Personality Change Following Unemployment

Christopher J. Boyce and Alex M. Wood

University of Stirling and University of Manchester

Michael Daly

University of Stirling

Constantine Sedikides

University of Southampton

Unemployment has a strongly negative influence on well-being, but it is unclear whether it also alters

basic personality traits. Whether personality changes arise through natural maturation processes or

contextual/environmental factors is still a matter of debate. Unemployment, a relatively unexpected and

commonly occurring life event, may shed light on the relevance of context for personality change. We

examined, using a latent change model, the influence of unemployment on the five-factor model of

personality in a sample of 6,769 German adults, who completed personality measures at 2 time points 4

years apart. All participants were employed at the first time point, and a subset became unemployed over

the course of the study. By the second time point, participants had either remained in employment, been

unemployed from 1 to 4 years, or had experienced some unemployment but become reemployed.

Compared with those who had remained in employment, unemployed men and women experienced

significant patterns of change in their mean levels of agreeableness, conscientiousness, and openness,

whereas reemployed individuals experienced limited change. The results indicate that unemployment has

wider psychological implications than previously thought. In addition, the results are consistent with the

view that personality changes as a function of contextual and environmental factors.

Keywords: unemployment, personality, personality change, well-being, five-factor model

Personality is most often viewed within the hierarchical five-

factor model (FFM; McCrae & Costa, 2008). The basic traits of

agreeableness, conscientiousness, extraversion, neuroticism, and

openness occupy the highest level of the personality hierarchy,

whereas other psychological characteristics (i.e., manifestations of

the basic traits) occupy lower levels. Given that the FFM was

partially motivated by biological considerations (McCrae et al.,

2000), there was an initial tendency to regard these traits as

relatively fixed, changing early in life through maturation but

becoming “set like plaster” at approximately the age of 30 (Costa

& McCrae, 1994; Srivastava, John, Gosling, & Potter, 2003).

Recent advances, however, have challenged the traditional “set

like plaster” perspective by demonstrating evidence of change

throughout the life-cycle stages (Lucas & Donnellan, 2011; Rob-

erts, Walton, & Viechtbauer, 2006a). Indeed, personality may be

as malleable as socioeconomic variables such as income or marital

status (Boyce, Wood, & Powdthavee, 2013; Osafo Hounkpatin,

Wood, Boyce, & Dunn, in press). Current debate now mostly

centers on the extent to which personality change is a function of

natural maturation processes versus events that occur throughout

life (Costa & McCrae, 2006; Roberts, Walton, & Viechtbauer,

2006b).

Some proponents of the FFM argue that most of the observed

personality changes are attributable to intrinsic maturation pro-

cesses brought about by genetic influences (McCrae & Costa,

2008). Such a perspective is bolstered by similarities in the way

traits appear to develop over the life cycle across diverse cultures

(McCrae et al., 1999, 2000). However, there is also a strong

environmental contribution to personality change (Kandler, 2012),

consistent with twin longitudinal studies that indicate that person-

ality change has both genetic and environmental components

(Bleidorn et al., 2010; Bleidorn, Kandler, Riemann, Spinath, &

Angleitner, 2009). In support of the role of environmental varia-

tion in personality change, commonly occurring events—such as

alterations in marital status (Specht, Egloff, & Schmukle, 2011),

marital and relationship quality (Neyer & Lehnart, 2007; Roberts

& Bogg, 2004; Watson & Humrichouse, 2006), retirement (Specht

et al., 2011), and experiences within the workplace (Roberts,

Caspi, & Moffitt, 2003)—have all been linked to personality

change.

However, many events that have been investigated in connec-

tion with personality change are normative, in the sense that they

This article was published Online First February 9, 2015.

Christopher J. Boyce and Alex M. Wood, Behavioural Science Centre,

Stirling Management School, University of Stirling, and School of Psy-

chological Sciences, University of Manchester; Michael Daly, Behavioural

Science Centre, Stirling Management School, University of Stirling; Con-

stantine Sedikides, Center for Research on Self and Identity, School of

Psychology, University of Southampton.

The authors would like to extend thanks to James Banks, Eamonn

Ferguson, and Nattavudh Powdthavee for helpful comments. The Eco-

nomic and Social Research Council (PTA-026-27-2665, ES/K00588X/1)

and the University of Stirling provided research support. The data used

here were made available by the German Institute for Economic Research

(DIW Berlin). Neither the original collectors of the data nor the Archive

bears any responsibility for the analyses or interpretations presented here.

Correspondence concerning this article should be addressed to Christo-

pher J. Boyce, Behavioural Science Centre, Stirling Management School,

University of Stirling, FK9 4LA. E-mail: [email protected]

2015, Vol. 100, No. 4, 991–1011 0021-9010/15/$12.00 http://dx.doi.org/10.1037/a0038647

991

occur at specific points in the life cycle that correspond with

age-graded social roles. As such, there may be alternative expla-

nations to personality changes. According to the model of person–

environment transactions (Roberts, Wood, & Caspi, 2008), con-

tinuous interactions between person and environment promote

both stability and change. Individuals may orient toward environ-

ments that match their personalities, but they will still face fluc-

tuations in the expectations placed upon them, by others and

themselves, both before and after they assume new roles. The

effect of normative events on personality change can therefore be

challenging to examine, because it is difficult to distinguish the

extent to which the experience (or anticipation) of an event pre-

cipitated personality change, whether the event itself happened to

co-occur with a natural process of personality maturation, or

whether personality change culminated in the event itself.

To minimize the conceptual and methodological problems as-

sociated with examining changes associated with normative roles,

it is more informative to explore the influence of non-normative

events on personality. We know, for example, that the use of

certain drugs (MacLean, Johnson, & Griffiths, 2011; Roberts &

Bogg, 2004), the experience of frightening or horrifying events

(Löckenhoff, Terracciano, Patriciu, Eaton, & Costa, 2009), and

involvement in intensive outpatient counseling (Piedmont, 2001)

can all initiate personality changes. However, although such find-

ings indicate personality changes as a result of contextual or

environmental factors, the relevant events are uncommon. In this

article, we examine changes in personality as a function of a

relatively major and commonly occurring non-normative life

event, namely, unemployment. In particular, we test whether,

relative to remaining employed, (a) unemployment precipitates

changes in basic personality traits, (b) this change depends on

unemployment duration, (c) the influence of unemployment on

personality differs by gender, and (d) unemployment-triggered

personality change endures following reemployment.

Personality Stability and Change

Debate on whether personality can and does change has been

hindered from lack of explicit definitions of personality. Indeed, a

good deal of disagreement has arisen by nonshared definitions of

the construct. This can be particularly problematic, if one under-

stands personality to represent the nonchanging aspects of the

person. In this case, personality change would be precluded by

terminological barriers or tautologies: If something is observed to

change, it can no longer be deemed “personality.” Moreover,

adopting a rigid definition of personality in terms of “unmitigated

stability” would lead to the unavoidable conclusion that changes

indicated by self-report measures of personality are inherently

meaningless, despite a vast literature documenting the reliability

and validity of such measures. Fortunately, personality psycholo-

gists are inclined to define personality more inclusively—for ex-

ample, as “the psychological component of a person that remains

from one situation to another” (A. M. Wood & Boyce, in press).

This definition implies a degree of temporal and cross-situational

stability, without which the construct would be viewed as a par-

ticular state arising in a particular situation, but does not preclude

substantive personality change over time.

