4 pages due in 48 hours.

Journal of Personality and Social Psychology Transactions Between Life Events and Personality Traits Across the Adult Lifespan Jaap J. A. Denissen, Maike Luhmann, Joanne M. Chung, and Wiebke Bleidorn Online First Publication, July 26, 2018. http://dx.doi.org/10.1037/pspp0000196

CITATION Denissen, J. J. A., Luhmann, M., Chung, J. M., & Bleidorn, W. (2018, July 26). Transactions Between Life Events and Personality Traits Across the Adult Lifespan. Journal of Personality and Social Psychology. Advance online publication. http://dx.doi.org/10.1037/pspp0000196

Transactions Between Life Events and Personality Traits Across the Adult Lifespan

Jaap J. A. Denissen Tilburg University

Maike Luhmann Ruhr University Bochum

Joanne M. Chung Tilburg University

Wiebke Bleidorn University of California at Davis

Life events refer to status changes in important demographic variables, such as employment or marital status. Life events offer an interesting opportunity for studying transactions between environmental changes and personality traits, which are of relevance for diverging theories about the role of environ- mental factors in life span personality development. Yet in spite of the potential importance of life events for personality development, nuanced and sufficiently powered longitudinal designs with frequent assessments of life events and personality traits are lacking. The current study aims to address this gap by examining the associations between different life events and personality trait change, using data from a large, nationally representative, and prospective longitudinal study. Results demonstrated a number of selection effects, indicating that personality traits affect the likelihood that individuals experience certain types of life events. Less frequently, results indicated average effects of life events on personality trait development, both in anticipation of a life event change as well as resulting from it. However, some of these event-related changes ran counter to the notion that personality maturity increases as a result of adopting mature social roles, like parenthood or paid employment. Furthermore, we found significant variation around average event-related trajectories, suggesting that individuals differ in their reactions to life events. Theoretical implications and recommendations for future research are discussed.

Keywords: adult development, Big Five personality traits, life events, longitudinal study, personality development

Supplemental materials: http://dx.doi.org/10.1037/pspp0000196.supp

Personality traits can be defined as relatively stable individual differences in affect, behavior, and/or cognition (Johnson, 1997). Over the past two decades, a large body of research on personality trait development in the Big Five framework has demonstrated substantial rank-order (Roberts & DelVecchio, 2000) and mean- level changes (Roberts, Walton, & Viechtbauer, 2006) across the life span, especially in young adulthood (Bleidorn, 2015; Roberts & Mroczek, 2008). The direction of personality trait changes in this age group is clearly positive (i.e., geared toward higher ma-

turity) as indicated by particularly strong mean-level increases in agreeableness, conscientiousness, and emotional stability (Blei- dorn et al., 2013). In the wake of these findings, there has been an increase in research examining the factors that drive personality trait development. Contextual theories of personality development emphasize the role of life events such as marriage or unemploy- ment. Such life events are usually associated with shifts in social roles and relationships and have been theorized to modify, inter- rupt, or redirect personality trajectories by altering people’s feel- ings, thoughts, and behavior (Bleidorn & Denissen, in press). Furthermore, the particularly high density of normative life events and role transitions during young adulthood (Rindfuss, 1991) may serve as a possible explanation for the accelerated pace of person- ality development during this life stage (Roberts & Davis, 2016).

Life events can be defined as “time-discrete transitions that mark the beginning or the end of a specific status” (Luhmann, Hofmann, Eid, & Lucas, 2012; p. 594). Thus, a life event can be understood as a categorical variable that indicates the occurrence of a qualitative change in life circumstances. Status changes typ- ically refer to shifts in demographic and socioeconomic status, such as marriage and employment. Although life events such as marriage have been thought of as influencing personality devel- opment, rigorous empirical studies of the impact of life events on

Jaap J. A. Denissen, Department of Developmental Psychology, Tilburg School of Social and Behavioral Sciences, Tilburg University; Maike Luhmann, Faculty of Psychology, Department of Psychological Method- ology, Ruhr University Bochum; Joanne M. Chung, Department of Devel- opmental Psychology, Tilburg School of Social and Behavioral Sciences, Tilburg University; Wiebke Bleidorn, Department of Psychology, Univer- sity of California at Davis.

Correspondence concerning this article should be addressed to Jaap J. A. Denissen, Department of Developmental Psychology, Tilburg University PO Box 90153, 5000 LE Tilburg, the Netherlands. E-mail: jjadenissen@ gmail.com

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

Journal of Personality and Social Psychology © 2018 American Psychological Association 2018, Vol. 1, No. 2, 000 0022-3514/18/$12.00 http://dx.doi.org/10.1037/pspp0000196

1

lasting personality trait change are rare (Bleidorn, Hopwood, & Lucas, 2018). The current study aims to address this gap by examining the associations between different life events and per- sonality trait change, using data from a large, nationally represen- tative, and prospective longitudinal study.

Life events offer an interesting opportunity for studying the impact of environmental changes on personality because they might be associated with consistent changes in social demands and expecta- tions, as well as associated behavioral routines and environmental demands and affordances (Belsky & Rovine, 1990; Bleidorn et al., 2016). Arguments in favor of the existence of consistent expectations are that events like childbirth greatly transform the life circumstances of individuals. In this example, behavioral routines change because new parents need to adapt to the physical, psychological, and social demands associated with the birth of a child. Specifically, parents are responsible around the clock and many adjust their work and social lives to care for their newborns. Additionally, parents must negotiate societal expectations for being a good parent, such as providing sound advice to children and modeling how to lead responsible lives. Ar- guments against the existence of consistent expectations is that the impact of events might be relatively heterogeneous. For example, parents execute this role in markedly different ways, and these dif- ferences are predicted by personality (Prinzie, Stams, Deković, Reijntjes, & Belsky, 2009). Also, it is possible that parenthood is not only associated with positive expectations, but also with new chal- lenges, the presence of which has been associated with decreases in socially desirable traits like self-esteem, particularly in new mothers (Bleidorn et al., 2016; van Scheppingen, Denissen, Chung, Tambs, & Bleidorn, 2017). These various possibilities beg the important ques- tion of whether or not shifts in social expectations and associated behaviors change personality. Various theoretical perspectives differ in their answer to this question, and we turn to them below.

Theoretical Relevance of Studying Transactions Between Personality and Experiencing Life Events

The study of the nature and shape of transactions between personality and life events has important theoretical implications. First, contextual theories emphasize the importance of life events as key triggers of personality change in adulthood. A prominent example of a contextual theory is Social Investment Theory (SIT; Roberts, Wood, & Lodi-Smith, 2005). SIT predicts that personality development in young adulthood can be explained by behavioral and psychological investments in new social roles that require mature functioning, such as the role of romantic partner, working professional, or parent. According to SIT, “experiences in these social roles, as well as expectations for role-appropriate behavior, contribute to changes in personality traits” (Lodi-Smith & Roberts, 2007, p. 81). For example, being successful in the work domain usually requires people to be on time, to dress and behave profes- sionally, and to invest time and energy in delivering goods or services, and so forth—all of these behaviors can be described as conscientious. Moreover, the theory assumes that taking on social roles can result in psychological investments, for example when individuals psychologically commit to a role (i.e., make it more central to their identity). This, in turn, can result in corresponding personality change. According to SIT, the combination of behav- ioral and psychological investments will, over time, lead to lasting changes in corresponding trait domains. It is important to note that

contextual theories do not ignore the relevance of genetic influ- ences. However, experiencing a life event can exert an environ- mentally mediated yet small influence on personality change, even after taking genetic factors into account (Kandler, Bleidorn, Ri- emann, Angleitner, & Spinath, 2012).

In a review of SIT, Lodi-Smith and Roberts (2007) conducted a meta-analysis of cross-sectional studies on the association between life events and Big Five personality traits in the domains of work, love, religion, and volunteering. To evaluate key premises of SIT, the authors distinguished between studies that had operationalized the experience of a life event as a mere change in role status (e.g., not married ¡ married) and those that had examined the psycho- logical investment into the new social role (e.g., commitment to spouse, marital satisfaction). The results of this meta-analysis indicated that personality maturity was largely associated with psychological investment in social roles, whereas only few asso- ciations with life event status were found. With regard to the latter, being married was associated with high levels of conscientiousness and emotional stability, and volunteering was associated with higher levels of conscientiousness. However, this study found no associations between personality maturation and work-related variables. Overall, the meta-analytic results provided initial sup- port for SIT, although Lodi-Smith and Roberts (2007) acknowl- edged that longitudinal evidence would be necessary to draw firm conclusions regarding the developmental links between life events and personality trait change (see below for a brief review).

Second, proponents of endogenous personality theories, like the Five Factor Theory (FFT; McCrae & Costa, 1999), stress the relevance of genetically determined biological influences on per- sonality trait change. Specifically, FFT assumes that biological factors exert a proximal influence on Big Five personality traits while environmental influences play no or only a negligible role. Proponents of this theory have explicitly argued against the influ- ence of nonbiological environmental factors on personality traits (e.g., Terracciano, 2014). According to FFT, personality traits might instead predispose individuals to experience certain life events or select themselves into particular environments.

The premise that personality affects the likelihood of experienc- ing certain life events has received general support. For example, there is evidence that suggests that individuals high on openness to experience are more likely to attend college than those low on openness to experience (Lüdtke, Roberts, Trautwein, & Nagy, 2011). These findings highlight the importance of taking into account selection effects when studying the associations between life events and personality traits. Conversely, however, relatively few studies have investigated the premise of FFT that life events have only a negligible effect on personality change. However, in one study, Costa, Herbst, McCrae, and Siegler (2000) concluded that life events had few effects on personality traits at midlife (measured at two time points). Specifically, the results indicated an effect of being fired from one’s job, which was associated with increased negative affect and decreased facets of conscientious- ness. The authors of this finding called for more intensive longi- tudinal designs to clarify “ambiguities regarding the time course of personality changes” (p. 377).

Third, research on the transactions between life events and personality is also relevant for more specific theories of life span development that highlight the influence of negative life events. For example, theories of adolescent personality development have

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

2 DENISSEN, LUHMANN, CHUNG, AND BLEIDORN

highlighted the possibility of a “disruption” of maturation follow- ing the experience of new psychological, biological, or social demands (Denissen, van Aken, Penke, & Wood, 2013; Soto & Tackett, 2015). Specifically, these theories state that the stress and changes associated with life transitions might lead to initial de- clines in adaptive traits such conscientiousness, agreeableness, or emotional stability. Other theories have stated that, after initial struggles and potential decreases in adaptive traits, individuals might experience growth in response to life events. For example, recent perspectives have emphasized that adversity may lead to character growth (Jayawickreme & Blackie, 2014). Similarly, the deficits-breeds growth perspective (Baltes, Staudinger, & Linden- berger, 1999) has argued that loss-based events, such as unem- ployment, might stimulate personal growth, and potentially induce change in broad personality traits. Because these theories have been somewhat less prominent in research on personality devel- opment (Specht et al., 2014), however, the present study focuses mainly on SIT and FFT when interpreting implications of trans- actions between life events and personality traits.

In sum, the study of the transaction between life events and personality traits has clear relevance for major theories of person- ality development, like SIT and FFT. Proponents of both theoret- ical traditions have called for more sophisticated longitudinal studies of personality change and life events. Longitudinal evi- dence for the impact of various life events on personality trait change, both before and after the event, would be consistent with strong versions of contextual theories (i.e., as opposed to weak versions, which emphasize that psychological commitment is a necessary ingredient). For example, it could be concluded that enough individuals commit to mature roles to set into motion personality maturation that would be detectible as mean-level changes, or that experiencing negative life events is relatively consistently associated with either disruption or (eventual) boost- ing of maturation. On the contrary, if there are no consistent effects of life events on personality change but mainly selection effects, this would be consistent with important predictions of endogenous accounts. It is important to note that the theories discussed above make predictions that are broader than just pertaining to life events, so the findings presented in this paper cannot be treated as evidence that can confirm or falsify entire theoretical edifices. We now review some of the longitudinal evidence regarding specific life events associated with personality change.

Specific Life Events as Longitudinal Predictors of Personality Development

As described above, theories like SIT suggest that different life events may have distinct effects on personality change. One im- portant distinction has been made between love- and work-related life events (Hogan & Roberts, 2004). Bleidorn et al. (2018) re- viewed prospective research on personality trait change in re- sponse to nine major life events that fell in either of these two broad domains. Their review suggested that the current state of research only allows for tentative conclusions concerning the impact of different life events on subsequent personality trait change. However, the most cohesive findings emerged for two life events that typically occur in young adulthood. Specifically, the transitions to the first romantic relationship has been found to increase emotional stability (Lehnart, Neyer, & Eccles, 2010;

Neyer & Asendorpf, 2001), but no effects have been found for the transition to marriage (Neyer & Asendorpf, 2001). Also, the tran- sition from high school to vocational education or work has been found to be associated with increased conscientiousness (Bleidorn, 2012; Lüdtke et al., 2011). Overall, preliminary evidence indicates that events related to both love and work are associated with changes in personality traits, but more longitudinal research is needed to identify which life events matter most and which traits are most strongly associated with particular life events (Bleidorn et al., 2018).

