty was the central systems can be distinguished in three respects
social question, in Sweden it was equality, and in concerning what and whom they address: rstly,
France it was family and fertility. Also the more whether they focus more on risks or more on
recent discussion about new risks emerging and adversity; secondly, which type of risk or adver-
challenging the existing welfare state institutions sity is addressed; and thirdly, which relational
focuses on developments in the labor market and groups are in the foreground when risks are
the family system. New risks in the labor market addressed. At all three levels, it is important to
reect the need for more exibility in the work- note that we should not only look at what is
force, resulting in job insecurities and the insuf- addressed but also what is not addressedas the
cient inclusion of the workforce in the labor negation of the relevance of risks or relational
market. New risks in the sphere of private life are groups (Leisering 2003: 217).
disruptive family events that threaten the role of To make the rst distinction of whether wel-
the family as a buffer against market failure fare states prefer to address the emergence of risk
(Esping-Andersen 1999). Again, countries differ or the buffering of adversity, DiPrete (2002) and
in what risks are given priority and what institu- Mayer (2005) describe risk management as a pre-
tional lters they choose to use to protect groups vention strategy that in the rst case inuences
against these risks (Hofcker et al. 2010). It is decisions and behavior with incentives or disin-
evident that a lot of attention in discussions about centives with the aim that risks should be avoided
the future of the welfare state is devoted to the (step 2 in Fig. 1). In the second case, we can
management of risk. speak of an adversity-oriented, mostly insurance-
Leisering (2003) proposed risk management based compensation strategy. We learn from
as an important part of welfare states agency DiPretes (2002) comparison among the US,
alongside education and old-age pensions. He Germany, and Sweden that Germany in general
concedes that this part has been largely neglected stands out for the rst strategy: risky decisions
by life course researchers (p. 213). According to are discouraged by the state and risky behaviors
Leiserings conceptualization, risk management should be avoided. In contrast, Sweden stands
comprises all measures of social assistance, out for buffering adversity while not trying to
social insurance, and social services meant to prevent risk. The US is often low on both dimen-
bridge lifes discontinuities and transitions. To do sions. Neither are risky events suppressed, nor
so, agencies of risk management react mainly to are negative consequences effectively buffered.
risky events in a short-term perspective, as situa- However, as the example of worker displacement
tional programs. Thus, they constitute only a par- shows, these distinctions are not completely valid
ticular part of comprehensive welfare state for any risk. The risk of worker displacement is
regimes. They are tied to specic institutions, very similar in all three societies, and only for
whereas education and the old-age pension sys- buffering negative consequences in the form of
tem form another part of governmental policy. poverty the expected pattern was found.
Nevertheless, risk management is said to shape From DiPretes (2002) comparison of life
the expectations of the citizens and thus secure course risks in different welfare state regimes,
the unity of the life course as a whole (ibid.). another component of risk management which is
How can risk management be eshed out as more on the side of risk orientation than adversity
a useful concept to capture inuences of the wel- orientation can be borrowed (p. 278): the extent
fare state on life course inequalities over and to which the opportunity for rapid recovery from
above specifying the single assistances, insur- adversity provided by counter-mobility events
ances, and services? As outlined above, cultural is institutionally facilitated or supported by
foundations of different welfare states set differ- specic policies such as reemployment after
ent preferences about which risks are perceived unemployment, upward occupational mobility
as salient and which are given priority over oth- after downward moves, or remarriage after
684 M. Diewald

family dissolution. The strong links between personal capabilities instead of simply assuming
education, training, and employment as well as that the person possesses them. In interdisciplin-
the high level of job stability and security in the ary life course research, the role of individual
German welfare state normally provide an excep- development in shaping life chances, the possible
tionally high level of predictability and continu- lifelong impact of early life phases, and the dis-
ity over the life course. However, if risk occurs, it covery of skill production (see Cunha et al.
may be especially difcult to correct for it and 2010) became driving forces in the production of
consequently, in such cases, lead to more down- life course theory and data. However, these devel-
ward mobility than is seen in less continuity-ori- opments have not yet found their way into com-
ented welfare states. parative welfare state research and the question
Such a longitudinal perspective focusing on of how risks of failure in skill development and
the duration of risk and possible interruptions of chances to block unwanted characteristics are
vicious circles is extremely important. As Schoon distributed across societies owing to welfare state
and Bynner (2003) conclude from existing stud- agency.
ies, a high-risk individual is characterized not so The discussion about the role of early indi-
much by exposure to a single situational risk fac- vidual development leads us immediately to the
tor but rather by a life history characterized by third question: whom a specic risk management
long durations of risk exposure or multiple favors or disadvantages, thus producing rela-
disadvantages. tional inequalities. If the role of general skills for
The rst step in risk development can be seen attainment is growing, then the question of
as a variant of the prevention strategy located ear- whether these competences are produced in the
lier in the chain of risks, which focuses on indi- systems of education and training or left to the
vidual development instead of behavioral initiative and ability of the parental family is of
incentives. Though the development of risky and paramount interest for the intergenerational
favorable characteristics has largely been transmission of advantage and disadvantage (see
neglected in comparative welfare state research, chapter Three Generation Studies:
it can be easily rooted in early work on the emer- Methodological Challenges and Problems by
gence of the modern welfare state. Mayer and Thornberry, this volume). The growing signi-
Mller (1986) in particular have argued that the cance of such competencies, which is not taken
modern welfare state, while regulating the indi- up by a corresponding governmental agency to
viduals decisions, has to address the responsibil- produce them within the formal systems of edu-
ity of the individual self. A rst question relates cation and training, may help to explain why the
to the importance of skill development and the direct link between social origin and destination
avoidance of adverse characteristics for the emer- is so persistent even if the transmission pathway
gence of risks. Over the last few decades, the dis- over education is controlled for (Jackson et al.
cussion about new risks and a redesign of the 2005). In other words, welfare states that focus
welfare state has fueled debates about the ever- on equalizing educational opportunities alone
growing importance of self-responsibility and may fail to reach their goal of an open society.
other cognitive and noncognitive skills for mini- The failure of governments to focus on general
mizing risk and maximizing personal opportuni- skills is predominantly a risk for the lower classes
ties (see, e.g., ORand 2001). Yet substantiated and the unemployed who do dispose these
empirical evidence is still missing, as it is for resources much less than higher social classes. In
international comparisons between welfare states other words it is a mechanism of social closure in
with respect rstly to the degree that different keeping considerable parts of the population
skills matter, and secondly, to which policies are away from the possibility of acquiring character-
used to produce these skills. istics relevant for achievement, and this is due to
The second question is how and how much the circumstances for which the disadvantaged can-
government invests in the production of these not be held responsible.
Life Course Risks and Welfare States Risk Management 685

Relational inequalities are not only created management systems address possible social and
with respect to skill development but also when emotional consequences even though they might
the welfare state shapes decisions and behaviors. impinge severely on the development of children
This becomes especially visible in the eld of in the household, which later on might inuence
work-life interference and the division of labor their mobility chances.
between the sexes. The welfare state agency is
formulated to be gender-neutral but in fact
ignores the existing gender differences in the 5 Summary and Outlook
division of labor. Thus, even policies designed to
support mothers employment prospects may in Taking together the evidence presented here about
fact undermine them. Instruments of work-fam- differences in welfare states risk management
ily reconciliation may even aggravate employer calls into question whether it is appropriate to
discrimination against women insofar as they limit this concept to a distinctive part of welfare
erode their attractiveness as employees and institutions that target adversity rather than risk.
restrict womens career mobility (Mandel and In contrast, I advocate a broader view on all poli-
Semyonov 2005). cies that inuence the life course for better or
A more complicated pattern arises if sex and worse. Examples were presented to demonstrate
class are jointly taken into consideration. It that different welfare state regimes and policies
becomes clear that, in addition, the inuence of can be differentiated along three lines: (1) whether
different family and labor market policies on they focus more on avoiding risk or on buffering
womens employment is markedly dependent on adversity; (2) which specic risks or adversities
their class position (Mandel and Shaley 2009): are focused on; and (3) which relational groups
whereas maternity leave policies and childcare are addressed and then favored or disadvantaged
facilities shelter lower-class women from with respect to risk. Differences between risk
employment risks, the same policies are rather a management systems based on these three lines of
handicap for women who compete with men for distinction form patterns that are greater than the
higher-class jobs, because they are perceived as impact of single policies. At the same time, they
risky and vulnerable instead of powerful. In addi- are more comprehensive than those suggested by
tion, extended maternity leaves and related poli- DiPrete (2002) insofar as they include a usually
cies are not as important because these women neglected part of risk emergence, namely skill
are, especially in liberal market societies, able to production, which is increasingly advocated as an
purchase necessary services in a low-wage labor efcient risk avoidance strategy.
market. This plea is backed by advances in life course
Finally, at the most basic level, risk manage- research stressing the pivotal role of cognitive
ment systems start with the question of who and noncognitive skills for the understanding of
should have access to (full) citizenship at all in risk chains, their emergence early in the life
the context of international migration. With the course, and their conceptual signicance for
increasing share of different groups of migrants understanding the emergence of new individual-
within the population, the legal status that should ized risks and corresponding ideas for a redesign
be granted to them and the degree to which wel- of the welfare state agenda. Though there are sev-
fare benet entitlements compared to natives dif- eral new studies that look at skill development in
fers and the maintenance of such is discussed an internationally comparative perspective, the
even in universalistic welfare state regimes. integration into a more comprehensive risk man-
Disruptive employment and family events agement framework is missing. Thus, a fuller
bear the risk of not only socioeconomic adversity account of life course concepts and general the-
but also social and emotional stress (see, e.g., ory of stratications seems more fruitful than the
Kalil and Wightman 2011). There is, however, no view on institutional differentiation for studying
comparative research to show how different risk welfare states risk management.
686 M. Diewald

References Heckman, J. J., & Borjas, G. (1980). Does unemployment


cause future unemployment? Denitions, questions,
and answers from a continuous time model of heteroge-
Avent-Holt, D., & Tomaskovic-Devey, D. (2014). A
neity and state dependence. Economica, 47, 247283.
Relational Theory of Earnings Inequality. American
Hofcker, D., Buchholz, S., & Blossfeld, H. (2010).
Behavioral Scientist, 58, 379-399.
Globalization, institutional lters and changing life
Breen, R., & Goldthorpe, J. H. (1997). Explaining educa-
course patterns in modern societies: A summary of the
tional differentials: Towards a formal rational action
results from the GLOBALIFE project. In R. Silbereisen
theory. Rationality and Society, 9(3), 275305.
& X. Chen (Eds.), Social change and human develop-
Cardona, A., & Diewald, M. (2014, May). Opening the
ment: Concept and results (pp. 101125). London:
black box of primary effects: Relative risk aversion
SAGE. http://dx.doi.org/10.4135/9781446252161.n5
and maternal time investments in preschool children.
Jackson, M., Goldthorpe, J. H., & Mills, C. (2005).
Paper presented at the spring meeting of the RC 28 in
Education, employers and class mobility. Research in
Budapest.
Social Stratication and Mobility, 23, 333.
Cunha, F., Heckman, J. J., & Schennach, S. M. (2010).
Kalil, A., & Wightman, P. (2011). Parental job loss and
Estimating the technology of cognitive and noncogni-
childrens educational attainment in black and white
tive skill formation. Econometrica, 78(3), 883931.
middle-class families. Social Science Quarterly, 92,
Diewald, M., & Faist, T. (2011). From heterogeneities to
5778.
inequalities: Looking at social mechanisms as an
Kaufmann, F.-X. (2012). European foundations of the
explanatory approach to the generation of social
welfare state. New York: Berghahn Books.
inequalities. Bielefeld: SFB 882 Working Paper Series
Leisering, L. (2003). Government and the life course. In
No. 1.
J. T. Mortimer & M. J. Shanahan (Eds.), Handbook of
Diewald, M., & Mayer, K. U. (2009). The sociology of the
the life course (pp. 205225). New York: Kluwer
life course and life span psychology: Integrated para-
Academic/Plenum Publishers.
digm or complementing pathways? Advances in Life
Mandel, H., & Semyonov, M. (2005). Family policies,
Course Research, 14(12), 514.
wage structures, and gender gaps: Sources of earnings
Diewald, M., Baier, T., Schulz, W., & Schunck, R. (2015).
inequality in 20 countries. American Sociological
Status attainment and social mobility How can genet-
Review, 70, 949967.
ics contribute to an understanding of their causes? In
Mandel, H., & Shalev, M. (2009). Gender, class, and vari-
K. Hank & M. Kreyenfeld (Eds.), Demographie.
eties of capitalism. Social Politics, 16, 161181.
Klner Zeitschrift fr Soziologie und
doi:10.1093/sp/jxp006.
Sozialpsychologie, Special issue 55. doi: 10.1007/
Mayer, K. U. (2005). Life courses and life chances in a
s11577-015-0317-6
comparative perspective. In S. Svallfors (Ed.),
DiPrete, T. A. (2002). Life course risks, mobility regimes,
Analyzing inequality: Life chances and social mobility
and mobility consequences: A comparison of Sweden,
in comparative perspective (pp. 1755). Palo Alto:
Germany, and the United States. American Journal of
Stanford University Press.
Sociology, 108, 267309.
Mayer, K. U. (2009). New directions in life course
Elder, G. H., Jr. (1995). The life course paradigm: Social
research. Annual Review of Sociology, 35, 413433.
change and individual development. In P. Moen, G. H.
doi:10.1146/annurev.soc.34.040507.134619.
Elder, & K. Lscher (Eds.), Examining lives in con-
Mayer, K. U., & Mller, W. (1986). The state and the
text. Perspectives on the ecology of human develop-
structure of the life course. In A. B. Sorensen, F. E.
ment (pp. 101139). Washington, DC: APA.
Weinert, & L. R. Sherrod (Eds.), Human development
doi:10.1037/10176-003.
and the life course: Multidisciplinary perspectives
Esping-Andersen, G. (1999). Social foundations of
(pp. 217245). Hillsdale: Lawrence Erlbaum
postindustrial economies. Oxford: Oxford University
Associates Publishers.
Press.
Nisbett, R. E., Aronson, J., Blair, C., Dickens, W., Flynn,
Ferraro, K. F., Shippee, T. P., & Schafer, M. H. (2009).
J., Halpern, D. F., & Turkheimer, E. (2012).
Cumulative inequality theory for research on aging
Intelligence: New ndings and theoretical develop-
and the life course. In V. L. Bengtson, D. Gans, N. M.
ments. American Psychologist, 67, 130159.
Putney, & M. Silverstein (Eds.), Handbook of theories
ORand, A. M. (2001). Stratication and the life course:
of aging (pp. 413434). New York: Springer.
Life course capital, life course risks, and social
Gangl, M. (2004). Welfare states and the scar effects of
inequality. In R. H. Binstock & L. K. George (Eds.),
unemployment: A comparative analysis of the United
Handbook of aging and social sciences (pp. 145162).
States and West Germany. American Journal of
San Diego: Elsevier.
Sociology, 109(6), 13191364.
ORand, A. M. (2003). The future of the life course: Late
Guo, G., & Stearns, E. (2002). The social inuences on
modernity and life course risks. In J. T. Mortimer &
the realization of genetic potential for intellectual
M. Shanahan (Eds.), Handbook of the life course
development. Social Forces, 80, 881910.
(pp. 693701). New York: Plenum.
Life Course Risks and Welfare States Risk Management 687