Aligned with this view, Mischel and Shoda (1995; Shoda &

Mischel, 1998) defined personality as the stable way in which

people behave within a given situation, such that people may have

stably different personalities in different situations (e.g., at work

vs. leisure). Personality indeed varies across social roles, with

higher variation across roles being linked to reduced authenticity

and impaired well-being (Bettencourt & Sheldon, 2001; Lenton,

Bruder, Slabu, & Sedikides, 2013; Sheldon, Ryan, Rawsthorne, &

Ilardi, 1997). Fleeson (2001, 2004) defined personality as the

average of personality expression across roles and situations, and

showed that personality expression varies continually such that a

person may score a “1” on extraversion one morning and a “7” the

next, depending on situational factors. At the same time, Fleeson

(2001, 2004) also demonstrated that individuals can be reliably

distinguished from one another by the mean point of their person-

ality expression distribution, which is to what people refer when

asked about their personality “in general.” Each of these perspec-

tives is compatible with definitions of personality as interindi-

vidual differences in either behavior or the propensity to behave

(Borghans, Duckworth, Heckman, & ter Weel, 2008; Eysenck,

1981).

Drawing consensus across these contemporary definitions, per-

sonality is regarded as a snapshot of a fluid process of individuals

engaging dynamically with their environments, expressing behav-

iors to varying degrees, but being differentiated by how they

typically feel, think, and behave—the “stable part of themselves”

(Gramzow et al., 2004; Hafdahl, Panter, Gramzow, Sedikides, &

Insko, 2000; Robinson & Sedikides, 2009). None of the perspec-

tives anticipates that personality remain completely stable over

time. Quite the converse: Were people to find themselves chron-

ically in a different life situation, they would (a) reliably exhibit

different characteristics in the new environment (Mischel &

Shoda, 1995; Shoda & Mischel, 1998), (b) have different mean

levels in distributions of personality expression (Fleeson, 2001,

2004), (c) and have stably different behavior propensities

(Gramzow et al., 2004; Hafdahl et al., 2000). Indeed, it is highly

plausible that living in new environments would precipitate per-

sonality change, given the adaptive advantage of adjusting flexibly

to one’s contextual circumstance; such an advantage would max-

imize the person–environment fit (Lewin, 1951; Magnusson &

Endler, 1977; Pervin, 1968).

These reflections on the nature of personality underlay our

expectation that personality would change following unemploy-

ment, particularly if the experience were prolonged (Reynolds et

al., 2010). Thus, unemployment, which represents a severe envi-

ronmental alteration that removes social contacts and restricts the

opportunity to engage in certain types of tasks, would likely enable

individuals to exhibit specific personality traits relevant to the new

unemployed situation, in line with Mischel and Shoda’s (1995;

Shoda & Mischel, 1998) definition of personality. Further, and

consistent with Fleeson (2001, 2004), the changes to an individu-

al’s life brought about by the experience of unemployment would

result in different mean levels of personality expression. It is also

reasonable to expect that the unemployment experience will per-

meate the individual’s life and help to instigate behavior change,

even within situations associated weakly with the work environ-

ment (e.g., during leisure activities or home stay). In all cases, the

unemployment experience is likely to give rise to stably different

ways of thinking, feeling, and behaving, which will precipitate

changes in personality.

992

BOYCE, WOOD, DALY, AND SEDIKIDES

The Psychological Effects of Unemployment

Unemployment has one of the strongest impacts on well-being

(d ⫽⫺0.38, McKee-Ryan, Song, Wanberg, & Kinicki, 2005),

with the impact often lasting beyond the period of unemployment

(Clark, Diener, Georgellis & Lucas, 2008; Clark, Georgellis, &

Sanfey, 2001; Daly & Delaney, 2013) and being comparable with

that of becoming disabled (Boyce & Wood, 2011b; Lucas, 2007)

or losing a spouse (Oswald & Powdthavee, 2008). However, much

less is known about how unemployment might shape personality.

The experience of unemployment is likely to bring considerable

and unexpected contextual fluctuation to an individual’s life, and,

potentially, to compromise the development of particular person-

ality traits. In accord with this notion, personality change has been

linked to other workplace variables (e.g., job satisfaction or status)

and counterproductive work behaviors (e.g., Roberts, 1997; Rob-

erts & Bogg, 2004; Roberts et al., 2003; Roberts, Walton, Bogg, &

Caspi, 2006; Scollon & Diener, 2006). Given that personality

maturates in normative ways across the life span (Lucas & Don-

nellan, 2011; Roberts et al., 2006a), we expect some change to take

place across the whole sample. However, we are specifically

interested in whether greater personality change occurs for those

who become unemployed. As such, we examine personality

change of the unemployed relative to the employed. Although

theorizing on how personality might change is not in found in

abundance, we build on this theory to offer several hypotheses

concerning whether and how the personality of the unemployed

(vs. employed) will change, while also ascertaining, where possi-

ble, precise forms of change (Pitariu & Ployhart, 2010).

Conscientiousness

Conscientiousness, which represents a tendency for individuals

to be goal focused (Barrick, Mount, & Strauss, 1993) and highly

motivated (Judge & Ilies, 2002), bears links with achievements

within the work environment. Hence, the experience of unemploy-

ment may curtail opportunities to express conscientious-type be-

havior. Conscientiousness is also positively linked to one’s eco-

nomic situation, such as wealth accumulation (Ameriks, Caplin, &

Leahy, 2003) or higher wages (Mueller & Plug, 2006; Nyhus &

Pons, 2005), and predicts fluctuations in life satisfaction following

income changes (Boyce & Wood, 2011a). Unemployment, then,

may cut off access to previously valued achievement goals, and

this may act as a catalyst for personality change. Consistent with

the theoretical expectation that unemployment will precipitate

changes in conscientiousness, both retirement and first-time entry

into employment have been associated with changes, negative and

positive, respectively, in conscientiousness (Specht et al., 2011).

Further, being in paid work has been linked with changes in

conscientiousness-related traits, such as increased social responsi-

bility (Roberts & Bogg, 2004). As a result of critical role of

conscientiousness in the workplace, we hypothesize that levels of

conscientiousness will be influenced by unemployment.

Hypothesis 1 (H1): The experience of unemployment (relative

to employment) will produce mean-level reductions in

conscientiousness.

Neuroticism

Unemployment may have an influence on neuroticism. Unem-

ployment is associated with high levels of stress (Frost & Clayson,

1991) and depression (Dooley, Prause, & Ham-Rowbottom, 2000).

Given that neuroticism entails stress and depression at the dispo-

sitional level (Widiger, 2009), it is likely that unemployment will

prompt higher neuroticism. Additionally, the work environment

provides a vital source of social support, which may dissipate

following unemployment (Atkinson, Liem, & Liem, 1986). Lack

of social support may result in loneliness (Heinrich & Gullone,

2006) and low self-esteem (Waters & Moore, 2002). In turn, lack

of social support and low self-esteem engender negative emotions,

cognitions, and behaviors (Cohen, Gottlieb, & Underwood, 2000;

Sedikides & Gregg, 2003). We therefore hypothesize that unem-

ployment will have a negative influence on neuroticism.

Hypothesis 2 (H2): The experience of unemployment (relative

to employment) will produce mean-level increases in

neuroticism.

Agreeableness, Extraversion, and Openness

Work, like many normative life events, can have a crucial

socialization influence (Roberts, 1997). The ability to interact

socially, convey ideas, and make compromises are typical aspects

of day-to-day activities within the workplace (Cohen et al., 2000).

Hence, the experience of unemployment may thwart the expres-

sion of socially oriented personality traits. However, given that

unemployment presents both new threats and new opportunities, it

is not entirely clear how unemployment might influence traits like

agreeableness, extraversion, and openness. For example, unem-

ployment may result in new social engagements. Contrastingly,

however, unemployed individuals may have fewer financial re-

sources, but more time to share with others. On a similar note,

openness may increase, as unemployment offers individuals the

opportunity to evaluate their lives and refocus on less material

outcomes (e.g., deepening relationships, appreciating aesthetics).

At the same time, unemployment could constrain the individual’s

ability for novel experiences (e.g., restaurant eating, travel) and

even beget perceptions of the world as distasteful and unfriendly.