Another distinction that can be made is between gain-versus loss-based events (this distinction is similar to the one between positive and negative life events, Specht, Egloff, & Schmukle, 2011). Gain-based events, such as marriage, parenthood, or enter- ing the workforce, are typically associated with new social roles and relationships. In contrast, loss-based events, such as widow- hood or divorce, involve the loss of a previously held social role or relationship. The latter events are sometimes categorized as “stressful life events,” which have been related to increases in depression (Kessler, 1997) and decreases in life satisfaction (Lu- cas, 2007). Implications for personality trait development are more difficult to derive and previous research has yielded mixed results concerning the effects of loss-based events on personality traits (see also Bleidorn et al., 2016). However, evidence from the research of Lüdtke et al. (2011) indicates that low levels of emotional stability and extraversion predict the frequency of ex- perienced negative life events in young adults. In turn, experienc- ing these life events further decreased emotional stability and extraversion. Therefore, it is possible that stressful life events correlate with temporary disruptions of personality maturation (Denissen et al., 2013; Hutteman et al., 2014). However, it is also possible that loss-based events are, over time, associated with growth in socially desirable personality traits (Baltes et al., 1999; Jayawickreme & Blackie, 2014).

In sum, a growing body of longitudinal research has found some evidence for the impact of life events on personality trait change. In particular, the most robust findings concern increases in con- scientiousness following work-related life events, as well as in- creases in emotional stability following transitions into romantic partnerships. For childbirth, a mix of positive and negative effects are found (for a recent demonstration, see van Scheppingen et al., 2016), and sometimes even negative changes are reported, such as decreases in self-esteem (Bleidorn et al., 2016; van Scheppingen et al., 2016) and increases in negative emotionality (Jokela, Kivi- mäki, Elovainio, & Keltikangas-Järvinen, 2009). However, it may be premature to draw firm conclusions regarding the impact of life events on personality development without additional longitudinal evidence that takes into account the complexities of the relevant developmental processes (Bleidorn et al., 2018; Luhmann, Orth, Specht, Kandler, & Lucas, 2014). We describe these complexities below.

Modeling the Shape of Event-Related Personality Change

Until recently, most longitudinal studies addressing the interplay between personality and life events have included two or three personality assessments over relatively long time intervals of 4 or more years (e.g., Specht et al., 2011). Although these studies can produce valuable information about the association between life

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

3LIFE EVENTS AND PERSONALITY TRAITS

events and personality traits over longer time intervals, they are not ideally suited to provide insight into how personality trait changes unfold before, during, and after the experience of a life event. More nuanced longitudinal designs with frequent assessments of life events and personality traits are necessary to examine the actual shape of personality development. In particular, the use of prospective designs with multiple assessment waves allows re- searchers to extend traditional two- or three-wave studies and address three important questions regarding mean-level change associated with life events.

Mean-Level Change

Life events might impact mean-level change, defined in this study as average changes over time in people who have experi- enced a certain event. The first important question that can only be tested with such prospective multiwave longitudinal designs is what the exact shape of event-related personality change is. If more measurement points are available, it becomes easier to dis- tinguish between different patterns, such as abrupt short-term or long-term shifts, versus more gradual linear changes (Luhmann et al., 2014). This differentiation is important for testing whether personality trait changes following a life event are long-lasting, or whether they rebound to preexisting levels after some time. In the subjective well-being literature, set-point theory (Lykken, & Tel- legen, 1996) predicts that well-being might decrease or increase following an event, but bounces back to its set-point after some time. Levels of personality traits that indicate maturity (e.g., con- scientiousness, emotional stability) might similarly decrease or increase abruptly following some event, but eventually change back to baseline levels. Such nonlinear or discontinuous trajecto- ries cannot be adequately described if only two measurements are available (although it is possible to use time following the event as a moderator of the pre–post event difference, which has been done by, e.g., Boyce, Wood, Daly, & Sedikides, 2015; van Scheppingen et al., 2016).

Anticipatory Change

A second important question is whether personality traits change in response to a life event, or whether mean-level changes already begin during the period before the event occurs. Theoret- ical perspectives acknowledge the possibility of anticipatory per- sonality changes before experiencing a life event. For example, SIT (Roberts et al., 2005) emphasizes psychological investments as the proximal mechanism of personality change. In the case of normative events that can be planned in advance (e.g., marriage, childbirth, work), such psychological investments and potentially related personality trait changes may already occur before the actual event (Roberts, O’Donnell, & Robins, 2004). Ignoring po- tential pre-event changes in traits might result in biased conclu- sions. For example, if anticipatory maturational increases are fol- lowed by decreases in maturity following the event, one might draw a false conclusion that the event is associated with a decrease in personality maturation. In reality, however, the event would have spurred a real but temporary boost in personality maturity. Having multiple pre- and post-event measures in a prospective design allows for an assessment of such anticipatory effects.

Individual Differences in Change

A third important question is whether life events are associated with individual differences in personality trait change. Some peo- ple may increase whereas others decrease in their personality traits in response to certain life events. These differences might cancel each other out, so that even in the absence of mean-level change, life events would still be associated with substantial personality change within individuals. Differences in event-related personality trait change can be statistically modeled as variance in the slope that is associated with experiencing a certain life event. Further- more, more sudden shifts in rank-order trait differences can be modeled by comparing test–retest correlation coefficients of indi- viduals who have experienced an event with those who have not. To enrich findings on event-related mean-level change with infor- mation about individual differences in change, the current study focuses on both analytic strategies to investigate whether there are individual differences in people’s personality trait change in re- sponse to life events.

The Current Study

The current study offers a unique glimpse into the associations between life events and personality development both in terms of the number of personality assessments (up to six) and the time resolution of these assessments (every 1 to 2 years). Moreover, the current study tracks the occurrence of life events using monthly assessments, which permits estimation of personality trait changes in the context of major gain- and loss-based life events in the domains of love and work with high temporal resolution. Multiple assessments are necessary for modeling the precise trajectory of change (which can be nonlinear) before, during, and after a life event (see also Bleidorn et al., 2016; van Scheppingen et al., 2016). Therefore, in the current study, we model separate slopes for anticipatory changes before experiencing the life event versus socializing influences after experiencing the event.

We focus on the “Big Five” trait domains of extraversion (E), agreeableness (A), conscientiousness (C), emotional stability (S), and openness to experience (O; John & Srivastava, 1999). Al- though there are many potential broader and narrower personality traits to consider, the Big Five traits represent a viable balance between conceptual breadth, descriptive fidelity, and generaliz- ability across samples and measures. In addition, we examined the impact of life events on life satisfaction (LS), defined as “a global assessment of a person’s quality of life according to his own chosen criteria” (Shin & Johnson, 1978, p. 478). There is a disagreement in the literature whether life satisfaction is a trait because it has some trait-like properties but is less stable than personality traits such as the Big Five (Anusic & Schimmack, 2016). Rather than taking a clear position in this debate, we also analyzed transactions between life events and life satisfaction to provide a benchmark perspective that helps gauging the results for the Big Five. Furthermore, our analytic models were novel because previous studies were based on a different data structure (i.e., life satisfaction measures with a yearly measurement interval). If our novel statistical framework replicates meta-analytic findings on the short- and long-term effects of various life events (Luhmann et al., 2012), it would boost confidence that it is capable of detecting changes in personality traits as well. Furthermore, detection of possible differences between the Big Five and life satisfaction

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

4 DENISSEN, LUHMANN, CHUNG, AND BLEIDORN

could inform perspectives that regard the former as less malleable and open to outside influences, when compared with the latter. For example, a review by Kandler, Zimmermann, and McAdams (2014) indicated that the Big Five traits are more stable and more influenced by genetic factors, when compared to self-related sche- mata, such as life satisfaction.

In the present study, we sought to examine a broad range of life events that reflect previous theory and research by crossing the work versus love distinction with the gain versus loss distinction. Specifically, we sampled two events from each of the following categories: work-gain (employment, volunteering), work-loss (un- employment, disability1), love-gain (marriage, childbirth), and love-loss (divorce, widowhood) and generated predictions regard- ing the ways in which personality traits interact with different life events. Based on recent reviews of longitudinal studies on life events and personality trait change and on life events and change in life satisfaction, we formulated the following hypotheses.

Hypothesis 1: Obtaining paid employment is associated with increases in conscientiousness (Bleidorn et al., 2018; Lüdtke et al., 2011).

Hypothesis 2: Experiencing unemployment, disability, di- vorce, and widowhood are associated with decreases in life satisfaction before and right after the events, and with in- creases thereafter (Lucas, 2007; Luhmann et al., 2012).

In addition, we derived hypotheses based on the ideas of leading theoretical accounts of personality development discussed above. Specifically, a narrow version of SIT would predict that all gain- based love and work events should be associated with increases in maturity-related personality traits.

SIT hypothesis: Entering work, marriage, childbirth, and vol- unteerism is associated with increases in agreeableness, con- scientiousness, and emotional stability.

By comparison, FFT would predict only selection effects of personality on life events, regardless of the nature of the events. This gives rise to the following hypothesis:

FFT hypothesis: Life events do not impact personality devel- opment (with the exception of life satisfaction) but personality traits predict the occurrence of life events.

In addition to these hypotheses, we explored the possibility of other transactions between life events and personality traits. For example, we systematically included random effects of time- varying predictors to model individual differences in personality change. Finally, we investigated whether rank-order stability is attenuated after an event transition, which would be expected if personality is indeed reorganized following an event transition, or whether existing personality differences are rather solidified (for a discussion of both possibilities, see Caspi & Moffitt, 1993).

Method

Participants

Data for this study came from the ongoing Longitudinal Internet Studies for the Social Sciences (LISS) panel, which follows a

representative sample of the Dutch population since 2008 with the most recent data collection completed in 2017 (Scherpenzeel, 2011; Van der Laan, 2009). The panel is based on a random sample of households drawn from the population register (for details, see Scherpenzeel, Das, Ester, & Kaczmirek, 2010). Of people with a usable address, 48% ended up as panel members. When recruiting refreshment samples, information about the asso- ciation between demographic variables (household type, age, eth- nicity) and response rates was used to oversample previously underrepresented groups (see https://www.lissdata.nl/lissdata/ about-panel/sample-and-recruitment). Participants who did not have a computer or Internet connection were provided with one so that they could complete the surveys. Because no new data were collected in the context of the present study and all the data are in the public domain, it was not necessary to obtain ethical approval from an institutional IRB. Note that Schwaba, Bleidorn, and col- leagues have used this dataset to study transactions between open- ness and cultural activities (Schwaba, Luhmann, Denissen, Chung, & Bleidorn, 2017), quantify individual differences in personality trait change (Schwaba & Bleidorn, 2017a), and investigate the effect of retirement on personality trait levels (Schwaba & Blei- dorn, 2017b). None of these previous publications overlap with the current study.

From a total pool of 21,377 participants, we included data from 13,040 participants with at least one personality assessment in the present study.2 The average age of participants was 44.5 years (SD � 17.7), and 46% of the participants were men. The first cohort started in 2008 and consisted of 6,769 participants. The panel is replenished regularly by adding new participants to the sample to counteract attrition (Lugtig, 2014). Table 1 specifies the year when such replen- ishment cohorts entered the sample (typically in even years), and also the years in which they were assessed again. The consequence for the statistical analysis is that only the starting 2008 cohort (the largest by far) could participate in the study for the maximum number of 9 years. For this cohort, the time resolution and statistical power was thus the largest.

Procedure

Panel members completed monthly surveys about their demo- graphic status, which we used to obtain information regarding the experience of the life events listed above. In addition to the monthly surveys, participants provided reports about their person- ality traits and life satisfaction every 1–2 years. From 2008 until 2013 and again in 2017, personality was assessed in May, but from 2014 onward, the timing of the assessments changed to November. Because we modeled time on a continuous scale, our statistical procedure was able to take this shift into account. If the survey was not completed in the focal months, a second opportunity was provided in the month thereafter (i.e., June and December; in 2008,

1 Disability certainly has implications on more life domains than only work (e.g., it might be related to difficulties in participating in public activities, stigmatization, etc.). The reason why we classified it as a work-related effect is that the disability status pertained to work, as one of the options to respond to the question about participants’ “primary occu- pation.”

2 Individuals with only one assessment were included because they contribute to the estimation of the intercept.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

5LIFE EVENTS AND PERSONALITY TRAITS

the second opportunity was offered in August). Second opportu- nity assessments constituted 8.2% of all measurements.

In addition to attrition, planned missingness was employed in the study design because the personality questionnaire did not always include Big Five measurements. The exact pattern of missingness depended on the year of sample entry (which defined a cohort; see Table 1 for an overview). For example, individuals from the 2008 cohort did not fill out the Big Five questionnaire in 2010 and 2016, and only few of them contributed data in 2012 and 2015.

Measures

Personality. The Big Five personality traits were assessed with the 50-item version of the IPIP Big-Five inventory (Goldberg, 1992). Example items included “Talk to a lot of different people at parties” (extraversion), “Sympathize with others’ feelings” (agree- ableness), “Pay attention to details” (conscientiousness), “Am re- laxed most of the time” (emotional stability), and “Have a vivid imagination” (openness to experience). Average internal consis- tencies (Cronbach’s alpha) across the 8 waves were satisfactory, ranging between .76 (openness) and .88 (emotional stability). None of the wave-specific reliabilities were lower than .74.

Life satisfaction. Life satisfaction was assessed with the five items of the Diener Satisfaction with Life Scale (Diener, Emmons, Larsen, & Griffin, 1985). An example item is “In most ways my life is close to my ideal.” Reliability across waves was very high, with an average alpha of .88.

Life events. Life events were derived from the monthly sur- veys. The variables that were used from this survey included questions about the participants’ primary occupation (14 options, including “paid employment,” “job seeker following job loss,” “job disability,” and “voluntary work”), marital status (5 options, including “married,” “divorced,” and “widowed”), and number of living-at-home children in the household (9 options). To check whether a child in the household was a newborn and not an older child that had moved back in, we checked the birth year of all household members and only coded a birth if the youngest member of the household was born in the year of the survey.