Reskin, B. (2003). Including mechanisms in our models Shanahan, M. J., Vaisey, S., Erickson, L. D., & Smolen, A.
of ascriptive inequality. American Sociological (2008). Environmental contingencies and genetic pro-
Review, 68, 121. pensities: Social capital, educational continuation, and
Roemer, J. E. (1998). Equality of opportunity. Cambridge, dopamine receptor gene DRD2. American Journal of
MA: Harvard University Press. Sociology, 114(Suppl), S260S286.
Schoon, I. (2006). Risk and resilience: Adaptations in chang- Tilly, C. (1998). Durable inequality. Berkeley: University
ing times. Cambridge: Cambridge University Press. of California Press.
Schoon, I., & Bynner, J. (2003). Risk and resilience in the Tomaskovic-Devey, D. (2014). The relational genera-
life course: Implications for interventions and social tion of workplace inequalities. Social Currents, 1,
policies. Journal of Youth Studies, 6(1), 2131. 5173.
Longitudinal Studies and Policy
for Children in Low- and Middle-
Income Countries: Evidence
from Young Lives

Paul Dornan

1 Introduction cohort study, set up to extend understandings of


childhood poverty in low- and middle-income
While cohort and panel studies have come to be countries. Low- and middle-income countries are
an established element of the social science infra- somewhat analogous to the UN Population Fund
structure of OECD countries, such data and life classication of less developed countries2 which
course approaches are much less common in low- covers 6.0 of the worlds 7.2 billion people
and middle-income countries. Yet such countries (UNPFA 2014). The grouping of low- and mid-
face many development challenges that longitu- dle-income countries is therefore broad and used
dinal approaches can help illuminate. There is here to identify countries which typically have
growing global interest in life course approaches less established data sources, lower access to sets
and these can help inform policy towards the of cohort or panel data, and research capacities to
achievement of the global Millennium create and use such evidence to inform policy.
Development Goals (MDGs) and the proposed The chapter is structured as follows. First, the
Sustainable Development Goal commitments.1 chapter identies inequalities between groups of
The aim of this chapter is rst to reect on children within countries, with socio-economic
policy implications from the Young Lives cohort inequalities clear between groups of children at
study, particularly for mitigating inequalities and, very young ages while inequalities by gender
second, to discuss the relationship between widen particularly through adolescence. Second,
research and policy. Young Lives is a four-country while inequalities emerge early there is also evi-
dence of change in individual childrens develop-
mental trajectories after the earliest period of life,
1
Following the Millennium Declaration, the worlds gov-
highlighting the combined importance of early
ernments agreed to a series of Minimum Development
Goals to improve human development by 2015 (includ- intervention and sustaining such investments.
ing, for example reducing extreme poverty, reducing pre- Third, the chapter identies interdependencies
ventable child mortality, and increasing school between domains of childrens development,
enrollment). MDG progress is reported by United Nations
whereby earlier good physical or cognitive test
(2014). At the time of writing in 2015, new Sustainable
Development Goals are being proposed, which will both
address the human development agenda, and combine this 2
The UN Population Fund use the term less developed to
with a stronger focus on sustainability.
encompass African countries, most of Asia (excepting
P. Dornan (*) Japan), Latin America and the Caribbean, plus Melanesia,
Department of International Development, University Micronesia and Polynesia (see UNPFA 2014). It is there-
of Oxford, 3 Manseld Road, Oxford OX1 3TB, UK fore somewhat analogous to the low- and middle- income
e-mail: paul.dornan@qeh.ox.ac.uk categorisation used by the World Bank.

Springer International Publishing Switzerland 2016 689


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_31
690 P. Dornan

scores are associated with later effects on childrens


development. The chapter then steps back from Box 1: Ethiopia, India, Peru and Vietnam:
research ndings, to discuss the research-to-policy National Stories with Global Resonances
process, identifying the contribution that life course Different national contexts frame experi-
studies can make in supporting better policy devel- ences of children, but there are also marked
opment, and some of the challenges in achieving similarities among these countries, includ-
that objective. The chapter concludes with key ing the rising importance of schooling,
implications both for policy and for the use of lon- improving infrastructure, but ongoing eco-
gitudinal evidence to inform policy development. nomic and social vulnerability.
Ethiopia. Recent decades have seen
economic growth, improvements in infra-
2 About the Young Lives Study structure, urbanization, and rising popula-
tion. The country has implemented rapidly
The Young Lives study has collected data on chil- increasing school enrollment but learning
dren, their households and communities in levels are often low. Despite social change,
Ethiopia, the former State of Andhra Pradesh3 in child marriage and female genital mutila-
India, Peru and Vietnam since 2002. The World tion still persist in some regions. Famine
Bank classies Ethiopia as a low-income coun- and food insecurity linked to drought has
try, India and Vietnam as lower-middle-income been a key concern, however in the past
countries and Peru as an upper-middle-income decade the Government implemented the
country.4 These countries demonstrate a range of Productive Safety Net Programme, which
historical experiences and current development provides support in return for work in food
pressures (see Box 1) and provide a microcosm insecure areas.
of circumstances faced in many developing coun- India. India has seen a rapidly rising
tries. In recent decades, many low- and middle- GDP per capita in recent years, but faces
income countries have been going through rapid the particular challenge of spreading eco-
technological, economic and social change for nomic and technological advance across
example with children spending longer periods in the population. The study sites are in South
school, later entry to the labor market, and later India. India faces concerning levels of
marriage and child bearing (Lloyd 2005; see also exclusion of women and marginalized
UNGA 2014). The individual experiences high- social groups. India has well-established
lighted in this chapter are being experienced public policies, including the pre-school
within a context of rapid social change. Integrated Child Development Service, and
The cohort study collected quantitative and Public Distribution System. In recent years
qualitative information from children, household India implemented key large-scale social
members, and in the school and community. The policy reforms such as the Right to
study was set up with a primary aim to under- Education Act and Mahatma Gandhi
stand the causes and consequences of child pov- National Rural Employment scheme,
erty and collects multi-purpose data on a range of which extends social protection in rural
different domains of life, including household areas.
poverty and socio-economic position, service Peru. Peru has the highest GDP per
capita of the four countries. The Shining
3
The State of Andhra Pradesh, divided in June 2014 into Path insurgency severely affected the coun-
the States of Andhra Pradesh, and Telangana. Since the try through the 1980s. The country contin-
data used here was collected prior to the division the old
ues to see high economic and social
State name of Andhra Pradesh is used.
4 inequality, including between Spanish-
Using gures from 2013, classications are available at
http://data.worldbank.org/about/country-and-lending-
groups#Low_income (continued)
Longitudinal Studies and Policy for Children in Low- and Middle-Income Countries 691

structed to be statistically representative of the


Box 1 (continued) country. The strength of the data therefore is to
speaking and indigenous communities in enable study within each country rst, and then
different parts of this diverse country. Key to consider similarities and differences in what
social policy interventions include the seems to matter for children at a country level
Juntos conditional cash transfer scheme across the study sites.
and pre-school interventions. However
there are concerns over the effectiveness of
schooling and other social policies in 3 How and When Inequalities
reducing inequalities. Are Formed Through
Vietnam graduated to middle-income Childhood
status after rapid progress following the
devastation of wars in the 1970s and before. In this rst section, the chapter begins to unpack
Social and economic progress has been fast life course evidence on how inequalities between
and comparatively equitable, with social children are established. At a global level, while
policy designed to reduce inequalities there has been considerable recognition of the
between groups. Disparities remain differences between countries, increasing atten-
between the marginalized and poorer eth- tion is being drawn to within country differ-
nic minority populations, and the more ences. A powerful criticism made of the
advantaged majority Kinh. Economic and Millennium Development Goals is that by typi-
infrastructure development has been fast cally measuring average progress in develop-
especially in urban areas, but in line with ment indicators, the MDGs encouraged a focus
the other countries, there are big differ- on where the quickest gains could be made,
ences between urban and rural areas. As a rather than on those most in need. Accordingly
lower-middle-income country, Vietnam faces this concern has motivated a focus on equity,
the challenge of maintaining economic notably from the United Nations Childrens
growth towards industrialization, with Fund (UNICEF 2010). Concerns over inequali-
developing the skill base of the country. ties in a range of social and economic indicators
have been picked up in debates about the pro-
posed global Sustainable Development Goals
(SDGs) and are supported by wider recognition
access, and indicators of childrens physical of economic inequality (e.g. Pickety 2013;
development, cognitive development and subjec- Ostry et al. 2014) and gender inequality (UN
tive wellbeing. The study design is summarized Statistics Division and UN Women 2014). Life
in Fig. 1, with further technical and sampling course analysis can identify when differences
detail available on the study website.5 emerge as children age and so help to identify
The four country design is an important fea- opportunities to mitigate such differences as
ture of Young Lives, since the use of similar early as possible. So what does life course anal-
instruments at the same age points across con- ysis suggest for when and how inequalities
texts enables us to test the broader relevance of between groups of children develop?
results by considering the similarity or difference Across the four countries there are relatively
of ndings across the four countries. The samples consistent patterns that emerge as to the systemic
are broadly representative of the population differences between social groups within coun-
groups (for example by urban/ rural location and tries. Across the four countries, household expen-
by ethnicity) within each country but are not con- diture or wealth levels, and access to basic
services (sanitation, water and electricity) tend to
5
www.younglives.org.uk be lower in rural areas, for those with less
692 P. Dornan

Fig. 1 Young lives study design

educated parents, and for those from disadvan- cussed in Sect. 4) than the least poor third
taged minority groups (with such groups often (Dornan and Pells 2014: p. 7). Differences
comprising the same children) (Cueto et al. 2011; between social groups in height trajectories per-
Galab et al. 2011; Le et al. 2011; Woldehanna sist through childhood (see Sect. 4). Inequalities
et al. 2011). Children from rural and poorer back- by socio-economic status in childrens perfor-
grounds also tend to have lower access to key ser- mance on simple cognitive tests are evident at
vices, including pre-school (Woodhead et al. 5 years, the rst time test data are available, and
2009). Households in rural communities also typically before children have started school.
tend to experience more risks such as drought or Such socio-economic related gaps tend to widen
ooding (Woodhead et al. 2013a, p. 13). The rst within each country between 5 and 8 years of
message to emerge is therefore how social and age (Woodhead et al. 2013a: p. 17). Differences
policy inputs to childrens development vary in physical development and cognition therefore
sharply between groups. appear very early in life, reinforcing the central
Inequalities in childrens outcomes by socio- importance of the early years to strong founda-
economic status, ethnicity, and location appear tions for later development.
while children are very young. In infancy, child While inequalities by socio-economic status
development inequalities can be measured using are somewhat similar across the four countries,
information on childrens physical development. there are important policy differences which
Examining such measures shows that across the affect learning during the school years. There is a
four countries at age 1, the poorest third6 were general trend across developing countries of ris-
between 1.7 (Ethiopia) and 3.1 times (Peru) more ing school enrollment. UNESCO estimates an
likely to be short for their age (i.e., stunted, dis- average 90 % net enrollment in primary school in
2011. This represents a generational shift for
6 many countries and has been particularly dra-
Ranked by household wealth using a measure which
summarises housing conditions, key service access and matic in Ethiopia where enrollment increased
ownership of consumer durables. from 37 % in 1999 to 87 % in 2011 (UNESCO
Longitudinal Studies and Policy for Children in Low- and Middle-Income Countries 693

2014: pp. 388390). However, within overall et al. 2013b). By contrast there is more encourag-
increased enrollment, important questions remain ing evidence of the effects of some schools in nar-
over differences in the effectiveness of education rowing existing test score gaps between ethnic
policies across countries. groups in Vietnam during primary school, linked
In the samples from Ethiopia and Andhra with a focus on all children reaching minimum
Pradesh, on average children tend to score more standards (Glewwe et al. 2014; Rolleston et al.
poorly on cognitive tests than in Peru and Vietnam 2013). Therefore, while education is a key policy
(Singh 2014; Rolleston 2014). Comparing the instrument to extend opportunities, performance
gain in cognitive test performance, while differ- varies across countries both in the level of learn-
ences between countries exist before school age, ing produced and in whether school systems rein-
the greatest widening between countries occurs force or disrupt existing social differences.
during the school years (where Vietnam pulls Through adolescence, school remains impor-
ahead), and seems linked with the different pro- tant but socio-economic pressures increase as
ductivity of school systems (Singh 2014). There families need for children to work grows, which
are also differences in how the school system then can affect when children stop schooling.
affects existing inequalities within the countries. Figure 2 examines socio-economic gaps in school
In Peru analysis links early socio-economic status enrollment rates between the poorest and least
with later opportunities to learn, evidence that the poor third of households. In Ethiopia at age 8 the
schooling system can reinforce existing inequali- large gap exists because poorer children typically
ties (Cueto et al. 2014). Similarly in Andhra enrolled later and the enrollment gap then reduces
Pradesh, increasing use of private schools in India by age 12 as children have entered the schooling
has been associated with widening inequalities by system. Across the four countries the gap widens
socio-economic status and gender (see Woodhead at older ages as poorer children leave school

Fig. 2 Education enrollment gaps by household socio- least poor third of households. The chart deducts the
economic status and age (Note: education covers pre- reported enrollment rate of the poorest third from the least
school, formal school and later vocational forms of poor third to show the difference in enrollment rate
education and training. Socio-economic status is mea- (Dornan and Pells 2014))
sured by household wealth, comparing the poorest and
694 P. Dornan

