Anda di halaman 1dari 18

The Macroeconomic Context of Job Stress

Author(s): Rudy Fenwick and Mark Tausig


Source: Journal of Health and Social Behavior, Vol. 35, No. 3 (Sep., 1994), pp. 266-282
Published by: American Sociological Association
Stable URL: https://www.jstor.org/stable/2137280
Accessed: 01-04-2019 10:45 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

American Sociological Association is collaborating with JSTOR to digitize, preserve and


extend access to Journal of Health and Social Behavior

This content downloaded from 113.210.65.127 on Mon, 01 Apr 2019 10:45:31 UTC
All use subject to https://about.jstor.org/terms
The Macroeconomic Context of Job Stress*

RUDY FENWICK
MARK TAUSIG
University of Akron

Journal of Health and Social Behavior 1994, Vol. 35 (September):266-282

Using data from the 1973-1977 Quality of Employment Panel Study, we test a
model that conceptually links research on macroeconomic causes of stress with
research on job structure causes of stress among employed workers. Results from
LISREL 7 (Joreskog and Sdrbom 1989) indicate that, while both macroeconomic
and job structure variables have significant cross-sectional and longitudinal effects
on stress, the macroeconomic effects are almost entirely indirect in their effect on
job structures. In particular, higher occupational unemployment rates increased
stress and lowered life satisfaction indirectly through reduced decision latitude and
increased job demands. Overall, results suggest that macroeconomic changes,
such as recessions, can affect individual stress because they lead to changes in
routine job structures that represent increased and continued exposure to stressful
conditions.

In this study we describe and test a model emphasizes the direct relationship between
that relates structural features of the macro- aggregate features of the economy and
economy to the everyday job experiences of aggregate indicators of mortality and morbid-
workers and to their individual feelings of ity. In contrast, a "work stress" literature
stress and related feelings of life satisfaction. documents a relationship between characteris-
In doing so we follow Pearlin's (1989) tics of jobs (which are viewed as stressors)
argument that the sociological study of the and the health of individual workers. This
stress process entails an examination of ". . . paper seeks to synthesize and go beyond these
the presence of similar types and levels of literatures by assessing the manner in which
stress among people who are exposed to macroeconomic structure and change affect
similar social and economic conditions individual levels of stress via effects on
(Pearlin 1989:242). working conditions that are themselves re-
Presently, in separate research traditions, lated to stress. We argue that macroeconomic
both macroeconomic structures and job struc- structure and change play an important role in
tures have been shown to have direct effects creating the concrete work environments
on stress. An "economic stress" literature which routinely affect worker exposure to
stressful work conditions.

*The data utilized in this paper were made


available by the Inter-University Consortium for
Political and Social Research. The data for the Economic Stress Research
Quality of Employment Panel Study, 1973-1977,
were originally collected by Robert P. Quinn and Research by Brenner (1976, 1984, 1987)
Graham Staines for the Employee Standards Marshall and Funch (1979) and especially
Administration of the U.S. Department of Labor.
Catalano, Dooley, and their associates (Cata-
The authors wish to thank Ralph Catalano, David
lano and Dooley 1977, 1979; Catalano,
Dooley, Ronald Kessler, Arnold Linsky, McKee
Dooley, and Jackson 1985; Dooley and
McClendon, James Wells, and the anonymous
reviewers for their comments and suggestions on Catalano, 1984; Catalano, Rook, and Dooley
an earlier version of this paper. Address correspon- 1986) has described a relationship between
dence to Rudy Fenwick, Department of Sociology, features of the macroeconomy (generally,
University of Akron, Akron, OH 44325-1905. unemployment rates) and indicators of stress-

266

This content downloaded from 113.210.65.127 on Mon, 01 Apr 2019 10:45:31 UTC
All use subject to https://about.jstor.org/terms
JOB STRESS 267

related poor health (psychiatric hospital Again, however, the mechanism for explain-
admissions, cardiovascular illness, mortality). ing individual outcomes is not empirically
The argument is derived from a human examined.
ecology approach in which changes in
unemployment rates, for example, increase
risk of exposure to negative work and Work Stress Research
financial-related events or reduce social
tolerance of deviant behavior. Either greater There is, however, a separate research
exposure to stressors or reduced tolerance literature on the relationship between work
lead to higher observed aggregate rates of structure and health that focuses on the
morbidity or mortality (Catalano 1989). individual. By and large this literature
However, studies conducted at an aggregate employs a stress-illness model that relates
level run the risk of committing an "ecologi- workplace stressors to health outcomes
cal fallacy" (Catalano and Dooley 1983) in (Holt 1982). Stressors examined include
which it might erroneously be assumed that those in the physical work environment,
persons experiencing greater morbidity are organizational properties or work, and those
also those who were exposed to the economic arising from job change and relationships
stressor (i.e., unemployment). with supervisors and co-workers. (House
To deal with this problem, individual level 1980; Selye 1976; Kahn 1974; Kasl 1978;
studies attempt to document a relationship Karasek 1979, 1989; Karasek and Theorell
between the direct experience of unemploy- 1990).
ment, which is seen as a stressor, and greater Two models of the etiology of stress in
physical and mental distress (Pearlin and the workplace are used to explain why these
Schooler 1978; see Liem and Rayman [1984] stressors arise (Baker 1985). The Person-
and Jahoda [1988] for summaries). At the Environment Fit model proposes that strain
individual level, however, the ecological develops when there is a discrepancy
context (unemployment rates) disappears. between characteristics of the work environ-
Most studies examine either the effects of ment and individual characteristics (French,
unemployment on the unemployed, or they Caplan, and Harrison 1982). A "good" fit
compare matched samples of employed and reduces the perception of strain among
unemployed workers. workers. The Job Demands-Control Model
Catalano et al. (1986) have attempted to argues that inadequate structural capacity
retain the ecological link to individual out- (e.g., control over use of time and decision-
comes by suggesting that workers' percep- making) needed to meet job demands leads
tions of employment security are affected by to perceptions of strain that are, in turn,
unemployment rates and that perceived em- predictive of physiological and psychological
ployment insecurity will be related to in- distress (Karasek, 1979; Karasek and Theo-
creased help-seeking. They suggest that in rell 1990). Empirical assessments of both
contracting economies management may seek the person-environment fit model and the
ways to increase productivity, thereby modi- job demands-control model indicate that
fying work structures. However, they do not occupation and work environment (i.e.,
empirically examine this hypothesis. Instead, demands-control/structural models) are the
they focus on the perceptions of workers- principal sources of observed variations in
perceived employment security- as these are stress rather than individual characteristics
affected by economic conditions (unemploy- (Baker 1985).
ment rates). Neither of these two models is specifically
Brenner (1987) suggests that, when mac- concerned with the origins of work structures
roeconomic conditions force a firm to reduce nor with the way in which the relationship
its labor force, remaining employees will between job demands, decision latitude, and
experience fear of employment loss and occupations are established (Johnson 1989).
destruction of careers as well as increased Moreover, the few studies that use longitudi-
work stress resulting from lower tolerance nal data do not assess the effect of changes in
for error, closer supervision by management work conditions on health (Karasek 1989),
(with loss of autonomy), demand for higher although House et al. (1986) did examine
levels of productivity, and possibly de- changes in "job stress" for their effects on
creases in wages or promotion opportunities. mortality.