As such, we do expect agreeableness, extraversion, and openness

to be influenced by unemployment, but we are uncertain of the

precise direction of influence; consequently, we adopt an explor-

atory approach.

Hypothesis 3 (H3): The experience of unemployment (relative

to employment) will produce mean-level changes in agree-

ableness (H3a), extraversion (H3b), and openness (H3c).

Influence of Unemployment on Personality as a

Function of Time Remaining Unemployed

Consistent with our earlier definitional considerations of per-

sonality (Fleeson, 2001, 2004; Mischel & Shoda, 1995; Shoda &

Mischel, 1998)—culminating in the conclusion that unemploy-

ment may give rise to stably different ways of thinking, feeling,

and behaving—we would expect the duration of unemployment

and whether reemployment took place to be differentially critical

for personality change. Distinctly stable ways of thinking, feeling,

993

UNEMPLOYMENT AND PERSONALITY

and behaving may prevail at various stages of the unemployment

experience. Personality change may therefore differ according to

whether individuals are short-term unemployed compared with

those who are long-term unemployed or transitioning between

short-term and long-term unemployment. For example, individuals

may be initially subject to personality change as they actively

search for new employment, but, after several years of failed

searches, may experience lack of motivation to continue pursuing

job leads (Kanfer, Wanberg, & Kantrowitz, 2001). This motiva-

tional burnout may still spark personality change, albeit different

from that of the initial “search” years. We therefore expect that the

impact of unemployment will depend on the number of years spent

unemployed and may develop in a nonlinearly fashion such that

larger changes will occur at various stages of unemployment. For

example, in the first year or two of unemployment, large person-

ality change may be evident, whereas in subsequent years, person-

ality may be stabilized at the newly formed level. Alternatively,

after a year or two of being out of work, individuals may learn to

engage more productively with the unemployment process, thus

being able to mitigate the initial personality change.

Hypothesis 4 (H4): The magnitude of the mean-level changes

in personality resulting from unemployment will be dependent

on the number of years that an individual has been unem-

ployed, such that a linear or nonlinear relation will be ob-

served between individuals at different years of unemploy-

ment and changes in their agreeableness (H4a),

conscientiousness (H4b), extraversion (H4c), neuroticism

(H4d), and openness (H4e).

Influence of Unemployment on Men’s and

Women’s Personalities

Unemployment may also have distinct personality implications

for men and women, owing to variability in thinking, feeling, and

behaving following the event. Different personality traits are val-

ued in the workplace for men and women; for example, agreeable-

ness is likely to be penalized in men but rewarded in women

(Mueller & Plug, 2006; Nyhus & Pons, 2005). Thus, to the extent

that individuals develop certain personality traits to achieve greater

workplace success, the absence of work may differentially disin-

centivize behavior patterns in the two genders. Further, men and

women may experience and cope with unemployment dissimilarly.

For example, men adopt a problem-focused orientation, and hence

are unlikely to seek social support, whereas women are symptom-

focused, and hence are likely to seek social support (Leana &

Feldman, 1991). As such, men may engage with the job search

process, whereas women may engage in socially oriented activities

(Kanfer et al., 2001). In addition, unemployment may present a

unique set of opportunities and threats across men and women that

vary according to the years spent unemployed. Some authors

(Forret, Sullivan, & Mainiero, 2010), for example, have speculated

that traditional gender roles could still be relevant to the experi-

ence of unemployment, with men viewing the experience as a

threat to their provider role and women viewing the experience as

a potential opportunity for child rearing. Hence, although we are

not in a confident position to ascertain precise patterns, we expect

gender differences in the way unemployment alters personality.

Hypothesis 5a (H5a): Men and women will exhibit different

mean-level changes in personality as a result of unemploy-

ment (relative to employment).

Hypothesis 5b (H5b): Men and women will exhibit different

nonlinear relations between years spent unemployed and per-

sonality change, such that the magnitude of the mean-level

changes in personality will vary differently for men and

women by the years they spend unemployed.

Unemployed-Triggered Personality Change and

Rebound Following Reemployment

Given that we anticipate unemployment to influence personality

change via the opportunity to express relevant traits, it is possible

that unemployment’s “impact” on personality will not be enduring.

Once an individual regains employment, the dynamic processes

that brought personality change in the first place will no longer

operate in the same way. Thus, within the new context of employ-

ment, further change may take place. However, because reemploy-

ment represents an absence of the unemployment context that

created change in the first place, it is possible that the reemploy-

ment context will foster psychological processes that result in

further change and may even return to preunemployment person-

ality levels. Personality change, then, may not be apparent in those

individuals who, although experiencing unemployment, subse-

quently become reemployed.

Hypothesis 6 (H6): Becoming reemployed will produce addi-

tional mean-level changes in personality (relative to remaining

employed).

Overview

Despite the strong theoretical case for expecting personality

change to accompany the experience of unemployment, there is a

dearth of relevant evidence. This is particularly surprising, given

that personality change has been linked to other momentous labor

market events, such as retirement or first-time entry into the labor

market (Specht et al., 2011). However, examining the influence of

unemployment on personality is methodologically much more

difficult than examining the influence of many other life events.

The latter events generally endure once they have occurred. For

example, individuals can enter the labor force for the first time

only once, and they typically enter retirement only once at the end

of their careers. As such, the influence on personality of starting

one’s first job or retiring can be determined by establishing

whether these events took place between two time points in which

personality was assessed. However, this is not the case with

unemployment, where individuals may enter into and out of un-

employment on multiple occasions and for varying temporal pe-

riods. Any results based simply on whether individuals experi-

enced some unemployment over the study period would be

confounded by potentially large subsets of those who had already

become reemployed, had experienced repeated periods of unstable

employment, or were experiencing long-term unemployment.

In this study, we therefore focus exclusively on unemployment,

a major non-normative life event, and differentiate between types

of unemployment experiences: becoming and remaining unem-

ployed versus becoming unemployed but being reemployed. We

994

BOYCE, WOOD, DALY, AND SEDIKIDES

also explore personality change differences by time spent in con-

secutive years of unemployment. We analyze longitudinal re-

sponses to questionnaires from a large sample in which all partic-

ipants were initially in employment. Participants completed

measures of personality at the first time point while in employment

and again 4 years later. We identify three subsets of participants:

those who became unemployed at various points over the 4-year

period and remained so until the end of the study, those who

became unemployed at some point over the study but regained

employment by the end of the study, and those who were in

employment in every time point in the study. After testing for

measurement invariance across the two personality-assessment

time points for each of the FFM traits, we use a latent change

model to compare relative differences in changes in the FFM traits

between these participants. We examine whether any impact of

unemployment on personality change (a) depends on how long

participants have been unemployed, (b) differs across men and

women, and (c) endures following reemployment.

Method

Participants and Procedure

We used the German Socio-Economic Panel Study (SOEP), an

ongoing longitudinal study of German households. The SOEP

began in 1984 with a sample of adult members from randomly

selected households in West Germany. Since 1984, the SOEP has

expanded to include East Germany and also added various sub-

samples to maintain a representative sample of the entire German

population (Wagner, Frick, & Schupp, 2007). The SOEP is one

of the primary socioeconomic data sets with which hundreds of

articles have been published.

The authors have used portions of

SOEP to answer different research questions in the following

published research articles: Boyce (2010); Boyce and Wood

(2011a, 2011b); and Boyce, Wood, and Brown (2010).

We focused on a subsample of SOEP participants who answered

questions on their personality in 2005 while still employed. The

employment status of these participants was recorded over for 4

years (2006, 2007, 2008, and 2009), and their personality was

assessed again in 2009. Our subsample consisted of 6,769 partic-

ipants (3,733 males, 3,036 females). Of these participants, 6,308

remained employed throughout this period (2005 to 2009). In an

effort to conduct a clean test of the effect of unemployment, we

separated the remaining 461 participants into two different groups:

those who experienced some unemployment but were reemployed

by 2009 (n ⫽ 251), and those who (a) had begun a phase of

unemployment between 2006 and 2009, and (b) were still in the

same phase of unemployment in 2009 (n ⫽ 210). Persons who

entered and exited multiple unemployment spells over this period,

yet were found to be unemployed in 2009, were excluded from our

subsample.