Status changes indicating life events were tracked monthly. Specifically, a dummy code was created to mark the occurrence of

a status change. For participants who experienced an event mul- tiple times, we only used the first occurrence to ensure the inter- pretability of our time indices (i.e., there needed to be a clear pre- and post-event distinction, which is muddled in case of repeat occurrences). Another reason to exclude repeat events is that these were rather rare for events such as widowhood and disability. To ensure that this decision did not bias the results, we repeated our analyses without the data of repeat event participants. This did not change any effect sizes.

Table 2 shows the sample sizes per personality assessment for each of the eight life events. Because only individuals with at least one personality assessment were included in the study, the first column equals the overall prevalence of the different life events in our sample. As can be seen in Table 2, the prevalence of the events varied substantially, with prevalence numbers ranging between 152 (widowhood) and 1,467 (transition to paid employment) par- ticipants who had experienced the events throughout the study period. Note that the sample size per subsequent assessment was smaller due to (planned and nonplanned) attrition. Specifically, between 61 (widowhood) and 415 (transition to paid employment) individuals completed all sixth personality assessments. The table also shows the age and gender of participants who experienced a life event change in our sample.

Statistical Analysis

The data had a hierarchical structure with measurement occa- sions (Level 1) nested within persons (Level 2) and households (Level 3). These data were analyzed using multilevel models for longitudinal data (e.g., Singer & Willett, 2003; Luhmann & Eid, 2012).

Time-invariant effects indicating group differences. We in- cluded gender (centered) as a time-invariant covariate. In addition, people who experience a life event might have certain character- istics before, during, and after the status transition. Time-invariant differences that predict who will experience an event (e.g., get married) represent selection effects that need to be accounted for in order to estimate potential socialization effects. To model these preexisting differences, we created a dummy variable called Event Selection, with 1 for participants who had ever experienced the transition, and with 0 for those who did not (i.e., participants who

Table 1 Sample Sizes for Different Cohorts, Per Assessment Year

Assessment year

Cohort 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Sum

2008 6769 5182 — 3893 234 3139 3135 169 — 2585 25106 2009 — 426 — 231 30 180 179 16 — 128 1190 2010 — — 1376 973 55 714 641 34 — 510 4303 2011 — — — 187 2 90 94 11 — 60 444 2012 — — — — 1153 836 781 39 — 524 3333 2013 — — — — — 172 126 9 — 75 382 2014 — — — — — — 1563 16 — 879 2457 2015 — — — — — — — 211 — 110 321 2016 — — — — — — — — — — 0 2017 — — — — — — — — — 1184 1184 Sum 6769 5608 1376 5284 1474 5131 6518 505 0 6055 38720

Note. The sums of each column refer to the number of participants that contributed personality data during each assessment year. The sums of each row refer to the number of personality assessments that have been contributed by each cohort by 2017.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

6 DENISSEN, LUHMANN, CHUNG, AND BLEIDORN

had never experienced the transition, neither before nor during the study).

Because of the relatively infrequent occurrence of major life events (many people experience them only once in their lives) and the fact that the average age of entry in the sample was around midlife (i.e., after the normative age for many of the covered events), many participants experienced the event transition already before their entry into the study. Our focus was on investigating event-related changes in participants who experienced the transi- tion while being a LISS panel member. To control for preexisting differences other than those captured by the above-described se- lection effects, an additional dummy variable was added. This dummy variable, called Transition Sample, is coded with 1 for participants who experienced the transition during the study, and with 0 for all other participants. Because of the way all our variables were coded, this transition variable reflects the contrast between those who experienced the transition during the study period versus those who experienced the transition prior to the study period.3

Time-variant effects indicating mean-level change. Linear age and quadratic age were included as a time-variant (Level 1) covariate. The linear age variable (grand-mean centered at age 48.4) was measured along a 5-year (60-month) metric to align with the coding of the linear event-related coefficients (see below). To ensure that the quadratic term did not deviate too much in range from the linear term, quadratic terms were divided by 10.4 Fur- thermore, we included a variable to indicate testing effects: Changes in personality scores that occur because of repeatedly asking the same questions in a longitudinal design. Testing effects are not typically modeled in personality development studies (see Baird, Lucas, & Donnellan, 2010, for an exception), but their inclusion is more common in studies on life span changes in intellectual abilities. In our study, we modeled such effects as linear changes as a function of the number of previous personality assessments (0 for the first assessment, 1 for the second assess- ment, etc.; see McArdle, Ferrer-Caja, Hamagami, & Woodcock, 2002, who used a similar approach).5

The coding of the time-variant event parameters was based on the recommendations by Luhmann and Eid (2012). Within-person changes in personality were modeled with four time variables (see Figure 1 for an illustration).

1. The Linear Anticipation variable (preLin) is a linear variable indicating the time leading up to the event. The metric of this variable was measured in 5-year (60 months) increments (the figures also used �5 years and �5 years as limits of their respective x axis).6 This variable is coded with negative values on all occasions preceding the event (e.g., �1 if a personality assessment was taken five years before the event) and with 0 on all occasions after the event. This variable indicates the rate of linear change in the outcome leading up to life event, over and above the effects of age and testing.

2. The Post-Event Year variable (firstYear) is a dummy variable with a value of 1 for any personality assessment that took place within one year after the event. This variable indicates a sudden but short-term change in the outcome, over and above the effects of age and testing (socialization effect).

3. The Post-Event Baseline Change variable (postD) is a dummy variable with a value of 1 for any personality

3 Note that these dummy variables are not perfectly orthogonal but allow us to test the specific contrasts we were interested in this study. The combination of the two dummy variables produced three groups (because by definition there were no participants who were part of the transition sample but did not experience the event).

4 The division by 10 ensured that coefficients would not become too small to display when using rounding at two decimals (see also footnote 5).

5 We are confident that our testing effect variable indeed constitutes an artefact, and not an interpretable developmental effect (i.e., related to event, age, or period). Different from samples with yearly measurements, the testing effect variable was almost independent from the preLin and postLin variables. Furthermore, we controlled for age in all analyses, ensuring the effect of assessment number did not simply reflect participants getting older during the survey. There also was no constant spacing between the assessments (because of shifting assessment months, and the inclusion of years in which no personality information was collected), and participants had their first, second etc. personality assessments in different years (see Table 1).

6 An important reason to choose five-year intervals was that coefficients would have become too small when using a monthly metric. Because converting months to five-year intervals only represents a linear transfor- mation, this choice did not have any consequences for the pattern of results.

Table 2 Sample Sizes for Each Event Across Personality Assessments, and Age and Gender at the First Assessment

Personality assessment Age/gender at the first assessment

Event 1 2 3 4 5 6 M age SD age % female

Mature adult role transitions Paid employment 1467 1353 1072 896 679 415 35.51 13.14 61 Childbirth 566 513 403 312 218 128 33.43 5.88 56 Marriage 557 517 420 360 267 171 37.65 12.53 54 Volunteering 253 241 206 181 145 81 56.54 17.55 70

Loss events Unemployment 787 746 626 550 431 285 43.51 12.34 55 Disability 260 239 199 180 144 105 48.74 12.63 58 Divorce 190 179 161 147 126 91 47.67 12.85 55 Widowhood 152 148 137 125 98 61 68.76 10.56 64

Note. The decline in numbers across assessments does not only reflect attrition but also the fact that only the 2008 cohort could contribute 5 data points. Note also that either 1 or 2 years could be between different assessments, depending on the cohort and assessment year (see Table 1).

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

7LIFE EVENTS AND PERSONALITY TRAITS

measurement that took place after the event. This vari- able indicates a sudden baseline change after the event. The baseline change can be interpreted as a permanent (for the duration of the study) shift in the outcome vari- able, over and above age effects, testing effects, and short-term changes.

4. The Linear Socialization variable (postLin) is a linear variable indicating the time since the event in 5-year (60 months) increments, the same metric used for the other timing variables. Specifically, this variable is coded with 0 on all pre-event occasions and with positive values on all occasions after the event (e.g., 1 if a personality assessment was collected five years after the event). This variable reflects linear change in the outcome after the event, over and above age effects, testing effects, short- term changes, and baseline shifts.

All timing variables were coded with zero at all occasions in the two comparison groups. These variables therefore indicated changes in the transition sample over and above those experienced by the comparison groups. The overall intercept reflects the pre- dicted level on the outcome variable for a person of average age and gender who has never experienced the transition, at the first measurement occasion. Because our models included a relatively

large number of (correlated) predictors, we used a formula pro- vided by Lefcheck (2012) to compute variance inflation factors (VIFs) for multilevel regression models. Our results (see Table S9 of the supplementary materials) indicate that only the VIF for the two unemployment prevalence effects were higher than 4. How- ever, even these values were lower than the cut-off value of 10 that is often used. Still, for unemployment we also ran a model that only included an event selection effect, and briefly report on the results.

Analyses were conducted separately for each of the eight life events and for each outcome variable, using the lme4 package in R (Bates, Mächler, Bolker, & Walker, 2014). Because of the large number of statistical tests involved, we wanted to decrease the likelihood of chance effects. Our statistical tests were dependent because the same outcome variable and events are used in multiple comparisons. We therefore used the “BY” method published by Benjamini and Yekutieli (2001), which accounts for dependency. Specifically, we took the p values related to the 8 events � 6 outcome variables � 5 coefficients that are relevant for our key results and adjusted these values by applying the R command “p.adjust” on the vector of uncorrected p values. For the p values belonging to the five control variables (gender, linear and qua- dratic age, testing effects, and sample), we conducted the same number of tests, so we applied the same correction. Because the

Figure 1. Example of the coding of timing coefficients based on monthly survey data for one participant. On the left side of the figure, the preliminary data structure is depicted, including the four monthly timing variables (preLin, firstYear, postD, and postLin), the personality assessment number (ass. #), and the score on a Big Five personality variable (B5). On the vertical time axis on the left (with months as the unit of time, running from down to up), the gray dots represent personality assessments, and the black bar represents the first month after the event. As can be seen, the values underlined in grey of the preliminary data matrix were taken to construct the final data structure.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

8 DENISSEN, LUHMANN, CHUNG, AND BLEIDORN

resulting p value cut-off was close to p � .001, we report 99.9% confidence intervals (z � 3.09) for all fixed effects. We also adjusted the p values of the random effects using the “BY” method.

To visualize the different trajectories and to probe how well the estimated models fit the (aggregated, see below) raw data, we created a graph from raw data for each event-trait combination that included at least one significant time-variant event parameter. We intentionally opted to use raw data to demonstrate the robustness and visibility (or lack thereof) of our effects. As a first step toward creating this graph, we ran regressions with the outcome variables as outcome and only the covariates (gender, age, and assessment number) as predictors, and then saved the residuals. To create a benchmark of effect size, we then divided scores by the standard deviation of the corresponding variable, thus creating standardized scores. In each graph, we depicted the personality trait or life satisfaction score averages for each month in relation to the event occurrence (to illustrate between-person variability, we also cre- ated additional spaghetti plots, which we report in Figures S9 –S16 of the supplementary materials). We also added two linear regres- sion lines to the plots: One for the data points before the event, and one for the data points after the event. This approach can visually highlight any of the coefficients that were uncovered in our more complex multilevel regressions. For example, in case we found a robust linear anticipation effect, our visualization approach would produce a tilted pre-event line running through the pre-event data points. Conversely, a robust linear socialization effect would pro- duce a titled post-event line. A robust post-event baseline change effect would be visible if the post-event line would be higher than the pre-event line. In cases when the post-event year effect was significant, a third linear regression line was added to the plot, based on the data points corresponding to the first 12 months after the event. Note that this represents a purely data-driven approach, which might provide a more faithful representation of the actual data than using predicted scores only.

Power to detect time-variant effects. Because power analy- sis using multilevel data is relatively complex (e.g., it depends on the number of Level 1 and Level 2 data points, the size of the ICC, the covariance between random effects, etc.), we decided to focus on the postYear effect on personality trait or life satisfaction changes following widowhood as the “weakest link” (in terms of the number of data points) of our ability to detect changes asso- ciated with life events. This decision was informed by the fact that widowhood was the least often experienced of all life events (see Table 2) and that the postYear effect had the least number of data points per individual (i.e., a maximum of one data point), as opposed to the other effects. For the power analyses, we used the SIMR package (Green & MacLeod, 2016), specifying an alpha level of .001. Note that although we focused on the postYear effect, in the power analysis this effect was estimated together with the other time-varying coefficients, to closely match our actual analytic approach (see OSF materials for syntax). The resulting power estimates were as follows: d � .10: 4.1%, d � .15: 20.9%, d � .20: 49.7%, d � .25: 78%, d � .30: 94.1%. This suggests that that our design was adequately powered to detect even small (d � .25) differences in the rarest event that we studied.

Individual differences in change. In the final models (cf. Tables 4 –11), the effects of the Level 1 time variables were modeled as fixed, meaning that no individual differences in the

individual trajectories were modeled. Of course, this assumption might not be realistic, because individuals differ in their reactions to life events. To test whether reactions differed between people who had experienced an event, we tested the significance of random effects indicating the extent of individual differences in reactions to life events. It was computationally impossible to model random effects for all four time-invariant predictors simul- taneously. Therefore, we inserted one random predictor at a time, and tested whether this resulted in a significant model fit improve- ment, using the “anova” command in R. To further illustrate individual differences in change, we have plotted distributions of the random effects for all relevant parameters, following an ap- proach by Doré and Bolger (2017). We report these distributions in the supplementary materials (Figures S1–S8).