This content downloaded from 113.210.65.127 on Mon, 01 Apr 2019 10:45:31 UTC
All use subject to https://about.jstor.org/terms
268 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

Macroeconomic Origins of Stressful Jobs METHODOLOGY

In our view, the economic stress and work Causal Model


stress approaches can be linked by arguing
that macroeconomic forces affect not only The causal model for assessing the effects of
exposure and vulnerability to life events macroeconomic and job structures on stress is
associated with economic change, particularly presented in Figure 1. Following the above dis-
increased unemployment, but also the level of cussion, we hypothesize that the causal rela-
stress resulting from the structure of and tionships are both cross-sectional and longitu-
changes in work roles. Stressful jobs are not dinal. Job structures are functions of
randomly distributed throughout the econ- macroeconomic and organizational structures
omy; rather, they are products of macroeco- and processes -as well as individual character-
nomic structures and forces such as the istics such as race and gender, which we will
economic sector and organizational structure treat as control variables in testing this model
of firms in which the jobs are located and, (see below for the discussion of variables and
perhaps most fundamentally, the economic measurement). In turn, stress measures are func-
markets in which firms and workers compete. tions of these macroeconomic and individual
In keeping with the focus of much of the variables, as well as the job structure measures.
previous economic stress literature (e.g., The Time 2 (T2) measures of stress and job
Brenner 1976, 1984; Brenner and Mooney structures are also functions of their respective
1983; Catalano and Dooley 1977, 1979, Time 1 (TI) measures. The macroeconomic and
1983), we emphasize in this paper the individual variables are assumed to be observed
importance of labor markets, especially as single item exogenous variables at both TI and
measured by unemployment rates. This em- T2.' This model is developed from both the eco-
phasis is also consistent with a basic assump- nomic stress and work stress literatures. Each
tion in much of the literature on work and literature emphasizes different causal paths to
inequality-that it is through labor markets stress. The economic stress literature (e.g., Bren-
that the other macroeconomic forces have ner and Mooney 1983; Catalano and Dooley
their effect on the job structures and rewards 1977, 1979; Catalano et al. 1986) emphasizes
of individual workers (Kalleberg and Berg the direct cross-sectional and longitudinal paths
1987:48-49; Kalleberg 1989). from the macroeconomic variables to stress,
Moreover, these arguments have a longitu- while the work stress literature (e.g., Karasek
dinal dimension in that changes in the market 1979, 1989; Karasek and Theorell 1990; House
positions of workers and firms increase or 1980; House et al. 1986; Kasl 1978) empha-
reduce the stressful characteristics of jobs. sizes the direct paths from job structures to stress
Most significantly, economic downturns (re- Our model assumes that both types of paths are
cessions) lead to deterioration of market significant, but also assumes that the effects of
situations for both. In these situations the macroeconomic variables on job structures are
survival of firms may be threatened, forcing significant, as are their indirect effects on stress
them to reduce production costs by laying off through job structures.
some workers and restructuring the jobs of
their remaining workers in order to maintain
levels of productivity and profits (Starrin et Data
al. 1989). Such restructuring may include
speed-ups or work overload due to reduced Data used to test this model are from the
staffing, reduced margins of error in output, 1973-1977 Quality of Employment Survey
and relaxation of rules governing worker Panel Study. This longitudinal study represents
safety and health (Brenner and Mooney a national probability sample of those in the
1983). In general, this restructuring is de- U.S. labor force 16 years of age or older, work-
signed to pass the increased vulnerability of ing 20 hours or more a week for pay, as of
the firm on to its workers. With higher 1973. Those respondents re-interviewed for the
unemployment rates and thus in a weaker 1977 survey also had to be in the labor force
market position, workers lack the ability to working 20 or more hours a week for pay. For
resist restructuring. For these workers, re- our analysis there were 830 such respondents.
structuring may mean jobs with increasingly Because of sample attrition-especially, respon-
stress-producing characteristics. dents exiting the labor force-and because of

This content downloaded from 113.210.65.127 on Mon, 01 Apr 2019 10:45:31 UTC
All use subject to https://about.jstor.org/terms
JOB STRESS 269

FIGURE 1. Causal Model Predicting Stress From Macroeconomic Job Structure and Control
Variables
TIME 1 TIME 2
(1973) (1977)

MACROECONOMIC MACROECONOMIC
STRUCTURE 1 AND STRUCTURE
CONTROL VARIABLES 2 VARIABLES1

JOB STRUCTURES3 JOB STRUCTURES3

STRESS4 S 4

Macroeconomic structures m
co-workers), economic sector
2 Control variables measured
and T2 is measured at T2.
3 Job structures measured at TI and T2 are job security, decision latitude, and job demands.
4 Stress and life satisfaction are measured at TI and T2.

new entrants into the labor force, the study ispercent at the time of the first wave
interviews in 1973 to 8.5 percent in 1975, and
not entirely representative of the U.S. labor force
in 1977. In particular, the sample overrepre- was still at 7.0 percent at the time of the
sents the percentage of respondents in higher- second wave interviews in 1977.
skilled occupations-e.g., professionals, man-
agers and administrators, and craft workers-
while underrepresenting those in lower-skilled Variables and Measurement
categories-e.g., non-farm labor and service
workers. The sample also overrepresents male Macroeconomic variables. As suggested
workers. above, the most fundamental macroeconomic
Despite this limitation, however, the panel variable potentially affecting job structures
data provide important information to test the and individual stress is the unemployment
model. The QES is one of the few national rate. Here, the unemployment rate is mea-
longitudinal studies of the American labor sured in each of ten broad occupational
force that provides in-depth information on categories used by the U.S. Census Bureau:
job structures as well as psychological (1) professional and technical; (2) managers
functioning. Moreover, it provides before and and administrators; (3) sales workers; (4)
after information on the effects on workers of clerical; (5) craft and kindred; (6) operatives,
the mid-1970s (1975-1976) recession, the except transport; (7) transport equipment
most severe (at the time) since the Great operators; (8) non-farm laborers; (9) service
Depression. Unemployment rose from 4.9 workers; and (10) farm workers (U.S. Bureau

This content downloaded from 113.210.65.127 on Mon, 01 Apr 2019 10:45:31 UTC
All use subject to https://about.jstor.org/terms
270 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