In all, our subsample comprised 6,308 individuals who re-

mained employed, 251 individuals who were unemployed but

became reemployed, and 210 individuals who were unemployed in

2009 for 1, 2, 3, or 4 years. One hundred seventeen of these 210

individuals had been unemployed for 1 year (their first year of

unemployment began in 2009), 41 had been unemployed for 2

years (their first year of unemployment began in 2008), 19 had

been unemployed for 3 years (their first year of unemployment

began in 2007), and 33 had been unemployed for 4 years (their first

year of unemployment began in 2006). In 2005, when all individ-

uals were employed, ages ranged from 17 to 61 years (M ⫽ 41.41,

SD ⫽ 10.45), and household income varied from €200 to €30,000

per month (M ⫽ 3107.53, SD ⫽ 1689.38, Mdn ⫽ 2786.09). Table

1 provides the means and standard deviations of the personality

variables at both time points by employment status. Table 2

provides the correlations of the personality variables, unemploy-

ment variables, and sociodemographic characteristics.

Measures

Employment status. Participants’ current employment status

is recorded in the SOEP as either in employment, retired, not

employed, in education, or unemployed. Given that we were

interested specifically in entry to unemployment from employ-

ment, we concentrate only on individuals who were recorded as

employed or unemployed throughout the study period. The not-

employed category included the subcategory of those who were

unemployed but also not looking for work. This subcategory,

though, would reflect inaccurately “individuals not in work but

wanting to work” (i.e., the unemployed), and thus we excluded

such participants from our analysis.

FFM personality measures. A shortened version of the Big

Five Inventory (Benet-Martínez & John, 1998) was administered

in both 2005 and 2009. This version, shown in the Appendix, was

developed specifically for use in the SOEP, in which space for

survey questions is severely limited (Gerlitz & Schupp, 2005).

Participants responded to 15 items (on a scale from 1 ⫽ does not

apply to me at all to 7 ⫽ applies to me perfectly scale), with three

items assessing each of the FFM domains of agreeableness (e.g.,

“has a forgiving nature”), conscientiousness (e.g., “does a thor-

ough job”), extraversion (e.g., “is communicative, talkative”), neu-

roticism (e.g., “worries a lot”), and openness (e.g., “is original,

comes up with new ideas”). The SOEP scale has comparable

psychometric properties to longer FFM scales. For example, using

different assessment methods, Lang, John, Lüdtke, Schupp, and

Wagner (2011) showed that the short-item scale produces a robust

five-factor structure across all age groups. Donnellan and Lucas

(2008) demonstrated that each of the scales contained in the SOEP

correlates highly (at least r ⫽ .88) with the corresponding subscale

of the full Big Five Inventory. In addition, Lang (2005) illustrated

that the retest reliability of the scale across 6 weeks is acceptable

(at least r ⫽ .75). In our sample in 2005 (2009), each of the

personality traits had the following Cronbach’s alphas: agreeable-

ness ⫽ .52 (.57); conscientiousness ⫽ .60 (.57); extraversion ⫽ .66

(.67); neuroticism ⫽ .61 (.59); openness ⫽ .60 (.63). After testing

for measurement invariance, we analyzed the FFM personality

variables as latent variables.

Gender. We used a binary variable (female) to denote

whether a participant was recorded as male (female ⫽ 0) or female

(female ⫽ 1). This variable was included as a main-effect variable

and also interacted with all of the unemployment variables to

establish whether there were gender differences in personality

change as a result of unemployment.

A list of publications using the SOEP may be found at http://www.diw

.de/en/diw_02.c.221182.en/publications_with_soep_data.html

995

UNEMPLOYMENT AND PERSONALITY

Covariates. We controlled for several third variables that

might account for the relation between unemployment and changes

to personality. Various life events may co-occur with unemploy-

ment (e.g., alterations to marital, disability, parental status), and,

given that such events have also been linked to changes in per-

sonality (Specht et al., 2011), any apparent effect of unemploy-

ment on these two variables may be because of the occurrence of

these events rather than the specific experience of unemployment.

To rule out this possibility, we controlled for alterations in events that

occurred between the two time points. Further, personality may have

different natural maturation rates by age or years of education (Lucas

& Donnellan, 2011), and these same factors may also be associated

with an increased likelihood of unemployment. We therefore also

controlled for age and education to rule out personality changes that

may have arisen from natural maturation rather than environment

variation. In some cases, there were missing values for education.

Given that this variable was not our main interest, and to avoid

excluding relevant participants, we recoded any missing values with

the full SOEP sample-wide means. We further included dummy

variables to indicate that a variable with a previously missing value

had been recoded with a sample-wide average. This practice ensured

that the inputted values had no effect on subsequent results.

Data Analytic Strategy

We used a latent change model to test whether unemployment

(including time spent unemployed and reemployment), compared

with lack of unemployment (i.e., employment), resulted in differ-

ences in mean-level personality change over a 4-year period (see

Figure 1). The latent change model assesses changes in unobserved

variables, so that both structural relations and measurement error

can be estimated simultaneously. The latent change model is based

on several testable assumptions (Allemand, Zimprich, & Hertzog,

2007). First, the relevant indicators (i.e., personality questionnaire

items linked to a specific trait) must load onto the latent factor of

interest at both time points (T0 and T1). Second, the extent to

which indicators load onto a latent factor must not vary over time.

This ensures that longitudinal change in the construct of interest

cannot be attributed to differences in how specific manifest indi-

cators link to latent variables from one testing occasion to another.

This assumption is tested by carrying out a set of analyses to

determine whether the measurements are factorially invariant.

Testing for Measurement Invariance

We took several steps to ensure that each personality trait

measure operated equivalently in 2005 (T0) and 2009 (T1), so that

observed changes in personality traits could be attributed to true

Table 1

Means and Standard Deviations of Aggregated Personality Variables at Each Time Point by Employment Status Across the Study

Variable

Agre. at T0 Agre. at T1 Cons. at T0 Cons. at T1 Extr. at T0 Extr. at T1 Neur. at T0 Neur. at T1 Open. at T0 Open. at T1

M SD M SD M SD M SD M SD M SD M SD M SD M SD M SD

Never unemployed (n ⫽ 6,308) 16.2 2.87 15.8 2.91 18.0 2.56 17.8 2.59 14.7 3.40 14.4 3.40 11.4 3.58 11.1 3.56 13.6 3.44 13.2 3.49

Unemployed for 1 year at T1 (n⫽ 117) 16.1 3.05 15.8 2.42 17.9 2.33 17.9 2.49 14.2 3.61 14.3 3.68 11.9 3.29 11.8 3.62 13.2 3.67 12.5 3.70

Unemployed for 2 years at T1 (n ⫽ 41) 15.7 4.01 16.5 2.73 18.2 2.61 17.4 2.45 14.4 3.43 14.0 3.70 11.7 4.13 12.1 3.59 12.6 4.95 13.0 4.26

Unemployed for 3 years at T1 (n ⫽ 19) 14.3 3.84 15.9 2.33 15.5 3.69 14.5 3.33 13.6 2.72 12.9 3.71 13.5 2.57 12.5 4.20 11.7 3.52 12.1 4.29

Unemployed for 4 years at T1 (n ⫽ 33) 16.1 3.26 14.7 3.29 18.6 2.57 16.7 2.95 13.8 3.33 13.1 2.83 12.9 3.42 12.6 3.94 13.2 3.36 11.2 2.75

Reemployed at T1 (n ⫽ 251) 15.3 3.27 15.3 2.82 17.6 2.61 17.5 2.77 14.6 3.72 14.1 3.66 11.4 3.62 10.8 3.55 13.6 3.50 12.9 3.08

Note. The scores for each of the five-factor model personality traits represent aggregated scores of the respective 3 items, following appropriate reverse coding as shown in the Appendix; N ⫽ 6,769.