Second, we checked whether the test–retest stability between two personality trait measurements differed between people who had experienced an event compared to participants who had not experienced the event. In these analyses, we predicted the outcome variables by their lagged precursors. We did so only for cases in which the time lag between both assessments was not longer than 24 months. In addition to adding the lagged score, we also in- cluded a dummy indicator that specified whether a life event had occurred between the two assessments. As a test of differential stability, we included an interaction between the status change dummy and the lagged personality trait or life satisfaction score. This interaction tests whether the test-retest correlation is different for individuals who have experienced an event-status change (see Mccrae, 1993, for a comparable approach).

Results

Descriptive Statistics

Table 3 presents a correlation matrix with the minimum (below the diagonal) and maximum (above the diagonal) of correlations between the dependent variables, across the separate waves. Fur- thermore, Table 3 presents the means and standard deviations for each separate wave.

Control Variables

As described above, we controlled for demographic variables as well as testing effects. Demographic effects were in line with previous research, with women reporting higher agreeableness and lower emotional stability, and personality trait levels increasing in terms of agreeableness, conscientiousness, and emotional stability. As can be seen in Tables 4 –11, the testing effect was statistically significant in all cases, with participants decreasing on all traits with every additional assessment, except for emotional stability, which increased over time. This illustrates that is important to control for testing effects when assessing the effect of life events on personality development. An overview of the effects for each of these control variables can be found in Tables 4 –11. Because the paper was not focused on these effects, we do not discuss them further.

Transactions Between Personality and Life Events

Paid employment. As can be seen in Table 4, regarding paid employment we found three selection effects: People who were

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

9LIFE EVENTS AND PERSONALITY TRAITS

employed scored higher on conscientiousness, emotional stability, and life satisfaction. The two latter effects were offset by coun- tervailing transition sample effects. Specifically, individuals who transitioned into paid employment during the study period were characterized by a lower emotional stability and life satisfaction compared with those who had experienced the transition prior to the study period. In anticipation of paid employment, we found that people increased in conscientiousness and openness to expe- rience. There were no socializing effects of the transition into paid employment on personality development that were limited to the first year after the transition. However, there was one sudden post-event baseline change effect: People who transitioned into work increased in emotional stability. This change was seemingly permanent, as there were no linear post-event changes. Effects for conscientiousness, emotional stability and openness are depicted in Figure 2. As can be seen, the anticipatory increases in conscien- tiousness and openness to experience, as well as the positive baseline change effect on emotional stability were visible in the raw data, even though they were not large in size (i.e., changes between | d | � .12 across the plotted 5 years). Together, these results indicated a number of selection effects (e.g., of high con- scientiousness, emotional stability and life satisfaction levels).

Regarding personality development, a number of rather subtle increases in socially desirable traits occurred both before people got their first paid job (linear increases in conscientiousness and openness) as well as thereafter (sudden increase in emotional stability).

Childbirth. As can be seen in Table 5, regarding childbirth there were selection effects, although they were modest in size. Specifically, individuals with children reported somewhat lower conscientiousness and openness levels and somewhat higher life satisfaction levels than individuals without children. Individuals in the transition sample had higher conscientiousness, emotional sta- bility, and life satisfaction levels than those who had experienced the transition prior to the study period. In terms of anticipatory effects, participants gradually increased in emotional stability and life satisfaction as they approached the birth of their child. There were no sudden socialization effects of childbirth that lasted only one year, but there was one post-event baseline change effect: Participants showed a decrease in conscientious after childbirth. Following childbirth, new parents linearly decreased in emotional stability, offsetting the previously experienced positive anticipa- tion effect on this trait. As can be seen in Figure 3, the effects on personality change were visible, but not very large in size (i.e., | d | � .31 across the plotted 60 months). Also, it can be seen that post-event reductions in conscientiousness, emotional stability, and life satisfaction did not result in values below zero, indicating that having children was never associated with below-average levels of maturity. Taken together, a mix of positive and negative selection effects were found (people with children were lower in conscientiousness and openness but higher in life satisfac- tion). In terms of development, we found positive anticipatory increases in emotional stability and life satisfaction). After childbirth, we found a sudden decrease in conscientiousness and linear decrease in emotional stability.

Marriage. Regarding marriage, there were several selection effects (see Table 6). Participants who ever reported to be married were higher in conscientiousness, emotional stability, and life satisfaction, and somewhat lower in openness. Participants who experienced the event during the study (transition sample), scored somewhat higher in openness than participants who had experi- enced the event prior to the study. There were no effects of marriage on change in Big Five personality traits, but individuals who anticipated to get married did gradually increase in life satisfaction (see Figure 4). Taken together, we found mainly positive selection effects suggesting that people who were high in conscientiousness, emotional stability and life satisfaction (but low in openness) were more likely to enter marriage during the study in period. In contrast, there were no effects on Big Five personality development, and only one (anticipatory) effect on life satisfac- tion, suggesting that life satisfaction increased as people ap- proached marriage.

Volunteering. We found no effects for the transition to vol- unteering. That is, neither Big Five nor life satisfaction predicted the transition into a volunteering position, and entering into a volunteering role did not affect the development of these charac- teristics.

Unemployment. As can be seen in Table 8, individuals who experienced unemployment were lower in life satisfaction than indi- viduals who never experienced this event (selection effect), which corresponded to a moderate to large effect size. In addition, there were

Table 3 Descriptive Statistics for All Dependent Variables, Across Measurement Waves

Variable E A C S O LS

Min/max correlations E .34 .13 .27 .38 .24 A .29 .32 .08 .34 .15 C .07 .26 .31 .24 .22 S .19 .01 .17 .20 .45 O .32 .23 .19 .13 .12 L .17 .08 .16 .37 .02

Year Means

2008 3.30 3.90 3.72 3.41 3.51 5.12 2009 3.28 3.88 3.69 3.42 3.49 5.09 2010 3.29 3.87 3.66 3.38 3.48 4.94 2011 3.25 3.85 3.69 3.45 3.46 5.05 2012 3.31 3.87 3.67 3.40 3.49 5.08 2013 3.24 3.85 3.71 3.49 3.45 5.05 2014 3.25 3.88 3.72 3.45 3.49 4.98 2015 3.27 3.85 3.53 3.35 3.58 4.93 2017 3.24 3.88 3.74 3.46 3.51 5.02

Year SDs

2008 .63 .49 .52 .68 .50 1.05 2009 .63 .49 .53 .66 .49 1.06 2010 .63 .49 .56 .66 .49 1.12 2011 .63 .49 .53 .67 .49 1.10 2012 .66 .48 .55 .68 .50 1.10 2013 .66 .51 .53 .69 .50 1.10 2014 .66 .51 .53 .70 .50 1.14 2015 .65 .54 .58 .73 .52 1.15 2017 .67 .52 .53 .70 .51 1.11

Note. E � extraversion; A � agreeableness; C � conscientiousness; S � emotional stability; O � openness to experience; and LS � life satisfac- tion. Regarding the correlations, the minimum correlation across all waves is displayed below the diagonal, and the maximum correlation is displayed above the diagonal.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

10 DENISSEN, LUHMANN, CHUNG, AND BLEIDORN

anticipatory effects of unemployment on emotional stability and life satisfaction: Participants decreased in these variables the closer they moved toward becoming unemployed. As can be seen in Figure 5, however, these anticipatory decreases were small, with changes cor- responding to | d | � .23 across the whole period of 60 months. Figure 5 also suggests a post-event baseline change in life satisfaction, but this effect did not reach statistical significance. Taken together, un- employed people were not only unhappier to begin with, they also decreased further in life satisfaction in anticipation of the effect, in addition to becoming less emotionally stable.7

Disability. As can be seen in Table 9, there were personality differences between participants who at one point reported being disabled and those who did not (selection effects). Specifically, people who reported being disabled were lower on extraversion, conscientiousness, emotional stability, and life satisfaction (with the latter effect being very large, d � �.90). Individuals who experienced the transition to disability during the study (transition sample) were somewhat higher in agreeableness and openness compared to those who had experienced the transition prior to the study. We found only one effect with regard to change, suggesting that people decreased in life satisfaction as they approached the event, corresponding to a small-to-moderate effect. As can be seen in Figure 6, the raw data were also suggestive of a postyear dip followed by a post-event rebound effect, but the corresponding coefficients did not meet our threshold of significance. All in all, we found a range of negative selection effects (particularly in terms of lower emotional stability and life satisfaction). In terms of development, however, disability impacted people’s life satisfac- tion in anticipation of the event but did not affect any of the Big Five traits.

Divorce. As can be seen in Table 10, individuals who reported being divorced at some point in their lives scored higher in

openness, and lower in life satisfaction compared to those who never experienced a divorce (event selection). In addition, people who made the transition sample were additionally characterized by lower emotional stability and life satisfaction levels compared to those who got divorced prior the study period. Effects of divorce on personality development reached significance in only two cas- es: There was a negative anticipatory effect on life satisfaction as participants moved to becoming divorced, and a countervailing (but somewhat smaller) linear socialization effect following the event. The effect on life satisfaction was moderate to large in size: The anticipatory decrease corresponded to d � �.63 across the plotted 60 months. Together, we found some selection effects, particularly of lower life satisfaction, but no effects of the event on change in the Big Five personality traits. Rather, a strong antici- patory decrease followed by an (incomplete) rebound in life sat- isfaction was found (see Figure 7).

Widowhood. As can be seen in Table 11, individuals who were widowed reported being lower in life satisfaction than indi- viduals who were never widowed, though effect was small, d � �.28. Widowhood was not associated with Big Five person- ality trait development but was related to changes in life satisfac- tion: People decreased before they became widowed, and also reported lower life satisfaction levels in the year following the event (see Figure 8). Overall, the effects of widowhood were

7 Because unemployment was the only event for which we found slightly problematic VIF values (see above), we ran additional models for unem- ployment in which the sampling bias factor was removed. Results were virtually unchanged when compared with our original findings. The only exception was that the negative event selection of emotional stability became statistically significant because of a reduction in standard error (the effect size remained virtually the same).

Table 4 Associations Between Paid Employment and Personality Development

E A C S O LS

Coefficient b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI

Intercept �.03 [�.09, .02] .06 [.00, .11] .02 [�.03, .08] �.15 [�.21, �.10] .07 [.02, .13] �.14 [�.19, �.09] Testing �.01� [�.02, �.00] �.03� [�.03, �.02] �.02� [�.03, �.01] .02� [.01, .03] �.01� [�.02, �.00] �.02� [�.03, �.01] Female gender .01 [�.04, .07] .65� [.60, .70] .19� [.14, .24] �.33� [�.38, �.28] �.17� [�.22, �.12] .07� [.02, .11] Age �.02� [�.03, �.01] .03� [.02, .03] .07� [.06, .07] .04� [.03, .05] �.04� [�.05, �.03] .02� [.01, .03] Age2 .03� [.01, .05] �.02� [�.04, �.00] �.08� [�.10, �.06] .01 [�.01, .03] �.04� [�.06, �.02] .03� [.01, .05] Transition sample �.05 [�.15, .05] .03 [�.07, .13] �.06 [�.16, .05] �.18� [�.29, �.08] .05 [�.06, .15] �.25� [�.35, �.15] Event selection .06 [�.01, .13] .01 [�.05, .07] .17� [.10, .23] .21� [.14, .27] .03 [�.04, .09] .19� [.13, .25] Linear anticipation .02# [�.07, .10] .03# [�.07, .13] .11�# [.01, .20] �.03# [�.12, .07] .12�# [.02, .21] �.09# [�.20, .02] Post-event year .02 [�.06, .09] .05 [�.04, .14] .03# [�.05, .12] �.04# [�.13, .04] �.00 [�.09, .08] .05# [�.05, .15] Post-event baseline

change .00# [�.07, .08] �.04# [�.13, .06] �.01# [�.10, .07] .11�# [.02, .19] �.02# [�.11, .07] .05# [�.05, .15] Linear socialization .00# [�.08, .08] .09# [�.01, .18] .08# [�.01, .17] �.05# [�.14, .04] .08# [�.00, .17] .03# [�.08, .13]

Note. Here and in Tables 5–11, E � extraversion; A � agreeableness; C � conscientiousness; S � emotional stability; O � openness to experience; and LS � life satisfaction. Regarding the coefficients, the intercept corresponds to a situation of a participant of average age and gender, who has never reported the event during the study. The event selection effect indicates the difference between individuals who at one point during the study experienced the event, vs. those who did not. The transition sample effect indicates the difference between individuals who experienced a status change during the study, vs. those who did not. The linear anticipation effect pertains to gradual changes as a function of the time before the status change. The post-event year effect indicates sudden changes in personality if the assessment took place within one year of the status change. The post-event baseline change effect indicates sudden and lasting (at least during the study period) changes. Finally, the linear socialization effect pertains to gradual changes as a function of the time after the status change. Fixed effects that are statistically significant are flagged with an asterisk (�), whereas significant random effects are flagged with a hashtag (#). The alpha level was corrected for false positive discoveries, using the “BY” method published by Benjamini and Yekutieli (2001), which depends on the number of considered p values. Separate corrections were applied to the control effects and the focal effects, and no significance was computed for the intercepts. To compute confidence intervals, a uniform z-value corresponding to an alpha level of p � .001 (two-sided) was used. In rare instances, confidence intervals therefore do not conform to the statistical significance flags that accompany the fixed effects.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

11LIFE EVENTS AND PERSONALITY TRAITS

limited to life satisfaction: People who experienced this event were less satisfied to begin with, decreased in anticipation of the event and decreased even more during the first year of widowhood, but rebounded thereafter.

