New Ideas in Psychology 39 (2015) 12e22

Contents lists available at ScienceDirect

New Ideas in Psychology

journal homepage: www.elsevier.com/locate/newideapsych

Trajectory-based methods in clinical psychology: A person centred

narrative approach
Kieron O'Connor a, *, Marie Robert b, Guilhe

me Pe
rodeau b, Monique Se
guin b
a
Research Centre, University Institute of Mental Health at Montreal, 7331 Hochelaga St., Montreal, Quebec H1N 3V2, Canada
b
Department of Psychoeducation and Psychology, Universit e du Qu
ebec en Outaouais, 283 Alexandre-Tach e Blvd, Gatineau, Quebec J8X 3X7, Canada

a r t i c l e i n f o a b s t r a c t

Article history: Life trajectories in clinical psychology research are often not treated as interactive trajectories but rather
Received 22 August 2014 as static transversal variables. But developmental pathways are often cumulative and conditional and
Received in revised form currently require sophisticated group-based modeling to tease out individual differences in trajectories.
25 June 2015
Clinical psychologists often require personal information on transitions and turning points in life which
Accepted 26 June 2015
Available online 20 July 2015
require eliciting information through qualitative life history approaches. A method is proposed for
identifying life events within the person's narrative and describing trajectories as event spaces likely to
reect end-point psychopathology.
Keywords:
Clinical psychology
2015 Elsevier Ltd. All rights reserved.
Trajectory-based methods
Group based models
Person centred constructions
Narratives
Event spaces

1. Introduction Several authors have indeed criticized the a-historic nature of

In clinical psychology early development, life events and bio-
graphical factors are recognized inuences on the development of
pathology, both theoretically and practically. However, there is the
question of translating such biographical context into a meaningful
account of life trajectories. Most attempts have either targeted: (1)
transversal designs following selected factors for usually specied
periods of time; (2) prospective and longitudinal studies which
have targeted risk and resilience through examining the impact of
multiple separate predictor variables. At the same time, human
events occur in sequences, are often cumulative and require con-
ditional calculation of cascade risk. Apart from the obvious
problem of separating cause and effect and disentangling prox-
imal from distal inuences, many questionnaire studies rely on
retrospective recall of general events which may be subsequently
tainted by current experience. Also life events are transactional in
nature and events require methods that address conditional
probabilities and pathways.
* Corresponding author.
E-mail addresses: kieron.oconnor@umontreal.ca (K. O'Connor), marie.robert@
uqo.ca (M. Robert), guilheme.perodeau@uqo.ca (G. Perodeau), monique.seguin@
guin).
uqo.ca (M. Se
the mainstream clinical psychology approach and suggested a
historically structured sampling method for comparing histories
which converge at nal points but diverge at important previous
bifurcation points. But the reality is that historical facts are usually
collected formulaically (e.g. Muller et al 2011; Smith 2009). A life
trajectory implies a continual interaction of person and environ-
ment in accordance with the construction of meaning and reality.
Unfortunately there is no current consensus on method or mea-
sures appropriate for a trajectory approach in clinical psychology.
Very few studies have specied trajectory as a continuous devel-
opment of a life sequence which necessarily involves transitions
and turning points not always accessible to quantitative measures.
The study of life trajectories within clinical psychology is still in
its infancy. Whilst there are numerous studies in the risk-resilience
mode, in the main, they are dealing with static stage dependent
variables, which, whilst contributing to pathological development,
do not capture the synergistic reciprocity essential to trajectories
point of view. Indeed such longitudinal ndings, even whilst sta-
tistically robust in terms of signicant regression coefcients, could
be misleading. For example, isolated ndings of child impulsivity or
maternal depression as determinants of behavioural conduct dis-
orders, do not allow for the interactive effect that environment may
contribute to child characteristics and the subsequent interaction

http://dx.doi.org/10.1016/j.newideapsych.2015.06.001
0732-118X/ 2015 Elsevier Ltd. All rights reserved.
 K. O'Connor et al. / New Ideas in Psychology 39 (2015) 12e22
that child coping behaviour may exert on maternal response. In this
sense, a key problem in cross-sectional risk studies is determining
causality. An example here is the well established nding of a link
between maternal over-protection and child anxiety (Turgeon,
O'Connor, Marchand, & Freeston, 2002). Several authors pose the
question: Is this cause or effect? Does an over-protective mother
produce a timid child, or does the timid child invoke over-
protection? Even longitudinal studies, if they follow only one or
two static variables, may miss out on this trajectorial philosophy.
One answer may be to follow the evolution of episodes of
overprotection.
In an attempt to look at the temporal process relationship be-
tween parental drinking, family function and child adjustment,
Keller, Cummings, Davies, and Mitchell (2008) found the impact
depended on parental gender, child gender, type of alcohol abuse.
Interestingly, in a footnote, the authors note the possibility of
reversing the temporal order of the model where the child's
behaviour predicts drinking and parental psychological problems,
which suggests complex pathways. Again the ambivalence of such
pathways maybe best captured qualitatively.
The key debates here are how to structure multiple causal in-
uences and whether equi-nality or multi-nality models are the
most appropriate to explain multiple causal inuences. Finally,
capturing the crucial inuence of life context in the form of
determining life events will most likely remain a puzzle with ap-
proaches which tend not to model dynamic interplay.
Latent curve analysis and growth modelling does offer a means
of quantifying trajectories by mapping sequential processes, but
again the statistical assumptions may not always make it appro-
priate for human behaviour. However sophisticated the models,
such designs may always be insufcient since they adhere to
inappropriate assumptions about human behaviour and human life
events. Statistical sophistication is not necessarily the only solution
and a paradigm shift which changes the way we view life sequences
in terms of personal transitions and turning points can lead to
adoption of more manageable approaches. A more person-centred,
contextual approach to life trajectory methods is proposed for
clinical psychology practice and research, including life history and
life course narrative approaches. A person centred approach could
accept from the start the necessity of conditional probability (if x
then y), and multidimensionality (e.g. thoughts and emotions in the
same event) in trajectory research, and could consider a pattern of
complex overlapping event spaces of life history rather than tem-
poral succession as the key metric for trajectories.
The new idea proposed here is that trajectory research begins
with narratives or qualitative interviews which capture transition
points and complex events which are then considered probable
event spaces leading to nal endpoints within a Bayesian model of
alternative possibilities. Firstly we illustrate the difculties of
quantitative approaches, supported by selective examples from
research in anxiety and depression, in addressing: the interaction
between temperament, life events and contextual interactions with
specicity of pathways; the inuence of cognitive factors and
coping; how individual differences in sequencing and mediation
determine the importance of predictors and the impact of stage of
development on predictors.
2. Methods
Transversal studies seek to identify specic precursors, risks and
resilience factors which relate to later development of psychopa-
thology. Such studies have focused on child temperament, trauma or
family environment as separate mechanisms. Periods of observation
are generally over short periods (5e7 yrs) and examining selected
explanatory risk factors (Luby, Belden, & Spitznagel, 2006) such as:
13
family dysfunction, behavioural inhibition, genetics, negative life
events and protective factors (Malkoff-Schwarts et al., 2000).
The problems with this transversal approach are two fold:
rstly, such a static picture failures to capture the synergy inter-
acting amongst distinct factors; secondly, in developmental terms,
one psychopathology (e.g., depression) may be the risk factor for
subsequent development problems. So ndings of contributions to
development even if proven strong predictors in regression equa-
tions cannot easily translate into causal hypotheses. These studies
have also mostly been termed back end research examining risk
factors subsequent to an established diagnosis, rather than front
end research following factors in the development of pathology.
This latter approach clearly has more implications for prevention.