T0 ⫽ the 1st time-point (2005); T1 ⫽ the 2nd time-point (2009); Agre. ⫽ Agreeableness; Cons. ⫽ Conscientiousness; Extr. ⫽ Extraversion; Neur. ⫽ neuroticism; Open. ⫽ openness.

We also note that changes to an individual’s income may precipitate

personality change (Powdthavee, Boyce, & Wood, 2011), but because the

losses in income that accompany unemployment are part of the causal

effect, we did not include income changes as a covariate in the main

analysis. Nevertheless, the inclusion of income change as a covariate did

not change our substantive conclusions, and there was little evidence that

the income change was related to personality change in the unemployed.

This finding is consistent with findings that the detrimental effect to

well-being goes far beyond the simple loss in income (Clark & Oswald,

2002).

996

BOYCE, WOOD, DALY, AND SEDIKIDES

differences or changes rather than changes in the psychometric

properties of the indicators over time. We examined the measure-

ment invariance of each of the Big Five Inventory traits, as these

models formed the basis for all subsequent latent change models.

Each trait was measured by three items at each time point. We

followed recommended practice for testing measurement invari-

ance, which suggests that it is necessary to establish configural,

metric, and scalar invariance prior to testing for latent change

(Bashkov & Finney, 2013; Vandenberg & Lance, 2000). To do

this, we (a) implemented a longitudinal confirmatory factor anal-

ysis model that systematically places a series of increasingly

restrictive equality constraints on specific parameters, and (b)

examined the impact of these restrictions on model fit.

We first identified a common model for each of the personality

measures across time points, which we derived from three indica-

tor variables for each of the five traits (Benet-Martínez & John,

1998). This configural model estimated the factor structure of the

traits across the two time points without placing any equality

constraints on the model. If the configural model demonstrates a

high level of fit, this will suggest that the basic factor structure with

the same pattern of fixed and freed loadings is invariant across

measurement occasions (Vandenberg & Lance, 2000). When config-

ural invariance is established, the configural model can be used as the

baseline model from which to evaluate changes in model fit associ-

ated with implementing tests of metric and scalar invariance. We

implemented several indices to gauge goodness of fit in order to test

the model data fit for each personality trait.

The chi-square value quantifies the extent to which sample and

fitted covariance matrices diverge, with a substantial discrepancy

suggesting a lack of fit and resulting in a large chi-square value.

The chi-square index of fit test is sensitive to sample size and

rejects the model in most instances in which large samples (N ⱖ

500) are used (Hayduk, 1988). The comparative fit index (CFI)

considers sample size and compares the fit of the hypothesized

model to a null model that assumes the included variables are

uncorrelated. CFI values range from 0 to 1, with values of ⱖ.95

recognized as indicative of good model data fit (Hu & Bentler,

1999). We also consider the root mean square error of approxi-

mation (RMSEA), which addresses how well a model “with un-

known but optimally chosen parameter values, [would] fit the

population covariance matrix if it were available” (Browne &

Cudeck, 1993). RMSEA values range from 0 to 1, with values

of ⱕ.06 indicating a good model fit (Hooper, Coughlan, & Mullen,

2008; Hu & Bentler, 1999).

We took a multistep approach, examining the change in model

fit that resulted from placing a logically ordered series of addi-

tional constraints on the initial configural or baseline model. Full

invariance was deemed to be supported when placing additional

constraints on the model did not produce a substantial change in

model fit. To evaluate whether a substantial change in model fit

occurred as a result of imposing additional equality constraints on

particular parameters, we examined the chi-square difference test

(⌬␹

), the CFI change (⌬CFI), the RMSEA goodness-of-fit statis-

tic, and the degree of overlap in RMSEA confidence intervals

between models. A nonsignificant chi-square difference test and a

small ⌬CFI (in which a decrease is no greater than .01) are

considered indicative of invariance (Cheung & Rensvold, 2002).

We contrasted the model fit of the configural or baseline model

with the fit associated with the measurement model, in which we

constrained factor loadings to be equal across measurement occa-

sions. This practice allowed a test of metric invariance or the

hypothesis that measurement weights are invariant between the

2005 and 2009 personality trait assessments. We then conducted a

test of scalar invariance by constraining the intercepts of the

manifest indicators on latent variables to be equal across measure-

ment occasions and by examining the change in model fit associ-

Table 2

A Correlation Matrix Showing the Association Between Aggregated Personality Variables, Unemployment Variables, and Key

Sociodemographic Characteristics

23456 7 8 9 10 11 12 13 14 15 16 17

1. Agre. at T0 .53

ⴱⴱ

.30

ⴱⴱ

.18

ⴱⴱ

.11

ⴱⴱ

.06

ⴱⴱ

⫺.13

ⴱⴱ

⫺.08

ⴱⴱ

.14

ⴱⴱ

.09

ⴱⴱ

⫺.00 .01 .01 ⫺.02 .15

ⴱⴱ

.02 .01

2. Agre. at T1 — .17

ⴱⴱ

.27

ⴱⴱ

.07

ⴱⴱ

.10

ⴱⴱ

⫺.08

ⴱⴱ

⫺.12

ⴱⴱ

.12

ⴱⴱ

.15

ⴱⴱ

.00 .00 ⫺.00 ⫺.01 .16

ⴱⴱ

.03

ⴱ

.02

3. Cons. at T0 — .52

ⴱⴱ

.19

ⴱⴱ

.12

ⴱⴱ

⫺.11

ⴱⴱ

⫺.07

ⴱⴱ

.17

ⴱⴱ

.07

ⴱⴱ

⫺.01 ⫺.01 ⫺.00 ⫺.01 .07

ⴱⴱ

.13

ⴱⴱ

⫺.10

ⴱⴱ

4. Cons. at T1 — .13

ⴱⴱ

.19

ⴱⴱ

⫺.06

ⴱⴱ

⫺.11

ⴱⴱ

.12

ⴱⴱ

.13

ⴱⴱ

⫺.02 ⫺.03 ⫺.04

ⴱⴱ

⫺.02 .10

ⴱⴱ

.08

ⴱⴱ

⫺.09

ⴱⴱ

5. Extr. at T0 — .63

ⴱⴱ

⫺.15

ⴱⴱ

⫺.11

ⴱⴱ

.36

ⴱⴱ

.26

ⴱⴱ

⫺.02 ⫺.01 .00 .02 .14

ⴱⴱ

⫺.07

ⴱⴱ

⫺.01

6. Extr. at T1 — ⫺.10

ⴱⴱ

⫺.16

ⴱⴱ

.27

ⴱⴱ

.34

ⴱⴱ

⫺.01 ⫺.01 ⫺.01 .01 .13

ⴱⴱ

⫺.08

ⴱⴱ

⫺.02

7. Neur. at T0 — .57

ⴱⴱ

⫺.07

ⴱⴱ

⫺.04

ⴱⴱ

.04

ⴱⴱ

.04

ⴱⴱ

.04

ⴱⴱ

.03

ⴱ

.19

ⴱⴱ

.03

ⴱⴱ

⫺.09

ⴱⴱ

8. Neur. at T1 — ⫺.04

ⴱⴱ

⫺.03

ⴱⴱ

.05

ⴱⴱ

.05

ⴱⴱ

.04

ⴱⴱ

.02

ⴱ

.18

ⴱⴱ

.02 ⫺.07

ⴱⴱ

9. Open. at T0 — .59

ⴱⴱ

⫺.03

ⴱ

⫺.03

ⴱ

⫺.02 .03

ⴱ

.08

ⴱⴱ

.03

ⴱ

.14

ⴱⴱ

10. Open. at T1 — ⫺.05

ⴱⴱ

⫺.05

ⴱⴱ

⫺.04

ⴱⴱ

.01 .08

ⴱⴱ

.04

ⴱⴱ

.18

ⴱⴱ

11. Unemp. at T0 — .85

ⴱⴱ

.65

ⴱⴱ

⫺.04

ⴱⴱ

⫺.01 0.01 ⫺.06

ⴱⴱ

12. Years Unemp. at T1 — .95

ⴱⴱ

⫺.03

ⴱ

⫺.01 .02 ⫺.05

ⴱⴱ

13. Years Unemp. at T1

(squared) — ⫺.02 ⫺.00 .02 ⫺.04

ⴱⴱ

14. Reemployed at T1 — .00 ⫺.10

ⴱⴱ

⫺.06

ⴱⴱ

15. Female — .01 ⫺.01

16. Age — .08

ⴱⴱ

17. Education (years) —

Note. The scores for each of the five-factor model personality traits represent aggregated scores of the respective three items, following appropriate reverse