Robustness checks. Because our models contain various novel features (e.g., monthly time resolution of life events; inclusion of testing effects; unequal intervals between measurements), we ran a series of models with only the post-event baseline change effect, in addition to the selection effect and the control variables. This ap- proach perhaps reflects most closely a cross-lagged regression ap- proach in which personality or life satisfaction change is predicted by a dummy indicating the presence of an event transition. Overall, these results (see Tables S9 –16 in the supplementary materials) suggested that ignoring the full spectrum of event-specific effects can lead to biased conclusions (particularly in the misidentification of anticipa- tory changes as post-event change). For example, the models includ- ing only a post-event change effect identified significant post-event increases in conscientiousness and openness following paid employ- ment, but these effects were identified in the main analyses as antic- ipatory linear changes.

Individual Differences in Change

Random variance around time-variant effects. We included random effects around each of the time-variant slope parameters, to account for individual differences in event-related change. As can be seen in Tables 4 –11, this random variance was significant in the majority of cases (see Figures S1–S8 for a graphical depic- tion of the distribution of random effects). The only exception was the postyear coefficient, which was often (but not always) nonsig- nificant. Another striking feature was that the random effect was almost always significant for life satisfaction (with only two ex- ceptions), whereas this was not always the case for personality traits (especially for conscientiousness and openness). Finally, random effects were more frequently significant for events like paid employment and unemployment, and less frequently for events like childbirth, volunteering, and widowhood.

Effects of life events on rank-order stability. We investi- gated the possibility that experiencing a transition between two personality assessments was associated with a reshuffling of rank- order differences of individual differences. In total, 6 outcomes �

Table 5 Associations Between Childbirth and Personality Development

E A C S O LS

Coefficient b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI

Intercept .00 [�.04, .05] .08 [.03, .12] .16 [.12, .21] �.06 [�.10, �.01] .14 [.10, .19] �.12 [�.16, �.07] Testing �.01� [�.02, �.00] �.02� [�.03, �.01] �.01 [�.01, .00] .02� [.02, .03] .00 [�.01, .01] �.02� [�.03, �.02] Female gender .01 [�.05, .06] .65� [.60, .70] .18� [.13, .23] �.35� [�.40, �.30] �.17� [�.22, �.12] .05� [.01, .09] Age �.02� [�.03, �.01] .02� [.01, .03] .06� [.05, .06] .04� [.03, .04] �.05� [�.06, �.04] .03� [.02, .04] Age2 .02� [.00, .04] �.03� [�.04, �.01] �.11� [�.13, �.09] �.02� [�.03, �.00] �.06� [�.07, �.04] .02� [.00, .03] Transition sample �.08 [�.24, .08] �.07 [�.23, .09] .23� [.07, .39] .22� [.06, .38] .12 [�.04, .29] .35� [.17, .52] Event selection .00 [�.06, .07] �.01 [�.07, .05] �.08� [�.15, �.02] .05 [�.01, .11] �.10� [�.17, �.04] .16� [.09, .23] Linear anticipation �.04 [�.21, .14] �.12# [�.33, .09] .06 [�.14, .25] .22�# [.02, .41] .06 [�.13, .26] .31�# [.09, .54] Post-event year �.03 [�.15, .10] .04 [�.10, .19] �.01 [�.15, .13] �.07 [�.21, .07] �.00 [�.14, .13] �.00 [�.16, .16] Post-event baseline

change .01 [�.11, .13] �.09# [�.24, .05] �.14�# [�.28, �.00] .00# [�.14, .14] �.09 [�.23, .05] �.03# [�.19, .12] Linear socialization �.04# [�.16, .08] .10# [�.05, .24] �.04# [�.17, .09] �.18�# [�.31, �.04] �.00 [�.14, .13] �.13# [�.29, .02]

Note. For an explanation of terms, see note to Table 4. � significant fixed effect, # significant random effect (after correcting p values for the rate of false positive discoveries).

Table 6 Associations Between Marriage and Personality Development

E A C S O LS

Coefficient b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI

Intercept �.04 [�.09, .02] .06 [.01, .12] .06 [.01, .11] �.08 [�.13, �.02] .22 [.16, .27] �.32 [�.38, �.27] Testing �.01� [�.02, �.00] �.02� [�.03, �.01] �.01� [�.02, �.00] .03� [.02, .03] �.00 [�.01, .00] �.02� [�.03, �.01] Female gender .01 [�.05, .06] .65� [.60, .70] .18� [.13, .23] �.35� [�.40, �.30] �.17� [�.22, �.12] .05� [.01, .10] Age �.02� [�.03, �.01] .02� [.02, .03] .05� [.05, .06] .03� [.02, .04] �.03� [�.04, �.02] �.01 [�.02, .00] Age2 .02� [.01, .04] �.03� [�.04, �.01] �.10� [�.11, �.08] �.01 [�.03, .00] �.07� [�.09, �.05] .05� [.04, .07] Transition sample �.03 [�.18, .13] .10 [�.06, .25] .03 [�.13, .18] �.00 [�.16, .16] .16� [.01, .32] �.13 [�.29, .03] Event selection .06 [�.01, .12] .01 [�.05, .07] .11� [.04, .17] .07� [.01, .13] �.19� [�.26, �.13] .46� [.40, .52] Linear anticipation �.02# [�.14, .11] .03# [�.12, .18] .06# [�.08, .20] .07# [�.07, .21] �.05# [�.19, .09] .21�# [.05, .38] Post-event year .03 [�.09, .15] �.02 [�.17, .13] �.01 [�.15, .13] �.04 [�.18, .10] .01 [�.13, .14] �.00# [�.16, .16] Post-event baseline

change �.00# [�.12, .11] �.09# [�.24, .05] �.01 [�.15, .12] .01# [�.13, .14] �.04# [�.18, .09] .03# [�.12, .19] Linear socialization .01# [�.12, .14] �.05# [�.20, .11] �.02 [�.16, .13] �.02# [�.16, .13] .04 [�.10, .19] .02# [�.15, .19]

Note. For an explanation of terms, see note to Table 4. � significant fixed effect, # significant random effect (after correcting p values for false positive discoveries).

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

12 DENISSEN, LUHMANN, CHUNG, AND BLEIDORN

8 events � 48 regression models were run. These models included personality or life satisfaction as an outcome, and lagged person- ality, life event occurrence during interval, and the interaction between lagged life satisfaction/personality and life event occur- rence as predictors. Because we wanted to demonstrate differential stability effects of event transitions, we were primarily interested in the interaction effects. The alpha level was set at p � .001 in all analyses. In only 2 cases, we found a significant difference. First, an effect was found of the transition to disability on life satisfac- tion rank-order stability (p � .0005): The subsample that experi- enced this transition had a stability of .56, whereas the remaining sample had a stability of .71. Second, an effect was found of the transition to widowhood on life rank-order satisfaction stability (p � .0001): The subsample that did experience this transition had a stability of .34, whereas the remaining sample had a stability of .71.

Discussion

The present study investigated the transactions between personality and loss versus gain-based life events in the domains of love and

work. A better understanding of such transactions is relevant for major theories of personality development. Specifically, contextual theories such as SIT propose that personality is an open system that is shaped by the experience of life events when they are associated with consistent behavioral scripts that individuals commit to (socialization effects). By comparison, endogenous theories such as FFT state that personality traits mainly determine who experiences a life event (selection effects). We used a prospective design with multiple as- sessments of personality and a stringent statistical framework to examine the reciprocal effects of life events and personality over time in a nationally representative sample of Dutch individuals. Below, we relate these findings back to our hypotheses, while also discussing unexpected results. We then review the current study’s theoretical and practical implications.

Hypotheses

Our first hypothesis was that obtaining paid employment would be associated with increases in conscientiousness. This hypothesis was partially supported: Participants who transitioned into paid employment increased in conscientiousness, yet these changes

Table 7 Associations Between Volunteering and Personality Development

E A C S O LS

Coefficient b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI

Intercept .00 [�.03, .04] .07 [.03, .10] .13 [.10, .17] �.03 [�.07, .01] .09 [.06, .13] �.02 [�.06, .01] Testing �.01� [�.02, �.00] �.02� [�.03, �.01] �.01� [�.02, �.00] .02� [.02, .03] �.00 [�.01, .01] �.02� [�.03, �.01] Female gender .01 [�.05, .06] .65� [.60, .69] .18� [.13, .23] �.35� [�.40, �.30] �.17� [�.22, �.12] .05� [.01, .10] Age �.02� [�.03, �.01] .02� [.02, .03] .06� [.05, .07] .03� [.02, .04] �.04� [�.05, �.03] .02� [.01, .03] Age2 .02� [.00, .04] �.03� [�.04, �.01] �.10� [�.12, �.09] �.02� [�.04, �.00] �.05� [�.07, �.04] .01 [�.01, .02] Transition sample .03 [�.29, .35] .14 [�.16, .44] �.14 [�.45, .17] �.15 [�.46, .16] .02 [�.29, .33] �.23 [�.54, .08] Event selection �.01 [�.24, .23] .09 [�.12, .30] .03 [�.19, .25] .12 [�.10, .35] .08 [�.15, .30] �.10 [�.32, .12] Linear anticipation .10# [�.08, .28] .17 [�.05, .39] .03 [�.17, .24] .17# [�.03, .37] �.06# [�.26, .14] �.05# [�.28, .19] Post-event year .11 [�.07, .28] .11 [�.10, .32] .04 [�.15, .24] .05 [�.14, .25] �.01 [�.21, .18] .05 [�.18, .28] Post-event baseline

change �.08# [�.25, .10] �.14 [�.36, .07] �.01# [�.21, .19] �.05 [�.25, .15] .08# [�.12, .27] .10# [�.13, .34] Linear socialization .11# [�.07, .29] .04 [�.18, .25] �.08# [�.28, .12] .06# [�.13, .26] �.04# [�.24, .16] .09# [�.14, .32]

Note. For an explanation of terms, see note to Table 4. � significant fixed effect, # significant random effect (after correcting p values for false positive discoveries).

Table 8 Associations Between Unemployment and Personality Development

E A C S O LS

Coefficient b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI

Intercept .00 [�.04, .04] .07 [.03, .10] .14 [.10, .17] �.02 [�.05, .02] .09 [.06, .13] �.00 [�.04, .04] Testing �.01� [�.02, �.00] �.02� [�.03, �.02] �.01� [�.02, �.00] .03� [.02, .04] �.00 [�.01, .00] �.01� [�.02, �.00] Female gender .01 [�.05, .06] .65� [.60, .70] .18� [.13, .23] �.35� [�.40, �.30] �.17� [�.22, �.12] .05� [.00, .09] Age �.02� [�.03, �.01] .02� [.02, .03] .06� [.05, .07] .03� [.02, .04] �.04� [�.05, �.03] .01� [.01, .02] Age2 .02� [.00, .04] �.02� [�.04, �.01] �.11� [�.12, �.09] �.02� [�.04, �.01] �.05� [�.07, �.04] �.00 [�.02, .02] Transition sample .04 [�.25, .34] �.01 [�.28, .26] .06 [�.22, .34] .03 [�.25, .31] �.00 [�.29, .28] .21 [�.07, .48] Event selection �.05 [�.32, .21] .04 [�.20, .29] �.12 [�.37, .14] �.21 [�.46, .05] .02 [�.24, .28] �.53� [�.78, �.29] Linear anticipation �.01# [�.11, .09] .06# [�.06, .18] �.04# [�.15, .08] �.14�# [�.25, �.03] .02# [�.09, .14] �.23�# [�.36, �.10] Post-event year .03# [�.07, .14] .01 [�.12, .13] .02 [�.10, .13] �.02# [�.14, .09] .04 [�.08, .15] �.05# [�.19, .08] Post-event baseline

change �.04# [�.14, .06] .05# [�.07, .17] .02# [�.10, .13] .05# [�.06, .17] .02# [�.09, .13] �.03# [�.17, .10] Linear socialization .06# [�.05, .17] �.02# [�.15, .11] .00# [�.12, .13] �.05# [�.17, .07] .06# [�.06, .18] .02# [�.12, .16]

Note. For an explanation of terms, see note to Table 4. � significant fixed effect, # significant random effect (after correcting p values for false positive discoveries).

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

13LIFE EVENTS AND PERSONALITY TRAITS

took place before the transition and were small in size. We also found that people who reported to have been in paid work were higher in conscientiousness to begin with. It might be that previous longitudinal studies with fewer assessment waves (e.g., Lüdtke et al., 2011) have interpreted these combined event selection and anticipation effects as socialization effect. Needless to say, sam- pling differences between the current study and other studies (e.g., in terms of age, country, historical period) might also account for the difference.

Our second hypothesis was that experiencing unemployment, disability, divorce, and widowhood would be associated with decreases in life satisfaction before and right after the events, possibly followed by a rebound in the years following the events. We indeed found that these events were associated with either anticipatory pre-event changes (this was found for all loss-based events, i.e., for unemployment, disability, divorce, and widowhood) and/or sudden changes in the first year after the event (in the case of widowhood). This replication of effects from the life-satisfaction literature boosts our confidence that our statistical models would also be able to detect event-related

changes in personality traits. The multitude of effects found for life satisfaction, in combination with the larger effect sizes of event-related change in this variable, seems to confirm the status of life satisfaction as a so-called surface characteristic. Surface characteristics are defined as shaped by the interplay between more stable traits (like the Big Five, which are some- times regarded as core traits) and environmental influences (Kandler et al., 2014). This is consistent with our finding that effect of life events on life satisfaction change were more likely than effects on the Big Five.