Fig. 3 Education enrollment gaps by gender and age deducts the reported enrollment rate of the girls and young
(Note: education covers pre-school, formal school and women from boys and young men to show the difference
later vocational forms of education and training. The chart in enrollment rate (Dornan and Pells 2014))

sooner. Gender differences, shown in Fig. 3, tend as girls graduate to secondary schools and as a
to be comparatively small during early and barrier pushing girls from the education system
middle childhood in Young Lives data but widen (Dornan and Pells 2014: p. 12). Also notable at
during adolescence with the pattern varying this older age and as children approach adulthood
across the countries (Dercon and Singh 2011). As is the growing importance of social reputations,
Fig. 3 demonstrates, the direction of bias in particularly for marriageability and how this
enrollment by gender varies between the coun- shapes decision-making by and for boys and par-
tries, underscoring the importance of understand- ticularly for girls. In Ethiopia for example, such
ing different life course trajectories within the future facing concerns affect whether young
country context. women undergo female genital mutilation
Increasing differences in trajectories of school (Boyden et al. 2013). In India, girls need to have
access associated with both socio-economic sta- learned the skills and retained the social reputa-
tus and gender are shaped, and punctuated, by tion to be a good wife (Morrow 2013). Social
key transitions. Identifying the transitions associ- pressures in adolescence add to disadvantages
ated with widening inequalities may provide associated with socio-economic status and pro-
intervention options for policy. Trajectories asso- duce differences by gender in young peoples
ciated with socio-economic status are frequently trajectories.
marked by transitions to work (for instance fol- In summary, this section has sought to identify
lowing family illness) as well as by exam failure general patterns of inequalities between social
pushing children out of school, each against a groups within countries, whilst recognizing
background of likely rising economic returns to wider differences between countries. Three broad
work with age, for boys particularly. Concerns of messages emerge. First, inequalities in childrens
sexual violence particularly facing older girls at, physical development and cognition are evident
or travelling to, school, were frequently reported from early ages by location, socio-economic sta-
Longitudinal Studies and Policy for Children in Low- and Middle-Income Countries 695

tus, parental education and ethnicity/caste status. Height-for-age averages across the Young
Second, school can have a moderating inuence Lives samples are comparatively low at -1 HAZ
on pre-existing inequalities, but in practice may across the countries between 1 and 8 years
well reinforce existing gaps, with different oppor- (Lundeen et al. 2013). This nding is consistent
tunities to learn. Schooling is marked by institu- with deprivations experienced in many low- and
tional gateways (such as transitions between middle-income countries, where lower food
schools or exams) that affect the chances of con- quantity and quality combine with greater expo-
tinued engagement. Third, across the countries sure of young children to disease, resulting in
differences by gender emerge later, widening high rates of under-nutrition. Stunting provides a
particularly as children near adulthood. The way summary measure of HAZ (dened as a child
in which gender inequities emerge is very depen- having a HAZ below 2). Across the four coun-
dent on the society in which young people are tries, lower wealth levels are consistently associ-
growing up, with opportunities shaped by social ated with higher stunting rates (Petrou and Kupek
norms and expectations of roles in adulthood. 2010). Figure 4 presents height-for-age in
Ethiopia for children at age 8, disaggregated by
poverty, location and gender. Zero on the chart is
4 The Extent of Resilience the reference population norm, and 2 the deni-
and Faltering in Physical tion of stunting. Average height-for-age are con-
Development Beyond sistently well below zero for all groups, but
Infancy particularly so for the poorest fth of children
and those in rural areas.
The previous section identies processes by Low height-for-age early in life has been asso-
which inequalities accumulate with age. A key ciated with a range of negative outcomes for chil-
benet of longitudinal data, with repeat observa- drens development, including increased chance
tions, is also to be able to test the extent of change of mortality, poor health, brain and cognitive
within the life course. To examine this, we con- development (UNICEF 2013). Analysis of the
sider the specic example of childrens height Young Lives cohorts has also linked stunting in
trajectories. Consensus has emerged in recent infancy with negative impacts on childrens cog-
years about the importance of the early stage of nition at 5 years (Le 2011; Sanchez 2009), and
childrens lives to later life (Grantham-McGregor demonstrated some evidence of a link between
et al. 2007). It has also been argued that early height in middle childhood and later aspirations,
nutritional deprivation during the rst 1,000 days self-efcacy and esteem in adolescence (Dercon
after conception leads to irreversible conse- and Sanchez 2013). The association between
quences for later life (UNICEF 2013). In low- socio-economic background and the chances of
and middle-income countries low height provides having a low height-for-age, and the associations
an important proxy for chronic under-nutrition. between height-for-age/stunting and childrens
Height is measured as height-for-age, i.e. con- wider development, highlight this as a key mech-
verted to express how far (in standard deviations) anism by which early inequalities are reinforced.
the childs height is from the reference popula- Therefore, there are important reasons to
tion (called height-for-age z-scores, HAZ). This improve child nutrition, and to reduce child stunt-
process allows the tracking of changes in height ing from the earliest point in life. It has often
trajectories with age. Recent insights from cohort been assumed that childrens height trajectories
data show there is change in such height trajecto- are comparatively xed at 2 years, with little
ries after the infancy period, which poses a ques- chance for later improvement (see UNICEF
tion of whether such changes are associated with 2014; Alderman and Walker 2014). However
other impacts on childrens development and of analysis of Young Lives cohort data suggests
what can be done to protect early gains, and sup- later change in height trajectories between 1 and
port recovery. 5, and between 5 and 8 years (Lundeen et al.
696 P. Dornan

Fig. 4 Variation in average height-for-age at age 8 in housing conditions, service access and durables.
Ethiopia (2009) (Note: poor/least poor is categorized Categorization is based on data collected when children
based on an indicator of household wealth (comprised of were aged 1, with HAZ measured at 8 years)

2013). Figure 5 demonstrates change between 1 ery. Height is important in itself but it is also used
and 5 years across the four countries. The height as a proxy for impacts on other aspects of chil-
of the bar is the percentage of children that were drens development (including health and cogni-
stunted, or not-stunted, in infancy. The divisions tion). It does not necessarily follow that recovery
within the bar show whether these children in height will extend to these other domains of
remained so by 5 years. Across all four countries life which may have been compromised by early
there is evidence of change, with both recovery deprivation, effects which may show up much
and faltering in growth trajectories. Analysis later in life (for example see Roseboom et al.
between 5 and 8 years also demonstrates change, 2001). Analysis has, however, found that across
though the level is smaller than occurs at the ear- the Young Lives countries that higher early HAZ
lier point. Analysis of the size of change shows (around 1 year) and HAZ gain (between 1 and
these can be quite large; between 1 and 8 years 8 years) were both associated with better reading,
those who recovered on average gained between mathematics and childrens receptive vocabulary
1.1 (Vietnam) and 1.9 (Ethiopia) height-for-age test scores at 8 years (Crookston et al. 2013). In
z-scores. Those who faltered lost between 1.5 other words both higher early HAZ and physical
(Ethiopia) to 0.9 (Peru) height-for-age z-scores recovery in HAZ were associated with better test
(Crookston et al. 2013: p. 2). Recalling that 0 is scores, suggesting cognitive gains associated
the central norm, and 2 the denition of stunt- with physical recovery.
ing, such average changes are substantial. If there are such wider gains for childrens
Evidence of such change raises questions both of development, a key question becomes whether or
whether children can be supported to recover not policy can support recovery. This question
from early deprivation, and whether other chil- can be informed by using multivariate techniques
dren can be helped not to falter in their growth, to model what predicts height recovery (using
to avoid losing initial gains. available child-level, household and community
While the evidence presented in Fig. 5 is data). While analysis of the factors associated
important, there are reasons for a cautious inter- with height recovery between 1 and 5 years high-
pretation of ndings of post infancy height recov- lights mothers height (suggesting the impor-
Longitudinal Studies and Policy for Children in Low- and Middle-Income Countries 697

Fig. 5 Percentage of children stunted at 1 year old, by whether or not they remain stunted at 5 years (Source: Lundeen
et al. 2013: p. 5, Table 3. Reproduced from Dornan and Woodhead 2015)

tance of biological inheritance or maternal protecting and building on the gains of the rst
health), this also shows that higher household thousand days of life, and to do so may bring
expenditure was predictive of improved height wider gains for those who experience early
gain, and provides some weaker evidence for the deprivation.
role of parental education and community health
infrastructure (Schott et al. 2013). Separate anal-
ysis for India identied that receipt of the midday 5 Taking Account of Childrens
meal school feeding program in drought-affected Multidimensional Lives
communities was linked with recovery (Singh
et al. 2012). So while more analysis is needed to While life course analysis demonstrates the
inform whether or not policy interventions can holistic nature of the experiences of and inu-
redress early deprivations, both household expen- ences on childrens lives, policy responses are
diture levels and public policy seem to be more typically focused on specic dimensions of
important. childrens lives learning, health, protection
In summary, life course analysis suggests that which are addressed through policy sectors. It
early circumstances matter critically, but also does not automatically follow that the best poli-
highlights change in childrens development after cies for improving education results will always
the earliest period of life. Neither the change in be found within the education system if, for
individual trajectories nor the factors associated example, it is poverty or illness that is keeping
with such change would have been apparent in children away from school, or hunger that pre-
cross sectional data. There is evidence of later vents them from learning. What then does the
change in height trajectories, and that such evidence highlight for how such domains of
improvements in height trajectories may lead to development interact?
cognitive test score gains. Investments beyond Longitudinal evidence enables analysis of
the very earliest period of life are consistent with these links between key domains throughout
698 P. Dornan

childrens development. Young Lives evidence Gutman and Schoon 2013; Dercon and Krishnan
contains information on childrens physical 2009, appendix 1). The analysis here includes
development, cognitive skills, and psychosocial measures that tested concepts including self-
development domains at different age points. esteem, self-efcacy and self-respect.
Figure 6 reports the conclusions of multivariate While not causal, this analysis gives an indica-
analysis that modelled whether earlier perfor- tion of where the associations between domains
mance on particular domains was then associated of childrens development are the strongest.
with effects on later performance on that domain Across all four countries, earlier height predicts
or others. As well as the listed domains, the anal- later height, and earlier cognitive test perfor-
ysis controls childs year of birth, maternal mance predicts later test performance. These are
schooling in years, and whether the child is grow- both the strongest statistical links, and the most
ing up in a rural area. The four-country design substantively important. By contrast, there is
allows us to see the similarity of patterns emerg- rather weak evidence (one country only) that ear-
ing across countries. The basis of the analysis lier psychosocial development independently
involves constructing measures of each domain, affects later psychosocial development. Given
using factor analysis of existing scales to produce this is multivariate (not descriptive) analysis, this
cognitive skill and psychosocial development is not to say that the same children did not report
measures from wider scales. For cognition these similar psychosocial development at 8 and
include vocabulary and math. Psychosocial 12 years, but rather that the source of such feel-
development is a particularly complex set of ings was not explained by earlier psychosocial
competencies to measure (see for example development, a nding consistent with these

Predicted by? Height-for-age at Cognitive skills at age 12 Psychosocial development at


age 12 age 12

Height-for-age Predicts later Predicts later cognitive Predicts later psychosocial


at age 8 height in all four skills in Ethiopia and development in Peru and
countries Vietnam Vietnam

Cognitive skills Do not predict Predicts later cognitive Predicts later psychosocial
at age 8 later height skills in all four countries development in Peru, India and
Ethiopia

Height-for-age at Cognitive skills at 15 Psychosocial development at


age 15 years 15 years

Height-for-age Predicts later Predicts later cognitive Does not predict later
at 12 height in four skills in Ethiopia, India, psychosocial development
countries Vietnam

Cognitive skills Predicts later Predicts cognitive skills in Predicts later psychosocial
at 12 height only in four countries development in India and Peru
Vietnam

Psychosocial Does not predict Predicts cognitive skills in Predicts later psychosocial
development later height Peru and Vietnam development only in Vietnam
at 12

Fig. 6 Life course interdependencies amongst developmental domains (Source: Sanchez 2013: p. 22, Table 9. All
results quoted are from multivariate regression analysis and are signicant at the 90 % condence interval or higher)
Longitudinal Studies and Policy for Children in Low- and Middle-Income Countries 699

competencies being malleable. Such evidence stature has been used as a proxy for school readi-
seems to suggest that while physical develop- ness (Woodhead et al. 2009). Stunting is also
ment and cognitive development appear to be, linked with greater risk of illness which then
somewhat, linearly produced through childhood, could reduce learning or attendance (Dewey and
psychosocial development appears to be less sta- Begum 2011). The central role which schooling
ble, potentially being continually open to inu- now occupies within childrens lives (Cameld
ence through childhood. and Tafere 2009; Boyden 2013) suggests links
There is also evidence of associations across between how well children do within school and
domains of development, as summarized in wider psychosocial indicators. Similarly, psy-
Fig. 6. Earlier height tends to predict later cogni- chosocial competencies (higher aspirations, self-
tive test performance, as is observed in the rela- esteem and so on) may plausibly affect how
tionship between very early height in infancy and children engage with the school and develop
cognition in early childhood (see Le 2011; cognitive skills or, relatedly, experiences of
Sanchez 2009) though these relationships are stigma or shaming may undermine how children
weaker in signicance and size than those found are able to engage with educational services
between earlier height and later height. There is (Dornan and Ogando Portela forthcoming).
also evidence at age 8 that height is associated While there are many possible reasons for why
with later psychosocial development (see also links exist between domains of development, a
Dercon and Sanchez 2013), but this is only found key point is simply that policy in one sector may
for the younger age point, and only in two coun- be supported (or undermined) by activities in
tries. There is less evidence that earlier cognitive another.
skills were associated with effects on later height In summary looking across different domains
(one country only, and for the older age points). of childrens lives highlights the interconnections
However there is more consistent evidence that of childrens development. Height is associated
cognitive skills were associated with later psy- with effects on learning; learning with effects on
chosocial development, and also that psychoso- psychosocial development and vice-versa. For
cial development is associated with improved policy, this reinforces conclusions that, for chil-
later cognitive skills (for the older age points, this dren, delivering the objectives of one policy sec-
is found in two countries). Such relationships tor requires inputs from beyond that sector,
across childrens development domains are sug- highlighting the importance of service delivery
gestive that there are both vicious and virtuous coordination/ integration (Woodhead et al. 2014).
cycles, whereby performance in one domain pro- Since it is typically the same groups of children
motes or undermines performance on other the poorest, those in rural areas, with less edu-
domains also. cated parents, and often minority groups who
Figure 6 demonstrates the holistic nature of are most likely to do less well in cognitive tests
childrens development, as domains interrelate. and who are more likely to be stunted, such
Identifying how such domains interrelate may be coordination and coverage between sectors are
instructive to identify policy intervention points. particularly important for reducing inequalities.
Height is assumed to be a proxy for wider depri-
vations to health, and so the links between height
and cognition may reect these wider depriva- 6 Closing the Gap Between
tions. The controls used in the analysis reported Life Course Studies
in Fig. 6 should minimize that effect, but since and Policy Debates
this is non-causal analysis we cannot rule out
such unobserved effects. Children being short In this nal substantive section, the chapter steps
for their age may affect how children were back from research ndings to consider the
treated in school. Stunting is linked with later context around research and of how this may
enrollment in Ethiopia, for example, where short affect the uptake of evidence. First, the section
700 P. Dornan