coding as shown in the Appendix; N ⫽ 6,769. T0 ⫽ the 1st time-point (2005); T1 ⫽ the 2nd time-point (2009); Unemp. ⫽ Unemployed; Agre. ⫽

Agreeableness; ⫽ Cons. ⫽ Conscientiousness; Extr. ⫽ Extraversion; Neur. ⫽ neuroticism; Open. ⫽ openness.

ⴱ

p ⬍ .05.

ⴱⴱ

p ⬍ .01.

997

UNEMPLOYMENT AND PERSONALITY

ated with placing these additional equality constraints on the model. In

the event of possible lack of scalar invariance for any of the person-

ality measures, we examined the modification indices in order to

determine which items resulted in fit decrement in the scalar invari-

ance models. We then allowed these items to vary freely and reex-

amined model fit as a test of partial scalar invariance.

Latent Change Models

To assess the extent to which mean-level personality change

took place as a function of unemployment, we constructed a latent

change model (Allemand et al., 2007; McArdle, 2009), as depicted

in Figure 1, for each personality trait separately. We explored the

effect of unemployment by including a dummy variable to indicate

that an individual was unemployed at T1 (“unemployed at T1”)

and by including variables to indicate the number of consecutive

years spent unemployed between 2005 and 2009 (years unem-

ployed; this variable took integer values ranging from 0 to 4), as

well as the quadratic of number of years spent unemployed (years

unemployed squared; this variable took integer values ranging

from 0 to 16). Inclusion of these variables enabled us to determine

0 11

T0 T1

i s

Years unemployed

Female x Unemployed at T1

Predictor 1

Unemployed at T1

Female

Female x Years unemployed

Predictor 2

Predictor k

Female x Years unemployed

squared

Female x Reemployed at T1

Reemployed at T1

Years unemployed

squared

Figure 1. Latent change model for analyzing the effects of unemployment on personality, accounting for initial

levels (i - intercept) and differences in personality change (s - slope). Each of the FFM personality traits were

analyzed separately with each trait being measured at both T0 (1st time-point - 2005) and T1 (2nd time-point

- 2009). with the item measurement residuals allowed to correlate over time. Factor loadings (a and b) and

measurement intercepts of the three items for each of the latent traits were constrained to be equal across time

points. In all cases, age, years of education, dummy variables missing values for education and to indicate

changes to marital status, changes to disability status, and changes to parental status were included as additional

predictors (Predictors 1 to k). The effect of unemployment was explored by including a dummy variable to

indicate that an individual was unemployed at T1 (“unemployed at T1”), and variables to indicate the number

of consecutive years spent unemployed since becoming unemployed (“years unemployed”) and the quadratic of

the number of years spent unemployed (“years unemployed squared”). We further included a dummy variable

to indicate those individuals who had experienced unemployment but had regained employment by T1

(“reemployed at T1”), with the missing dummy being employed across all years of the study. To discern whether

there were important personality change differences between men and women, we additionally included a gender

dummy variable (female), which we interacted with all the unemployed variables. The results of all of these

models are contained in Table 5.

998

BOYCE, WOOD, DALY, AND SEDIKIDES

whether unemployment and years spent unemployed related to

personality. We further used a dummy variable to indicate those

participants who had experienced unemployment but had regained

employment by T1 (“reemployed at T1”). This would enable us to

establish whether personality change through unemployment re-

mains following reemployment. To discern whether there were

personality change differences between men and women, we ad-

ditionally interacted all the unemployed variables with the gender

variable (female), and included the gender dummy variable and

these interactions in each of our personality models.

All effects in the models were in relation to those experiencing

no unemployment, and therefore this group represented the miss-

ing dummy coded group in each of the analyses. We adjusted for

age, as well as for years of education, and any changes to partic-

ipants’ marital, disability, and parental statuses. The latent model

consisted of an intercept factor (i) and a latent slope factor (s). The

latent intercept factor reflected differences that already existed

between participants at the first time point (T0). Hence, any

significance on the intercept variable for predictor variables in

each of the models would suggest preexisting differences in per-

sonality before the commencement of unemployment. The latent

slope factor reflected differences in mean-level change between

participants from the first to the second time point (T1). If our

primary predictor variables in each model explained a significant

portion of variance in the slope variable, this would suggest that

the mean-level change in a given personality trait is contingent on

unemployment.

Significance on the slopes of any of the unemployed variables

(unemployed at T1, years unemployed at T1 [linear or squared])

would denote that unemployed participants experience mean-level

personality changes relative to those who remain employed. Spe-

cifically, a significant negative slope coefficient on the unem-

ployed at T1 dummy variable in the conscientious model would

support H1 that conscientious reduces following unemployment.

Similarly, a negative significant slope coefficient on the unem-

ployed at T1 dummy variable in the neuroticism model would

support H2—that neuroticism reduces following unemployment.

Given the uncertainty of the direction of change for agreeableness,

extraversion, and openness, significant slopes in any direction for

the remaining models on the unemployed at T1 dummy variable

would offer support for H3. Specifically, significance on the years

unemployed at T1 (either linear or squared) slopes for agreeable-

ness, conscientiousness, extraversion, neuroticism, or openness

would constitute support for H3a, H3b, H3c, H3d, and H3e,

respectively. Significant slopes on the years unemployed at T1

(either linear or squared) variables across each of the models

would suggest that personality change were dependent upon the

year of unemployment, and therefore offer support for H4. Spe-

cifically, significance on the agreeableness, conscientiousness, ex-

traversion, neuroticism, or openness years-unemployed-at-T1 (ei-

ther linear or squared) slopes would constitute support for H3a,

H3b, H3c, H3d, and H3e, respectively. Significance on the slopes

of any of the gender interaction variables in any of the models

would indicate gender differences in personality change follow-

ing unemployment, yielding support for H5. More specifically,

significance on any of the gender interactions with unemployed

at T1 would indicate support for H5a, whereas significance on

any of the gender interactions with either of the years unem-

ployed at T1 slopes would support H5b. Significance on any of

the slopes of the reemployed at T1 variable would indicate that

unemployment resulted in sustained changes to personality,

offering support to H6—that following reemployment person-

ality does not return to preunemployment levels. We estimated

all the models using AMOS 19.