The aim of most front end studies is to improve predictions of
later diagnostic status, severity and treatment outcome. Again,
developmental studies bring into question premature diagnoses
where often early psychopathology is uid. For example, early life
events seem linked to both depression and anxiety. The existence of
distinct trajectories may suggest the need for developmentally
tailored intervention strategies independent of diagnosis. Devel-
opmental research brings into question categorical diagnoses of, for
example, anxiety and depression. In developmental terms, anxiety
and depression may represent a common dimension of distress,
namely negative reactivity. For example core internalizing prob-
lems can be operationalized as the underlying common factor in
anxiety and depression with two main pathways: the rst through
temperament mediated by child emotionality mid-childhood; the
second through early contextual risk factors, namely family
adversity before ve years of age. Whereas trajectory research
supports a transdiagnostic approach, such models do not currently
explain how multinality and divergent trajectories occur (Nolen-
Hoeksema & Watkins, 2011). Phillips, Hammen, Brennan, Najman,
and Bor (2005), for example, used logistic regression to predict
group status from different early adversity, in a prospective
exploration of the specicity of early childhood adversities as
predictors of anxiety and depressive disorders in adolescents. But
child adversities such as poor parenting, marital discord, mental
illness and abuse, were predictive of a broad range of disorders.
Phillips et al. (2005) did nd that anxious adolescents experienced
more adversities than depressed ones. However, specicity remains
elusive and as the authors note, the explanation for the association
remains unclear. Also, the pathways to depression remain incon-
clusive. The causes may be more proximal, or anxiety may be a
precursor to depression, or there may be multiple pathways to
depression. Furthermore, variables may be important at one age
but not at another. Also, different variables may have a different
variance rate.
2.1. Individual differences in person-centred life trajectory
Smith (2009) considers the fundamental task is to understand
why different individuals progress along different life trajectories,
and what forces increase the likelihood of successful versus non-
successful life pathways. So trajectories should focus on person-
centred rather than variable-centred approaches. In the person-
centred approach, the aim is to classify individuals into distinct
groups and categories based on individual response patterns. One
attempt to overcome these dilemmas of non-linearity, differential
progression and person-centred approaches, has been the adoption
of growth mixture models. This can be accomplished using latent
trajectory classes, which allow different groups of individuals to
vary around different means.
Traditional growth modeling techniques (for example, hierar-
chical linear modeling) assume that participants belong to a single
homogeneous population and do not capture inter-individual or
14 K. O'Connor et al. / New Ideas in Psychology 39 (2015) 12e22
intra-individual change over time. Finite growth mixture models
are powerful models for mapping outcomes in apparently homo-
geneous groups which, however, differ in underlying processes.
There has been a move away from the interpretation of trajectory
groups as distinct entities and instead towards a view of trajectory
groups as providing approximations for a more complex reality
through the identication of clusters of individuals following tra-
jectories over time.
In not insisting on treating a single population, Latent Growth
Mixture Models (LGMM) can identify heterogeneous patterns of
responding (or trajectory classes) that represent diverse pop-
ulations (Muthe n, 2004). The mean growth curves of these distinct
populations can then be modeled independently, allowing for wide
exibility and at the same time precision in locating the various
person-centred ways that people exhibit change across time. This
exibility results from simultaneous modeling of latent continuous
variables (e.g. trends) and latent categorical variables (trajectory
class assignment). In addition, LGMM considers other covariates
that can either inuence or predict the formal trajectory classes,
allowing it to distinguish people in different trajectory classes.
However these studies reveal the difculties in trying to capture
variables in person centred trajectories.
For example studies using Growth Mixture Models (GMM) and
Latent Curve Analysis (LCA) have shown indeed that trajectories
differ between individuals, some showing fast, some slow rises
times in anxiety symptoms (Crocetti, Klimstra, Keijsers, Hale, &
Meeus, 2009). Duchesne, Larose, Vitaro, and Tremblay (2010)
identied four trajectory groups in childhood anxiety: low, low
increasing, high declining and high anxiety. M'Bailara, Cosnefroy,
Vieta, Scott, and Henry (2013) identied four trajectories of
symptom change in bipolar disorder based on cognitive, emotional
and psychomotor assessment: including: persistent inhibition (no
recovery), transient inhibition (slow recovery), transient activation
(linear recovery) and over-activation (fast recovery). The problem is
the stability of the categories which were independent of bipolar
status and the assumption of randomly distributed variables.
Colman, Ploubidis, Wadsworth, Jones, and Croudace (2007) using
LCA reported distinct proles of depressive and anxious symp-
tomatology over the life course and associated with neuro-
developmental markers. However, the authors note LCA assumes
independence of variables.
Optimal individual matching analysis is required to explore
complex life trajectories. This analysis enables researchers to take
into account the order, chronology and duration of states during a
life trajectory and multiple life dimensions are combined with one
another using a multi-channel approach of optimal matching
(Gauthier, Widmer, & Dame, 2010). This methodological approach
is designed to reveal the overall structure associating various life
trajectories (e.g. co-residence, occupational and intimacy trajec-
tories) without inferring a causal order. This order is difcult to
decide a priori, as those dimensions belong to interacting elds,
that is, phenomena that are interconnected rather than occurring
independently (Gagadinho, Ritschard, Studer, & Mller, 2008).
This technique allows trajectories nonetheless to be classied
into distinct categories preserving individual information. Ge,
Lorenz, Conger, Elder, and Simons (1994), looking at the timing of
distinct trajectories using LGMM, located the time point when a rise
in depressive symptoms in girls takes place. deRoon-Cassini, Rusch,
Mancini, and Bonanno (2010) explored the individual trajectories of
clients who do and do not develop post-traumatic stress disorder.
Nagin and Odgers (2010) reported on the use of group-based
trajectories permitting the identication of both group and
distinct trajectories. The group-based methodology is responsive to
calls for the development of person-based approaches to devel-
opment. But such models still require averaging and aggregating
distinct changes and variability over time.
Haviland, Nagin, Rosenbaum, and Tremblay (2008) conclude:
Making causal inferences with non-experimental data is fraught
with ambiguity in almost all problem contexts. In developmental
studies much of this ambiguity takes a particular form, namely, the
prior trajectory of the behaviour under study predicts both the
likelihood of the turning-point event and the future trajectory of
the behaviour (Pg 453). The prior trajectory of behaviour maybe
best revealed by examining person centred interactions with
events.
A key factor inuencing future development may be coping
ability.
2.2. Coping trajectories
Once psychological problems have developed, they may be
maintained or resisted by different coping factors. For health re-
searchers, resilience is an interactive concept which refers to the
capacity for successful adaptation, the ability to rebound after
encountering difculties (Rutter, 2006). It includes a sense of self-
esteem or self-condence, humour and a problem-solving atti-
tude (Connor & Davidson, 2003). A number of personal maintaining
psychological characteristics have been found to play an important
role, in particular, self-efcacy beliefs, cognitive distortions,
dysfunctional attributions, immature defence mechanisms, and
dysfunctional coping strategies (Carr, 1999). Coping strategies are
assumed to be consciously and deliberately used methods for
regulating negative emotions or to manage situations in which
there is a perceived discrepancy between stressful demands and
available resources. A distinction is made between problem- and
emotional-focused coping strategies or between functional and
dysfunctional strategies (Fields & Prinz, 1997). Active ways of
coping such as problem-solving, cognitive distraction, self-calming,
and asking for help from others, contribute to better adaptation and
decrease depressive and anxiety symptoms, in contrast to the more
passive ways of coping such as avoidance and social isolation
(Compas, Malcarne, & Fondacaro, 1988).
It is important to factor in the development and contribution of
strategies coping with adversity which requires autoregressive
modelling of sequential coping. An important element is recovery
style and whether people use covering up or sealing over,
avoidant strategies or adopting active integration of strategies for
problem-solving. These strategies can change over time but sealing
over has been identied as associated with negative self-evaluation
and more insecure identity. Interestingly those with distinct coping
styles may be drawn to different services. Sealing over may lead to
seeking low key non-confrontational services whereas approaches
may more likely seek services which actively engage the person
(Tait, Birchwood, & Trower, 2004).