of the Census 1978). Each respondent was ness, dissatisfaction, and "alienation" among
assigned the rate of unemployment for his workers (Blauner 1964; Hodson 1984; Spen-
or
her occupational category that had existed ner 1983), and to increasing intraorganiza-
two years prior to the times of the surveys - tional conflict (Lincoln 1978). We measure
1971 rates for the 1973 survey and 1975 rates size as the natural log of the number of em-
for the 1977 survey. This was done to account ployees working at the respondents' work-
for the lagged effects of unemployment rates sites. The natural log is used instead of the
on stress-related outcomes that have been absolute number because the effects of orga-
found in prior research, with two years nizational size on structure, attitudes, and be-
appearing to be about the optimal lag time havior generally has been found to best fit the
(Brenner and Mooney 1983; Brenner 1987. log function (Stolzenberg 1978).
Other macroeconomic variables that affect Job Structures. Three job structures are an-
job structures and thus the potential for stressalyzed as intervening variables in our model.
include the economic sector of firms, the Each is measured identically at Ti and T2. Job
presence of unions, and the size of the security is measured by a single item asking the
workplace. The economic sector of the firms respondent to evaluate the "security" of their
for which respondents worked was deter- current job. The other job structures, decision
mined by the Census Bureau's 1970 Industry latitude and job demands, are measured indi-
Code. Firms were assigned to one of three rectly through multiple indicators (see Appen-
sectors: (1) periphery, (2) core, or (3) state dix A for the individual items used). These mea-
(see Hodson 1978). These sectors differ as to sures are derived from the job demands-control
the level of concentration and centralization model developed by Karasek (1979).2 Decision
of capital, the organization of production, and latitude is defined by Karasek as the "working
market situations. In the periphery, produc- individual's potential control over tasks and .
tion is organized for low-profit, competitive conduct during the working day" (1979:289-
markets. Firms are vulnerable to the large 90), and includes items which reflect evalua-
number of competitors and to rapid market tions of various aspects of the respondent's job,
changes. Thus, job structures lack stability such as the degree to which they feel the job
and security (Lincoln and Kalleberg 1985). In requires high levels of skill, they learn new
contrast, in the core, production is organized things, and the work is nonrepetitious and cre-
for oligopolistic, high-profit markets, while inative, as well as the degree to which the job
the state sector it is organized by the state allows for freedom and ability to have input
itself; i.e., goods and services are produced about how the job is done. The job demands
directly by the public sector or by private dimension measures "the psychological stres-
capital under contract with the state. Both sors involved in accomplishing workload"
sectors reflect protected market situations (1979:291), and includes respondents' evalua-
with stable and secure job structures, such as tions of how much their jobs require working
job ladders and internal labor markets, that hard and fast, whether they are allowed time to
help reduce worker vulnerability to economic finish the tasks, and whether or not they are
change (Lincoln and Kalleberg 1985). subject to conflicting job demands. Our analy-
Likewise, unions may protect workers fromsis focuses on these dimensions of job struc-
some of the stressful job effects of economictures primarily because they are concerned with
change through collective bargaining agree- the social, rather than the physical aspects of
ments and grievance procedures (Brenner jobs (Karasek 1979), and because they have
1987). Union membership is measured by been found to be remarkably consistent dimen-
whether or not the respondent was a member sions with consistent effects on outcomes
of a labor union or employee association at (Karasek 1979, 1989; Karasek and Theorell
the time of the interviews. 1990).
Larger firms, because of their greater eco- Stress Outcomes. Our dependent variables,
nomic resources and control over markets, may stress and satisfaction, are also measured
also reduce their employees' vulnerability to indirectly through multiple-and identical-
economic change (Edwards 1979). Paradoxi- indicators at TI and T2. Stress measures the
cally, however, organizational (workplace) sizepsycho-physiological symptoms associated
increases the bureaucratic structure of jobs with anxiety, such as sweating hands, pound-
(Stolzenberg 1978), and is related to loss of ing heart and lack of energy (Mirowsky and
job control, increasing isolation, powerless- Ross 1989). Items used in the stress construct

This content downloaded from 113.210.65.127 on Mon, 01 Apr 2019 10:45:31 UTC
All use subject to https://about.jstor.org/terms
JOB STRESS 271

are modified versions of those originally and T2. (Correlations are not presented, but
developed for the Health Opinion Survey are available from the authors upon request.)
(MacMillan 1957). While questions have We also present the unemployment rates for
been raised about the validity of such each occupational category used to describe
measures as indicators of specific psycholog- this component of macroeconomy. (Because
ical disorders (Dohrenwend and Dohrenwend every respondent in a category is assigned the
1969; Wheaton 1978), they are used here to same rate, there are no standard deviations
indicate symptoms of general distress, not and no t-tests.) As indicated by the t-test
specific disorders. Satisfaction measures re- results, these data reveal a number of
spondents' general feelings about their significant changes between 1973 and 1977.
present (at the time of the interview) lives. The overall unemployment rate increases by
Eight items using the semantic differential almost 50 percent, while respondents move
format are used to probe feelings that one's into jobs in the core and out of the periphery,
life is interesting, enjoyable, full, and worth- and out of unions. Job security declines while
while (or the opposite). These items are job demands increase. Respondents' stress
derived from the work of Gurin, Veroff, and increases while their life satisfaction de-
Feld (1960) and Campbell et al. (1976). The creases. Overall, except for the increase in
satisfaction construct is sometimes thought to core sector employment, these changes are
be a dimension of well-being (Campbell et al. representative of the trend toward greater
1976), and therefore the opposite pole from economic uncertainty that characterized the
stress (Mirowsky and Ross 1989). However, 1970s (Blumberg 1980; Bluestone and Harri-
in this analysis satisfaction will be treated son 1982).
simply as an indicator of the extent to which
respondents' expectations match their experi-
Causal Relationships
ence. When the match is poor-i.e., when a
respondent is dissatisfied-it means they feel The heart of our analysis is exploring the
life is miserable, empty and discouraging, if causal relationships-total, direct, and indi-
not stressful (see Appendix A). rect effects-among these macroeconomic
Control Variables. We have included age, (and individual) job structure and stress/
gender, race (White/non-White), and years of satisfaction variables as depicted in Figure 1.
education to control for the effects of First, we analyze the overall (total) effects of
individual characteristics on job structures macroeconomic (and individual) variables on
and stress (Hankin 1990; Kessler and Neigh- stress and satisfaction using LISREL 7
bors 1986; Mirowsky and Ross 1989). These maximum likelihood procedures (applied to
variables are measured exogenously at T1. covariance matrices) to estimate a linear
We have also included a measure of job structural equation model which omits the
change between TI and T2 to control for the intervening job structure variables (Jbreskog
effects of occupational mobility on job and Sorbom 1989).
structure change, since changing job struc- Macroeconomic Effects. Table 2 presents
tures are partly due to respondents changing the unstandardized coefficients from simulta-
jobs. By controlling for this individual neous equations based on this model. The
mobility we are able to measure the pure first two equations present the Tl (1973)
effects on and by changes in job structures per cross-sectional results. Among macroeco-
se. This measure is dichotomous: it groups nomic variables, 1971 occupational unem-
those who either changed employers or who ployment rates have significant but modest
changed jobs within their old firms versus effects on both 1973 satisfaction and stress,
those working in the same jobs in the same and these effects are in the directions
firms. It is measured exogenously at T2. expected: higher unemployment rates reduce
satisfaction and increase stress. Also, as
expected, satisfaction declines with work-
ANALYSIS place size and is greater among state sector
workers. Among the individual control vari-
Table 1 presents the means and standard ables, female workers have greater stress than
deviations for the variables in our model males and stress declines with years of
along with the results of t-tests for mean education.
differences in their measurements between TI The next two equations present the longitu-

This content downloaded from 113.210.65.127 on Mon, 01 Apr 2019 10:45:31 UTC
All use subject to https://about.jstor.org/terms
272 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

TABLE 1. Means and Standard Deviations for Variables in the Model (N = 830)