discusses some of the potential for longitudinal and Court 2005; UK Cabinet Ofce 2013;
approaches to inform key global policy debates. Newman 2014), the challenges to move
Second the section discusses some of the key towards evidence informing policy are greatest
challenges to linking research and policy. Third where capacity is least (Crivello and Murray
the section nishes off by discussing potential 2012; Porter and Feinstein 2014; Newman
ways in which the process of research can affect et al. 2012). Evidence is only one input into
its relevance and uptake in policy debates. decision-making (Newman 2014) and will also
be inuenced by political interest, ideology
First, we have already noted the lack of life course and feasibility considerations (see also Laub,
evidence within low- and middle-income coun- chapter Life Course Research and the Shaping
tries, but there is increasing interest within of Public Policy this volume). The timescales
international organizations in life course per- by which research is produced tend to be lon-
spectives and in increasing the availability of ger than those in which policy decisions are
panel and cohort data. Life course approaches taken. Not all research is necessarily commu-
are well embedded in many international orga- nicated in a way that can easily be picked up by
nizations thinking. The World Health policy audiences. But in low- and middle-
Organization has used both life cycle and life income countries such challenges are increased
course approaches to identify, for example, by the typically lower capacity to create and
how earlier life events shape non-communica- use data (for example see IEAG 2014). An
ble diseases such as heart disease or stroke additional difference is that while in OECD
(Aboderin et al. 2002) and ageing processes countries the state will itself be a key research
(Stein and Moritz 1999). This approach, inu- funder, in developing countries key funders
enced by epidemiology, seeks to identify how may be external donors, raising issues of how
earlier physical and social risk factors impact to ensure national ownership of research. Such
later health. Both UNDP and UNICEF have research scarcities are likely to be particularly
recognized life course approaches as helpful to acute for longitudinal research, given its
policy and programming interventions to sup- greater complexity (Lawlor et al. 2009;
port human development across age phases McKinnon and Campbell 2011; Scott and
(Malik 2014; Banati and Alexander 2012). Mariotti 2014). Indeed, where resources are
UNICEF has brought together those working particularly tight, it may be easier to make the
on cohort and panel studies to identify how case for specic evaluation studies of existing
such studies can contribute to better policies for policy approaches, with tangible, shorter term,
children in low- and middle-income countries benets, than approaches such as cohort or
(UNICEF 2014). The debate over Sustainable panel studies which take longer to pay off.
Development Goals (SDGs), has brought with Third, greater attention needs to be given to the
it proposals for a data revolution to support research process itself, in order to overcome
development objectives (UNSG 2014). But for some of the typical barriers between research
such a data revolution to extend beyond and policy (Dhaliwal and Tulloch n.d.). As
improved monitoring and towards evaluating evidence is only one of several inputs into
and informing policy choices, longitudinal as decision-making, it needs to be produced and
well as cross-sectional approaches will be provided in a way that is most effective.
needed. While capacity for life course studies Researchers need to be mindful and respon-
may be tight, building greater space for longitu- sive to the wider context of likely interest, or
dinal approaches in low- and middle-income demand, potential stakeholders and time win-
countries is a key element of designing more dows of policy interest surrounding decision-
effective policy approaches. making (Crivello and Murray 2012; Start and
Second, while the case for evidence-based policy Hovland 2004). Where there is potential pol-
making is constant across countries (Sutcliffe icy interest in research topics, strategies to
Longitudinal Studies and Policy for Children in Low- and Middle-Income Countries 701

support policy engagement include involving households to support children are important.
policy actors early and throughout the research In the data reviewed, inequalities between
process (for example in choices over specic boys and girls emerge strongly in youth and
research questions, through advisory groups, adolescence, shaped by wider norms prevalent
more informal channels and early dissemina- within society.
tion of ndings before formal publication). A Second, as well as showing inequalities are evi-
Young Lives study of barriers to research dent early in life, longitudinal evidence of the
uptake in Ethiopia and India highlighted barri- development of height trajectories also shows
ers to overcome, including that decision mak- change with age. Such analysis shows again the
ers may both be overwhelmingly busy and not particular advantage of cohort approaches since
be equipped with skills to interpret complex change is inherently underestimated within
research. This study suggests the importance cross sectional data, and which would therefore
of tools such as evidence briefs which suc- mask the correlates and implications of such
cinctly communicate key messages (Crivello change. Physical recovery and faltering is asso-
and Murray 2012). Strategies to provide good ciated with socio-economic circumstances,
quality research evidence in accessible and identifying a process by which change over the
timely ways to policy-making cycles increase early life course may widen inequalities. While
the chances of such evidence being made use the chapter emphasizes that more analytic work
of in policy debates. is needed on post-infancy recovery and falter-
ing, these initial ndings highlight an important
addition to messages about early intervention:
early is best but sustained policy intervention is
7 Concluding Discussion needed to secure early gains and such an
approach may also support some recovery.
Debates about the proposed Sustainable Third, the chapter highlights a disjuncture: chil-
Development Goals, have highlighted the need drens lives are multidimensional, with inter-
for data to lighten the way for policy (UNSG dependencies between domains of childrens
2014). This context, is therefore one of the recog- lives, whereas public policy approaches are
nition of the importance of evidence to inform typically organized in policy sectors to
policy. In this concluding discussion we summa- address health, education or protection con-
rize and reect on the implications of the ndings cerns. While such approaches are administra-
reviewed, and the potential contribution of longi- tively efcient, they may not be as effective as
tudinal studies to this agenda in low- and middle- would be combining efforts across sector
income countries. areas in securing gains in child wellbeing.
Approaches within the early childhood eld
First, the chapter identied increasing attention towards greater integration give some indica-
to inequalities within countries. The particular tion of both the potential gains, but also the
power of longitudinal studies is to link earlier delivery challenges of greater integration
circumstances and later outcomes, and so such (Britto et al. 2014). Approaches to greater
studies have much to contribute in considering integration span from joint delivery of ser-
how policy may mitigate the development of vices on the ground to joint planning of ser-
inequalities. The conclusions from the studies vices nationally. Given that the most
reviewed highlight that both early socio- marginalized children are also most likely to
economic and locational-related inequalities be multiply deprived, with potentially com-
in childrens development are observable in pounding results, improving the synergies
infancy and early childhood. As such, social between service areas is likely to be particu-
policies, such as social protection and early larly important in helping reach the most mar-
childhood interventions, which can help ginalized children.
702 P. Dornan

Finally, the chapter identies the need not only to diabetes: Key issues and implications for policy and
research. Geneva: World Health Organization.
consider research ndings, but also the process
Alderman, H., & Walker, S. (2014). Enhancing resilience
of research generation, in order that such evi- to nutritional shocks. 2020 conference brief 17.
dence is most useful for policy development. Washington, DC: International Food Policy Research
In doing so the chapter identied commonali- Institute.
Banati, P., & Alexander, G. (Eds.). (2012). The structural
ties across countries in the circumstances in
determinants of child well-being. Florence: UNICEF
which research is inuential on policy deci- Ofce of Research.
sion-making. Beyond effective and simple Boyden, J. (2013). Were not going to suffer like this in
communication of often complex research the mud: Educational aspirations, social mobility and
independent child migration among populations living
ndings, the chapter emphasized the need for
in poverty. Compare: A Journal of Comparative and
choices over research questions to be informed International Education, 43(5), 58600.
by an understanding of policy needs and inter- Boyden, J., Pankhurst, A., & Tafere, Y. (2013). Harmful
ests, and for the involvement of policy makers traditional practices and child protection contested
understandings and practices of female child mar-
within the process of research to help ensure
riage and circumcision in Ethiopia. Oxford: Young
research supply meets policy demand. Lives.
Britto, P., Yoshikawa, H., van Ravens, J., Ponguta, L.,
While capacity challenges are real, the chapter Reyes, M., Oh, S., Dimaya, R., Nieto, A., & Seder, R.
(2014). Strengthening systems for integrated early
ends with a couple of points looking to the future.
childhood development services: A cross national
There is considerable interest from international analysis of governance. Annals of the New York
organizations in life course approaches and a rec- Academy of Sciences, 1308, 245255.
ognition these can improve policy and program- Cameld, L., & Tafere, Y. (2009). Children with a good
life have to have school bags: Diverse understand-
ming impact. Considerable efforts are being
ings of well-being among older children in three
made across research programmes to create Ethiopia communities. Oxford: Young Lives.
research in ways which are accessible for busy Crivello, G., & Murray, H. (2012). Why strengthening the
and non-research trained policy makers to use. linkages between research and practice is important:
Learning from Young Lives. Oxford: Young Lives.
Finally while the pure aim of evidence-based
Crookston, B., Schott, W., Cueto, S., Dearden, K., Engle,
policy seems nave in politically orientated P., Georgiadis, A., Lundeen, E., Penny, M., Stein, A.,
policy-making, such improved approaches to the & Behrman, J. (2013). Post-infancy growth, school-
research process and communication show how ing, and cognitive achievement: Young Lives.
American Journal of Clinical Nutrition. doi:10.3945/
the research community can improve the chances
ajcn.113.067561.
of evidence playing a greater role. Cueto, S., Escobal, J., Penny, M., & Ames, P. (2011).
Tracking disparities: Who gets left behind? Initial
Acknowledgements I would like to acknowledge the ndings from Peru, round 3 survey report. Oxford:
involvement of Young Lives participants on whose experi- Young Lives.
ence this analysis is based. I would also like to thank Cueto, S., Guerrero, G., Leon, J., Zapata, M., & Freire, S.
Kirrily Pells, and chapter reviewers for thoughtful com- (2014). The relationship between socioeconomic sta-
ments. Young Lives is funded from by UK aid from the tus at age one, opportunities to learn and achievement
UK Department for International Development (DFID) in mathematics in fourth grade in Peru. Oxford Review
and co-funded by the Netherland Ministry of Foreign of Education, 40(1), 5072.
Affairs from 2012 to 2014 and by Irish Aid from 2014 to Dercon, S., & Krishnan, P. (2009). Poverty and the psy-
2015. The views expressed are those of the author. They chosocial competencies of children: evidence form the
are not necessarily those of, or endorsed by, Young Lives, Young Lives sample in four developing countries.
the University of Oxford, DFID or other funders. Children Youth and Environments, 19(2), 138163.
Dercon, S., & Sanchez, A. (2013). Height in mid child-
hood and psychosocial competencies in late child-
hood: Evidence from four developing countries.
References Economics & Human Biology, 11(4), 426432.
Dercon, S., & Singh, A. (2011). From nutrition to aspira-
Aboderin, I., Kalache, A., Ben-Shlomo, Y., Lynch, J. W., tions and self-efcacy: Gender bias over time among
Yajnik, C. S., Kuh, D., & Yach, D. (2002). Life course children in four countries. Oxford: Young Lives.
perspectives on coronary heart disease, stroke and
Longitudinal Studies and Policy for Children in Low- and Middle-Income Countries 703

Dewey, K., & Begum, K. (2011). Long-term conse- McKinnon, R., & Campbell, H. (2011). Systematic review
quences of stunting in early life. Maternal and Child of birth cohort studies in South East Asia and Eastern
Nutrition, 7(3), 518. Mediterranean. Journal of Global Health, 1(1), 5971.
Dhaliwal, I., & Tulloch, C. (n. d.). From research to pol- Morrow, V. (2013). Troubling transitions? Young peoples
icy: Using evidence from impact evaluations to inform experiences of growing up in poverty in rural Andhra
development policy. Cambridge, MA: Abdul Latif Pradesh, India. Journal of Youth Studies, 16(1),
Jameel Poverty Action Lab/Massachusetts Institute of 86100.
Technology. Newman, K. (2014). What is the evidence on the impact of
Dornan, P,. & Ogando Portela, M. (forthcoming). Do feel- research on international, development. London:
ings of shame undermine childrens development? DFID.
Oxford: Young Lives. Newman, K., Fisher, C., & Shaxson, L. (2012). Stimulating
Dornan, P,. & Woodhead, M. (2015). How inequalities demand for research evidence: What role for capacity
develop through childhood. Florence: UNICEF. building? IDS Bulletin, 43(5), 1724.
Dornan, P., & Pells, K. (2014). From infancy to adoles- Ostry, J., Berg, A., & Tsangarides, C. (2014).
cence: Preliminary ndings from round 4. Oxford: Redistribution, inequality, and growth. Washington,
Young Lives. DC: International Monetary Fund.
Galab, S., Kumar, V., Reddy, P., Singh, R., & Vennam, U. Petrou, S., & Kupek, E. (2010). Poverty and childhood
(2011). The impact of growth on childhood poverty in undernutrition in developing countries: A multi-
Andhra Pradesh: Initial ndings from India. Oxford: national cohort study. Social Science & Medicine,
Young Lives. 71(7), 13661373.
Glewwe, P., Krutikova, S., & Rolleston, C. (2014). Do Pickety, T. (2013). Capital in the twenty-rst century.
schools reinforce or reduce learning gaps between Cambridge, MA: Harvard University Press.
advantaged and disadvantaged students? Evidence Porter, S., & Feinstein, O. (2014). Demand for and supply
from Vietnam and Peru. Oxford: Young Lives. of evaluations in selected Sub-Saharan African coun-
Grantham-McGregor, S., Cheung, Y., Cueto, S., Glewwe, tries. Johannesburg: Graduate School of Public and
P., Richter, L., & Strupp, B. (2007). Developmental Development Management, University of the
potential in the rst 5 years for children in developing Witwatersrand.
countries. The Lancet, 369, 6070. Rolleston, C. (2014). Learning proles and the skills gap
Gutman, L., & Schoon, I. (2013). The impact of non- in four developing countries: a comparative analysis of
cognitive skills on outcomes for young people. schooling and skills development. Oxford Review of
London: Institute of Education. Education, 40(1), 132150.
Independent Expert Advisory Group Secretariat (IEAG). Rolleston, C., James, Z., Pasquier-Doumer, L., & Tran
(2014). A world that counts mobilising the data revo- Ngo Thi Minh Tam. (2013). Making progress: Report
lution for sustainable development. Report prepared at of the Young Lives school survey in Vietnam. Oxford:
the request of the United Nations Secretary-General, Young Lives.
www.undatarevolution.org Roseboom, T., van der Meulen, J., Ravelli, A., Osmond,
Lawlor, D., Andersen, A., & Batty, G. (2009). Birth cohort C., Barker, D., & Bleker, O. (2001). Effects of prenatal
studies: Past, present and future. International Journal exposure to the Dutch famine on adult disease in later
of Epidemiology, 38, 897902. life: An overview. Molecular and Cellular
Le, T. D. (2011). Height and cognitive achievement of Endocrinology, 185(2001), 9398.
Vietnamese children. World Development, 39(12), Sanchez, A. (2009). Early nutrition and cognitive achieve-
22112220. ment in pre-school children in Peru. Oxford: Young
Le, T. D., Nguyen, T., Nguyen, V. T., Mai, T. H., & Vu, Lives.
T. T. T. (2011). How do children fare in the new mil- Sanchez, A. (2013). The structural relationship between
lennium? Initial ndings from Vietnam. Oxford: nutrition, cognition and non-cognitive skills: Evidence
Young Lives. from four developing countries. Oxford: Young Lives.
Lloyd, C. (Ed.). (2005). Growing up global, the changing Schott, W., Crookston, B., Lundeen, E., Stein, A.,
transitions to adulthood in developing countries. Behrman, J., & the Young Lives Determinants and
Washington, DC: The National Academies press. Consequences of Child Growth Project Team. (2013).
Lundeen, E., Behrman, J., Crookston, B., Dearden, K., Periods of child growth up to age 8 years in Ethiopia,
Engle, P., Georgiadis, A., Penny, M., & Stein, A. India, Peru and Vietnam: Key distal household and
(2013). Growth faltering and recovery in children aged community factors. Social Science & Medicine, 97,
18 years in four low and middle-income countries: 278287.
Young Lives. Public Health Nutrition, 17(9), Scott, L., & Mariotti, C. (2014). A place for panel data in
21312137. the data revolution. London: Overseas Development
Malik, K. (Ed.). (2014). Safeguarding human progress: Institute.
Reducing vulnerabilities, building resilience. Singh, A. (2014). Emergence and evolution of learning
New York: UNDP. gaps across countries: Panel evidence from Ethiopia,
India, Peru and Vietnam. Oxford: Young Lives.
704 P. Dornan