Parental behaviour and child rearing practices contribute to the
child's development of a perception and appraisal of the life events
and ways of coping with stressful situations (Kortlander, Kendall, &
Panichelli-Mindel, 1997). Overprotective, anxious, or aggressive,
disapproving parents provoke low coping in children. Clinically
depressed and anxious children use less efcient coping strategies
that eventually increase their problem (Rapee, Schniering, &
Hudson, 2009). There is evidence that authoritarian parenting is
signicantly associated with obsessional symptoms (Timpano,
Keough, Mahaffey, Schmidt, & Abramowitz, 2010).
Individual differences in coping interact with differences in
temperament and reactivity to stress.
2.3. Reactivity to stress
The effects of life events are cumulative and interactive. The risk
 K. O'Connor et al. / New Ideas in Psychology 39 (2015) 12e22
and protection factors are not so much mediated or moderated by
other variables as conditional on temperament, group membership,
stage of development, or attributes such as gender. The issue here is
that the temperamental factors rather than discrete life events,
traumas, experiences determine the development of a reactivity to
stress which then may condition response to subsequent life events
and coping with such events, which may in turn protect from other
problems.
Compas et al. (2006) present a model for the development of
depression involving a complex series of relationships between life
events, negative and positive affectivity and stress-related vari-
ables, to the extent that coping strategies may be effective or not
depending on the level of affectivity. This dependence, in turn,
depends on whether the person is female and excessively reactive
to stressful events and whether anxious or depressive symptoms
develop rst. In other words, the same events may produce distinct
reactions, even opposite reactions and eventual pathologies,
depending on temperament, sex or age. But whether reactions
develop may depend crucially on trajectory, in the sense that an
interacting series of previous events determine whether a nal
event or diagnosis arrives.
According to the tripartite model, high negative affectivity (NA),
dened as unpleasurable arousal with elevated subjective distress,
is a non-specic factor related to both anxiety and depression.
However, low positive affectivity (PA), characterized by a lack of
pleasurable engagement with the environment, a lack of sociability
and enthusiasm is specically associated with depression. In a test
of a tripartite model in the reactivity to stress approach, Fox,
Halpern, Ryan, and Lowe (2010) speculated that high NA may
relate to anxiety and physiological reactivity, whilst low PA may
relate to depression. In a second example, Compas, Connor-Smith,
and Jaser (2004) presented a framework for the development of
depression involving a complex series of relationships between
temperament and stress-related variables. According to Compas
et al., individual differences in temperament may facilitate or
restrict both automatic and voluntary responses to stress. Children
with high NA, for instance, may be limited in the ability to regulate
their emotions in response to stress, potentially leading to
increased risk for depression. Even when youth do engage in coping
strategies, their impact on the development of depressive symp-
toms may be a function of temperamental variation. For example,
cognitively reframing a situation to focus on the positive may
protect children with high PA from experiencing depression; hav-
ing low PA may interfere with the effectiveness of this coping
strategy and prevent children from reaping its protective benets.
However, individual differences depended on medium or high
levels of affectivity. Temperamental characteristics may moderate
and be moderated by stress-sensitivity. The stress-sensitivity can
be operationalized either as NA reactivity or over reaction to small
daily hassles or stresses both of which may indicate a liability to-
wards depression or anxiety.
Although there may be a role for general underlying factors like
stress-sensitivity exposure to life events may in turn contribute to
stress sensitivity. Wickers et al. (2009), for example, measured
childhood adversity on the basis of a child trauma questionnaire. A
multi-level analysis took into account the variability associated
with each level of nesting, according to life events. There was a
signicant association between daily stress and both childhood
adversity and negative life events. The effects of adult life adverse
event exposure were explained by previous childhood adversity
and stress sensitivity. The ndings are consistent with other
studies emphasizing the contribution of early environment and
adult stress sensitivity. Progressively larger negative reactions are
induced by stressors of similar magnitude, indicating a
sensitization.
15
2.4. Longitudinal and developmental psychopathology approaches
A few studies testify to the complex longitudinal relations be-
tween temperamental and contextual risk factors and their effect
on, for example, internalizing symptoms. How to measure or test
such complex interplay is a challenge when investigating medi-
ating, reciprocal, and dynamic relations between several variables.
A variable that serves as a stressor in one relation might turn into a
mediator in another relation. Different mediators have a shorter or
longer variation and time course. An exploratory approach to the
relations between the specic variables would e when put together
in one model e examine all relations across development at the
same time. Rutter, Kim-Cohen, and Maughan (2006) have pointed
out that there is a need for identifying mediating processes, insofar
as we know too little to develop an all-encompassing model of
continuities in childhood and onwards. But it is doubtful if medi-
ation models can reveal the extent of conditional interaction. Psy-
chopathologies have multiple causes which interact with each
other and change over time (Smith, 2009). Researchers have moved
away from discrete factors to look at transitional processes like
tolerance of adversity, and to examine dynamic aspects such as
non-linear trajectory curves.
Ge, Natsuaki, and Conger (2006) suggest longitudinal pathways
and trajectories take a curvilinear pattern but this varies depending
on gender. The authors urge also considering age as a unit of
analysis. They note the importance of using interlocking trajec-
tories of life events and symptoms. Their research raises the
question of how life events should best be categorized. Conditional
models are designed to explain individual variability by relating
trajectories to multiple exploratory variables and selected cate-
gories. Duchesne et al. (2010) attempted to address some of these
problems using a multinomial logistic regression and noted the
presence of multiple trajectories of anxiety and the heterogeneous
nature of development.
Attempts have combined discrete models with continuous ones,
in mixed models which allows creating discrete classes or groups,
even when the distribution is continuous. Colman et al. (2007)
attempted to clarify life-course variations in the experience of
anxiety and depression. Using a typological model, they identied
six distinct longitudinal courses from 13 to 53 yrs. But such latent
class analysis still assumes group response patterns underpin a
typological summary. The statistical modelling risks losing indi-
vidual developmental variation. Individual contextual factors may
determine vulnerability.
Several researchers have noted that there is a need to promote
studies on contextual and relational factors which are person
centred and how they mediate the environment with regard to
child development (Luthar & Brown, 2007; Zucker, 2006). To
inform on risk and resilient pathways, it is important to focus on
what makes individual children vulnerable. Are there sensitive
periods for a contextual impact on child mental health, and what
are the mediating relations between different predictors?
Cummings, Davies, and Campbell (2000) described interacting risk
factors across different environments that mutually affect child
adjustment, and specically adolescent depression. Family envi-
ronment, child characteristics and everyday psychological func-
tioning mutually interacted and covaried with each other.
Finally emphasising the difculty pinpointing general factors in
an interactive reciprocal eld, Karevold, Rysamb, Ystrom, and
Mathiesen (2009) looked at the specic developmental period
where children are most exposed to risk from early infancy to early
adolescence. In other words, the authors examined both predictors
and pathways (Kendler, Gardner, & Prescott, 2002, 2006; Jo reskog
& So rbom, 1993). Outcome analysis comprised a higher order fac-
tor model, including four observed variables (child- and parent-
16 K. O'Connor et al. / New Ideas in Psychology 39 (2015) 12e22
reported indices of anxiety and depression), two rst-order factors
(anxiety and depression), and a second order factor (core inter-
nalizing problems). The predictor and mediator part of the model
included all ve predictor variables observed at four time points.
The basic model allowed cross-time, within-variable paths, to
represent the stability of the variables. But no general risk factors
predicted child anxiety, although internalizing problems featured
as a general latent factor mediating other variables. Bayer, Hiscock,
Ukoumunne, Wake, and Price (2008) reported early temperamental
and environmental factors both contributed to later anxiety, and
Mian, Wainwright, Briggs-Gowan, and Carter (2011) also reported
that temperament moderated environmental risks (such as
depressed or anxious parent) in the later development of anxiety.
The picture becomes even more complicated if we wish to look at
complex issues such as development of identity (Crocetti et al.,
2009).