1973 1977 t-value

Variables X SD X SD

Age (years) 37.59 12.03


Femalea .29 .46
Non-Whitea .10 .30
Education (years) 12.53 2.96
Job Changea .47 .50
Unemployment Rate 4.42 2.70 6.49 3.90 15.45***
(1971, 1975)
Professional and Technicalb 2.90 3.70
Managers and Administratorsb 1.60 3.00
Sales Workersb 4.30 5.80
Clericalb 4.80 6.60
Craft and Kindredb (operatives, 4.70 8.30
except transport ) 7.60 14.70
Transportation Equipment
Operatorsb 4.70 8.50
Non-farm Laborersb 10.80 15.60
Service Workersb 6.30 8.60
Farm Workersb 2.60 3.50
Size (natural log) 4.28 2.10 4.34 2.05 .68
Corea .59 .48 .65 .48 3.30***
Statea .12 .32 .11 .31 - 1.18
Uniona .32 .47 .28 .45 - 2.93**
Job Security 3.28 .94 3.19 .90 - 2.42*
Decision Latitudec 2.91 .64 2.94 .55 1.73
Job Demandsc 2.71 .52 2.88 .51 8.22***
Stressc 1.59 .51 1.84 .52 12.40***
Satisfactionc 5.89 .98 5.67 .98 - 6.28***

a Mean represents percentage of respondents in the particular category.


b Occupation-specific unemployment rates are from the U.S. Bureau of the Census (1978).
c Simple additive scales were constructed for these variables in this table since LISREL 7 does not compute means
and standard deviations for latent constructs.
* p ? .05; ** p < .01; *** p .001.

dinal results for T2 (1977). By including The Role of Job Structures. The above
(controlling for) Ti measures for both macro- results point out that overall unemployment
economic and stress/satisfaction in these rates and changes in those rates affect the
equations, the effects of the macroeconomic stress and satisfaction of employed individu-
variables measured at T2 on the T2 stress/ als. But to what extent are these effects
satisfaction variables are the equivalent of the mediated by the structure of jobs and/or the
effects of change per se net of their initial level of insecurity that high unemployment
levels (Kessler and Greenberg 1981:9-11). levels produce? To answer we again use
The most noteworthy of these longitudinal LISREL 7 techniques to estimate a linear
effects involve T2 (1975) unemployment rates structural equation model which includes
on T2 (1977) satisfaction and stress. These these job structure measures as intervening
results indicate that increases in occupational variables-the full model depicted in Figure
unemployment rates between 1971 and 1975 1. Results are presented in Table 3.
led to reductions in respondents' life satisfac- The first three equations in Table 3 present
tion and increases in their levels of stress in the 1973 cross-sectional effects of macroeco-
1977, over their initial pre-recession levels of nomic and individual variables on job struc-
1973. However, no other macroeconomic tures. Overall, these exogenous variables
variable has effects on T2 satisfaction or explained substantially more variance in
stress, and among individual control vari- respondents' decision latitude (R2 = .370)
ables, only age has an effect-positive-on than in job security (R2 = .056) or job
satisfaction. T2 satisfaction and stress are also demands (R2 = .044). All variables except
and most significantly products of their initial core sector employment had significant ef-
TI measures. fects on decision latitude, while no macroeco-

This content downloaded from 113.210.65.127 on Mon, 01 Apr 2019 10:45:31 UTC
All use subject to https://about.jstor.org/terms
JOB STRESS 273

TABLE 2. Unstandardized Coefficients from Simultaneous Equations Predicting Satisfaction and


Stress from Macroeconomic and Control Variables, 1973 and 1974.

Dependent Variables

Satisfaction 73 Stress 73 Satisfaction 77 Stress 77


Independent Variable (1) (2) (3) (4)

Age .002 .001 .007*** .002


(.002) (.002) (.002) (.002)
Female - .038 .187*** - .049 .047
(.079) (.042) (.066) (.039)
Non-White - .040 .089 - .113 .010
(.139) (.073) (.114) (.067)
Education (years) - .002 - .029*** .002 - .004
(.014) (.007) (.013) (.007)
Unemployment 71 -.018* .013* -.010 -.011
(.009) (.006) (.015) (.010)
Size 73 -.053** -.003 .004 .001
(.017) (.009) (.019) (.011)
Core 73 .103 .018 .119 -.013
(.083) (.044) (.083) (.049)
State 73 .264* -.021 -.046 -.034
(.133) (.070) (.157) (.092)
Union 73 -.052 -.037 .023 -.002
(.081) (.042) (.083) (.049)
Job Change .041 .027
(.060) (.035)
Unemployment 75 -.026** .008*
(.010) (.004)
Size 77 -.020 .006
(.019) (.01 1)
Core 77 -.067 -.076
(.084) (.049)
State 77 .213 -.084
(.156) (.091)
Union 77 .019 -.022
(.085) (.050)
Satisfaction 73 .491*** -.009
(.041) (.074)
Stress 73 .019 .639***
(.020) (.063)

R 2 .026 .063 .320 .448

* p . .05; ** p ? .01;
Note: Numbers in parentheses are standard errors; x2= 2460.65, with 890 d.f.; GFI = .883; N = 830.

nomic variable had significant effects on job but also job demands. Female and non-White
demands. As expected, high 1971 unemploy- respondents, not surprisingly, had lower job
ment rates are related to lower job security security and decision latitude.
and decision latitude, while workplace size Equations (4) and (5) are the structural
presented an apparent trade-off between equations for 1973 satisfaction and stress with
desirable job structures: increasing job secu- the 1973 job structure variables included.
rity but reducing decision latitude. Employ- These equations are almost identical, both in
ment in either core or state sectors increased overall variance explained in the dependent
security, and state sector employees also had variables and in the independent variables
greater decision latitude. Union members had which had significant effects. Decision lati-
lower decision latitude than non-union em- tude increased life satisfaction and lowered
ployees, which was perhaps a reflection of the stress, while job demands reduced satisfaction
standardization (or rigidity) of work rules in and increased stress. Job security increased
unionized worksites (Kochan, Katz, and satisfaction, but had no effect on stress. With
McKensie 1986; Edwards 1979). Among the these job structure variables included in the
individual control variables, both age and equations, the significant overall effects of
years of education increased decision latitude, macroeconomic variables found in Table 2

This content downloaded from 113.210.65.127 on Mon, 01 Apr 2019 10:45:31 UTC
All use subject to https://about.jstor.org/terms
274 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

TABLE 3. Unstandardized Coefficients from Simultaneous Equations Predicting Satisfaction and


Stress from Macroeconomic, Job Structure, and Control Variables, 1973 and 1977.