Table 3

Fit Indices for Testing Measurement Invariance for the Personality Variables From 2005 (T0) to 2009 (T1)

Model ␹

df ⌬␹

⌬df CFI ⌬CFI RMSEA RMSEA 90% CI

Agreeableness

Configural invariance 48.0

ⴱ

5 — — .994 — .036 .027, .045

Metric invariance 49.0

ⴱ

7 1.0 2 .994 .000 .030 .022, .038

Scalar invariance 199.3

ⴱ

10 150.3

ⴱ

3 .973 ⫺.021 .053 .047, .059

Partial scalar invariance 151.0

ⴱ

9 102.0

ⴱ

2 .980 ⫺.014 .048 .042, .055

Conscientiousness

Configural invariance 36.9

ⴱ

5 — — .996 — .031 .022, .040

Metric invariance 48.6

ⴱ

7 11.7

ⴱ

2 .995 ⫺.001 .030 .022, .038

Scalar invariance 123.9

ⴱ

10 75.3

ⴱ

3 .987 ⫺.008 .041 .035, .048

Extraversion

Configural invariance 34.5

ⴱ

5 — — .997 — .030 .022, .040

Metric invariance 51.3

ⴱ

7 16.8

ⴱ

2 .996 ⫺.001 .031 .023, .039

Scalar invariance 170.1

ⴱ

10 118.8

ⴱ

3 .987 ⫺.009 .049 .042, .055

Neuroticism

Configural invariance 17.8

ⴱ

5 — — .999 — .019 .010, .030

Metric invariance 29.6

ⴱ

7 11.8

ⴱ

2 .997 ⫺.002 .022 .014, .030

Scalar invariance 386.6

ⴱ

10 209.7

ⴱ

3 .957 ⫺.040 .075 .068, .081

Partial scalar invariance 74.3

ⴱ

9 44.7

ⴱ

2 .993 ⫺.04 .033 .026, .040

Openness

Configural invariance 9.1 5 — — 1.000 — .011 .000, .022

Metric invariance 13.2 7 4.1 2 1.000 .000 .011 .000, .021

Scalar invariance 120.6

ⴱ

10 107.4

ⴱ

3 .988 ⫺.012 .040 .034, .047

Note. ⌬␹

⫽ chi-square difference; ⌬df ⫽ degrees of freedom difference; N ⫽ 6,769.

ⴱ

p ⬍ .01.

999

UNEMPLOYMENT AND PERSONALITY

Results

First, we tested for measurement invariance within each of the

latent personality traits across time points. We then examined,

using latent change models, the extent to which unemployment

influences mean-level change in all FFM personality traits.

Measurement Invariance Model

The configural model simultaneously estimated the baseline

factor models for each of the FFM personality traits for responses

from 2005 and 2009. The analysis of this model showed chi-square

values ranging from 9.1 to 40.8, CFI values ranging from .994 to

1.000, and RMSEA values ranging from .011 to .036 (see Table 3).

Taken together, these fit indices indicate that the configural model

fits the data very well, and that the basic model structure is

invariant across measurement occasions. Having established con-

figural invariance, we then tested for metric invariance by restrict-

ing factor loadings to be equal in 2005 and 2009. The metric

invariance model also fit the data very well, and imposing equality

constraints on the factor loadings produced little change in the chi

square (⌬␹

ranging from 1.0 to 16.8), or CFI values (⌬CFI

ranging from ⫺.002 to .000), thus suggesting that measurement

weights were invariant across time points (see Table 3).

Our final scalar invariance model showed that constraining the

intercepts of all corresponding items to be equal across measure-

ment occasions had some impact on the fit indices. This model

yielded evidence that the measurement intercepts for the person-

ality traits agreeableness (⌬␹

⫽ 209.7, p ⬍ .01; ⌬CFI ⫽⫺.040;

RMSEA ⫽ .075) and neuroticism (⌬␹

⫽ 150.3, p ⬍ .01;

⌬CFI ⫽⫺.021; RMSEA ⫽ .053) varied across time points.

Overall these fit indices represented a moderately well-fitting

model.

Restricting measurement intercepts to be equal produced a de-

cline in fit across the remaining personality traits (i.e., conscien-

tiousness, extraversion, openness; ⌬␹

ranging from 75.3 to 150.3;

⌬CFI ranging from ⫺.021 to ⫺.008), as shown in Table 3. How-

ever, the goodness-of-fit indices suggested that the model data fit

for these traits remained high (i.e., CFI ⱖ .95, RMSEA ⱕ .06), and

the RMSEA confidence intervals showed overlap in the majority

of cases, implying that the hypothesis of scalar invariance may not

be rejected. In all, tests of measurement invariance demonstrated

that the personality traits conscientiousness, extraversion, and

openness are characterized by full configural and metric invari-

ance, as well as a moderate degree of evidence of scalar invari-

ance. The results of these tests allowed us to proceed with the

latent change analysis.

Given that both agreeableness and neuroticism did not satisfy all

of the fit criteria, we explored the issue of scalar invariance in

these measures further. We examined modification indices (MIs)

to assess which items resulted in a fit decrement in the scalar

invariance models for both measures. For agreeableness, the

thresholds for the item “forgives” varied markedly over time

(MI ⫽ 32), and thus relaxing the equivalence constraint on this

item and allowing it to vary freely over time substantially im-

proved the model fit (⌬␹

⫽ 102.0, p ⬍ .01; ⌬CFI ⫽⫺.004;

RMSEA ⫽ .033). For neuroticism, the thresholds for the item

“worry a lot” varied markedly over time (MI ⫽ 214). Relaxing the

equivalence constraint on this item and allowing it to vary freely

over time substantially improved the model fit (⌬␹

⫽ 44.7, p ⬍

.01; ⌬CFI ⫽⫺.014; RMSEA ⫽ .048). In our main analysis, we

estimated latent change models assuming scalar invariance for

both agreeableness and neuroticism. However, the improvement in

model fit after relaxing the equivalence constraints suggests that a

partial invariance intercept model would be acceptable (Chun-

gkham, Ingre, Karasek, Westerlund, & Theorell, 2013). Thus, for

agreeableness and neuroticism, we also carried out latent change

models that allowed the items with the highest MI to vary freely so

that we could account for the partial scalar invariance. This en-

abled us to determine whether partial scalar invariance had a

substantive influence on our results.

Latent Change Models

First, we determined the unconditional means and variances for

both the intercept and slope terms, that is, estimates without any

controls or predictors for each of the FFM traits. As shown in

Table 4, all of the unconditional means and variance were signif-

icant for both the intercept and slope terms (except the slope term

on neuroticism).

We then examined the extent to which unemployment influ-

enced each of the FFM traits using latent change models (see Table

5). For each trait, we explored the effect of unemployment on

personality by (a) including a dummy variable in order to indicate

that a participant was unemployed at T1 (unemployed at T1), and

(b) including the number of consecutive years spent unemployed

(years unemployed) and the quadratic of the number of years spent

unemployed (years unemployed squared), to further test for un-

employment duration effects. We further included a dummy vari-

able to index those participants who had experienced unemploy-

ment but had regained employment by T1 (reemployed at T1). All

changes in the model were made relative to the missing dummy,

which represents participants employed across all years of the

study.

To discern gender differences, we additionally interacted all the

unemployed variables with our gender variable (female). Each

model included intercept and slope factors that we allowed to vary

according to all of the included predictors. The missing dummy

was the group who was employed across all years of the study,

with resulting intercept and slope coefficients always made rela-

tive to this group. We adjusted for age, years of education, and any

changes to participants’ marital, disability, and parental statuses.

All of the intercept terms were statistically significant (p ⬍ .01),

and an observation of the association between unemployment and

the intercept differences across all models in Table 5 suggests that,

prior to unemployment, there were very few personality differ-

ences between those who remained employed and those who

eventually became unemployed. Only with openness was there

evidence that initial levels for women predicted subsequent unem-

ployment. There was no evidence of a selection effect for the

remaining personality traits for either men or women.

We detail, in the portion of Table 5 titled “Slopes for each

personality trait,” the extent to which unemployment precipitated

There was some evidence to suggest that agreeableness and neuroti-

cism were not scalar invariant. Thus, we also carried out latent change

models that allowed the most problematic item in each of the scales

(highest MI) to vary freely. We found no substantive differences in our

results.