2.5. Reciprocal perceptions on development
Additionally in trajectory research, there may be different ac-
counts of life experiences from different sources. There is the child's
perspective, the parents or authorities, and their interaction. One
theory that attempts to predict and explain major personality or
mental-health-related consequences with perceived parental
acceptance and rejection is parental acceptance-rejection theory
(PART) conceptualized by Rohner, Rohner, and Rool (1984) and
Rohner (1999). This aspect of the PART is viewed from two per-
spectives: from the subjective perception of a child (such as often
presented in research) or parents and as objectively observable
behaviour. Accepting parents are described as those who show
their love and affection towards children and induced the child to
feel loved and accepted. Rejecting parents in the PART are those
who dislike, disapprove of, or resent their children, which man-
ifested in two principal ways: in the form of hostility and aggres-
sion and in the form of parental indifference and neglect.
Vuli
c-Prtori
c and Macuka (2006), for example, compared the
interaction of contextual (perception of family) and personal
(coping strategies) factors. Different conceptions of the associations
between the child's early temperament, the family environment
and subsequent internalizing problems have been reported. For
instance, one view holds that parent and child factors are not
separable in parent reports such that parent reports would simply
reect the parent-child relationship (Foreman & Henshaw, 2002).
Difcult child temperament could inuence parents' response to
signals and vice versa. Another view recognizes the importance of
both reciprocal and transactional relationships within families,
with the child's characteristics inuencing his or her own devel-
opmental course as a function of the complex interplay between
the parents (Goodman & Gotlib, 1999).
2.6. Clinical psychological consideration
The preceding review emphasized the importance of individual
differences, both in terms of multicausal and multinality path-
ways. Such progressions inevitably involve different conditional
probabilistic accounts and identifying transition points and are
multicontextual including not only immediate life and social con-
texts but cultural contexts. There has been a move away from
standardization towards pluralisation of trajectories in mental
illness. Trajectories of individuals with severe psychiatric problems
are diverse with different outcomes (Harding, 1988; Mller, Sapin,
Gauthier, Orita, & Widmer, 2011). There are wide differences in
pluralisation across countries, gender and class. Mapping rela-
tionship trajectories, for example, takes the form of a retrospective
survey in episode format which allows construction of a
sequential representation that each individual was experiencing at
each age. Stulz, Thase, Klein, Manber, and Crits-Christoph (2010)
looked at symptom trajectories in those treated by cognitive
behaviour therapy (CBT) and pharmacotherapy and identied three
subgroups with different rates of recovery, depending on combi-
nation of treatments and baseline anxiety. Stulz et al. used a latent
growth model to look at differential effects of treatment for chronic
depression. The authors identied three distinct patient subgroups
who showed substantial heterogeneity in their treatment. Further
baseline characteristics permitted correct allocation of 61% of the
sample to trajectory subgroups.
The overall conclusions of these studies reveal that these are
different classes of trajectories, that the classes develop at different
rates, interact differentially with life events and may have different
baseline characteristics. Focussing on specic variables and their
interaction is essential. But these variables differ at different stages
and clinical psychologists are interested in the processes of devel-
opment, their degree and rates rather than an absolute event. It
often seems the very factors determining trajectories are not easily
identiable as variables but rather as personal transitions, changes
and reverses in life. Turning points involve abrupt and substantial
change from one state to another (Black, Holditch-Davis, & Miles,
2009).
Rutter, 1996 critically reviews the concept of turning points
(apparent changes in early adult life of psychological life trajec-
tories) in relation to developmental psychopathology. Turning
points have been associated with a wide range of experiences
including those over which the individual has no control and
those subject to individual choice and actions. Changes in life
trajectory may also reect genetic programming or internal bio-
logical alterations rather than external happenings. The author
concludes that turning points represent a heterogeneous range of
lasting changes in psychological functioning, and their investiga-
tion may shed light on developmental processes. Since personal
turning points do not represent a simple mechanism, there is no
particular process responsible for them. They are nevertheless
crucial for an understanding of developmental continuities and
discontinuities.
One attempt to overcome this limitation on examining the
processes of development in trajectory research involves identi-
fying personal experiences which may lead on to cognitive ap-
praisals and coping mechanisms. Careau, O'Connor, Turgeon, and
Freeston (2012), in a series of studies, have tried to model the
relationship between child experiences and the development of
beliefs and appraisals in adulthood, which may mediate obses-
sional symptomatology. In a more recent paper, Careau, O'Connor,
Turgeon, and Freeston (2014) explored the Affect Infusion Model
(AIM) as an integrative framework for understanding cognitive
reactivity on the basis of the preferential reliance on different levels
(or strategies) of processing and the conditions of their use. In order
to explore the dynamics of cognitive reactivity in obsessive
compulsive disorder (OCD), this study provided an analysis of the
longitudinal, naturalistic, covariation between transient mood
states and appraisals in eight individuals with obsessional rumi-
nation. Participants completed diaries recording mood-states and
appraisals over the course of a cognitive behaviour therapy (CBT)
for OCD. Based on the sequential analysis of conditional probability,
contingency magnitude estimations of covariations revealed
important rates of cognitive reactivity in ve participants. Indi-
vidual case analyses and aggregated correlational analyses revealed
that the results of the covariation analyses may be closely related to
the participants' neutralization proles. Careau et al. (2015) linked
adoption of neutralizing strategies and information processing
styles to the importance accorded to type of emotional input. So
again reactions to one variable may chain on to behavioural
 K. O'Connor et al. / New Ideas in Psychology 39 (2015) 12e22
reactions which may feedback to emotional coping and stress (Stulz
et al., 2010).
2.7. Narrative additions to trajectory research
We have noted the importance of the perception on stress, of
personal evaluations of events and interactions, and the impor-
tance of perceived coping. Narrative approaches lead to a
constructionist approach rather than a realist approach which fo-
cuses on events and narratives that the person perceives as
important. Most importantly these narratives and events are
complex and dynamic. The person can recall with the aid of a life-
history approach key events from childhood. A narrative approach
ts well with a dynamic person-based unit of trajectory analysis,
talking of events, transition phases, turning points and generally
human rather than statistical signicance. Promising results have
come from qualitative/narrative approaches. Such accounts have
specically targeted specic transitional steps in the attempts to
understand the contextual aspects of psychological processes.
de Guzman, Llantino, See, Villanueva, and Jung (2008), for
example, probed the life trajectories of aged working professors
and charted points where the respondents felt successful, fullled
or anxious. Their self-esteem coincided with success of their stu-
dents and contribution to academic environments. Thomas and
Hall (2008) adopted a narrative approach to following the pace of
a healing trajectory in survivors of child abuse. These authors
identied four types of redemption narratives associated with
successful adaptation. Similarly, Se guin, Renaud, Lesage, Robert,
and Turecki (2011) used a life course method in suicide research
to quantify burden over time and identied two distinct life tra-
jectories of early and late onset differentially exposed to adversity.
Robert, Se guin, and O'Connor (2013) examined pathways and
transition points in post-adolescent adaptation and found that
emerging new experiences can redirect life trajectories.
Nelson (2010) adapted the life history calendar methodology
(LHC) for qualitative assessment, using a semi-structured interview.
The LHC approach captures the processes of engagement and
disengagement in activities at specic times. It looks at attitudes,
aspirations, interpretations and explorations of life transitions and
major events that would not be captured by objective events.
Constructive activity captures the role of human agency. The LHC
can be completed by interview or by the person. It is reliable and
includes a matrix dealing with event timing and domains and
listing key reference events (birth, marriage). Both calendar and
memories activate recall through multiple pathways: sequentially
within a life theme and across parallel themes involving simulta-
neous or sequential events. The LHC can be adapted to open ended
questioning and to elicit narratives. The qualitative approach co-
constructs time and event cues with the person, creates rapport
and elicits in depth narratives. Because interviewers do not know
the difcult periods, allowing respondents to discuss their
sequence of events captures the narrative of emotionally sensitive
events, physical stickers and markers on the calendars give a per-
sonal visual sense to LHC. An extension of the narrative approach is
the life course theory approach. The life course approach as well as
emphasizing trajectories on paths of change in the developmental
process (van Geert, 1994) also looks at the impact of transitions and
the entry points of new states.