As stated in the introduction, transactions between life events and personality are relevant to theoretical perspectives like FFT and SIT. It must, however, be kept in mind that these theories have a much broader scope and conclusions about their relative merit must take additional sources of evidence into account.

Findings regarding our first theoretical hypothesis have most direct relevance for SIT, which (in a strong version) predicts that gain-based love and work events would be associated with increases in agreeableness and conscientiousness, and emo- tional stability. One pattern that would be consistent with this

Table 9 Associations Between Disability and Personality Development

E A C S O LS

Coefficient b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI

Intercept .02 [�.02, .05] .07 [.03, .10] .15 [.11, .18] .01 [�.03, .04] .10 [.06, .14] .03 [�.01, .06] Testing �.01� [�.02, �.00] �.02� [�.03, �.01] �.01� [�.02, �.00] .03� [.02, .03] �.00 [�.01, .01] �.02� [�.02, �.01] Female gender .01 [�.05, .06] .65� [.60, .70] .18� [.13, .23] �.34� [�.39, �.29] �.17� [�.22, �.12] .06� [.01, .10] Age �.02� [�.03, �.01] .02� [.02, .03] .06� [.05, .07] .03� [.03, .04] �.04� [�.05, �.03] .02� [.01, .03] Age2 .02� [.00, .03] �.03� [�.04, �.01] �.11� [�.12, �.09] �.03� [�.04, �.01] �.05� [�.07, �.04] �.00 [�.02, .01] Transition sample .11 [�.16, .38] .21� [�.04, .47] .21 [�.05, .48] .09 [�.17, .35] .22� [�.04, .48] .11 [�.15, .36] Event selection �.31� [�.47, �.15] �.03 [�.17, .12] �.32� [�.47, �.17] �.63� [�.79, �.48] �.15 [�.30, .01] �.90� [�1.04, �.76] Linear anticipation �.05# [�.23, .13] .12# [�.09, .34] �.06# [�.27, .14] �.14# [�.34, .05] �.05# [�.25, .15] �.38�# [�.61, �.15] Post-event year �.04# [�.21, .13] �.04# [�.24, .17] �.05 [�.24, .14] �.07# [�.26, .12] .03 [�.15, .22] �.16# [�.38, .05] Post-event baseline

change �.02# [�.19, .15] �.10# [�.31, .10] �.04# [�.23, .15] �.09# [�.28, .10] �.02# [�.21, .17] �.09# [�.31, .13] Linear socialization .10# [�.06, .27] .03# [�.18, .23] .05# [�.14, .24] .09# [�.10, .28] .05# [�.14, .23] .16# [�.06, .37]

Note. For an explanation of terms, see note to Table 4. � significant fixed effect, # significant random effect (after correcting p values for false positive discoveries).

Table 10 Associations Between Divorce and Personality Development

E A C S O LS

Coefficient b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI

Intercept �.00 [�.04, .04] .06 [.02, .10] .14 [.10, .18] �.03 [�.06, .01] .08 [.04, .12] .03 [�.01, .07] Testing �.01� [�.02, �.00] �.02� [�.03, �.01] �.01� [�.02, �.00] .03� [.02, .03] �.00 [�.01, .01] �.02� [�.03, �.01] Female gender .01 [�.05, .06] .65� [.60, .70] .18� [.13, .23] �.35� [�.40, �.30] �.17� [�.22, �.12] .05� [.01, .10] Age �.02� [�.03, �.01] .02� [.02, .03] .06� [.05, .07] .03� [.02, .04] �.04� [�.05, �.03] .02� [.01, .03] Age2 .02� [.00, .04] �.02� [�.04, �.01] �.11� [�.12, �.09] �.02� [�.04, �.00] �.05� [�.07, �.03] �.00 [�.02, .01] Transition sample �.22 [�.49, .05] �.14 [�.40, .12] .03 [�.23, .30] �.27� [�.54, �.01] �.06 [�.33, .20] �.36� [�.62, �.09] Event selection .04 [�.07, .14] .08 [�.02, .17] �.08 [�.18, .02] �.02 [�.11, .08] .13� [.03, .23] �.42� [�.52, �.33] Linear anticipation �.05# [�.23, .13] �.00 [�.22, .21] �.05 [�.25, .15] �.19# [�.39, .01] �.03 [�.23, .17] �.63�# [�.87, �.40] Post-event year �.01 [�.21, .19] .14 [�.10, .39] .11 [�.12, .34] .08 [�.15, .30] .06 [�.16, .29] �.08# [�.34, .18] Post-event baseline

change .14# [�.06, .34] .14 [�.10, .39] �.12 [�.35, .11] .02# [�.21, .25] .02 [�.20, .25] .18# [�.08, .45] Linear socialization �.12# [�.32, .09] �.09# [�.34, .16] .12 [�.11, .36] .12# [�.12, .35] .05 [�.18, .28] .28�# [.01, .55]

Note. For an explanation of terms, see note to Table 4. � significant fixed effect, # significant random effect (after correcting p values for false positive discoveries).

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

14 DENISSEN, LUHMANN, CHUNG, AND BLEIDORN

predicted pattern is that of maturation increases following a corresponding status transition. This is only partly what we found. We did find that people increased somewhat in emo- tional stability following the transition to employment. We did not find, however, that either agreeableness or conscientious- ness increased following any of the gain-based life event tran-

sitions. Rather, conscientiousness actually decreased after childbirth. Overall, the strong version of SIT was therefore only partially supported in the current study. This conclusion may cast the meta-analytic findings of Lodi-Smith and Roberts (2007) in a different light. For example, in their study, marriage was associated with higher levels of conscientiousness and

Table 11 Associations Between Widowhood and Personality Development

E A C S O LS

Coefficient b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI b 99.9% CI

Intercept .00 [�.03, .04] .07 [.04, .11] .13 [.10, .17] �.03 [�.07, .00] .10 [.06, .13] �.02 [�.06, .02] Testing �.01� [�.02, �.00] �.02� [�.03, �.02] �.01� [�.02, �.00] .03� [.02, .03] �.00 [�.01, .01] �.02� [�.03, �.01] Female gender .01 [�.05, .06] .65� [.60, .70] .18� [.13, .23] �.35� [�.40, �.30] �.17� [�.22, �.12] .06� [.02, .11] Age �.02� [�.03, �.01] .03� [.02, .03] .06� [.05, .07] .03� [.02, .04] �.04� [�.05, �.03] .02� [.02, .03] Age2 .02� [.01, .04] �.02� [�.04, �.01] �.10� [�.12, �.08] �.02� [�.04, �.00] �.05� [�.07, �.04] .02� [.00, .04] Transition sample .08 [�.23, .38] .11 [�.18, .39] .03 [�.27, .32] �.17 [�.47, .12] �.00 [�.30, .29] �.16 [�.46, .14] Event selection �.07 [�.21, .08] �.10 [�.23, .03] �.09 [�.23, .05] .06 [�.08, .20] �.01 [�.15, .13] �.28� [�.42, �.14] Linear anticipation �.02 [�.22, .18] .03 [�.20, .27] �.08 [�.31, .14] �.03 [�.25, .20] �.10 [�.32, .12] �.37�# [�.62, �.11] Post-event year .03 [�.19, .26] .13 [�.14, .41] .12 [�.14, .38] �.13 [�.39, .13] .06 [�.20, .31] �.43�# [�.72, �.13] Post-event baseline

change �.09 [�.31, .14] �.06 [�.33, .22] �.14 [�.39, .12] �.04# [�.29, .21] �.09 [�.35, .16] �.02# [�.31, .28] Linear socialization .12 [�.13, .38] .00 [�.30, .31] .13 [�.15, .42] .10 [�.18, .39] .12 [�.16, .40] .18# [�.15, .51]

Note. For an explanation of terms, see note to Table 4. � significant fixed effect, # significant random effect (after correcting p values for false positive discoveries).

Figure 2. Figure depicting personality trait levels, as a function of the number of months relative to the transition, of individuals who experienced the transition to paid employment during the study period. Personality trait levels were standardized before inclusion in the graph. The dots represent average trait levels for different months, across all participants with a corresponding personality assessment during that month. The color of the dots indicates the number of underlying data points, with darker dots representing more cases than lighter dots. Two linear slopes were fitted to the data: One before the event, and one after event. In case the post-event year effect was significant, a third slope, corresponding to the first 12 months after the event, was also estimated. Only traits that were significantly associated with event timing are included in the composite graph.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

15LIFE EVENTS AND PERSONALITY TRAITS

emotional stability, but our results suggest that these correla- tions might reflect selection effects, that is, individual differ- ences that were already in place before the transition into the event.

Another possibility suggested by SIT is that the anticipation of a major life event would trigger corresponding psychological commitments in individuals, resulting in maturational changes before the occurrence of an event. We found partial support for this possibility. Specifically, emotional stability increased as individ-

uals approached childbirth, and conscientiousness slightly in- creased in anticipation of paid employment. No anticipation effects were found for marriage or volunteering, however. Moreover, none of the gain-based events were associated with anticipatory increases in agreeableness, even though this trait is also regarded by maturity-related traits in SIT.

FFT states that life events would not impact personality devel- opment but personality traits would instead predict the occurrence of life events. Partial support was found for this hypothesis as the Big Five traits indeed predicted the occurrence of some life events (Ozer & Benet-Martínez, 2006). Some gain-based events were predicted by traits that are theoretically and intuitively related to these events. For example, conscientious individuals were more like to end up in paid employment, and conscientious and emo- tionally stable individuals were more likely to marry (the latter individuals were also lower in openness to experiences). Work settings as well as marriage requires individuals to behave in a structured and orderly fashion, something that might come more naturally to individuals who are high in conscientiousness (who might therefore select themselves into work or marriage, and/or might be more likely to be selected by employers or potential romantic partners). For loss-based events, the pattern of findings was heterogeneous. Open individuals were more likely to experi- ence divorce, and people low in extraversion, conscientiousness, and emotional stability were more likely to experience disability. In total, around two thirds of all significant effects pertaining to the Big Five were selection effects.

That said, about a third of all significant effects indicated that life events were associated with change in the Big Five, either in

Figure 3. Figure depicting personality trait levels, as a function of the number of months relative to the transition, of individuals who experienced the transition to childbirth during the study period. For an explanation of the graph, see Figure 2 caption.

Figure 4. Figure depicting personality trait levels, as a function of the number of months relative to the transition, of individuals who experienced the transition to marriage during the study period. For an explanation of the graph, see Figure 2 caption.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

16 DENISSEN, LUHMANN, CHUNG, AND BLEIDORN

anticipation or in response to experiencing an event. This finding runs counter to strong versions of the FFT (although it is also possible that anticipatory changes reflect endogenous change, in turn predicting the subsequent selection of a life event). Because effects were often relatively small, however, our findings can be construed as consistent with its prediction that Big Five traits are not greatly affected by environmental demands. We think, how- ever, that more research is needed into why certain exceptions to this pattern occurred (e.g., why people become more conscientious and open in anticipation of work, and why emotional stability first increases and then decreases around the experience of childbirth), which would enhance theory development.8

Regarding SIT, our findings qualify the ways in which events might be associated with personality maturation. Regarding the latter, our data demonstrate that personality development is not likely a result of the mere occurrence of a life event, but might still (as SIT predicts) depend on people’s psychological experiences when undergoing major role transitions. This might also explain why some anticipatory increases took place before the event oc- curred (i.e., people increased in emotional stability before child- birth, and increased in conscientiousness and openness before paid

employment). One possible explanation consistent with SIT is that people already were committing to these roles before the actual event transition took place. It should be noted, however, that people may differ in how strongly committed they are to a role (e.g., as a worker, partner, or parent), and may thus be more or less psychologically invested in that new role. To the extent that commitment moderates the impact of role transitions on behavior and personality change, we would expect individual differences in people’s trajectories to depend on their level of commitment.

Unexpected Effects

In addition to results that pertained to our hypotheses, we found a number of unexpected effects. We will offer some speculations regarding three unexpected effects, while stressing the fact that they should be replicated in future research.

Anticipation of work. In the anticipation of the transition to paid employment, participants increased in openness. Possibly, individuals become more open to be better prepared for entering a setting in which they have to learn new skills, experience new roles, and so forth. More research is needed to substantiate this explanation, however.

Changes associated with childbirth. Before giving birth, in- dividuals increased in emotional stability but decreased again thereafter. The finding of anticipatory growth is important, as ignoring the prechildbirth anticipatory increase would paint a much bleaker picture of this life event. Instead, the postbirth decrease seems to represent a reversal to previously held levels before anticipating the change. Possibly, individuals become more stable and happy because they have positive expectations regard- ing having children, and planning this event might be emotionally rewarding. After childbirth, these positive expectations might be disappointed, explaining the post-event decreases in emotional stability. Alternatively, individuals are more likely to decide to have children during a period of social-emotional improvements, in which case the postbirth decrease would represent a regression to the mean. More research is needed to substantiate either possi- bility.

8 We thank an anonymous reviewer for this suggestion.

Figure 5. Figure depicting personality trait levels, as a function of the number of months relative to the transition, of individuals who experienced the transition to unemployment during the study period. For an explanation of the graph, see Figure 2 caption.

Figure 6. Figure depicting personality trait levels, as a function of the number of months relative to the transition, of individuals who experienced the transition to disability during the study period. For an explanation of the graph, see Figure 2 caption.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

17LIFE EVENTS AND PERSONALITY TRAITS

As already mentioned, an unexpected finding was that consci- entiousness levels showed a sudden decrease after childbirth. Al- though this finding is incompatible with SIT, it is more compatible with self-regulation theories if one assumes that the parenting of young children is a very stressful period (as is also suggested the decreases in well-being in Figure 3). This increased level of stress might disrupt young parents’ resources to act in a conscientious fashion (Denissen et al., 2013; Hutteman et al., 2014). That said, other stressful events (e.g., unemployment, divorce or widowhood) were not associated with maturity decreases as would be expected as a result of resource depletion. Therefore, the overall level of support for the self-regulatory perspective must be regarded as inconclusive. In no case, however, were loss-based events associ- ated with increases in personality maturity, which disconfirms deficits-breeds growth perspectives (Baltes et al., 1999).