Dependent Variables

Security 73 Decision 73 Demand 73 Satisfaction 73 Stress 73


Independent Variable (1) (2) (3) (4) (5)

Age .003 .005*** .002* - .003 .000


(.003) (.001) (.001) (.003) (.002)
Female - .145* - .276*** .010 .190* .147***
(.072) (.038) (.038) (.081) (.044)
Non-White - .590*** - .200** - .049 .145 - .018
(.128) (.061) (.068) (.136) (.023)
Education (years) - .002 .030*** .020** - .002 - .018
(.013) (.006) (.007) (.019) (.014)
Unemployment 71 -.036* -.061*** -.011 -.028 .005
(.017) (.009) (.009) (.019) (.006)
Size 73 .042** -.046*** -.008 -.024 -.001
(.016) (.008) (.008) (.017) (.009)
Core 73 .201** -.057 -.011 .128 .021
(.077) (.036) (.040) (.080) (.043)
State 73 .332** .117* -.036 .135 -.047
(.124) (.058) (.065) (.130) (.070)
Union 73 .036 -.094** -.037 .005 -.034
(.074) (.035) (.039) (.078) (.042)
Job Change

Unemployment 75

Size 77

Core 77

State 77

Union 77

Security 73 .080* .006


(.032) (.017)
Decision 73 .771*** -.336***
(.115) (.112)
Demand 73 -.254*** .136*
(.056) (.063)
Satisfaction 73

Stress 73

Security 77

Decision 77

Demand 77

R2 .056 .370 .044 .143 .113

Dependent Variables

Security 77 Decision 77 Demand 77 Satisfaction 77 Stress 77


Independent Variable (6) (7) (8) (9) (10)

Age .002 - .002 - .005* .006* .001


(.003) (.002) (.002) (.003) (.002)
Female .084 - .106** - .033 .004 - .046
(.072) (.038) (.039) (.081) (.044)
Non-White -.340** -.014 - .074 .119 - .011
(.122) (.051) (.061) (.115) (.068)

(continued)

This content downloaded from 113.210.65.127 on Mon, 01 Apr 2019 10:45:31 UTC
All use subject to https://about.jstor.org/terms
JOB STRESS 275

TABLE 3. (continued)

Dependent Variables

Security 77 Decision 77 Demand 77 Satisfaction 77 Stress 77


Independent Variable (6) (7) (8) (9) (10)

Education (years) - .014 .001 .015* .006 .009


(.012) (.005) (.006) (.012) (.007)
Unemployment 71 .006 .012 .014 - .008 .012
(.019) (.008) (.009) (.018) (.011)
Size 73 -.022 .007 -.013 -.013 -.011
(.020) (.008) (.010) (.019) (.011)
Core 73 -.052 .018 -.064 .088 -.014
(.088) (.037) (.044) (.083) (.049)
State 73 - .157 - .001 - .084 - .078 - .060
(.166) (.069) (.082) (.155) (.091)
Union 73 -.137 .001 .035 .047 .009
(.088) (.037) (.044) (.082) (.048)
Job Change -.184** .003 -.029 .042 .017
(.063) (.026) (.031) (.059) (.034)
Unemployment 75 -.018 -.025*** .010* -.010 .004
(.010) (.004) (.005) (.010) (.004)
Size 77 .039* -.033*** .010 .001 .014
(.019) (.008) (.010) (.019) (.011)
Core 77 .074 -.018 -.021 -.066 .071
(.089) (.037) (.044) (.083) (.048)
State 77 .239 .067 -.012 .166 .067
(.164) (.068) (.081) (.153) (.089)
Union 77 .066 -.042 .011 .020 -.021
(.090) (.037) (.044) (.094) (.053)
Security 73 .177*** .002 .021 .010 .012
(.030) (.012) (.014) (.027) (.016)
Decision 73 .265** .455*** .060 .227* -.178**
(.100) (.055) (.046) (.113) (.067)
Demand 73 -.088 .024 .419*** -.215* .111
(.092) (.041) (.067) (.114) (.068)
Satisfaction 73 .475*** .018
(.042) (.021)
Stress 73 - .022 .628***
(.074) (.063)
Security 77 .038 .003
(.024) (.014)
Decision 77 .537*** -.349***
(.122) (.077)
Demand 77 - .422*** .163*
(.113) (.069)
R2 .085 .476 .258 .372 .506
* p c .05; ** p c .01; * p c .001.
Note: Numbers in parentheses are standard errors; x2= 6097.02 with 2755 d.f.; GFI = .847; N = 830.

"wash out": none had significant direct no lagged effects of TI macroeconomic


effects. Gender was the only exogenous variables on T2 job structures. However, T2
variable with direct effects: females had (1975) unemployment rates were related to
greater stress, but also greater satisfaction decreased decision latitude and increased job
than males. This last effect is net the higher demands in 1977, while T2 size was related
insecurity and lower decision latitude among to higher job security but lower decision
female workers-job structures which, in latitude. These effects are consistent with
turn, were related to lower satisfaction. those at Ti, and since they are net of their Ti
Equations (6) through (8) present the levels, they can be interpreted as the effects
longitudinal results for T2 (1977) job struc- of changes in these macroeconomic variables
tures regressed on Ti and T2 macroeconomic on changes in job structures.
variables, as well as their own Ti measures Among the individual control variables,
and the control variables. These results show non-Whites had lower job security in 1977,

This content downloaded from 113.210.65.127 on Mon, 01 Apr 2019 10:45:31 UTC
All use subject to https://about.jstor.org/terms
276 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

while females had lower decision latitude. results in Table 4, which presents the
Years of education increased job demands, unstandardized coefficients by decomposing
while age reduced demands. Except for this the total effects of the model's exogenous
last relationship, which is opposite T1, the variables into their direct (Table 3) and
longitudinal effects of the individual variables indirect effects. These results show that,
were similar to their cross-sectional effects. when decomposed, the significant effects of
Respondents who changed jobs between Ti macroeconomic variables on stress and satis-
and T2 reported lower job security-not faction were, without exception, indirect, and
surprisingly-but job change had no effects in the directions expected. (Indeed, some of
on other job structures. Each of the T2 job the indirect effects are larger than the total
structures is also affected by its Ti measure, effects due to trivial, nonsignificant direct
but by no other TI job structure. As at TI, the effects in directions opposite those of the
model is better at explaining variance in indirect effects). The effects of 1971 and
decision latitude (R2 = .476) than in job 1975 unemployment rates on 1973 and 1977
demands (R2 = .258) or job security (R2 = stress and satisfaction, respectively, were
.085). entirely indirect through their effects on job
The longitudinal structural equations for structures; i.e., higher unemployment rates
1977 satisfaction and stress are presented in led to more "dissatisfying" and "stressful"
(9) and (10). Overall, the model explains a job structures. Likewise, increasing size led
substantial amount of variance in each to more dissatisfying jobs at both times, while
measure (R2 = .372 and .506, respectively). state sector employment in 1973 led to more
However, none of the macroeconomic or satisfying jobs in 1973. State employment in
control variables had significant effects, with 1973 also had a lagged indirect effect on
the exception of age, which marginally satisfaction in 1977, as did 1971 unemploy-
increased satisfaction. Decision latitude and ment rates. These lagged effects, however,
job demands, on the other hand, had were not only through job structure variables
consistently significant effects. Both initial at Ti and T2, but also through Ti stress and
levels (1973) and especially subsequent satisfaction.
changes (1977) in decision latitude were Indirect effects were also important for the
positively related to changes in satisfaction individual control variables. Education had
and negatively related to changes in stress significant indirect effects on satisfaction and
(1977). The 1973 level of job demands was stress at Ti and satisfaction at T2, despite
negatively related to changes in satisfaction. insignificant total effects. Likewise, the
Changes in demands (1977) were also nega- effects of age and race on Ti satisfaction
tively related to changes in satisfaction and were indirect despite the lack of significant
positively related to changes in stress (1977). total effects. On the other hand, the effect of
These 1977 measures were also strongly age on T2 satisfaction was direct. Gender
related to their own 1973 measures, but there differences had both significant direct and
were no cross-lagged effects between the two indirect effects. Females were per se (direct
measures. effect) more satisfied than males in 1973, but
Decomposition of Indirect Effects. To- were employed in less satisfying jobs (indi-
gether, results from Tables 2 and 3 are rect effect). As a result, there were no overall
consistent with our argument that the primary
(total effect) gender differences in satisfac-
effects of macroeconomic structures and tion. At the same time, females experienced
processes on the stress and satisfaction of more stress, both directly and through the
workers are indirect via their effects on job structures of their jobs. By 1977, however,
structures. Macroeconomic variables, espe- gender differences in stress were a product of
cially unemployment rates, had significant gender differences in job structures.
overall cross-sectional and longitudinal ef- Although this decomposition does not
fects on stress/satisfaction (Table 2). They specify which, if any, of the specific job
also had significant effects on job structures structures was most important in mediating
(Table 3). However, once these job structures macroeconomic and individual effects, re-
were introduced as intervening variables, the examination of the direct effects in Table 3 is
direct effects of macroeconomic variables on suggestive of the important role of decision
stress/satisfaction disappeared (Table 3). latitude. Overall, decision latitude was the job
These arguments are strengthened by the structure most determined (largest explained