From a constructionist point of view, narratives reveal how a
person makes sense of events and contextualizes events in terms of
the person's experience and permit various modalities to gain
expression. Russell and van den Broek (1992) have noted that the
structure of narratives require events, relations and contingencies.
Narratives capture the transitions, reactions to events and the
relational structure of events. Events are grounded in the
17
experience and adaptive context of the person. In a narrative, one is
not concerned with veridical perception. Also narratives occupy
space in that the events make up a cohesive space rather than
constitute separate points spread over time.
There are many different types of narrative analysis, depending
on the aspect of the narrative to be captured (Riessman, 1993). From
a constructionist point of view the narrative serves several pur-
poses: it reveals how the person makes sense of events; and how he
or she has represented the event to themselves (through language
and metaphor); at the same time, it reveals the client's point of
view and is persuasive employing rhetoric and reason; and it
contextualizes the problem in surrounding events (White &
Epstein, 1990).
The main advantage of a narrative approach is that it does not
treat observations as stand-alone thoughts units, judged according
to their orientation bias or compared with veridical perception, nor
does it reduce them to products of hypothetical schemas imposing
an embedded core network of perceptions on reality. A narrative
approach grounds thoughts in the experience and adaptive context
of the person. The unit becomes the voice(s) of the person intro-
duced within a narrative, to account for the problem, discuss its
development, elaborate on its implications or relate it to other as-
pects of life alongside other problems, non-problems and other
events. Zimmerman and Dickerson (1994) note that narratives
evolve through time and are uid and capture recursive patterns.
The following abbreviated narrative describes a childhood
experience.
I was fooling around with my younger brother who I always
tried to protect. I think we were playing soldiers and he hit me
hard which was unexpected and I reacted badly and got very
angry which surprised me. I guess it was the beginning of my
seeing that I couldn't always trust people!
Here the person describes a transition point with marked re-
actions and emotions which dene this experience which is likely
carried over to other similar situations. This event was distinct from
other events where the person describes for example a sense of
belongingness in a group and pleasant emotions.
At the same time this repertoire of event spaces allows us to see
how the person did not react but may have reacted given their
sensitivity to subtle changes in the episode context. Narrative lled
event spaces may reveal a pattern of reactions and although event
spaces are mutually exclusive the spaces may overlap and contain
complex multimodal reactions such as perception, emotion,
behaviour, which leaves the space open to be dened by a complex
theme rather than connecting isolated variables. So for example a
series of experiences reporting difculties in relating and feelings
and perceptions of isolation might be considered a theme of
alienation dened by person centred content and interactions. A
future episode may be likely to elicit a similar theme. So similar
themes carry over different events and represent one pattern of
experience e possibly distinct from other patterns in the same
repertoire. A select number of complex themes could typify this
person's reaction over a range of situations.
In the case of a hoarder who hoards garbage and objects she
nds in the street, she recalls a deprived childhood; her parents
separated when she was ve and she was passed around among
relatives. She felt abandoned and she lost contact with her siblings.
She was bullied at school because of her teeth and traumatized by a
re. She was addicted to drugs, and in addition was diagnosed with
anxiety sensitivity and borderline personality. She has trouble
maintaining relationships, keeping a job. She avoids contact and
social activities but tends to lurch from one crisis to another. What
is her trajectory? At the individual case level, what were her
18 K. O'Connor et al. / New Ideas in Psychology 39 (2015) 12e22
turning points? She recalls her parents separating around her fth
birthday. She remembers distinctly the raised voices and feeling
she was to blame. She remembers being an outsider among rela-
tives, and she remembers the occasion when she had to ask for a
present. She remembers several occasions when she felt excluded
from events or treated like a stranger. At the same time she
remembered specic incidents where she was good at coping and
problem-solving and helping others out.
The person repeats several complex narratives which she recalls
from childhood. These form discrete event spaces which she per-
ceives as key events and transition points in her life trajectory. We
can see from the ow of events in Fig. 1 that whereas there are
several possibilities at the beginning, certain events and self-
perceptions seem enhanced over the years, with other positive or
resilient factors conditioning the progression and her reection on
the problem. These are her perceptions and constructions of life
events. A more objective account might seek to measure the nature
and impact of specic events, and perhaps look at underlying
temperament. But the narrative life history approach provides
event spaces and transition points revealed by the person to be
important in their life history and forming a necessary complex and
right or
Fig. 1. Schema combining qualitative event spaces, critical incidents, transition points, expressed as conjunctions/dysjunctions leading onto to multiple nal pathways.
conditional portrait of personally signicant factors. Reactions in
the event space are conditioned by the perception of the event but
form a series of conditional probabilities of how the person may
react to similar events in the future.
This event approach is distinct from a mixed method approach
where qualitative and quantitative are combined. A problem with
the mixed methods approach is that subjectivity and objectivity are
not exclusive in the sense that quantied variables are often self-
report, but there is a tendency for numerical methods to take
precedence over narrative approaches. The idea of dividing
research into separate categories and creating a mixed method, is
erroneous, not because methods should be kept separate but
because they should not have been divided at the outset (Gorard,
2007).
In the constructed event approach the only data are the con-
structed events which may be complex or thematic. Some con-
structions repeat themselves whereas other possibilities may occur
once and are not afterwards reected in subsequent constructed
events. For example, the person is able to relate a whole series of
events from different points in her life where she felt excluded.
These events marked her because partly she was not expecting
 K. O'Connor et al. / New Ideas in Psychology 39 (2015) 12e22
them but at the same time they conrmed her difculty with
people. The events were reported as separate transition points but
the events nally seemed to form a composite pattern revealing her
identity as a strong but excluded person needing to nd her own
well-protected space. Such an individual person centred trajectory
encompassing her emotions, perceptions and behaviour in the
same event belies an analysis which treats interacting factors as
isolated variables which does not capture their embeddings.
Consequently the current dominant discourse concerns are avoid-
ance of people, a love of objects, a perfectionism, a feeling of
possession of her house, a sense of loss of people but a selective
coping ability, which lead on to the various diagnosable problems
of schizotypy, borderline personality, hoarding and some depres-
sion. In her constructed trajectory of events, a psychotic, anxiety
and other personality problems seem less likely.
From a variable centred point of view, there were a number of
cumulative events. But from an experiential qualitative point of
view, she remembers distinctly how her mother valued objects
over her, how she was reprimanded severely for breaking an object,
how she felt abandoned and decided that by increasing the number
of objects around her, she would feel security. She remembers the
turning point at which she felt a profound feeling of security sur-
rounded by objects. She remembered feeling objects could not hurt
her, they were solid, dependable. She remembers when at school
she knew she couldn't trust others. Her best friend betrayed her
over some trivial event and she recalls this incident marked her
view of humans. Some of these transition points came as revela-
tions in recounting her narrative, other points from gradual insight,
and she was aware that the development of her resilience and risk
factors processes grew at different rates. Her sense of rejection and
loss grew faster than her realization of her good coping skills.
Hence the coping tended to t around her identity as different and
rejected. The individual variation in experience becomes informa-
tion not easily quantiable but revealed in a qualitative narrative
whose key events form independent thematic event spaces whose
likely contributions can then be calculated (see Table 1a).
The personal event approach seems to overcome some of the
limitations noted above amongst quantitative methods by being
person-centred, respecting individual trajectories, and the timing
of such trajectories, considering the complex interplay of processes
in anyone reaction. The question is to what extent can constructed
events which are certainly a mixture of fact and ction predict
future events? In fact, by their nature narratives lead to connect-
edness of events over time. Stories involve the plotting of events
over time but yet the narratives themselves can be encapsulated in
event spaces. Early events are constructed as an event space rather
than a series of discrete isolated time points interacting together.