Singh, A., Park, A., & Dercon, S. (2012). School meals as UNICEF. (2013). Improving child nutrition: The achiev-
a safety net: An evaluation of the midday meal scheme able imperative for global progress. New York:
in India. Oxford: Young Lives. UNICEF.
Start, D., & Hovland, I. (2004). Tools for policy impact: A UNICEF. (2014). International symposium on cohort and
handbook for researchers. London: Overseas longitudinal studies in low and middle income coun-
Development Institute. tries (Meeting Report). Florence: UNICEF.
Stein, C., & Moritz, I. (1999). A life course perspective of United Nations (UN). (2014). The millennium develop-
maintaining independence in older age. Geneva: ment goals report 2014. New York: United Nations.
World Health Organization. Woldehanna, T., Guidisa, R., Tafere, Y., & Pankhurst, A.
Sutcliffe, S., & Court, J. (2005). Evidence-based policy- (2011). Understanding change in the lives of poor
making: What is it? How does it work? What relevance children: Initial ndings from Ethiopia. Oxford:
for developing countries? London: Overseas Young Lives.
Development Institute. Woodhead, M., Ames, P., Vennam, U., Abebe, W., &
UK Cabinet Ofce. (2013). What works: Evidence centres Streuli, N. (2009). Equity and quality? Challenges for
for social policy. London: HM Government early childhood and primary education in Ethiopia,
UN General Assembly (UNGA). (2014). Status of the India and Peru. The Hague: Bernard van Leer
convention on the rights of the child report of the Foundation.
secretary-general. New York: UN. Woodhead, M., Dornan, P., & Murray, H. (2013a). What
UN Secretary General (UNSG). (2014). The road to dig- inequality means for children: Evidence from Young
nity by 2030: Ending poverty, transforming all lives Lives. Oxford: Young Lives.
and protecting the planet. New York: UN. Woodhead, M., Frost, M., & James, Z. (2013b). Does
UN Statistics Division & UN Women. (2014). Millennium growth in private schooling contribute to Education
development goals gender chart. New York: UN. for All? Evidence from a longitudinal, two cohort
UNESCO. (2014). Teaching and learning: Achieving study in Andhra Pradesh, India. International Journal
quality for all, education for all global monitoring of Educational Research, 33(1), 6573.
report. Paris: UNESCO. Woodhead, M., Feathersone, I., Bolton, L., & Robertson,
UNPFA. (2014). The power of the 1.8 billion, adolescents, P. (2014). Early childhood development: Delivering
youth and the transformation of the future, State of the intersectoral policies, programmes and services in low
Worlds Population 2014. New York: United Nations resource settings. Oxford: Health & Education Advice
Population Fund. & Resource Team.
UNICEF. (2010). Progress for children: Achieving the
MDGs with equity. New York: UNICEF.
Lags and Leaps: The Dynamics
of Demography, Economy
and Policy and Their Implications
for Life Course Research

Angela M. ORand and Amie Bostic

1 Introduction continue as fundamental to the future of life


course research across the life span.
Two decades ago Matilda White Riley proposed While today the challenge of this provocative
that the idea of structural lag best characterized view remains, it must also be framed by a global
the relationships of social structures, policies and perspective on population aging and be focused
culture to population aging in the U.S. She and on the entire life course, not just the older years.
her colleagues (Riley et al. 1994) argued that At least three related demographic patterns coin-
social institutions, organizational arrangements cide to present this new challenge. First, life
and cultural stereotypes were lagging behind expectancy has doubled around the world over the
demographic change, especially the extension of last 200 years, with remarkable increases (i.e. a
life expectancy (including healthy life expec- near doubling) in life expectancy in the last half
tancy), the gender revolution, and medical- century in the developing world (Uhlenberg
scientic knowledge about population aging 2009). These leaps in population aging along with
from birth until death, among other trends. Unlike the longer-developing demographic transitions in
earlier times, most people live to be old, with life advanced economies present new challenges for
expectancies reaching 85 or more among the social policy, especially social welfare (including
healthiest subgroups. Yet, social institutions have education) and employment policies that lag
not evolved to incorporate these groups in main- behind demographic change, and raise new ques-
stream life. Similarly, the lives of young adults tions for social science. Second, as life expec-
have become more unpredictable and vulnerable tancy has generally increased across countries,
to changes in economic institutions that no lon- gender differences persist in life expectancy while
ger accommodate easy entry into lifetime work the shapes of gendered lives reect both remark-
or assure pay-offs to education. Finally, womens able changes and continuing inequalities. Fertility
lifetime patterns of labor force participation still decline and the expansion of educational opportu-
do not t well with institutional arrangements nities have improved and extended womens lives,
that continue to challenge their childcare, educa- but longer lives continue to be associated with
tional, and work demands. Rileys argument disability, loneliness and the higher risks for pov-
brought renewed attention to the dynamics erty. Third, migration patterns across countries
among demography, economy and policy, which and regions have differentiated the cohort-compo-
sitions of many advanced societies especially in
A.M. ORand (*) A. Bostic
race/ethnicity/nativity mix and now challenge
Duke University, Durham, NC, USA intergenerational and inter-cohort relationships
e-mail: angela.o.rand@duke.edu and the social policies that regulate them.

Springer International Publishing Switzerland 2016 705


M.J. Shanahan et al. (eds.), Handbook of the Life Course, Handbooks of Sociology
and Social Research, DOI 10.1007/978-3-319-20880-0_32
706 A.M. ORand and A. Bostic

These major population trends have been account for growing variations in life transitions
accompanied by transformations in the global and the pace of aging, including both older and
economy. The twenty-rst century global econ- younger cohorts in advanced countries but also
omy is characterized by a division of labor and a extending to developing countries. The second
system of nancial control that are redening focuses on the ascendance of market institutions,
employment and family security across societies especially nancial institutions, in the mainte-
and creating a new risk environment, both for nance of individual and household nancial and
individuals and their families, and for national health decision-making. Financial and health lit-
policies. This environment brings greater uncer- eracy are new life course risks which pervade the
tainty to the conduct of daily lives and to short- life course. Third, increased migration levels
and long-term policymaking, respectively. The over recent decades, especially to some advanced
ascendance of market institutions has reorganized countries, are changing the demographic compo-
the global labor market and redened work and sitions of successive cohorts who face changing
work careers across countries. It is also imposing educational opportunities and workplace protec-
pressures on public policymaking across coun- tions. The life course implications of migration
tries to privatize the nancing and the delivery of are now matters of speculation, which life course
educational and other welfare services and to nar- research should address in the future as data
row the eligibility for these services across the life permit.
span through such mechanisms as educational
vouchers and means-testing for income support
within national populations. Population aging is 2 The Macrosocial
independently exerting nancial pressures on Foundations of the Life
public policymaking and, interdependently with Course
global economic restructuring, having an impact
on educational and employment policies affecting The dynamics among global economy, social
all age groups. The most recent economic crisis policy and demography are complex, variable
starkly illuminated what has otherwise been a and usually asynchronous across countries with
long developing restructuring of economic life different development histories, demographic
with implications for the economic independence compositions, and policy legacies. Each trend has
of individuals and families and for sustainable had unique recent effects on the development of
public policies with relevance for the life course the life course that are very briey considered
from childhood through old age. As such, these below. These unique effects can be countervail-
large-scale trends are having on-the-ground ing within contextsyielding leaps and lags. For
impact on institutional arrangements that affect example, globalization processes pressure
educational opportunity, job security, health care employers to reorganize their workforces through
eligibility and access, and individual and family reduction and/or relocation, while demographic
well-being across the life span. pressures such as population aging pressure
This chapter will identify selected areas of life employers to retain workers and governments to
course research that should be motivated by these extend retirement ages to relieve welfare state
considerations. The selection is far from exhaus- budgets (Blossfeld et al. 2011). Similarly, eco-
tive, but illustrates some directions for the future nomic shocks that erupt and spread in the global
of research in this area. Three areas for future economy place pressures on families to limit fer-
research are proposed. The rst is a more concep- tility, which in turn affects the prospective sizes
tual proposal to guide life course research in gen- of future tax bases to support long-term welfare
eral. It focuses on the life course as a continuous policy planning to support aging populations.
manifold process with diverse temporality in a These complex dynamics, arguably, do not sup-
population. Age-graded, phasic constructions of port the expectation of a convergence in the con-
the life course are treated as too restrictive to struction of the twenty-rst century life course as
Lags and Leaps: The Dynamics of Demography, Economy and Policy and Their 707

such. Everyone will not age in the same way or at employees and employers with compensation
the same pace. On the other hand, clearly predict- packages for a majority of public/private sector
able specic patterns of divergence or heteroge- workers that promised job security, income main-
neity in the future life course within and across tenance after retirement with dened benets,
countries are also problematic. and health and worker compensation coverages
while employed. However, employers have
Economic Restructuring Three aspects of eco- retreated from the provision of these packages at
nomic globalization with life course inuences accelerating rates since 1980, even for fulltime
are reviewed below. First, the global division of workers. Nonstandard employment arrangements
labor differentially distributes work across such as part-time, contract, and temporary jobs
employment sectors and differentially concen- have grown dramatically and particularly in
trates nancial, manufacturing, service, retail, unstable sectors in low end services and retail
and research and development sectors across across countries (e.g. Kalleberg 2009). This
countries and regions resulting in the displace- workplace trend is shifting the full responsibility
ment of once relatively stable manufacturing and for protection against income loss and health
related employment sectors from advanced econ- decline to individuals and their families from an
omies to developing countries (Berg and early age (Hacker 2006). This reects the indi-
Kalleberg 2001). This has been occurring over vidualization of risk which was once based on
four decades, and has been punctuated by reces- contracts with intermediary organizations, pri-
sions which have become more frequent, and marily employers in liberal market economies to
recently more severe. Job loss in the U.S. in the cover pensions and health insurance. Evidence
recent Great Recession reached more workers at reveals that subsidiary processes are also being
all educational levels; was the highest since the implemented in countries with traditionally more
1980s; and was followed by lower rates of reem- generous welfare systems (ORand et al. 2011).
ployment and lower earnings after reemployment Arguably, in the U.S., lower-level governmental
partly as a result of part-time employment fol- bodies (e.g. U.S. state governments) are also
lowing earlier fulltime employment (Farber pressured to redistribute responsibility for uni-
2011, using the Displaced Workers Surveys, versal elementary and secondary education
19842010). Also, the typical gap in U.S. job loss through mechanisms that encourage individual
rates between older and younger workers nar- choice (e.g. charter schools, school vouchers,
rowed to bring younger workers closer to their home schooling). Population aging is making
older counterparts in vulnerability to involuntary long-term pension and healthcare promises by
job loss (Farber 2011). Meanwhile, the accompa- employers and the provision of adaptive educa-
nying decline in job growth has negatively inu- tional systems by the public sector unsustainable
enced rst-job entry rates for young workers under previous arrangements.
across educational levels in advanced countries A third critical component of global economic
and increased the likelihood of longer durations restructuring is the ascendance of a neoliberal
of part-time work, underemployment and unem- policy model associated with a concentration of
ployment in the early career (Blossfeld et al. the nancial sector (Tomaskovic-Devey and Lin
2005). The implications for the future careers of 2011). Financialization of the global economy
these cohorts are not fully known and the impact centralizes power in nancial institutions such as
of the perceptions of the future by these cohorts banks, investment rms, and insurance organiza-
has gone unstudied. tions as opposed to manufacturing, retail or pub-
lic institutions. The former institutions are more
The second related trend associated with responsive to short-term economic cycles and the
global restructuring is the contraction of the rapid distribution of abstracted market data and
industrial-era employment relationship charac- not embedded in traditional employment or gov-
terized by long-term, fulltime contracts between ernment relationships. Traditional pensions are
708 A.M. ORand and A. Bostic

disappearing and being replaced by investment term demographic transition characterized by


vehicles (e.g., 401ks, IRAs) that expose individu- new family or household forms that have dis-
als to equities and bonds which carry consider- placed the nuclear family of the twentieth century
able risks. Importantly, they require a level of in advanced societies. Variable patterns of timing
nancial literacy for their purchase and manage- and sequencing of educational completion, work
ment that is not equally distributed in popula- entry, marriage or cohabitation and fertility have
tions. And they require the capacity and developed in advanced economies and appear to
motivation to plan far into the future. be spreading to newly developing economies in
Hence, the ascendance of new risks is an distinctive ways (Lestaeghe 2010). The extent to
emergent aspect of everyday life characterized by which these patterns are responding to global
growing uncertainties about the support of local economic change and variations in the policy
and national institutions for life course chal- environment deserve more attention in compara-
lenges and the increased vulnerability of individ- tive perspective. The major hypothesis about the
uals across the life course to remote shocks. spread of the second demographic transition is
Risks to health and well-being are not new (Graff cultural: that elements of modernization, espe-
1995). Infectious diseases and poverty were risks cially higher levels of education for women and
that prevailed across the twentieth century and the diffusion of values related to autonomy and
were addressed by collective institutional solu- individualization are principle drivers of this
tions in public health and Social Security. The trend.
newer individualization or devolution of risk has Third, the increased migration of populations
spawned pervasive uncertainty and the spread of from less developed to more developed regions is
risk aversion across economic and policy accelerating rapidly within and across countries.
sectors. The impact of migration on the life courses of
migrating groups and on the policy environments
Demographic Change Three aspects of variable of receiving (host) countries or regions is among
demographic change reect both the independent the major issues of the twenty-rst century. The
force of collective and cumulative individual economic and political arrangements for receiv-
decision-making in response to the daily material ing and integrating non-native populations lag
conditions of life and populations responses to behind the pressures of immigrating populations.
economic forces and policy environments. Added to these institutional issues are the
reactions of native cohorts that vary in their cul-
Population aging is occurring across coun- tural tolerance of new immigrant groups. Even
tries, although it varies in its historical leg- societies like the U.S., which has a history of
acy (Poterba 2014). The delay or limitation of immigration, are ill-prepared in the new eco-
fertility is the principle mover of population nomic and cultural environment to easily absorb
aging. The second factor in this trend is the exten- new groups.
sion of life expectancy through economic devel-
opment, improved environmental conditions, and Social Policy Twentieth century policies with
medical innovations that contain or mitigate historical relevance for the construction of the
infectious, chronic and disabling diseases that life course and protection against life course risks
increase mortality. This is inuencing the demo- are being challenged by globalization and popu-
graphic composition of societies and producing lation changes outlined earlier (Sainsbury 2012).
inter-cohort differences in composition, differen- The devolution of risk in social policies is a dom-
tial exposures to economic shocks across the life inant trend across countries, including those with
span and potentially competing cohort interests historically generous welfare states and corporat-
in policies that insure against life course risks ist economies. The devolution of risk refers to the
with age. erosion of inclusive social insurance structures
The second demographic transition refers to a and their replacement with neoliberal selective,
recent leap and complex extension of the longer- privatized and individualized nancial mecha-
Lags and Leaps: The Dynamics of Demography, Economy and Policy and Their 709