This content downloaded from 113.210.65.127 on Mon, 01 Apr 2019 10:45:31 UTC
All use subject to https://about.jstor.org/terms
JOB STRESS 277
TABLE 4. Total, Dire
Variables on 1973 an

Dependent Variables

Satisfaction 73 Stress 73
Independent Total Direct Indirect Total Direct Indirect
Variable (1) (2) (3) (4) (5) (6)
Age .002 - .003 .005** .001 .000 .001
(.002) (.003) (.002) (.002) (.002) (.001)
Female - .038 .190* - .228*** .187*** .147*** .041**
(.079) (.081) (.041) (.042) (.044) (.016)
Non-White - .040 .145 - .185** .089 - .018 .107
(.139) (.136) (.58) (.073) (.023) (.073)
Education (years) - .002 - .018 .016** - .029*** .001 - .030***
(.014) (.014) (.006) (.007) (.002) (.007)
Unemployment 71 -.018* .028 -.046*** .013* .005 .008*
(.009) (.019) (.009) (.006) (.006) (.004)
Size 73 -.053** -.024 -.029*** -.003 .001 -.004
(.017) (.017) (.008) (.009) (.009) (.004)
Core 73 .103 .128 -.024 .018 .021 -.004
(.083) (.080) (.032) (.044) (.043) (.012)
State 73 .264* .135 .129* -.021 -.047 .027
(.133) (.130) (.052) (.070) (.070) (.019)
Union 73 -.052 .005 -.057 -.037 .034 -.003
(.081) (.078) (.032) (.042) (.042) (.012)
Job Change

Unemployment 75

Size 77

Core 77

State 77

Union 77

Dependent Variables

Satisfaction 77 Stress 77

Independent Total Direct Indirect Total Direct Indirect


Variable (7) (8) (9) (10) (11) (12)
Age .008** .006* .002 .003 .001 .002
(.003) (.003) (.002) (.002) (.002) (.002)
Female - .067 .004 - .071 - .166*** - .046 - .120***
(.073) (.081) (.052) (.043) (.044) (.034)
Non-White (.091) .119 -.028 .065 - .011 .076
(.128) (.115) (.081) (.075) (.068) (.054)
Education (years) .003 .006 - .004 - .022** - .009 - .013*
(.013) (.012) (.008) (.008) (.007) (.006)
Unemployment - .010 .012 - .022* - .009 - .008 - .001
(.016) (.018) (.011) (.011) (.011) (.008)
Size 73 -.022 -.013 -.008 .004 -.011 -.007
(.020) (.019) (.012) (.012) (.011) (.008)
Core 73 .169 .088 .082 .026 .014 .040
(.091) (.083) (.050) (.053) (.049) (.033)
State 73 .086 -.078 .164* -.040 -.060 .020
(.167) (.155) (.084) (.098) (.091) (.055)
Union 73 -.003 .047 -.050 -.027 .009 -.036
(.090) (.082) (.048) (.053) (.048) (.032)
Job Change .049 .042 .007 .027 .017 .010
(.060) (.059) (.020) (.035) (.034) (.012)
(continued)

This content downloaded from 113.210.65.127 on Mon, 01 Apr 2019 10:45:31 UTC
All use subject to https://about.jstor.org/terms
278 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

TABLE 4. (continued)

Dependent Variables

Satisfaction 77 Stress 77

Independent Total Direct Indirect Total Direct Indirect


Variable (7) (8) (9) (10) (11) (12)

Unemployment 75 -.026** -.010 -.017*** .008* -.004 .012***


(.010) (.010) (.004) (.004) (.004) (.003)
Size 77 -.021 .001 -.021** -.006 .014 -.009
(.019) (.019) (.008) (.011) (.011) (.006)
Core 77 -.064 -.066 .002 .076 .071 .005
(.084) (.083) (.028) (.049) (.048) (.016)
State 77 .215 .166 .050 .082 .067 .016
(.156) (.153) (.052) (.091) (.089) (.031)
Union 77 .020 .045 -.025 -.021 -.011 -.010
(.090) (.084) (.028) (.053) (.049) (.017)

* p < .05; ** p < .01; *** p ? .001.


Note: Direct and indirect effects may not add up to total effect because of rounding. Numbers in parentheses are
standard errors.

variance and most direct effects) by macro- tures that represent increased and continued
economic and individual variables and, in exposure to stressful conditions. This would
turn, had significant effects on satisfaction appear to be a more general situation applying
and stress at both times. Job demands, in to all workers. Whether consciously or not,
contrast, while also having consistent effects firms appear to pass any increased uncertainty
on satisfaction and stress, was not predicted in the marketplace to their workers in the
as well by the model-especially the macro- form of increased job demands, decreased
economic variables. Job security was pre- decision latitude, and increased job insecu-
dicted by a number of variables but had only rity. Because of the deteriorating labor market
one significant effect itself-on Ti satisfac- situation (i.e., higher unemployment) work-
tion. ers become less able to resist this restructur-
ing. Thus, even if individuals are not laid off
or subject to negative financial events, their
CONCLUSION well-being is affected by changes in day-to-
day exposure to stressful conditions because
In this paper we have attempted both to their job structures become more stressful.
clarify the relationship between macroeco- These findings are precisely in line with
nomic structure and individual stress and to Pearlin's (1989) suggestion that the sociolog-
illustrate one way in which the sociological ical study of stress must explicitly incorporate
study of stress can be usefully developed. the effects of economic and institutional
Clarification required linking together- context on individual health. In the case at
theoretically and methodologically-argu- hand we use elements of the macroeconomic
ments from the "economic stress" and "job context to explain the job structures that
stress" literatures, a link heretofore not individuals may find stressful. Given that
explicitly made. In much of the previous work is a central role for most adults, the
"economic stress" literature, the link between capacity to explain how the general macro-
macroeconomic change and individual stress, economic context affects the job structure
mortality, and morbidity was made through context and, in turn, individual health,
actual or potential life events, such as demonstrates the way in which sociological
unemployment and related financial events structures affect health and affirms the utility
(Brenner 1976, 1984; Catalano and Dooley of studies specifically aimed at describing this
1977, 1979, 1983; Catalano, Rook, and process.
Dooley 1986). Although illuminating, this Our findings also suggest that we have not
link does not present the entire picture. Our exhausted the potential utility of this ap-
results indicate that macroeconomic changes proach. For example, the independent effects
also affect individual stress because they lead of gender on various elements within our
to changes in routine, day-to-day job struc- model clearly suggest that we need to account