Each narrative reporting a separate episode reveals a unique con-
struction and positioning of events, but the pattern can be thematic
and represent systematic patterns for any individual. A select
number of complex themes could typify this person's reaction over
a range of situations. The themes could be combined into a likeli-
hood distribution where a future episode will likely induce one
Table 1a
Gives objective events and comparable subjective constructions of events which lead to the dominant discourse, but could lead to alternative constructions of events.
Age (yrs.) Objective events
5 Parental separation
5e10 Adopted by family passed from relative to relative
Authorisation parenting
10e12 Bullied at school due to appearance
13 Trauma in re
14e16 Dependence on drugs
18e20 Success at school, Professional qualications
21
Abandonment of people in favour of material success
19
theme rather than another. The likelihood of a theme manifesting
itself in future situations could be updated on the basis of Bayesian
type prior and posterior likelihood (see O'Connor, 2015). The
presence or absence of themes characterising one episode rather
than another could be dened by correspondence analysis
(Greenacre, 2007).
The event space is complex and thematic and dened qualita-
tively as for example The child who felt abandoned in family
interaction but developed a forceful identity with good coping
skills. This complex theme could form a latent category which
instead of isolated or general features may predict behaviours in
future episodes. This experiential grouping could replace an ab-
stract reied grouping for group based trajectory methods. Other
critical incidents can always be identied to support or nuance a
dominant theme surrounding an event until a complex narrative
comprising, turning points, personal themes and constructed in-
cidents becomes the most likely event space leading to the nal
endpoint. This example is illustrated in Fig. 1. Elements of early
themes most likely reecting a current dominant theme are given
together with endpoints. The example is individual but content
analysis could focus on the frequency of terms across discourses or
the implied themes across discourses.
Alternatively discourse analysis could reveal repeatable re-
lations between events, people and actions from similar trajec-
tories. So, for example, a discourse about the person feeling rejected
in an adoption family relates her emotion of rejection over time but
within a specic series of events: a party, an outing; a family get
together, where the same pattern of her being treated separately
and apart from others repeated itself. If we look at what seem to be
constructions of events which were problematic, we certainly nd
a pattern where she recalls emotions of being abandoned, isolated
and events (physical rejection by others) and reactions (judging
herself abnormal) which we can nd in the current dominant
discourse (I feel detached and I avoid people and I love objects and
give them names and keep well inside my house). But when
focussing again on related events, she remembers that a teacher
once made her favourite, she shared experiences listening to music
with others, the friend who betrayed her apologized and didn't
want to lose her as a friend. Plus she recalled numerous instances of
her helping out people and being appreciated. In re-telling the
constructed experience the person may be able to identify
neglected aspects of lived experience that are not available in the
dominant discourse but become available through the retelling of
events. This retelling could lead to a therapeutic repositioning
which might eventually lead to dislodging the dominant discourse
(see Table 1b).
3. Discussion
The literature reveals a bewildering array of variables for in-
clusion in trajectory models. Life events, stress, temperament are
important but inherently these factors interact and produce
distinct cumulative effects at distinct stages and groups. Further,
Constructed events
Recalls incident of parents ghting I am to blame
I mage of being outside family events I am strange
Rigid life I must have rules
Sense of betrayal over a friend revealing a secret I can't trust people
Sense of relief at surviving I am a survivor
Recourse to escape people I will live in my own world
Sense of coping by self I have no need of others
Sense of security and safety around objects I'm Ok with my belongings
20 K. O'Connor et al. / New Ideas in Psychology 39 (2015) 12e22
Table 1b
The universe of events ts into a bounded event space where independent constructions coalesce into a thematic narrative.
Dominant construction of the world possible discourse
I see myself as a strong but abnormal person
who is best not to trust others but stay close to
myself and belongings and organize myself with little
contact with others to upset me.
In Table 1b past Independent constructions come together in the dominant discourse, but importantly they permit construction of a possible
alternative springing from the same past constructions which fuelled the dominant discourse. Where the dominant discourse is dysfunctional
an alternative needs to be rendered feasible and credible for the client.
the interaction is not passive but depends on the person's
perception and on the way he or she copes. It may depend on
general sensitivities, learned experiences, subtle adversities, ex-
periences and self-beliefs. Is it not time to abandon the practice of
isolated variables as single effects and dene units rather as pro-
cesses, dening events, transition points, meaningful to the
person?
Most studies have adopted linear analysis of variance models,
hierarchical modelling or latent curve analysis. While the models
underlying hierarchical modelling and latent curve analysis can
reveal important communalities, they assume a population distri-
bution of trajectories. Modelling individual differences requires
that assumptions be made about the distribution of trajectory pa-
rameters as a multivariate or multinomial distribution.
Other complications arise when we consider that the same
factor at distinct stages of development may have distinct effects
and that the behaviour at one stage may itself be an outcome due to
prior experiences rather than a predictor itself. One interesting
nding covering a number of pathologies is that up to a certain
point, precursors of anxiety, depression and psychosis seem to
share a common thread of negative affect and life adversity. This
raises the debate of multi-nality or common pathways. Are there
many routes to the same problem? Or do all pathologies share a
common pathway with bifurcation or transition points on the way,
whereby peculiar inuences create the climate for depression or
psychosis?
The question then is what methods are appropriate for a tra-
jectory analysis? One recourse has been to attempt further and
further sophistication of path analyses and LGMM to attempt to
map and weight multivariate contributing variables. However,
there are problems with such approaches if the pathways include
the use of nested designs, where variables are conceived a priori as
nested within each other and explorations ordered according to
short-term, medium- or long-term contributions. However, most
promising seem methods which adhere more strictly to the prin-
ciple of trajectory, that is events which are sequential, dependent,
non-linear, interactive and reciprocal. These statistical models
eschew conventional variance assumptions and employ event-
based methods, likelihood or probit models to assess the co-
occurrence of events. Narrative approaches may be useful here
since they t a stream of events through a time-based stream of
consciousness, and permit at least a qualitative appreciation of the
syntax of multi-level events and experiences. These methods dis-
cussed here also highlight micro-management of life trajectories in
clinical practice.
Already clinical psychology has introduced cognitive subtleties
into the reciprocal nature of trajectories. Other cognitive variables
might be meta-cognition, personality traits, beliefs, subtle action
tendencies. There is also evidence that we can change organization
of our trajectories for therapeutic purposes. Palgi and Shmotkin
(2010) in a study of holocaust survivors found adopting a time
trajectory, according to perceived nearness or farness of the event
improves psychological health. As illustrated in Table 1b it may be
Alternative
I see myself as strong and independent person
although I need be careful with others, others
can be helpful and I have projected from having
friends and contacts as well as being true to myself.
possible to reposition a dominant narrative through generating
alternatives.
Risk and resilience are not static and unchangeable but dynamic,
adaptive and malleable. Different interventions may be successful
at different ages and may involve cognitive and metacognitive
strategies for dealing with apparently unresolvable environmental
issues. Clinical thinking needs to go beyond simple facts and iso-
lated attributes towards thinking in terms of conditional and cu-
mulative probabilities and contextual factors. Achieving this may
not require advanced quantitative methods but rather more
discerning qualitative approaches detailing a pattern of event
spaces.
Narrative and other qualitative methods may be indispensable
to track psychological factors in the trajectory equation. What does
the person rather than the experimenter see as a turning point or as
a resilience factor? On the other hand charting the development of
cognitive and psychological characteristics of a subtle concept such
as identity requires mapping not only self-perceptions but peer
perceptions, parenting, and most of all, the person's role
experience.