nisms for the management of life course risks in Educational participation, marriage, parenthood,
education, health, and income maintenance. A labor market entry and even retirement can occur
major result is unequal access to life course and recur over the life course at times not pre-
resources that yield cumulative patterns of dicted by earlier models. Health and illness over
inequality. the life course add another temporal dimension
that has been discovered to be less age-graded
Financialization and individualization of than earlier believed. The macro-level trends just
retirement are most apparent in liberal market discussed have contributed to the changing and
economies where private pension mechanisms variable temporality of lives. But, more compre-
have existed for decades, but these are spreading hensive data collection has also contributed to
to previous social democratic and corporatist this observation.
welfare states that can no longer sustain acceler- The availability of large representative, matur-
ated population aging and volatile global eco- ing samples followed over several decades has
nomic shocks. The erosion of the retirement made possible ner observations of unfolding
institution is proceeding across advanced econo- lives in the U.S. The introduction and mainte-
mies and its construction in developing econo- nance of similar and sometimes highly compara-
mies is being shaped by global economic forces ble panel surveys in other countries are now
(ORand 2011). enriching our understanding of life course pro-
cesses under sometimes highly variable condi-
tions. Moreover, the increasing collection of
3 Exemplary Life Course biodemographic, biomedical, and administrative
Processes for Future (e.g. clinical and death les; school-level records;
Research satellite geographic information) data that are
linked to individual survey les brings the eld
This section of the chapter will propose examples closer to examining lives holistically. Finally, new
of major life course processes predicted to con- data collection technologies such as Ecological
tinue to be inuenced, sometimes in contradic- Momentary Assessments (Cain et al. 2009) that
tory ways, by the dynamics among economic use smartphones and other mobile devices present
change, demography and elements of social pol- opportunities to conduct deeper studies of the
icy. Notable demographic and economic changes impacts of proximate environments (such as poor
(leaps) observable at the micro and macro levels or ethnically-concentrated neighborhoods; com-
in the last few decades are featured and social munity patient delivery systems) on life course
policies that enable or challenge these changes processes as they unfold.
(lags) are considered in their associations. The life course is a manifold cumulative phe-
Selected problematic features of these changes nomenon consisting of intertwining processes
are identied as directions for future research. associated with human development from birth to
death, that include biological, cognitive, and
social transitions that are interdependent and
3.1 The Life Course cumulative in their impact with age and inu-
as a Continuous Manifold enced by environmental factors such as macro-
Process with Diverse economic structures and social policies and
Temporality in a Population meso-level contexts such as family, school,
neighborhood and geographic conditions. As
Longitudinal research over the last two decades such, life course mechanisms include several
has revealed considerable heterogeneity in the components: selection; interdependent cumula-
temporality of lives that deviates from the strict tive impact; relative agency within contexts; and
age-graded, phasic model that guided social sci- pervasive heterogeneity and inequality along the
ence research over most of the twentieth century. way (Chart1).
710 A.M. ORand and A. Bostic

Selection Prior/initial traits, contexts or conditions, especially at extremes,


have persistent direct and indirect impact on later outcomes

Cumulative impact Successive events/transitions are conditioned by earlier


ones leading to greater continuity than discontinuity

Agency within contexts Selected environments provide more choices than others
among options that influence later outcomes

Heterogeneity/ The cumulative biographies across environments yield


inequality increasing heterogeneity and inequality in cohorts

Chart 1 Life course mechanisms

Strong selection processes pervade the life pressure for the continuation of (or return to) for-
course, from the earliest environments to suc- mal education well into the adult years brought
cessive ones. Gene-environment interactions on by technological changes in the workplace as
are possible from conception and gestation and well as by economic shocks that affect job oppor-
continue throughout life in latent and manifest tunities and job security for workers across
ways. Childhood environments provide forma- ages (Blossfeld et al. forthcoming). Both of these
tive contexts, especially family conditions trends are directly affecting the temporal organi-
related to social and economic advantage or zation of the life course, i.e. the timing and
disadvantage and childhood health, that shape sequencing of once more tightly scheduledand
the childs capacity to enter and move through genderedtransitions among young adults
the educational system. The educational sys- across educational, family and work roles. They
tem, in turn, is perhaps the most powerful insti- also reect cumulative and agentic processes
tutional environment outside the family to operating to move individuals across different
inuence life chances and well-being over the successive contexts within which different
life course for the majority of the U.S. popula- choices present themselves.
tion. The observed empirical impact of educa-
tional attainment on cumulative adult outcomes The Gendered Life Course Signicant gender
ranging from marriage and marital stability to cross-overs have been observed in recent decades:
occupational, earnings and wealth attainment (1) women across many countries have equiva-
and security, to health behaviors and to risks for lent if not higher relative educational achieve-
morbidity and mortality is so widely docu- ment than men and nearly equivalent labor force
mented as to have achieved the logical status of participation rates, yet occupational segregation
a truism. and wage inequality persist; (2) womens house-
In spite of the robustness of the association of hold arrangements have become more variable
education with later life outcomes, the selective with implications for lifelong advantage or disad-
role of education in the life course is still prob- vantage; and (3) women now engage in more
lematic and will continue to invite further analy- risky health behaviors such as smoking and obe-
ses. Two cross-national phenomena associated sity, while mens health behaviors have relatively
with global economic and demographic factors improved with selective divergent consequences
that invite such investigation are presented below. for both. These changes reect selective, cumula-
The rst relates to the recent cross-over of rela- tive, and agentic shifts that are yielding inter-
tive educational attainment levels placing women cohort variations in work, family and health
ahead of men in education in recent cohorts behaviors among women with consequences in
across many countries. The second is the growing adulthood that invite interrogation.
Lags and Leaps: The Dynamics of Demography, Economy and Policy and Their 711

The gender cross-over in educational attain- marriage, childbearing without marriage, etc., after
ment represents a leap in recent history. In the rst leaving school at higher median ages than in
U.S. context womens and mens college comple- the past. Meanwhile, the process of labor market
tion rates have risen since 1970; mens from 20 to entry from rst completed schooling has grown
27 % and womens from 14 to 36 %, resulting in more varied and less orderly for women and men
current comparative rates of completion favoring across advanced economies (Buchmann and Kreisi
women by 8 %. Women also complete the major- 2011) and job mobility following rst job has
ity of masters degrees (60 %) and more than half increased, especially in the U.S. (Royalty 1998). In
of doctoral and professional degrees (DiPrete and short, advances for some women in education and
Buchmann 2013). Also, women are now a major- work may be offset by new uncertainties associated
ity of tertiary level students in the U.S., Latin with economic restructuring, especially if social
American and Western Europe (excluding policies do not provide safety nets.
Germany). Explanations lean towards changing Finally, another fruitful area of research
national-level gender ideologies and school cul- should address health outcomes related to these
tures at the elementary and secondary levels that cumulative pathways. Projections of life expec-
retain masculine identity norms among boys tancy among women and men suggest some nar-
(McDaniel 2009; DiPrete and Buchmann 2013). rowing of the traditional average female
These explanations contain elements of selection advantage, in part as a result of changing health
and agency. Meanwhile these studies, and others, behaviors such as smoking (Preston et al. 2014)
reveal improvements in, but also persistence of, but also arguably as a result of generally increased
gender earnings differences favoring men cumu- stressful lifestyles related to work-family balance
latively over the career. and economic instability. Women are smoking
Different levels of educational achievement more and the increase in obesity among them has
set women and men on different life course been especially dramatic in most recent decades.
schedules and cumulative tracks through adult- But considerable heterogeneity in these behav-
hood. The most advantaged track is a sequence of iors also exists based on education and cumula-
educational (college) completion, followed by tive disadvantage. Even region of residence,
employment and then family formation. However, which is highly associated with economic and
the least advantaged track is associated with demographic concentration and differential pub-
lower educational attainment, early childbearing, lic policies towards health and income support,
and unstable work and marriage histories (Sawhill has unique effects on health behaviors and out-
2014). The latter is especially cumulatively dis- comes (Boardman et al. 2005).
advantageous for women who carry the heavier
childbearing/rearing burdens. Early childbearing Extended Work-Education Sequences Across
has negative consequences for the health and Adulthood Some studies suggest that both the
well-being of women and their children, espe- youngest and oldest workers have been espe-
cially in countries whose welfare policies are less cially impacted by economic globalization and
generous and based on employment (workfare) the rise of precarious work (Blossfeld et al. 2005;
policies (Sainsbury 2012). Blossfeld et al. 2011). The most recent unem-
Meanwhile, the majority fall in between the ployment statistics from Eurostat (2014) report
most advantaged and disadvantaged, have less the youth unemployment rate (ages 1524) as
clearly predictable paths and considerable uncer- double that of other groups and as high as 50 %
tainty about economic security (Kalleberg 2009; in some countries (e.g. Italy). Of course, low fer-
Hacker 2006). Among these majority populations, tility rates in Europe have made the size of this
the early adult life course is reecting this uncer- group relatively small such that the ratio of
tainty in a complex mix of schedules emerging unemployed youth to all unemployed is less dra-
across countries including the U.S. with some sub- matic. Nevertheless, the cumulative life course
groups falling within Lestaeghes putative second implications for these future adult populations
demographic transition: cohabitation, postponed require some monitoring.
712 A.M. ORand and A. Bostic

One response to unemployment is to remain tract embedded in all twentieth century welfare
in, or return to, school. By 2011, the OECD regimes. In this era, the work career may be
(2011) reports that approximately 26 % of 2029 increasingly a punctuated and disordered
year olds in OECD and G20 countries were sequence of work, education, unemployment,
enrolled in part-time or full-time education, up underemployment, and retirement.
8 % since 1995. However, these unemployment
effects of the recent global recession spread to
mid-adult working populations across employ- 3.2 Financial and Health Literacy
ment sectors and across countries. Adult enroll- Are New Life Course Risks
ment in formal coursework increased as well
with increases in enrollments among 3040 year The individualization of risk and increased
olds reaching double digit levels in many coun- responsibility for independent management of
tries (e.g. Australia 12 %; Israel 21 %; U.S. 17 %; expected and unexpected life course challenges
Chile 29 %). In the U.S. enrollment by adults in requires new forms of knowledge and understand-
post-secondary education has increased steadily ing in the conduct of daily life. Individuals are
over that last three decades, in part as a response increasingly confronted with risk assessments and
to job insecurity (Elman and ORand 2002), or as choices in nancial planning and in health (includ-
a catchup strategy to complete earlier postponed ing end-of-life) decisions. Besides a fundamental
credentials (especially among women and minor- sense of the futurea future time perspective
ities) (Elman and ORand 2004; Jacobs and King framed by projected time horizonsdecision-
2002; Maralani 2011), or because within-rm making in both areas has two other similar cogni-
training programs have been severely reduced in tive requirements: some level of numeracy and
response to economic restructuring (Elman and some basic conceptual understandings specic to
Weiss (forthcoming). Even older age groups each situation. Beginning with the idea of future
exhibit nontrivial participation rates in work- time perspectives, some scholars argue that future
related education: one-third of adults 4554 and planning of the life course somehow goes against
one-fourth of adults aged 5564 have annually our nature (see Carstensens (2009) argument that
enrolled in one or more work-related courses nature abhors a 401 k) and, following this
since 1990 (Levesque et al. 2008). assumption, proposes that it is time to stop refer-
Finally, while ages 62 and 65 continue to act ring to old age and instead to long life both in
as major thresholds for retirement for signicant popular and academic discourse. Carstensens
proportions of workers, those working later than larger body of work has been distinguished by her
the normal full retirement age for Social Security nearly unique interest as a life course researcher
has steadily increased over the decade with more on changes in temporal orientations with age,
and more working beyond age 70 (ORand 2011). which argues that older persons view the future as
Similarly, a new pattern of post-retirement limited and consequently prioritize more immedi-
employment, called unretirement is also ate, emotionally meaningful goals while younger
emerging (see Maestas 2010). Still another pat- persons view the future as more open-ended and
tern represented by one-fourth of the population prioritize instrumental goals to navigate the future
over age 50 is one of a mix of unemployment, (Carstensen 2006 for summary). Still, she proba-
underemployment, and disability in the years bly does not deny that variations in instrumental
leading up to retirement (ORand 2011; ORand future time perspectives probably exist within age
and Hamil-Luker 2011). Accordingly, the indus- groups as well as across them.
trial life course with an ordered sequence of Long traditions of research on differences in
education-work-retirement may no longer reect future time perspectives can be found especially
actual work lives. Employment security for the in psychology and in economics. In psychology,
majority of workers was the hallmark of indus- the famous marshmallow studies started by
trial societies, and the linchpin of the social con- Robert Mischel in the 1960s established that
Lags and Leaps: The Dynamics of Demography, Economy and Policy and Their 713