This content downloaded from 113.210.65.127 on Mon, 01 Apr 2019 10:45:31 UTC
All use subject to https://about.jstor.org/terms
JOB STRESS 279

for the way in which other macroeconomic 2. We have used the exact items Karasek (1979)
conditions (e.g., occupational segregation) used to measure decision latitude and job
differentially affect exposure and vulnerabil- demands. However, where Karasek con-
ity to stressors that originate in the macro- structed his measures by computing respon-

economy. Given the large body of literature dents' mean scores on each individual item,

on gender, work, and health, consideration of we have estimated the relationships between

macroeconomic structure will clearly extend individual items and unobserved constructs by
using the maximum likelihood technique in
our ability to explain the relationship between
LISREL 7. However, despite the difference in
gender and well-being. Linking the economic
measurement techniques, our results support
stress and job stress literatures in this way
those of Karasek: items assumed to measure
will allow us to reexamine numerous other
decision latitude do, in fact, load significantly
observations about the relationship between
on that construct, while those assumed to
work and health.
measure job demands load significantly on
that construct (see Appendix A). Karasek also
NOTES found that the interaction between these
dimensions-a variable he termed job strain-
1. T2 macroeconomic and control variables are as-
had significant effects on his measures of
sumed to be correlated with, rather than caused
stress (Karasek 1979, 1990). We did not test
by, macroeconomic and control variables at T 1.
for this interaction effect in our LISREL
This assumption is necessitated by the categor-
ical measurement of some of these variables - model, since it violates LISREL's assumptions

core and state economic sectors and union mem- of linear measurement. However, OLS regres-
bership (see discussion of variable measurement sion equations using this job strain measure to
in the text) -and because such variables violate predict stress and satisfaction were not signif-
assumptions of normality of distribution and con- icantly different from equations using the
stant variance when treated as dependent, or en- separate measures of decision latitude and job
dogenous, variables (Joreskog and Sdrbom 1989). demands.

This content downloaded from 113.210.65.127 on Mon, 01 Apr 2019 10:45:31 UTC
All use subject to https://about.jstor.org/terms
APPENDIX
Measurement Model for Latent Variables in the Models.

19

S
Latent Variable and Observed Variable Lambda

Security (Eta,, Eta6)


I. "The job security is good." 1.oooa *0
Decision Latitude (Eta2, Eta7)
I . "My job requires that I keep learning new things."
2. "My job requires a high level of skills." b
3. "My job requires that I do the same things over and o
4. "My job requires that I be creative." 1.5
5. "I have the freedom to decide what I do on the job
6. "It is basically my own responsibility to decide how my job get
7. "I have a lot of say about what happens on my jo
8. "I am given a lot of freedom to decide howlI do my own w
Job Demands (Eta3, Eta8)
I . "My job requires that I work very fast." 1.o
2. "My job requires that I work very hard."
3. "I have too much work to do everything well
4. "I have enough time to get the job done. " b
5. "I am not asked to do excessive amounts of work.",
6. "I never seem to have enough time to get everything done on m
7. "I am free from the conflicting demands that other people mak
Satisfaction (Eta4, Eta9)C
"Here are some words and phrases which we would like you to use to describe how yo
feel about your present life. ..
I. "Boring . . . interesting" 1.oooa .
2. "Enjoyable . . . miserable " .988
3 . "Useless . . . worthwhile" .758
4 . "Friendly . . . lonely" b 1.042
5 . "Full . . . empty, 1.059 .04
6 . "Discouraging ..,. hopeful" .843
7 . "Disappointing . . . rewarding" b1.0
8 . "Brings out the best in me . . . doesn't give me a chan
Stress (Eta5, Eta10)
"...how often each (physical condition) has happened to you in the last year"
I . "Trouble breathing or shortness of breath" 1
2. "Pains in my back or spine" .835
3. "Becoming very tired in a short time" 1.
4. "Feeling my heart pounding or racing"
5. "Hands sweating so that they feel damp and clamm
6. "Feeling nervous or fidgety and tense" 1.
7. "Poor appetite" b.677 .0
8. "Never have any pep or energy" .21
aFixed coefficient.
bReversed cod

This content downloaded from 113.210.65.127 on Mon, 01 Apr 2019 10:45:31 UTC
All use subject to https://about.jstor.org/terms
JOB STRESS 281