Anticipatory decreases in emotional stability before unemployment. In the case of unemployment, individuals de- creased in emotional stability before experiencing the event. This mirrored similar decreases in life satisfaction before the event, which have already been documented in the life-satisfaction liter- ature (e.g., Lucas, 2007; Luhmann et al., 2012). It has been argued that these decreases may be explained by the fact that, in many cases, the event does not come unexpected. For example, individ- uals’ job security might be threatened before they actually lose their job. It is a novel finding that the anticipatory changes before unemployment generalizes to changes in emotional stability.

Summary and Implications

Overall, some conclusions can be drawn regarding transactions between life events and personality traits. First of all, prevalence effects accounted for almost two thirds of all significant effects pertaining to the Big Five, indicating that personality traits and life events are intertwined because the occurrence of certain life events is related to preexisting personality trait levels. This is consistent with behavioral genetic research that shows that life events are partly heritable (Kandler et al., 2012). Other than selection effects, person- ality sometimes did change as a function of life events, but these

events were mostly anticipatory in nature. This is a novel finding, because previous research has not been able to zoom in on these effects using high-resolution data. These anticipatory changes were mostly of the type that emotional stability increased in anticipation of gain-based events (like childbirth), and decreased in anticipation to unemployment (Lucas, Clark, Georgellis, & Diener, 2004).

We also found some support for SIT such as increases in emotional stability related to work and childbirth but overall the evidence must be considered mixed. Future research is needed to explain this inconsistent pattern of results. For example, future research should measure individual role experiences re- lated to different events (see Denissen, Ulferts, Lüdtke, Muck, & Gerstorf, 2014, for a similar approach in the domain of job roles). Moreover, according to a broader version of SIT, it is not the transition to certain events per se that is associated with personality trait changes, but qualitative differences in how these transitions are mastered and invested in. This possibility did obtain some support in the finding that there were abundant yet unexplained between-person differences in event-related personality change. For future studies, it would be important to measure the degree of success of relevant transitions, as well as participants’ psychological commitment to various social role (e.g., using an identity interview; Klimstra & Denissen, 2017) or behavioral investment in trait-consistent activities (Schwaba et al., 2017), as moderators of event-related personality change.

Overall, we found some support for the premises of FFT, as far as they pertain to the overall pattern of transactions between personality and life events. As already stated, many significant transactions pertained to selection effects predicted by that theory. Specifically, of 40 possible selection effects, 11 (27.5%) were statistically significant. As stated above, it can be argued that linear anticipation effects are ambiguous, so that only 3 genuinely event-contingent effects remained (out of a total possible number of 120 effects; i.e., 2.5%). Proponents of FFT might retort that the corresponding changes, as displayed in Figures 2– 8, were rather minor in size. Furthermore, judging from these figures, personality trait levels sometimes returned to their pre-event levels—as can be seen, for example, when

Figure 7. Figure depicting personality trait levels, as a function of the number of months relative to the transition, of individuals who experienced the transition to divorce during the study period. For an explanation of the graph, see Figure 2 caption.

Figure 8. Figure depicting personality trait levels, as a function of the number of months relative to the transition, of individuals who experienced the transition to widowhood during the study period. For an explanation of the graph, see Figure 2 caption.

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

18 DENISSEN, LUHMANN, CHUNG, AND BLEIDORN

inspecting the effect of childbirth on changes in emotional stability in Figure 2. Finally, of all 7 significant change effects (i.e., including the linear anticipation effects), 4 pertained to emotional stability. As we demonstrated above, some of the emotional stability findings seemed to be driven by life satis- faction changes. This would seem to fit McCrae and Costa’s (1999) idea that life satisfaction represents a so-called charac- teristic adaptation that is more malleable to environmental influences, as compared to the Big Five dimensions, which are referred to as core traits in FFT (for a meta-analysis of stability estimates of personality and life satisfaction, see Anusic & Schimmack, 2016).

In spite of all this, it should be noted that life events did have effects on personality trait development even after implementing some very strict statistical controls, which runs counter to the theory. Furthermore, sometimes changes appeared more perma- nent. For example, when it came to childbirth, parents permanently (at least for the duration of the study) decreased in conscientious- ness. Finally, the anticipation effects were also consistent with contextual theories. More specific operationalizations of FFT in combination with biologically and genetically informative designs are called for in order to evaluate the relevance of our findings for the core tenets of this theory.

Future research should look into biological and/or environ- mental mediators of personality change following life events. The FFT hypothesis indicates that all personality changes as- sociated with life events should be biologically based. Although this is plausible when it comes to childbirth (with its associated hormonal changes), this is an open question regarding other transitions. For example, transitions such as entering into paid employment or unemployment could affect stress levels and corresponding biological markers, like cortisol. More research is needed to establish the biological correlates of event-related changes. Moreover, studies should investigate the specific mechanisms by which personality traits influence the likelihood that people end up experiencing certain life events. For exam- ple, it might be that people with certain personality levels differentially evaluate the benefits of various event transitions (Wood, Gardner, & Harms, 2015). For example, people high in openness to experience might have more liberal attitudes to- ward divorce, explaining their higher likelihood to go through this transition. Furthermore, generally happy individuals might have a more optimistic outlook on life and would thus be more confident when it comes to marrying a romantic partner.

Strengths and Limitations

The present study has many strengths that set it apart from previous studies on the effects of life events on personality change (e.g., it uses a large and representative prospective longitudinal design with a fine-grained resolution to measure various types of change effect). Still, this study suffers from some important limitations. First, although we used a large and representative sample, statistical power was still a limiting factor. For example, only 61 individuals who experienced the transition to widowhood during our study contributed the max- imum of 6 possible personality assessments. Even though our study was adequately powered to detect linear effects (see above), more subtle or nonlinear (e.g., quadratic or cubic)

effects would require larger sample sizes and even more assess- ment waves. More generally, it is clear from the current data that effects of life events on personality development were relatively small and subtle, at least when compared to effects on life satisfaction. Also, we did not investigate the effect of repeat events (e.g., whether becoming unemployed multiple times has effects that are different from being unemployed only once). We hope that larger cohort studies will become available to conduct these analyses, or that data from different cohort stud- ies can be pooled to increase power to examine such effects.

Furthermore, it should be noted that we only studied the effects of life events from a variable-centered perspective. For example, it might be that a life events have the greatest impact on traits that are particularly salient for an individual, for example the trait with the lowest score in an individual’s personality profile (i.e., this would indicate the existence of a relative weak spot that becomes more pronounced after the transition; see footnote 8). Also, it is possible that life event changes have an effect on constellations of traits or affect an individual’s personality type (Asendorpf & van Aken, 1999). For example, it might be that after a certain life event, individuals change from a resilient type to an undercontrolled type (Meeus, van de Schoot, Klimstra, & Branje, 2011). Future research is needed to investigate these possibilities.

Finally, another limitation is that our study did not provide information regarding individuals’ idiographic experience of life transitions. Gaining insight into the ways people approach and master life transitions is an important task for future research. For example, it may be that the dissolution of a marriage has different effects on spouses’ well-being and personality depending on whether it was perceived as a good marriage, who initiated the break-up, and so forth. Finding the courage to leave an abusive marriage might have empowering consequences, whereas being abandoned by a spouse in a (seemingly) good marriage might have more negative effects. Consistent with this, we found that event- related personality change varied substantially between people, as indicated by an abundance of random slope effects. A more in- depth examination of the experiences people have during major life transitions would be also important to test more circumscribed predictions. For example, SIT states that the psychological com- mitment to adult social roles such as work, marriage, would predict increases in personality maturation to a greater extent than the mere occurrence of these events. Furthermore, key processes out- lined by theories about the disruption of maturity would require measures of self-regulation resources and long-term goals. Such measures should be examined in to future studies.

Conclusion

The current study employed high-resolution longitudinal data to delineate the effect of life events on personality trait change across the adult life span. Our results indicate that the impact of life events differed a great deal between traits and events. Our findings did not perfectly square with a particular theoretical approach to personality development, although the endogenous perspective generally received some support by the finding that selection effects were more frequent than socialization effects. Adding to this overall impression, post-event socialization effects tended to be small and were often not long-lasting. Interestingly, the current study found some anticipatory effects that were not identified in

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

19LIFE EVENTS AND PERSONALITY TRAITS

previous studies, such as the finding that emotional stability in- creased during the time leading up to childbirth. Understanding how transactions between personality traits and life events are mediated by biological, psychological and social processes is an important task for future research.

References

Anusic, I., & Schimmack, U. (2016). Stability and change of personality traits, self-esteem, and well-being: Introducing the meta-analytic stability and change model of retest correlations. Journal of Personality and Social Psychology, 110, 766 –781. http://dx.doi.org/10.1037/pspp0000066

Asendorpf, J. B., & van Aken, M. A. (1999). Resilient, overcontrolled, and undercontrolled personality prototypes in childhood: Replicability, predic- tive power, and the trait-type issue. Journal of Personality and Social Psychology, 77, 815– 832. http://dx.doi.org/10.1037/0022-3514.77.4.815

Baird, B. M., Lucas, R. E., & Donnellan, M. B. (2010). Life satisfaction across the lifespan: Findings from two nationally representative panel studies. Social Indicators Research, 99, 183–203. http://dx.doi.org/10 .1007/s11205-010-9584-9

Baltes, P. B., Staudinger, U. M., & Lindenberger, U. (1999). Lifespan psychology: Theory and application to intellectual functioning. Annual Review of Psychology, 50, 471–507. http://dx.doi.org/10.1146/annurev .psych.50.1.471

Bates, D., Mächler, M., Bolker, B., & Walker, S. (2014). Fitting linear mixed-effects models using lme4. arXiv:1406.5823 [Stat]. Retrieved from http://arxiv.org/abs/1406.5823

Belsky, J., & Rovine, M. (1990). Patterns of marital change across the transition to parenthood: Pregnancy to three years postpartum. Journal of Marriage and the Family, 52, 5–19. http://dx.doi.org/10.2307/352833

Benjamini, Y., & Yekutieli, D. (2001). The control of the False Discovery Rate in multiple testing under dependency. Annals of Statistics, 29, 1165–1188.

Bleidorn, W. (2012). Hitting the road to adulthood: Short-term personality development during a major life transition. Personality and Social Psychology Bulletin, 38, 1594 –1608. http://dx.doi.org/10.1177/01461 67212456707

Bleidorn, W. (2015). What accounts for personality maturation in early adulthood? Current Directions in Psychological Science, 24, 245–252. http://dx.doi.org/10.1177/0963721414568662

Bleidorn, W., Buyukcan-Tetik, A., Schwaba, T., van Scheppingen, M. A., Denissen, J. J., & Finkenauer, C. (2016). Stability and change in self- esteem during the transition to parenthood. Social Psychological and Personality Science, 7, 560 –569. http://dx.doi.org/10.1177/19485 50616646428

Bleidorn, W., & Denissen, J. J. A. (in press). Personality and life transitions in early adulthood. In O. P. John & R. W. Robins (Eds.), Handbook of personality: Theory and research. New York, NY: Guilford.

Bleidorn, W., Hopwood, C. J., & Lucas, R. E. (2018). Life events and personality trait Change. Journal of Personality, 86, 83–96. http://dx.doi .org/10.1111/jopy.12286

Bleidorn, W., Klimstra, T. A., Denissen, J. J. A., Rentfrow, P. J., Potter, J., & Gosling, S. D. (2013). Personality maturation around the world: A cross-cultural examination of social-investment theory. Psychological Science, 24, 2530 –2540. http://dx.doi.org/10.1177/0956797613498396

Boyce, C. J., Wood, A. M., Daly, M., & Sedikides, C. (2015). Personality change following unemployment. Journal of Applied Psychology, 100, 991–1011. http://dx.doi.org/10.1037/a0038647

Caspi, A., & Moffitt, T. E. (1993). When do individual differences matter? A paradoxical theory of personality coherence. Psychological Inquiry, 4, 247–271. http://dx.doi.org/10.1207/s15327965pli0404_1

Costa, P. T., Jr., Herbst, J. H., McCrae, R. R., & Siegler, I. C. (2000). Personality at midlife: Stability, intrinsic maturation, and response to life

events. Assessment, 7, 365–378. http://dx.doi.org/10.1177/107319110 000700405

Denissen, J. J. A., van Aken, M. A., Penke, L., & Wood, D. (2013). Self-regulation underlies temperament and personality: An integrative developmental framework. Child Development Perspectives, 7, 255– 260. http://dx.doi.org/10.1111/cdep.12050

Denissen, J. J. A., Ulferts, H., Lüdtke, O., Muck, P. M., & Gerstorf, D. (2014). Longitudinal transactions between personality and occupational roles: A large and heterogeneous study of job beginners, stayers, and changers. Developmental Psychology, 50, 1931–1942. http://dx.doi.org/ 10.1037/a0036994

Diener, E., Emmons, R. A., Larsen, R. J., & Griffin, S. (1985). The satisfaction with life scale. Journal of Personality Assessment, 49, 71– 75. http://dx.doi.org/10.1207/s15327752jpa4901_13

Doré, B., & Bolger, N. (2017). Population- and individual-level changes in life satisfaction surrounding major life stressors. Social Psychological & Personality Science. Advance online publication.