An important mission in clinical psychology is to develop
comprehensive methods that explain how different individuals
progress along different life trajectories. All this can be very
beguiling for the clinical psychologist faced with taking a patient's
history. Asking the client to tell me about yourself, in an open
question, may seem satisfactorily non-judgemental, but is the
person likely to represent key personal events in an interactive and
systematic enough way to catch their cumulative impact? In
highlighting certain events, the person may already be positioning
themselves for the interview. There have been attempts to develop
trajectory-based interviews (e.g. Se guin, DiManbro, & Chawky,
2010) but more specialized approaches are required where the
person recounts in detail life events and their reactions to them.
Although group-based latent models go some way to individu-
alizing trajectories, they are still bound by statistical constraints,
where these assumptions can be met and the groups satisfactorily
dened according to meaningful clinical criteria, pathways and
particularly treatment pathways can be mapped. But otherwise
recourse to qualitative and narrative approaches seems appropriate
and constructed themes can form the basis for group based models.
We know retrospective recall is a biased affair subject to
contextual mood and selective recall. The selective recall may
further be biased by complaint behaviour. However, reporting on
concrete experiences and transitions and turning points is likely to
be more accurate than abstract recall. The life calendar method of
Nelson (2010) offers a key to anchoring recall in key landmark
events, rather like squares on a game board. The person is likely to
recall their detailed experiences around these key points. One could
then calculate the most likely events that were concurrent with the
turning point, including cognitive and emotional reactions to
outside events, and identify the a posteriori likelihood of the early
patterns reported as early transition points reecting current pat-
terns (as illustrated in Fig. 1).
 K. O'Connor et al. / New Ideas in Psychology 39 (2015) 12e22 21

References Kendler, K. S., Gardner, C. O., & Prescott, C. A. (2002). Toward a comprehensive
developmental model for major depression in women. American Journal of
Psychiatry, 159, 1133e1145.
Bayer, J. K., Hiscock, H., Ukoumunne, O. C., Wake, M., & Price, A. (2008). Early
Kendler, K. S., Gardner, C. O., & Prescott, C. A. (2006). Toward a comprehensive
childhood aetiology of mental health problems: a longitudinal population-
developmental model for major depression in men. American Journal of Psy-
based study. Journal of Child Psychology and Psychiatry, 49(11), 1166e1174.
chiatry, 163, 115e124.
Black, B. P., Holditch-Davis, D., & Miles, M. S. (2009). Life course theory as a
Kortlander, E., Kendall, P. C., & Panichelli-Mindel, S. M. (1997). Maternal expecta-
framework to examine becoming a mother of a medically fragile preterm infant.
tions and attributions about coping in anxious children. Journal of Anxiety
Research in Nursing and Health, 32(1), 38e49.
Disorders, 11, 297e315.
Careau, Y., O'Connor, K., Turgeon, L., & Freeston, M. H. (2012). Childhood experiences
Luby, J. L., Belden, A. C., & Spitznagel, E. (2006). Risk factors for preschool depres-
and adult beliefs in OCD: evaluating a specic aetiological model. Journal of
sion: the mediating role of early stressful life events. Journal of Child Psychology
Cognitive Psychotherapy: An International Quarterly, 26(3), 236e256.
and Psychiatry, 47, 1292e1298.
Careau, Y., O'Connor, K. P., Turgeon, L., & Freeston, M. H. (2014). Cognitive reactivity
Luthar, S. S., & Brown, P. J. (2007). Maximizing resilience through diverse levels of
in obsessional rumination: a naturalistic and longitudinal investigation. Journal
inquiry: prevailing paradigms, possibilities and priorities for the future.
of Experimental Psychopathology, 5(4), 425e456.
Development and Psychopathology, 19, 931e955.
Carr, A. (1999). The handbook of child and adolescent clinical psychology. London:
Malkoff-Schwarts, S., Frank, E., Anderson, B. P., Hlastala, S. A., Luther, J. F.,
Routledge.
Sherrill, J. T., et al. (2000). Social rhythm disruption and stressful life events in
Colman, I., Ploubidis, G. B., Wadsworth, M. E. J., Jones, P. B., & Croudace, T. J. (2007).
the onset of bipolar and unipolar episodes. Psychological Medicine, 30,
A longitudinal typology of symptoms of depression and anxiety over the life
1005e1016.
course. Biological Psychiatry, 62, 1265e1271.
M'Bailara, K., Cosnefroy, O., Vieta, E., Scott, J., & Henry, C. (2013). Group-based
Compas, B. E., Beckjord, E., Agocha, B., Sherman, M. L., Langrock, A., Grossman, C. I.,
trajectory modeling: a novel approach to examining symptom trajectories in
et al. (2006). Measurement of coping and stress responses in women with
acute bipolar episodes. Journal of Affective Disorders, 145, 36e41.
breast cancer. Psycho-Oncology, 15(12), 1038e1054.
Mian, N. D., Wainwright, L., Briggs-Gowan, M. J., & Carter, A. S. (2011). An ecological
Compas, B. E., Connor-Smith, J., & Jaser, S. S. (2004). Temperament, stress reactivity,
risk model for early childhood anxiety: the importance of early child symptoms
and coping: implications for depression in childhood and adolescence. Journal
and temperament. Journal of Abnormal Child Psychology, 39, 501e512.
of Clinical Child and Adolescent Psychology, 33, 21e31.
Mller, N. S., Sapin, M., Gauthier, J.-A., Orita, A., & Widmer, E. D. (2011). Pluralized
Compas, B. E., Malcarne, V. L., & Fondacaro, K. M. (1988). Coping with stressful
life courses? An exploration of the life trajectories of individuals with psychi-
events in older children and young adolescents. Journal of Consulting and
atric disorders. International Journal of Social Psychiatry, 58(3), 266e277.
Clinical Psychology, 56, 405e411.
Muthe n, B. (2004). Latent variable analysis: growth mixture modeling and related
Connor, K. M., & Davidson, J. R. (2003). Development of a new resilience scale: the
techniques for longitudinal data. In D. Kaplan (Ed.), Handbook of quantitative
Connor-Davidson Resilience Scale (CD-RISC). Depression and Anxiety, 18(2),
methodology for the social sciences (pp. 345e368). Newbury Park, CA: Sage
76e82.
Publications.
Crocetti, E., Klimstra, T., Keijsers, L., Hale, W. W., III, & Meeus, W. (2009). Anxiety
Nagin, D. S., & Odgers, C. L. (2010). Group-based trajectory modeling in clinical
trajectories and identity development in adolescence: a ve-wave longitudinal
research. Annual Review of Clinical Psychology, 6, 109e138.
study. Journal of Youth Adolescence, 38, 839e849.
Nelson, I. A. (2010). From quantitative to qualitative: adapting the life history cal-
Cummings, E. M., Davies, P. T., & Campbell, S. B. (2000). Developmental psychopa-
endar method. Field Methods, 22(4), 413e428.
thology and family process. New York: Guilford Press.
Nolen-Hoeksema, S., & Watkins, E. R. (2011). A heuristic for developing trans-
Duchesne, S., Larose, S., Vitaro, F., & Tremblay, R. E. (2010). Trajectories of anxiety in
diagnostic models of psychopathology. Explaining multinality and divergent
a population sample of children: clarifying the role of children's behavioural
trajectories. Perspectives on Psychological Science, 6(6), 589e609.
characteristics and maternal parenting. Development and Psychopathology, 22,
O'Connor, K. P. (2015). A constructionist clinical psychology for cognitive behaviour
361e373.
therapy. London and New York: Routledge.
Fields, L., & Prinz, R. J. (1997). Coping and adjustment during childhood and
Palgi, Y., & Shmotkin, D. (2010). The predicament of time near the end of life: time
adolescence. Clinical Psychology Review, 17, 937e976.
perspective trajectories of life satisfaction among the old-old. Aging & Mental
Foreman, D. M., & Henshaw, C. (2002). Objectivity and subjectivity in postnatally
Health, 14(5), 577.
depressed mothers' perceptions of their infants. Child Psychiatry Human
Phillips, N. K., Hammen, C. L., Brennan, P. A., Najman, J. M., & Bor, W. (2005). Early
Development, 32, 263e275.
adversity and the prospective prediction of depressive and anxiety disorders in
Fox, J. K., Halpern, L. F., Ryan, J. L., & Lowe, K. A. (2010). Stressful life events and the
adolescents. Journal of Abnormal Child Psychology, 33(1), 13e24.
tripartite model: relations to anxiety and depression in adolescent females.