children differed in their future time perspectives (compound interest, ination, stock diversica-
by their relative willingness to defer gratication tion) in several surveys (starting with the HRS in
by accepting postponed greater rewards instead 2004) and nd that approximately one-third of the
of immediate lesser rewards (e.g. Mischel et al. population answers correctly to all three ques-
1989). Most of this work has been carried out on tions; three-fourths correctly answer the ination
children and adolescents (ORand and Ellis question; two-thirds, compound interest; and half,
1974). In economics, the interest has been in the stock diversication. In the case of the third ques-
study of preferences for rewards where the tion, one-third of respondents answer Dont
repeated nding is that the tendency is to dis- Know. More sophisticated aspects of nancial
count the value of later rewards, and to do so literacy have also been identied. One is the
increasingly with the length of the delay. When understanding of the structure of mortgages and
risks, or uncertainty in the rewards, are added mortgage payment schedules, clearly evident by
to the situation the tendency towards risk aver- its poignant absence during the recent Great
sion is prevalent; study subjects respond more to Recession brought on, in part, as a result of wide-
the probability of loss than to the probability of spread misunderstanding of these structures and
gain (Kahnemann and Tversky 1979). schedules. Another is the understanding of asset
While these literatures are substantial and pricing, and the widespread ignorance of the
more complex than this summary can reect, the inverse relationship between bond prices and
point for our purposes here is that given these interest rates (Lusardi and Mitchell 2014). The
general repeated ndings, new demands for indi- latter has special relevance to portfolios of many
vidual planning in matters of health and wealth aging investors who are confronted with changing
across the life course must confront the observed the mix of bonds and equities in their efforts to
central tendency to resist such planning and the avoid outliving their savings.
patterns of heterogeneity in the population to do The recognition of the new demand for nan-
so from childhood forward. Yet, choices have to cial literacy is not limited to U.S. popula-
be made. Adults who need to plan for their retire- tions (Organization of Economic Cooperation and
ments may also be managing the cost of educa- Development 2005). Studies conducted in other
tion for their children (or for themselves) and countries nd low levels of nancial literacy, but
have to navigate the world of educational loans, variations across countries in the familiarity of
choices between loans, and trade-offs between these concepts based on national economic expe-
retirement accounts and educational savings riences and policies. For example, German and
accounts or loans. The seemingly odious burden Dutch respondents understand ination better
of nancial life course choices is a paramount than the other two concepts; the Japanese under-
feature of modern life. stand deation; and Russians scored lowest of all
Numeracy and the specic conceptual require- in the questions. The urgency of these observa-
ments of nancial and health literacy are the other tions was recognized by the OECD Programme
cognitive elements of these life course risks. In for International Student Assessment (PISA)
the case of nancial literacy, the capacities to which added nancial literacy questions in 2012
understand and to calculate such constructs as to its annual educational assessment of 15 year
compound interest, ination and risk diversica- olds (Lusardi and Mitchell 2014).
tion in investment portfolios are fundamental Numeracy and conceptual knowledge vary in
skills for investment (retirement) planning. the health domain as well. The Institute of
Accounting for interest and ination are also criti- Medicine (2004) denes health literacy as the
cal for the management of debt (i.e. the costs of capacity to obtain, process and understand basic
borrowing, making minimum payments, incur- health information and services needed to make
ring fees for late payments and over-limit expen- health decisions. Research in this area suggests
ditures, etc.). Lusardi and Mitchell (2013, 2014) that health literacy is the product of both individ-
developed measures of some of these phenomena ual capacities and motivations and the resources
714 A.M. ORand and A. Bostic

provided from multiple environments, including some of the mechanisms through which educa-
the health system and other sectors involved with tion inuences health and wealth outcomes. But,
insurance and related supports of health (see some evidence regarding nancial literacy sug-
Baker 2006 for review). Some aspects of medical gests that other contexts also serve to transmit
literacy have been well-known for decades, knowledge that lends itself to greater literacy.
among them the literacy associated with patient Workplace experiences appear to be important
communication of symptoms as well as under- for nancial literacy (McArdle et al. 2009) espe-
standing written medical documents and pre- cially to the extent that workers are covered by
scription labels and complying with prescriptions and educated or motivated to understand their
and directives. pensions. Workers covered under traditional
These core long-term problems have been dened benet (DB) pensions had little under-
identied as related to basic vocabulary and the standing of their pensions other than their entitle-
numeracy associated with following prescribed ments upon retirement. Workers in occupations
treatment regimes. For these reasons clinical tests covered by dened contribution (DC) plans that
of medical literacy have been developed to require more year-to-year decision-making about
emphasize vocabulary and numeracy, e.g. The their plans have, on average, greater nancial lit-
Rapid Assessment of Literacy in Medicine eracy than their older counterparts, but still vary
(REALM) and the Test of Functional Health widely in their levels of understanding (Helppie
Literacy in Adults (TOFHLA). However, more et al. 2010). Because younger workers have been
sophisticated and long-term issues of medical lit- less motivated to participate in DC plans, pension
eracy have emerged. Faced with technical, proba- legislation in the 2000s introduced mandatory
bilistic diagnosis and prognosis information enrollment in DC plans with opt-out options.
regarding health, choices are presented to patients However, the behavioral inertia of aging workers
who often bring little, if any, prior probabilistic not to make changes in their investment mixes
knowledge and understanding to make health once they are enrolled is another indicator of
care decisions. Decisions regarding health insur- nancial illiteracy that has led to devastating
ance coverage are equally challenging and often losses during economic downturns or market
beyond reach. Relatedly, the emergence of pre- dislocations.
ventive medicine is centrally a life course issue. A nal reference here must be made to the role
As in the case of nancial literacy, choices (many of non-cognitive factors or what some economists
life style or health behavior choices) regarding call soft skills (Heckman and Kautz 2012;
health maintenance and disease prevention now Shanahan et al. 2014) on life course outcomes.
pervade the life course, not just the older years. Personality matters for economic achievement
Not coincidentally, the Educational Testing and health maintenance. Conscientiousness and
Service has developed a Health Activities self-control, specically, are persistent individual
Literacy Scale (HALS) that measures prose, traits that appear to be particularly important. An
numeracy, and knowledge in ve areas (health exemplary study in this regard has demonstrated
promotion, health prevention, disease prevention, the long-term impact of measured childhood con-
health care and maintenance, and systems scientiousness and self-control on adult health.
navigation). The Dunedin Longitudinal Study has recently
By all accounts, a key correlate of nancial reported that credit scores are signicantly associ-
and health literacy is level of education (years of ated with cardiovascular disease risk, and that
schooling). The robust correlations between level both are directly inuenced by human capital fac-
of education and economic and health outcomes tors that include educational attainment, cognitive
across the life course are perhaps the most well- ability and self-control (Israel et al. 2014). It fol-
established in the epidemiological literature. But lows from these ndings that both nancial and
the correlations are not sufcient explanations. health literacy are probably linked to non-cogni-
Financial and health literacies may represent tive as well as cognitive factors.
Lags and Leaps: The Dynamics of Demography, Economy and Policy and Their 715

3.3 Life Course Implications destinations countriestend to be positively


of Migration selected, that is they are healthier, more talented
as measured by educational and occupational
Migration across countries and regions for eco- backgrounds, and intrinsically more agentic by
nomic and political reasons is growing. According choosing to migrate. Consequently, it is argued
to the UN Population Fund, approximately3 % of that they can catch up to and, in some cases, over-
the worlds population in 2010 lived outside of take the native population (in educational attain-
their countries of origin. International migrants ment or in earnings, as examples). However,
are typically dened by organizations that docu- catching up and overtaking are not the typical
ment these movements as persons living for 1 experiences of all U.S. immigrant groups. In the
year or longer in countries other than the ones in U.S. context the debate over the Hispanic para-
which they were born. They include foreign dox confronts this issue directly. The Hispanic
workers, international students, refugees and paradox is based on observations over two
their families. According to the Pew Research decades of the apparent health advantages among
Center, the general pattern of migration is over- Hispanic groups compared to their white coun-
whelmingly to the U.S. and other wealthy coun- terparts. This was initially (and intuitively) con-
tries (69 %) by a growing share of persons born sistent with the selectivity hypothesis. However,
in middle-income nations such as India and closer scrutiny of the elements of selection
Mexico (60 %), as classied by the Pew Research among Hispanic groups in the recent decade has
Center (2013).1 We add, however, that immi- uncovered the diverse experiences of different
grants to the U.S. from Asian countries come Hispanic groups and the role of immigrant cohort
from higher income groups in their native survival underlying the paradox. When the
countries. Hispanic population is differentiated into the
The scale of this phenomenon alone is wor- subgroups of Mexicans, Puerto Ricans, Cubans,
thy of serious consideration by social scien- and other Hispanic groups (Central and South
tists. However, from the vantage point of life Americans) different patterns emerge. When
course research in the context of population mortality is the outcome of interest, only foreign-
aging patterns, the long- and short-term inux born Mexican and non-Puerto Rican and non-
of younger adult immigrants introduces new Cuban Hispanics exhibit an advantage. And in
patterns of heterogeneity and inequality. the case of Mexicans this advantage is signi-
Selection and survival of immigrant groups is a cantly attributable to the return migration, or the
growing focus of attention in life course loss, of less healthy migrants from the study
research, though facing analytical challenges (Palloni and Arias 2004).
associated with causality. This selection-sur- These kinds of observations have raised cau-
vival focus, which arguably can be considered tions regarding standard research designs and
a supply-side approach, tends to ignore modes of data analysis in immigration research
demand side factors that vary widely across across countries. Patterns of return migration and
countries and have independent implications other bases of attrition are now taken more seri-
for both selection and survival. ously (Bulatao and Anderson 2004). Research
U.S. and European studies suggest that immi- designs that depend on data collection on vari-
grantsthose migrants who seek to settle in their ables measuring immigrant traits only and ignore
information about the origin communities and
1
According to the World Bank, India is also dened as a families of immigrants are being challenged. The
middle-income economy, albeit a lower-middle-income obstacles to pursuing alternative designs to col-
economy. The World Bank classies a country as lower lect data on both sides of the border are daunting,
income if the GNI falls below $1,045. We follow the World
but successful examples exist. Such examples
Banks guidelines, which have also been adopted by the
Luxembourg Income Study (http://www.lisdatacenter.org/ include the Mexican Family Life Survey
working-papers/country-income-classications/). (MxFLS) and the Mexican Health and Aging
716 A.M. ORand and A. Bostic

Study (MHAS). The MxFLS includes a number during an economic downturn constrains immi-
of socioeconomic, health and fertility indicators grant incorporation into the economy); language
over time in Mexico during the rst wave and fol- barriers (immigrants arriving in host communi-
lows respondents in subsequent waves including ties where the language of their home country is
those who emigrated to the U.S. (http://www. spoken benet from more support for integra-
ennvih-mxfls.org/english/introduccion.html ). tion); sector (or region) of employment (employ-
The MHAS focuses on health outcomes among ment sectors and regions with higher demand for
Mexican-born respondents both in Mexico and workers who match these needs provide earlier
the U.S. offering multiple comparisons regarding opportunities for integration); cohort category
selection and outcomes (http://www.mhasweb. (immigrants admitted as workers as opposed to
org/StudyDescription.aspx). reunited family members or political refugees
The more focused study of destination con- have an advantage); and tenure in destination
texts of host communities as sources of heteroge- country (longer tenure facilitates integration and
neity and inequality is gaining currency. Besides achievement). These contingencies are nontrivial
the economic opportunities and social costs of insofar as they can differentiate immigrant expe-
this demographic behavior, some environments riences and life course outcomes. For example,
into which migrants arrive present cultural- apropos the earlier review in this essay about
political obstacles related to discrimination, womens educational advantage, immigrant
racialization and social exclusion. These experi- women, who now represent about half of all
ences can have negative effects on life course immigrants in advanced countries, benet more
outcomes. Different immigrant groups arrive from educational attainment, while men benet
with varying levels of education and nancial and more from overcoming language barriers (Adsera
medical literacy that select them into different and Chiswick 2007).
cumulative pathways of integration and mobility The global perspective that this essay has
in their new home countries. However, most lack adopted requires that some attention be paid to
nation-specic institutional knowledge, leading some national-level factors of the major host
to social exclusion. The lengthy, costly and chal- countries besides their global economic posi-
lenging process of becoming a citizen usually tions. One category of national level factors is
limits the level of immigrant political incorpora- cultural and represents variations in how sym-
tion. Immigrants, therefore, often lack the same bolic ingroup-outgroup boundaries serve to
rights as natives and face greater challenges to frame attitudes towards immigrants. A recent
confronting anti-immigrant sentiment and poli- study, using the European Social Survey of 21
cies through traditional means (like voting). This countries on this specic topic deserves attention
likely has spillover effects, particularly economic by life course researchers (Bail 2008). An item in
effects. For example, immigrants are typically the survey asked respondents to evaluate the
less incorporated into political institutions such desirability of a hypothetical immigrant on six
as the welfare state. Both visa type and country of traits: race/ethnicity; religion; language; culture
origin is profoundly inuential on access to the (commitment to the host countrys way of life);
welfare state. This lack of access has been linked education; and occupation. Using a fuzzy-set
to higher rates of poverty among immigrants analysis based on the distances of specic coun-
(Sainsbury 2012). try averages from all-country averages on each
Supply-side characteristics are variously asso- item, three congurations of countries emerged:
ciated across studies with a number of alternative one is represented mainly by southern and east-
contextual (demand-side) contingencies, includ- ern European countries with the most recent
ing cohort quality (immigrant groups vary in edu- immigration inux and exhibited relatively stron-
cational, occupational and class backgrounds ger religious, racial, educational and occupation
giving some relative advantage from entry); eco- boundaries; the second represented mainly the
nomic cycle of host country upon arrival (arrival core of Western Europe with historically longer
Lags and Leaps: The Dynamics of Demography, Economy and Policy and Their 717

immigration histories from Africa, Asia and the migration patterns, just as it has relocated jobs
Middle East and exhibited stronger language, once part of the manufacturing sectors of
culture and educational boundaries and weaker advanced countries to middle-income and devel-
racial and religious boundaries; the third repre- oping countries. However, it is not clear yet how
sented Scandinavian countries and exhibited rel- the increased frequency and severity of economic
atively weaker boundaries on all items. Timing of crises like the Great Recession will inuence
immigrations, origins of immigrants and other migration and host country policies.
factors varied in importance for the symbolic
boundaries differentiating the three congura-
tions. Citizenship policies of the host country 4 Summary and Conclusions
congurations did vary. Arguably, these are the
institutional extensions of cultural preferences. Our charge as authors in this volume is to think
They include the importance of laws regulating about fruitful questions for life course research
antidiscrimination, naturalization, family that lie ahead. This essay proposes that the life
reunion, long-term residence, and labor market course be placed within the larger global context
inclusion (Bail 2008). While this study did not to understand the forces of history and social
intend to go further in examining the impact of change that bear upon it. The dynamics of social
these laws on immigrant outcomes, evidence change are characterized by an asynchrony of
from other policy analyses have. Sainsbury leaps and lags, which reect rapid forces of social
(2012) has examined immigrant rights across change met by institutional-cultural forces that
welfare states. Adding criteria for inclusive ver- resist or hamper change, respectively. In the
sus restrictive immigration policies to Esping- twenty-rst century the life course is caught in
Andersens widely applied typology (1999), she the rough currents generated by these leaps and
has among the most detailed comparative analy- lags. Population aging and global nancialization
ses on this subject. Among her ndings are (1) a are the major accelerating forces for change over
dualization is evident across welfare states in recent decades. Persistent policy legacies across
which immigrants have lower access to major countries and regions lag behind these changes.
insurance and pension benets primarily as a These dynamics raise new questions for life
result of lower earnings; (2) immigrant house- course research.
holds have access to unemployment and family The life course itself is not a rigid institution
benets and social assistance due to their precari- that is immutably ordered by age or social status.
ous status in the employment systems; and (3) the Instead, it is perhaps best characterized as a
immigrant-native poverty gap varies across wel- dynamic and cumulative manifold phenomenon
fare states closely following the Esping-Andersen that develops with time within diverse contexts of
classications; the widest gaps are in liberal social change as individuals must respond to their
regimes and narrowest in social democratic. material and cultural conditions in the conduct of
A global perspective for studying the immi- their lives. The manifold characterization stems
grant life course that links immigrant characteris- from the idea that lives are lived across simulta-
tics with host country cultures and institutions is neously experienced domains, i.e. health, family,
overdue and becoming more feasible. Cross- education, work, leisure (among others), that are
national and comparable survey data that include interdependent and mutually inuential. The
or target immigrant groups and that are collected cumulative characterization captures the patterns
and linked to contextual events and factors are of path dependency of the life course which is
the standard to aspire to. Similarly, as biomedical comprised of chains of sequentially contingent
data are added, the triangulation of these data transitions. Early life conditions provide oppor-
sources brings the study of the immigrant life tunities and constraints that select individuals
course to a higher level. Finally, long-term eco- into cumulative manifold pathways through life,
nomic restructuring has contributed to these in which earlier resources and choices inuence
718 A.M. ORand and A. Bostic