Goldberg, L. R. (1992). The development of markers for the Big-Five factor structure. Psychological Assessment, 4, 26 – 42. http://dx.doi.org/ 10.1037/1040-3590.4.1.26

Green, P., & MacLeod, C. J. (2016). SIMR: An R package for power analysis of generalized linear mixed models by simulation. Methods in Ecology and Evolution, 7, 493– 498. http://dx.doi.org/10.1111/2041-210X.12504

Hogan, R., & Roberts, B. W. (2004). A socioanalytic model of maturity. Journal of Career Assessment, 12, 207–217. http://dx.doi.org/10.1177/ 1069072703255882

Hutteman, R., Bleidorn, W., Keresteš, G., Brković, I., Butković, A., & Denissen, J. J. A. (2014). Reciprocal associations between parenting challenges and parents’ personality development in young and middle adulthood. European Journal of Personality, 28, 168 –179. http://dx.doi .org/10.1002/per.1932

Jayawickreme, E., & Blackie, L. E. R. (2014). Post-traumatic growth as positive personality change: Evidence, controversies and future direc- tions. European Journal of Personality, 28, 312–331. http://dx.doi.org/ 10.1002/per.1963

John, O. P., & Srivastava, S. (1999). The Big-Five trait taxonomy: History, measurement, and theoretical perspectives. In L. A. Pervin & O. P. John (Eds.), Handbook of personality: Theory and research (Vol. 2, pp. 102–138). New York, NY: Guilford Press.

Johnson, J. (1997). Units of analysis for the description and explanation of personality. In R. Hogan & S. Briggs (Eds.), Handbook of personality psychology (pp. 73–93). San Diego, CA: Academic Press. http://dx.doi .org/10.1016/B978-012134645-4/50004-4

Jokela, M., Kivimäki, M., Elovainio, M., & Keltikangas-Järvinen, L. (2009). Personality and having children: A two-way relationship. Jour- nal of Personality and Social Psychology, 96, 218 –230. http://dx.doi .org/10.1037/a0014058

Kandler, C., Bleidorn, W., Riemann, R., Angleitner, A., & Spinath, F. M. (2012). Life events as environmental States and genetic traits and the role of personality: A longitudinal twin study. Behavior Genetics, 42, 57–72. http://dx.doi.org/10.1007/s10519-011-9491-0

Kandler, C., Zimmermann, J., & McAdams, D. P. (2014). Core and surface characteristics for the description and theory of personality differences and development. European Journal of Personality, 28, 231–243. http:// dx.doi.org/10.1002/per.1952

Kessler, R. C. (1997). The effects of stressful life events on depression. Annual Review of Psychology, 48, 191–214. http://dx.doi.org/10.1146/ annurev.psych.48.1.191

Klimstra, T. A., & Denissen, J. J. A. (2017). A theoretical framework for the associations between identity and psychopathology. Developmental Psychology, 53, 2052–2065. http://dx.doi.org/10.1037/dev0000356

Lefcheck, J. (2012, December 28). Dealing with multicollinearity using VIFs. [Web log post]. Retrieved from https://jonlefcheck.net/2012/12/ 28/dealing-with-multicollinearity-using-variance-inflation-factors/

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

20 DENISSEN, LUHMANN, CHUNG, AND BLEIDORN

Lehnart, J., Neyer, F. J., & Eccles, J. (2010). Long-term effects of social investment: The case of partnering in young adulthood. Journal of Personality, 78, 639 – 670. http://dx.doi.org/10.1111/j.1467-6494.2010 .00629.x

Lodi-Smith, J., & Roberts, B. W. (2007). Social investment and personal- ity: A meta-analysis of the relationship of personality traits to investment in work, family, religion, and volunteerism. Personality and Social Psychology Review, 11, 68 – 86. http://dx.doi.org/10.1177/10888683 06294590

Lucas, R. E. (2007). Adaptation and the set-point model of subjective well- being does happiness change after major life events? Current Directions in Psychological Science, 16, 75–79. http://dx.doi.org/10.1111/j.1467-8721 .2007.00479.x

Lucas, R. E., Clark, A. E., Georgellis, Y., & Diener, E. (2004). Unem- ployment alters the set point for life satisfaction. Psychological Science, 15, 8 –13. http://dx.doi.org/10.1111/j.0963-7214.2004.01501002.x

Lüdtke, O., Roberts, B. W., Trautwein, U., & Nagy, G. (2011). A random walk down university avenue: Life paths, life events, and personality trait change at the transition to university life. Journal of Personality and Social Psychology, 101, 620 – 637. http://dx.doi.org/10.1037/a0023743

Lugtig, P. (2014). Panel attrition: Separating stayers, fast attriters, gradual attriters, and lurkers. Sociological Methods & Research, 43, 699 –723. http://dx.doi.org/10.1177/0049124113520305

Luhmann, M., & Eid, M. (2012). Studying reaction to repeated life events with discontinuous change models using HLM. In G. D. Garson (Ed.), Hierarchical linear modeling: Guide and applications (pp. 273–289). Thousand Oaks, CA: Sage.

Luhmann, M., Hofmann, W., Eid, M., & Lucas, R. E. (2012). Subjective well-being and adaptation to life events: A meta-analysis. Journal of Personality and Social Psychology, 102, 592– 615. http://dx.doi.org/10 .1037/a0025948

Luhmann, M., Orth, U., Specht, J., Kandler, C., & Lucas, R. E. (2014). Studying changes in life circumstances and personality: It’s about time. European Journal of Personality, 28, 256 –266. http://dx.doi.org/10 .1002/per.1951

Lykken, D., & Tellegen, A. (1996). Happiness is a stochastic phenomenon. Psychological Science, 7, 186 –189. http://dx.doi.org/10.1111/j.1467- 9280.1996.tb00355.x

McArdle, J. J., Ferrer-Caja, E., Hamagami, F., & Woodcock, R. W. (2002). Comparative longitudinal structural analyses of the growth and decline of multiple intellectual abilities over the life span. Developmental Psy- chology, 38, 115–142. http://dx.doi.org/10.1037/0012-1649.38.1.115

McCrae, R. R. (1993). Moderated analyses of longitudinal personality stability. Journal of Personality and Social Psychology, 65, 577–585. http://dx.doi.org/10.1037/0022-3514.65.3.577

McCrae, R. R., & Costa, P. T., Jr. (1999). A Five-Factor theory of personality. In L. A. Pervin & O. P. John (Eds.), Handbook of person- ality: Theory and research (pp. 139 –153). New York: Guilford Press.

Meeus, W., Van de Schoot, R., Klimstra, T., & Branje, S. (2011). Person- ality types in adolescence: Change and stability and links with adjust- ment and relationships: A five-wave longitudinal study. Developmental Psychology, 47, 1181–1195. http://dx.doi.org/10.1037/a0023816

Neyer, F. J., & Asendorpf, J. B. (2001). Personality-relationship transac- tion in young adulthood. Journal of Personality and Social Psychology, 81, 1190 –1204. http://dx.doi.org/10.1037/0022-3514.81.6.1190

Ozer, D. J., & Benet-Martínez, V. (2006). Personality and the prediction of consequential outcomes. Annual Review of Psychology, 57, 401– 421. http://dx.doi.org/10.1146/annurev.psych.57.102904.190127

Prinzie, P., Stams, G. J. J. M., Deković, M., Reijntjes, A. H. A., & Belsky, J. (2009). The relations between parents’ Big Five personality factors and parenting: A meta-analytic review. Journal of Personality and Social Psychology, 97, 351–362. http://dx.doi.org/10.1037/a0015823

Rindfuss, R. R. (1991). The young adult years: Diversity, structural change, and fertility. Demography, 28, 493–512. http://dx.doi.org/10 .2307/2061419

Roberts, B. W., & Davis, J. P. (2016). Young adulthood is the crucible of personality development. Emerging Adulthood, 4, 318 –326. http://dx .doi.org/10.1177/2167696816653052

Roberts, B. W., & DelVecchio, W. F. (2000). The rank-order consistency of personality traits from childhood to old age: A quantitative review of longitudinal studies. Psychological Bulletin, 126, 3–25. http://dx.doi .org/10.1037/0033-2909.126.1.3

Roberts, B. W., & Mroczek, D. (2008). Personality trait change in adult- hood. Current Directions in Psychological Science, 17, 31–35. http://dx .doi.org/10.1111/j.1467-8721.2008.00543.x

Roberts, B. W., O’Donnell, M., & Robins, R. W. (2004). Goal and personality trait development in emerging adulthood. Journal of Per- sonality and Social Psychology, 87, 541–550. http://dx.doi.org/10.1037/ 0022-3514.87.4.541

Roberts, B. W., Walton, K. E., & Viechtbauer, W. (2006). Patterns of mean-level change in personality traits across the life course: A meta- analysis of longitudinal studies. Psychological Bulletin, 132, 1–25. http://dx.doi.org/10.1037/0033-2909.132.1.1

Roberts, B. W., Wood, D., & Lodi-Smith, J. (2005). Evaluating Five Factor Theory and social investment perspectives on personality trait develop- ment. Journal of Research in Personality, 39, 166 –184. http://dx.doi .org/10.1016/j.jrp.2004.08.002

Scherpenzeel, A. (2011). Data collection in a probability-based internet panel: How the LISS panel was built and how it can be used. Bulletin of Sociological Methodology. Bulletin de Methodologie Sociologique, 109, 56 – 61. http://dx.doi.org/10.1177/0759106310387713

Scherpenzeel, A., Das, J. W. M., Ester, P., & Kaczmirek, L. (2010). ‘True’ longitudinal and probability-based internet panels: Evidence from the Netherlands. Social and Behavioral Research and the Internet, 77–103.

Schwaba, T., & Bleidorn, W. (2017a). Individual differences in personality change across the adult life span. Journal of Personality. Advance online publication. http://dx.doi.org/10.1111/jopy.12327

Schwaba, T., & Bleidorn, W. (2017b). Personality trait development across the transition to retirement. Journal of Personality and Social Psychology. Advance online publication. http://dx.doi.org/10.1037/pspp0000179

Schwaba, T., Luhmann, M., Denissen, J. J. A., Chung, J. M., & Bleidorn, W. (2017). Openness to experience and culture-openness transactions across the lifespan. Journal of Personality and Social Psychology. Advance online publication. http://dx.doi.org/10.1037/pspp0000150

Shin, D. C., & Johnson, D. M. (1978). Avowed happiness as an overall assessment of the quality of life. Social Indicators Research, 5, 475– 492. http://dx.doi.org/10.1007/BF00352944

Singer, J. D., & Willett, J. B. (2003). Applied longitudinal data analysis: Modeling change and event occurrence. Oxford, England: Oxford Uni- versity Press.

Soto, C. J., & Tackett, J. L. (2015). Personality traits in childhood and adolescence: Structure, development, and outcomes. Current Directions in Psychological Science, 24, 358 –362. http://dx.doi.org/10.1177/ 0963721415589345

Specht, J., Bleidorn, W., Denissen, J. J. A., Hennecke, M., Hutteman, R., Kandler, C., . . . Zimmermann, J. (2014). What drives adult personality development? A comparison of theoretical perspectives and empirical evidence. European Journal of Personality, 28, 216 –230. http://dx.doi .org/10.1002/per.1966

Specht, J., Egloff, B., & Schmukle, S. C. (2011). Stability and change of personality across the life course: The impact of age and major life events on mean-level and rank-order stability of the Big Five. Journal of Personality and Social Psychology, 101, 862– 882. http://dx.doi.org/10 .1037/a0024950

Terracciano, A. (2014). Pregnancy, marriage, and quitting school nurture personality development? Commentary on Bleidorn, Klimstra, Denis-

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

21LIFE EVENTS AND PERSONALITY TRAITS

sen, Rentfrow, Potter, and Gosling (2013). Psychological Science, 25, 1049 –1050. http://dx.doi.org/10.1177/0956797613515003

Van der Laan, J. (2009). Representativity of the LISS panel. (Discussion Paper 09041). The Hague/Heerlen, the Netherlands: Statistics Nether- lands.

van Scheppingen, M. A., Denissen, J. J. A., Chung, J. M., Tambs, K., & Bleidorn, W. (2017). Self-esteem and relationship satisfaction during the transition to motherhood. Journal of Personality and Social Psychology. Advance online publication. http://dx.doi.org/10.1037/pspp0000156

van Scheppingen, M. A., Jackson, J. J., Specht, J., Hutteman, R., Denissen, J. J. A., & Bleidorn, W. (2016). Personality trait development during the

transition to parenthood: A test of social investment theory. Social Psychological and Personality Science, 7, 452– 462. http://dx.doi.org/10 .1177/1948550616630032

Wood, D., Gardner, M. H., & Harms, P. D. (2015). How functionalist and process approaches to behavior can explain trait covariation. Psycholog- ical Review, 122, 84 –111. http://dx.doi.org/10.1037/a0038423

Received May 2, 2017 Revision received January 25, 2018

Accepted February 27, 2018 �

T hi

s do

cu m

en t

is co

py ri

gh te

d by

th e

A m

er ic

an P

sy ch

ol og

ic al

A ss

oc ia

ti on

or on

e of

it s

al li

ed pu

bl is

he rs

. T

hi s

ar ti

cl e

is in

te nd

ed so

le ly

fo r

th e

pe rs

on al

us e

of th

e in

di vi

du al

us er

an d

is no

t to

be di

ss em

in at

ed br

oa dl

y.

22 DENISSEN, LUHMANN, CHUNG, AND BLEIDORN