Rapee, M. M., Schniering, C. A., & Hudson, J. L. (2009). Anxiety disorders during
Journal of Adolescence, 33, 43e54.
childhood and adolescence: origins and treatment. Annual Review of Clinical
Gagadinho, A., Ritschard, G., Studer, M., & Mller, N. S. (2008). Mining sequence data
Psychology, 5, 335e365.
in R with the TraMineR package: A user's guide. Geneva: University of Geneva.
Riessman, C. K. (1993). Narrative analysis. Newbury Park, CA: Sage.
Gauthier, J.-A., Widmer, E., & Dame, C. (2010). Multichannel sequence analysis
Robert, M., Se guin, M., & O'Connor, K. (2013). Pathways to adulthood for foster care
applied to social science data. Sociological Methodology, 40(1), 1e28.
adolescents: integrating the insights of developmental psychology and social
van Geert, P. (1994). Dynamic systems of development: Change between complexity
work. International Journal of Humanities and Social Science, 3(7), 132e140.
and chaos. New York: Harvester Wheatsheaf.
Rohner, R. P. (1999). Acceptance and rejection. In D. Levinson, J. Ponzetti, &
Ge, X., Lorenz, F. O., Conger, R. D., Elder, G. H., Jr., & Simons, R. L. (1994). Trajectories
P. Jorgensen (Eds.), Encyclopedia of human emotions (Vol. 1, pp. 6e14). New York:
of stressful life events and depressive symptoms during adolescence. Develop-
Macmillan Reference.
mental Psychology, 30(4), 467e483.
Rohner, E. C., Rohner, R. P., & Rool, S. (1984). Perceived parental acceptance-rejection
Ge, X., Natsuaki, M. N., & Conger, R. D. (2006). Trajectories of depressive symptoms
and children's reported behavioral disposition. Journal of Cross-Cultural Psy-
and stressful life events among male and female adolescents in divorced and
chology, 11, 221e225.
nondivorced families. Development and Psychopathology, 18, 253e273.
deRoon-Cassini, T. A., Rusch, M. D., Mancini, A. D., & Bonanno, G. A. (2010). Psy-
Goodman, S. H., & Gotlib, I. (1999). Risk for psychopathology in the children of
chopathology and resilience following traumatic injury: a latent growth
depressed mothers: a developmental model for understanding mechanisms of
mixture model analysis. Rehabilitation Psychology, 55(1), 1e11.
transmission. Psychological Review, 106, 458e490.
Russel, R. L., & Van den Brock, P. (1992). Changing narrative schema in Psycho-
Gorard, S. (2007). Mixing methods is wrong: An everyday approach to educational
therapy. Psychotherapy, 29(3), 344e354.
justice. London: British Educational Research Association.
Rutter, M. (1996). Connections between child and adult psychopathology. European
Greenacre, M. (2007). Correspondence analysis in practice (2nd ed.). London:
Child and Adolescent Psychiatry, 5, 4e7.
Chapman & Hall/CRC.
Rutter, M. (2006). Implications of resilience concepts for scientic understanding.
de Guzman, A. B., Llantino, M. L. S., See, V. C. L., Villanueva, R. F. P., & Jung, Y. (2008).
Annals of the New York Academy of Sciences, 1094, 1e12.
Horizons, not boundaries, are the language of retirement: sense of fullment,
Rutter, M., Kim-Cohen, J., & Maughan, B. (2006). Continuities and discontinuities in
fears, and life trajectories of faculty extendees in the Philippines. Educational
psychopathology between childhood and adult life. Journal of Child Psychology
Gerontology, 34(9), 749e762.
and Psychiatry, 47, 276e295.
Harding, C. M. (1988). Course types in schizophrenia: an analysis of European and
guin, M., DiManbro, M., & Chawky, N. (2010). Questionnaire sur les trajectoires de
Se
American studies. Schizophrenia Bulletin, 14, 633e643.
vie. Montre al: Que bec: Premie
re Publication. January 2002.
Haviland, A., Nagin, D. S., Rosenbaum, P. R., & Tremblay, R. E. (2008). Combining
guin, M., Renaud, J., Lesage, A., Robert, M., & Turecki, G. (2011). Early adversity as
Se
group-based trajectory modeling and propensity score matching for causal
risk factor for suicide among youth and young adult suicide: a comparative
inferences in nonexperimental longitudinal data. Developmental Psychology, 44,
study of life trajectory. Journal of Psychiatric Research, 45(7), 863e870.
422e436.
rbom, D. (1993). LISREL 8 user's reference guide. Chicago: SSI. Smith, G. T. (2009). Why do different individuals progress along different life tra-
Joreskog, K. G., & So
jectories? Perspective in Psychological Sciences, 4(4), 415e421.
Karevold, E., Rysamb, Ystrom, E., & Mathiesen, K. S. (2009). Predictors and path-
Stulz, N., Thase, M. E., Klein, D. N., Manber, R., & Crits-Christoph, P. (2010). Differ-
ways from infancy to symptoms of anxiety and depression in early adolescence.
ential effects of treatments for chronic depression: a latent growth model
Developmental Psychology, 45, 1051e1060.
reanalysis. Journal of Consulting and Clinical Psychology, 78(3), 409e419.
Keller, P. S., Cummings, E. M., Davies, P. T., & Mitchell, P. M. (2008). Longitudinal
Tait, L., Birchwood, M., & Trower, P. (2004). Adapting to the challenge of psychosis:
relations between parental drinking problems, family functioning, and child
personal resilience and the use of sealing-over (avoidant) coping strategies.
adjustment. Development and Psychopathology, 20, 195e212.
British Journal of Psychiatry, 185, 410e415.
22 K. O'Connor et al. / New Ideas in Psychology 39 (2015) 12e22
Thomas, S. P., & Hall, J. M. (2008). Life trajectories of female child abuse survivors
thriving in adulthood. Qualitative Health Research, 18(2), 149e166.
Timpano, K. R., Keough, M. E., Mahaffey, B., Schmidt, N. B., & Abramowitz, J. (2010).
Parenting and obsessive compulsive symptoms: implications of authoritarian
parenting. Journal of Cognitive Psychotherapy: An International Quarterly, 24(3),
151e164.
Turgeon, L., O'Connor, K. P., Marchand, A., & Freeston, M. (2002). Recollections of
parent-child relationships in patients with obsessive-compulsive disorder and
panic disorders. Acta Psychiatrica Scandinavica, 105, 1e7.
Vuli
c-Prtoric, A., & Macuka, I. (2006). Family and coping factors in the differentia-
tion of childhood anxiety and depression. Psychology and Psychotherapy: Theory,
Research and Practice, 79, 199e214.
White, M., & Epston, D. (1990). Narrative Means to Therapeutic Ends. New York:
W.W.Norton.
Wickers, M., Schrijvers, D., Geschwind, N., Jocobs, N., Myin-Germeys, I., Thiery, E.,
et al. (2009). Mechanisms of gene-environment interactions in depression:
evidence that genes potentiate multiple sources of adversity. Psychological
Medicine, 39, 1077e1086.
Zimmerman, J. L., & Dickerson, V. C. (1994). Using a narrative metaphor: implica-
tions for theory and clinical practice. Family Process, 33, 233e245.
Zucker, R. A. (2006). Alcohol use and the alcohol use disorders: a devel-
opmentalebiopsychosocial systems formulations covering the life course. In
D. Cicchetti, & D. J. Cohen (Eds.), Developmental psychopathology (2nd ed., pp.
620e656). Hoboken, NJ: Wiley.