later ones. Individuals do make active choices, suggest that educational attainment may not con-
especially in contexts where multiple and often tinue to have the same putative effects on the
competing options exist. These choices are con- health and mortality of women going forward.
strained by opportunities but also driven by Similarly, another recent change in the tempo-
cumulatively developing identities and personali- ral organization of the life course is that across
ties. Therefore, agency in the life course also countries younger populations are remaining in
contributes to heterogeneity over time. educational institutions longer and older popula-
These dynamics are evident across many tions are returning to school at later ages. Global
aspects of the life course, but three are selected restructuring has delayed opportunities for the
for consideration here as foundational for future youngest workers and displaced (or threatened to
research. The rst is diversity in the temporal displace) older workers whose skill-sets do not
organization of lives. Longitudinal survey data match the new landscape of work. Hence, the tri-
linked to biological markers, administrative phasic organization is replaced by extended
records, ecological momentary assessments, and work-education sequences across the life course.
other innovation data collection strategies is The second aspect of the life course that
revealing the long, broad, and deep contours of requires future attention relates to new life course
lives that produce heterogeneity across contexts. risks in the form of increased demands for indi-
Observations over decades of lives that capture vidual and family decision-making based on
events and conditions related to health, economic knowledge that is not widespread in the general
resources, educational achievements, family population and only moderately associated with
transitions, work careers, and other life contexts educational attainment. Financial and medical
reveal that the life course is not organized by literacy have colonized household decision-
strict schedule or biosocial script. The pace of making on a daily basis. The nancialization of
aging varies widely as diverse life conditions and the global economy pervades daily decision-
personalities interact over time. making regarding the management of credit, con-
The timing and duration of education in lives sumption, debt, saving and investing.
offers a critical lens on this phenomenon. Yet, studies in the U.S. and elsewhere reveal
Educational attainment, usually measured as startlingly low levels of nancial literacy. Basic
years of schooling, is the usual suspect to pre- knowledge of compound interest, ination and
dict diverse life course outcomes. However, the equity investments is restricted in most cohorts to
conceptualization of education lags behind a minority, although younger cohorts maturing in
observed patterns of learning and does not fully the new environment are more informed on some
capture the cognitive and non-cognitive aspects of these matters. In this climate, rates of indebt-
of social life. The timing and sequencing of edu- edness and bankruptcy have increased in a trend
cation, family and work do not conform to the preceding but dramatically illuminated by the
twentieth century tri-phasic model. And, status recent Great Recession. Life course research
categories such as gender and race-ethnicity points to both cognitive and non-cognitive factors
often have greater variations within, than differ- that inuence nancial literacy. Medical literacy
ences from, comparison groups. The gendered is also implicated in this trend since medical
life course, for example, is changing rapidly. The technologies and limited sources for nancing
spread of educational opportunities to women medical interventions require individuals to make
appears to account for changes in patterns of important life-and-death decisions for which
childbearing, marriage and workoften referred most are not equipped. Faced with technical
to as the second demographic transition. diagnostic information regarding health, choices
However, dramatic increases in educational are presented to patients who bring little, if any,
attainments among women across societies are prior knowledge or understanding to health care
associated with global job uncertainties and new decisions. Decisions regarding health insurance
stresses, including risky health behaviors, which coverage are equally challenging and often
Lags and Leaps: The Dynamics of Demography, Economy and Policy and Their 719

beyond reach. Policies for improving nancial Blossfeld, H.-P., Buchholz, S., & Kurz, K. (Eds.). (2011).
Aging populations, globalization, and the labor mar-
and medical literacy lag far behind the pressures
ket: Comparing late working life and retirement in
for decision-making in these areas. modern societies. Cheltenham: Edward Elgar.
Finally, the third major trend(s) that should Blossfeld, H.-P., Kilpi-Jakonen, E., de Vilhena, D. V., &
motivate future life course research is migration Buchholz, S. (forthcoming). Adult learning in modern
societies. London: Edward Elgar.
and immigration and their life course implica-
Boardman, J. D., Saint Onge, J. M., Rogers, R. G., &
tions. Migration across countries and regions for Denney, J. T. (2005). Race differences in obesity: The
political and economic reasons is a major global impact of place. Journal of Health and Social
phenomenon. Selective, contextual and agentic Behavior, 46, 229243.
Buchmann, M. C., & Kriesi, I. (2011). Transition to adult-
factors initiate the migration process in the life
hood in Europe. Annual Review of Sociology, 37,
course and are consequential for later life, and 481503.
subsequent intergenerational well-being. Besides Bulatao, R. A., & Anderson, N. B. (Eds.). (2004).
the economic opportunities and social costs of Understanding racial ethnic differences in health in
late life: A research agenda. Washington, DC:
this demographic behavior, the institutional
National Academies Press.
environments into which migrants arrive present Cain, A. E., Depp, C. A., & Jeste, D. V. (2009). Ecological
cultural-political obstacles related to discrimina- momentary assessment in aging research: A critical
tion, racialization and social exclusion. These review. Journal of Psychiatric Research, 43(11),
987996.
experiences have negative effects on the life
Carstensen, L. L. (2006). The inuence of a sense of time
course, but go unmeasured (and perhaps some- on human development. Science, 30, 19131915.
times remain unmeasurable). Within the frame- doi:10.1126/science.1127488.
work we propose here, it is also plausible to Carstensen, L. L. (2009). A long bright future: An action
plan for a lifetime of happiness, health and nancial
consider the reciprocal impact of migration and
security. New York: Public Affairs.
immigrants on their new political and economic DiPrete, T. A., & Buchmann, C. (2013). The rise of
environments, especially as their relative share women: The feminine advantage in education and
of the total population increases. These patterns what it means to American schooling. New York:
Russell Sage Foundation.
provide the opportunities to examine both the lag
Elman, C., & ORand, A. M. (2002). Perceived job inse-
of policies and, alternatively, their potential curity and entry into work-related education and train-
transformations in the face of demographic ing among adult workers. Social Science Research,
recomposition (though probably over long peri- 31(1), 4976.
Elman, C., & ORand, A. M. (2004). The race is to the
ods) and resulting life course outcomes.
swift: Socioeconomic origins, adult education and
wage attainment. American Journal of Sociology,
110(1), 123160.
Elman, C., & Weiss, F. (forthcoming). Adult educational
References participation and implications for employment in the
US context. In H.-P. Blossfeld, E. Kilpi-Jakonen, D. V.
Adsera, A., & Chiswick, B. R. (2007). Are there gender de Viilhena, & S. Buchholz (Eds.), Adult learning in
and country of origin differences in immigrant labor modern societies. Cheltenham: Edward Elgar.
market outcomes across European destinations. Esping-Andersen, G. (1999). The social foundations of
Journal of Population Economics, 20, 495526. post-industrial economies. New York: Oxford
Bail, C. (2008). The conguration of symbolic boundaries University Press.
against immigrants in Europe. American Sociological Eurostat. (2014). Unemployment statistics. Epp.eurostat.
Review, 73, 3759. ec.europa.eu/statistics_explained/index.php/.
Baker, D. W. (2006). The meaning and the measure of Downloaded August 2014.
health literacy. Journal of General Internal Medicine, Farber, H. (2011). Job loss in the great recession:
21, 878883. doi:10.1111/j.1525-1497.2006.00540.x. Historical perspective from the displaced workers sur-
Berg, I., & Kalleberg, A. (Eds.). (2001). Sourcebook of vey, 19842010 (Working Paper 17040). National
labor markets: Evolving structures and processes. Bureau of Economic Research.
New York: Kluwer Academic/Plenum Publishers. Graff, H. J. (1995). Conicting paths: Growing up in
Blossfeld, H.-P., Klijzing, E., Mills, M., & Kurz, K. America. Cambridge, MA: Harvard University Press.
(Eds.). (2005). Globalization, uncertainty and youth in Hacker, J. (2006). The great risk shift: The assault on
society. London: Routledge. American jobs, families, health care and retirement
720 A.M. ORand and A. Bostic

and how you can ght back. New York: Oxford ORand, A. M. (2011). The devolution of risk and the
University Press. changing life course in the United States. Social
Heckman, J. J., & Kautz, T. D. (2012). Hard evidence for Forces, 90(1), 116.
soft skills. Labour Economics, 19, 451464. ORand, A. M., & Ellis, R. A. (1974). Social class and
Helppie, B., Karpinos, K. A., & Willis, R. J. (2010). social time perspective. Social Forces, 53, 5362.
Occupational learning, nancial knowledge, and ORand, A. M., & Hamil-Luker, J. (2011). Late employ-
accumulation of retirement wealth (MRRC Working ment careers, transitions to retirement and retirement
Paper 237). Michigan Retirement Research income in the U.S. In H.-P. Blossfeld, S. Buchholz, &
Consortium. Ann Arbor: University of Michigan. K. Kurz (Eds.), Aging populations, globalization
Institute of Medicine. (2004). Health literacy: A prescrip- and the labor market (pp. 283307). Cheltenham:
tion to end confusion. Washington, DC: National Edward Elgar.
Academies Press. Organization of Economic Cooperation and Development.
Israel, S., Caspy, A., Belsky, D. W., Harrington, H., (2005). Improving nancial literacy: Analysis of
Hogan, S., Houts, R., Ramrakha, S., Sanders, S., issues and policies. Washington, DC: Organization of
Poulton, R., & Moftt, T. (2014). Credit scores, car- Economic Cooperation and Development.
diovascular disease risk, and human capital. Organization of Economic Cooperation and Development.
Proceedings of the National Academy of Sciences, (2011). Education at a glance: Access to education,
111(48), 1708717092. participation and progression. Washington, DC:
Jacobs, J. A., & King, R. B. (2002). Age and college com- Organization of Economic Cooperation and
pletion: A life history analysis of women aged 1544. Development.
Sociology of Education, 75(3), 211230. Palloni, A., & Arias, E. (2004). Paradox lost: Explaining
Kahnemann, D., & Tversky, A. (1979). Prospect theory: the Hispanic adult mortality advantage. Demography,
An analysis of decision under risk. Econometrica, 41, 385415.
47(2), 263292. Pew Research Center. (2013). Changing patterns of global
Kalleberg, A. L. (2009). Precarious work, insecure work- migration and remittances: More migrants in U.S. and
ers: Employment relations in transition. American other wealthy countries; more money to middle
Sociological Review, 74(1), 122. income countries. http://www.pewsocialtrends.org/
Lestaeghe, R. C. (2010). The unfolding story of the sec- les/2013/12/global-migration-nal_12-2013.pdf
ond demographic transition. Population and Poterba, J. M. (2014). Retirement security in an aging
Development Review, 36(2), 211251. population. American Economic Review, 104(5), 130.
Levesque, K., Laird, J., Hensley, E., Choy, S. P., Cataldi, Preston, S. H., Stokes, A., Mehta, N. K., & Cao, B. (2014).
E. F., & Hudson, L. (2008). Career and technical edu- Projecting the effects of changes in smoking and obe-
cation in the United States: 19902005, NCES 2008- sity on future life expectancy in the U.S. Demography,
03. Washington, DC: U.S. Department of Education. 51, 2749.
Lusardi, A., & Mitchell, O. S. (2013). Older adult debt Riley, M. W., Kahn, R. L., & Foner, A. (Eds.). (1994). Age
and nancial frailty (Michigan Retirement Research and structural lag: Societys failure to provide mean-
Center Working Paper 2013-291). Ann Arbor: ingful opportunities in work, family and leisure.
University of Michigan. New York: Wiley.
Lusardi, A., & Mitchell, O. S. (2014). The economic Royalty, A. B. (1998). Job-to-job and job-to-
importance of nancial literacy: Theory and evidence. nonemployment turnover by gender and education
Journal of Economic Literature, 52(1), 544. level. Journal of Labor Economics, 16, 392343.
Maestas, N. (2010). Back to work expectations and real- Sainsbury, D. (2012). Welfare states and immigrants
izations of work after retirement. Journal of Human rights: The politics of inclusion and exclusion.
Resources, 45(3), 718748. New York: Oxford University Press.
Maralani, V. (2011). From GED to college: Age trajecto- Sawhill, I. V. (2014). Generation unbound: Drifting into
ries of nontraditional educational paths. American sex and parenthood without marriage. Washington,
Educational Research Journal, 34(4), 10581090. DC: Brookings Institution Press.
McArdle, J. J., Smith, J. P., & Willis, R. (2009). Cognition Shanahan, M. J., Hill, P. L., Roberts, B. W., Eccles, J., &
and economic outcomes in the health and retirement Friedman, H. S. (2014). Conscientiousness, health and
survey (NBER Working Paper 15266). Boston: aging: The life course personality model. Developmental
National Bureau of Economic Research. Psychology, 50(Special Issue), 14071425.
McDaniel, A. (2009). Cross-national gender gaps educa- Tomaskevic-Devey, D., & Lin, K.-H. (2011). Income
tional expectations: The inuence of national level dynamics, economic rents, and the nancialization of
gender ideology and educational systems. Chicago: the U.S. economy. American Sociological Review,
University of Chicago Press. 76(4), 538559.
Mischel, W., Shoda, Y., & Rodriguez, M. L. (1989). Delay Uhlenberg, P. (Ed.). (2009). International handbook of
of gratication in children. Science, 244, 933938. population aging. New York: Springer.

Anda mungkin juga menyukai