REFERENCES wend. 1969. Social Status and Psychological


Disorder. New York: Wiley and Sons.
Baker, Dean B. 1985. "The Study of Stress at
Dooley, David and Ralph Catalano. 1984. "The
Work." Annual Review of Public Health 6:367-
81. Epidemiology of Economic Stress." American
Journal of Community Psychology 12:387-409.
Blauner, Robert. 1964. Alienation and Freedom:
Edwards, Richard. 1979. Contested Terrain: The
The Factory Worker and His Industry. Chicago:
University of Chicago Press. Transformation of the Workplace in the Twenti-
eth Century. New York: Basic Books.
Bluestone, Barry and Bennett Harrison. 1982. The
Deindustrialization of America. New York: French, John R.P., Jr., Robert D. Caplan, and R.
Basic Books. Van Harrison. 1982. The Mechanisms of Job
Blumberg, Paul. 1980. Inequality in an Age Stress and Strain. New York: Wiley and Sons.
of Decline. New York: Oxford University Gurin, Gerald G., Joseph Veroff, and Sheila Feld.
Press. 1960. Americans View Their Mental Health.
New York: Basic Books.
Brenner, M. Harvey. 1976. Estimating the Social
Costs of Economic Policy: Implications for Hankin, Janet R. 1990. "Gender and Mental
Mental and Physical Health, and Criminal Illness." Pp. 183-201 in Research in Commu-
Aggression. Report to the Congressional Re- nity and Mental Health, vol. 6, edited by J.
search Service of the Library of Congress Joint Greenley. Greenwich, CT: JAI Press.
Economic Committee of Congress. Washington, Hodson, Randy D. 1978. "Labor in Monopoly,
D.C. Competitive, and State Sectors of Production."
Politics and Society 8:429-80.
_ 1984. "Estimating the Effect of Eco-
nomic Change on National Mental Health and _ 1984. "Corporate Structure and Job
Social Well-Being." Prepared by the subcom- Satisfaction: A Focus on Employer Characteris-
mittee on economic goals and intergovernmen- tics." Sociology and Social Research 69:22-49.
tal policy of the Joint Economic Committee, Holt, Robert R. 1982. "Occupational Stress." Pp.
U.S. Congress. 419-444 in Handbook of Stress, edited by L.
. 1987. "Relation of Economic Change to Goldberger and S. Breznitz. New York: Basic
Swedish Health and Social Well-Being, 1950- Books.
1980. " Social Science and Medicine 25:183-95. House, James S. 1980. Occupational Stress and
and Anne Mooney. 1983. "Unemployment the Mental and Physical Health of Factory
and Health in the Context of Economic Workers. Ann Arbor, MI: Institute for Social
Change." Social Science and Medicine 17: Research Report Series.
1125-38. , Victor Strecher, Helen L. Metzner, and
Campbell, Angus, Philip E. Converse, W. Ward, Cynthia A. Robbins. 1986. "Occupational
and L. Rogers. 1976. The Quality of American Stress and Health among Men and Women in the
Life. New York: Russell Sage. Tecumseh Community Health Study." Journal
Catalano, Ralph. 1989. "Ecological Factors in of Health and Social Behavior 27:62-77.
Illness and Disease." Pp. 87-101 in Handbook Jahoda, Marie. 1988. "Economic Recession and
of Medical Sociology, fourth edition, edited by Mental Health: Some Conceptual Issues." Jour-
H. Freeman and S. Levine. Englewood Cliffs, nal of Social Issues 44:13-23.
NJ: Prentice Hall. Johnson, Jeffrey. 1989. "Control, Collectivity, and
Catalano, Ralph and David Dooley. 1977. "Eco- the Psychological Work Environment." Pp.
nomic Predictors of Depressed Mood and 55-74 in Job Control and Worker Health, edited
Stressful Life Events." Journal of Health and by S.L. Sauter, J.J. Hurrell, Jr., and C.L.
Social Behavior 18:292-307. Cooper. Chichester, West Sussex, England:
and David Dooley. 1979. "The Economy Wiley and Sons.
as Stressor: A Sectoral Analysis." Review of JTreskog, Karl G. and Dag Sorbom. 1989. LISREL
Social Economy 37:175-87. 7: A Guide to the Program and Applications.
and David Dooley. 1983. "Health Effects Chicago: SPSS, Inc.
of Economic Instability: A Test of Economic Kahn, Robert L. 1974. "Conflict, Ambiguity, and
Stress Hypothesis." Journal of Health and Overload: Three Elements of Job Stress." Pp.
Social Behavior 24:46-60. 47-61 in Occupational Stress. Edited by A.
, David Dooley, and Robert Jackson. 1985. McLean. Springfield, IL: Thomas.
"Economic Antecedents of Help-Seeking: Re- Kalleberg, Arne. 1989. "Linking Macro and Micro
formulation of Time-Series Tests." Journal of Levels: Bringing the Workers Back into the
Health and Social Behavior 26:141-52. Sociology of Work." Social Forces 67:582-92.
, Karen Rook, and David Dooley. 1986. and Ivar Berg. 1987. Work and Industry:
"Labor Markets and Help Seeking: A Test of theStructures, Markets, and Processes. New York:
Employment Security Hypothesis." Journal ofPlenum.
Health and Social Behavior 27:277-87. Karasek, Robert A. 1979. "Job Demands, Job
Dohrenwend, Bruce P. and Barbara S. Dohren- Decision Latitude, and Mental Strain: Implica-

This content downloaded from 113.210.65.127 on Mon, 01 Apr 2019 10:45:31 UTC
All use subject to https://about.jstor.org/terms
282 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

tions for Job Redesign." Administrative Science in Communities." Psychological Reports 3:325-
Quarterly 24:285-306. 29.
. 1989. "Control in the Workplace and Its Marshall, James R. and Donna P. Funch. 1979.
Health-related Aspects." Pp. 129-59 in Job "Mental Illness and the Economy: A Critique
Control and Worker Health, edited by and Partial Replication." Journal of Health and
S.L. Sauter, J.J. Hurrell, and C.L. Cooper. Social Behavior 20:259-72.
Chichester, West Sussex, England. Wiley and Mirowsky, John and Catherine E. Ross. 1989.
Sons. Social Causes of Psychological Distress. New
and Thres Theorell. 1990. Healthy Work: York: Aldine de Gruyter.
Stress, Productivity, and the Reconstruction of Pearlin, Leonard I. 1989. "The Sociological Study
Working Life. New York: Basic Books. of Stress." Journal of Health and Social
Kasl, Stan V. 1978. "Epidemiological Contribu- Behavior 30:241-56.
tions to the Study of Work Stress." Pp. 3-48 in and Carmi Schooler. 1978. "The Structure
Stress at Work, edited by C.L. Cooper and R.
of Coping." Journal of Health and Social
Payne. New York: Wiley-Interscience.
Behavior 19:2-21.
Kessler, Ronald C. and David F. Greenberg. 1981.
Selye, Hans. 1976. Stress in Health and Disease.
Linear Panel Analysis: Models of Quantitative
Woburn, MA: Butterworth.
Change. New York: Academic Press.
Spenner, Kenneth. 1983. "Deciphering Prome-
and Harold W. Neighbors. 1986. "A
theus: Temporal Change in the Skill Level of
New Perspective on the Relationships Among
Work." American Sociological Review 48:824-
Race, Social Class, and Psychological Dis-
37.
tress." Journal of Health and Social Behavior
Starrin, Bengt, Bertil Lunberg, Bosse Angelow,
27:107-15.
and Hans Wall. 1989. "Unemployment, Over-
Kochan, Thomas A., Harvey C. Katz, and Robert
time Work, and Work Intensity." Pp. 261-75 in
B. McKensie. 1986. The Transformation of
Unemployment, Poverty, and Quality of Work-
American Industrial Relations. New York: Basic
ing Life: Some European Experiences, edited by
Books.
B. Starrin, P. G. Svensson, and H. Winters-
Liem, Ramsay and Paula Rayman. 1984. "Per-
spectives on Unemployment, Mental Health, berger. Berlin: World Health Organization,

and Social Policy." International Journal of European Regional Office and the European

Mental Health 13:3-17. Center for Social Welfare Training and Re-
Lincoln, James R. 1978. "Community Structure search, Education Service.
and Industrial Conflict: An Analysis of Strike Stolzenberg, Ross M. 1978. "Bringing the Boss
Activity in SMSA's. " American Sociological Back In: Employer Size, Employee Schooling,
Review 43:199-220. and Socioeconomic Achievement." American
and Arne L. Kalleberg. 1985. "Work Sociological Review 43:813-28.
Organization and Work Commitment: A Study U.S. Bureau of the Census. 1978. Statistical
of Plants and Employees in the U.S. and Abstracts of the United States. Washington,
Japan." American Sociological Review 50:738- D.C.: U.S. Government Printing Office.
60. Wheaton, Blair. 1978. "The Sociogenesis of
MacMillan, A.M. 1957. "The Health Opinion Psychological Disorder: Reexamining the Causal
Survey: Technique for Estimating Prevalence of Issues with Longitudinal Data." American
Psychoneurotic and Related Types of Disorder Sociological Review 43:383-403.

Rudy Fenwick is associate professor of sociology and Director of the Public Policy Certificate Program,
University of Akron, Akron, Ohio. His current research interests include the effects of economic markets
on organizational and job structures, and the political and legislative origins of market restructuring in the
United States and Canada.

Mark Tausig is associate professor of sociology, University of Akron, Akron, Ohio. His current research
involves identifying the effects of various forms of job insecurity on mental health. He is also studying
social influence processes in the social networks of injection drug users.

This content downloaded from 113.210.65.127 on Mon, 01 Apr 2019 10:45:31 UTC
All use subject to https://about.jstor.org/terms

Anda mungkin juga menyukai