1000

BOYCE, WOOD, DALY, AND SEDIKIDES

mean-level changes in personality. Significance on the slope terms

for any of the unemployed variables, including the gender inter-

action terms, indicates that the mean-level personality changed

relative to the group of participants who remained employed. We

hypothesized that unemployment would produce mean-level de-

creases in conscientiousness (H1) and increases in neuroticism

(H2). We also hypothesized there would be mean-level changes in

an unspecified direction in agreeableness, extraversion, and open-

ness (H3a to H3c). Further, we hypothesized that these mean-level

changes would be dependent upon the year of unemployment (H4a

to H4e) and would differ by gender (H5a/H5b). Overall, we observed

mean-level changes across the sample in all of the FFM personality

traits except neuroticism. These effects were small and can be taken

to reflect the process of personality development that typically takes

place across the life span (Lucas & Donnellan, 2011; Roberts et al.,

2006a). Importantly, however, we observed significant slope effects

that are large in magnitude for the unemployed variables across

agreeableness (Model 1), conscientiousness (Model 2), and openness

(Model 5), suggesting that unemployment precipitated a substan-

tial amount of additional mean-level personality change relative to

the group of participants who remained employed. Our results lent

partial support to hypotheses H3, through the effect on openness

(H3c). Our results offer no support for H1 or H2. Mean-level

changes in both agreeableness and openness, however, depended

on the number of years unemployed (statistical significance on

years unemployed at T1 and/or years unemployed squared at T1),

giving partial support to H4, through H4a and H4e. As indicated

by the gender interaction terms, there were mean-level change

differences between men and women in both conscientiousness and

openness, as well as nonlinear differences between unemployment

duration and agreeableness, conscientiousness, and openness. Thus,

we obtained partial support for both H5a and H5b.

We present the

significant results separately for men and women in Figure 2 to

illustrate how the mean-level of personality would be expected to

change as participants experience a spell of unemployment of at

least 4 years. Each point estimate is derived from the mean-level

change in personality for each unemployed period according to

Table 5. The overall pattern therefore only reflects an implied

trajectory of personality change. Dashed lines represent the back-

ground change in personality of participants who remained em-

ployed throughout the study.

Unemployment’s influence on FFM personality traits. The

upper-left graph depicts the agreeableness results for men who

exhibit a significant trend, and the upper-right graph depicts the

results for women. For men, the graph implies that agreeableness

increases in the first 2 years of unemployment by approximately

0.25, whereas those who never became unemployed (represented

in the graph by the dashed line) experienced decreases in agree-

ableness of 0.05. Thus, the relative difference in change between

these two groups is approximately 0.30. However, because of the

nonlinear influence, after 2 years, agreeableness levels of the

unemployed men begin to diminish and, in the long run, are lower

than that of the group who remained in employment. For women,

we find large reductions in agreeableness with each year of un-

employment. Specifically, each additional year of unemployment

results in reductions in agreeableness, and, after 4 years of unem-

ployment, agreeableness is approximately 0.40 lower than before

unemployment.

The middle-left graph indicates that the longer men spent un-

employed, the larger their reduction in conscientiousness. After 4

years of unemployment, their conscientiousness dropped by more

than 0.60. Relative to the group who remained employed through-

out, this represents a change of approximately 0.50. For women in

the middle-right graph, there is evidence of a nonlinear influence

of unemployment on personality, with increases in the early and

late stages of unemployment but reductions in the medium term.

Relative to those in the employed group, who do not experience

any significant reductions in conscientiousness, unemployed

women in the second and third years of unemployment are ap-

proximately 0.40 lower in conscientiousness.

The bottom-left graph indicates how openness changes with

respect to the years spent unemployed in men, and the bottom-right

graph indicates that of women. Unemployed men manifest approx-

imately similar openness levels in the first year of unemployment,

but the results imply that openness may increase by approximately

0.10. However, with increasing years of unemployment, men start

to decrease in openness to more than 0.40 lower than they were

before unemployment. Women, on the other hand, show sharp

reductions in openness and are approximately 0.70 lower than their

preunemployment openness levels in the second and third years of

unemployment. However, in the fourth year, openness begins to

The full model results, including the coefficients from the covariates,

are available upon request.

Table 4

Unconditional Intercept and Slope Means and Variances in Personality From 2005 (T0) to 2009

(T1) Using Latent Change Models

Agreeableness Conscientiousness Extraversion Neuroticism Openness

Intercept b ⫽ 2.863 b ⫽ 6.312 b ⫽ 5.114 b ⫽ 3.408 b ⫽ 4.755

SE ⫽ .155

ⴱ

SE ⫽ .010

ⴱ

SE ⫽ .016

ⴱ

SE ⫽ .098

ⴱ

SE ⫽ .015

ⴱ

Intercept variance b ⫽ .220 b ⫽ .468 b ⫽ 1.151 b ⫽ 0.623 b ⫽ .682

SE ⫽ .016

ⴱ

SE ⫽ .018

ⴱ

SE ⫽ .039

ⴱ

SE ⫽ .032

ⴱ

SE ⫽ .030

ⴱ

Slope b ⫽⫺.063 b ⫽⫺.065 b ⫽⫺.138 b ⫽⫺.021 b ⫽⫺.117

SE ⫽ .007

ⴱ

SE ⫽ .010

ⴱ

SE ⫽ .013

ⴱ

SE ⫽ .012 SE ⫽ .012

ⴱ

Slope variance b ⫽ .127 b ⫽ .349 b ⫽ .573 b ⫽ .323 b ⫽ .368

SE ⫽ .011

ⴱ

SE ⫽ .017

ⴱ

SE ⫽ .026

ⴱ

SE ⫽ .020

ⴱ

SE ⫽ .022

ⴱ

Note. The models contain no predictor variables; N ⫽ 6,769.

ⴱ

p ⬍ .01.

1001

UNEMPLOYMENT AND PERSONALITY

Table 5

Effects of Consecutive Years of Unemployment on the Level (Intercept) and Change (Slope) in Personality From 2005 (T0) to 2009 (T1) Using Latent Change Models and

Controlling for the Occurrence of Other Nonrelated Life Events and Demographic Characteristics

Dependent variable at T1

Model 1 Model 2 Model 3 Model 4 Model 5

Agreeableness Conscientiousness Extraversion Neuroticism Openness

bSE ␤ bSE␤ bSE ␤ bSE ␤ bSE␤

Intercept 3.121 .137

ⴱⴱ

6.278 0.014

ⴱⴱ

4.946 .022

ⴱⴱ

3.484 .101

ⴱⴱ

4.708 .019

ⴱⴱ

Unemployed at T1 .037 .267 .013 .379 .373 .096 ⫺.730 .567 ⫺.119 ⫺.145 .471 ⫺.030 .708 .477 .150

Years unemployed at T1 ⫺.066 .284 ⫺.050 ⫺.404 .398 ⫺.221 .719 .605 .255 .271 .503 .121 ⫺.810 .509 ⫺.372

Years unemployed at T1 (squared) .013 .059 .033 .075 .083 .138 ⫺.146 .126 ⫺.172 ⫺.025 .104 ⫺.037 .156 .106 .239

Reemployed at T1 ⫺.032 .050 ⫺.012 ⫺.027 .070 ⫺.008 .033 .106 .006 .133 .088 .030 .147 .089 .034

Female .188 .016 .189

ⴱⴱ

.079 .020 .057

ⴱⴱ

.375 .031 .176

ⴱⴱ

.389 .027 .230

ⴱⴱ

.092 .026 .056

ⴱⴱ

Female ⫻ Unemployed at T1 ⫺.049 .417 ⫺.011 ⫺.770 .584 ⫺.129 .025 .887 .003 .105 .738 .014 ⫺1.545 .747 ⫺.218

ⴱ

Female ⫻ Years Unemployed at T1 .015 .452 .008 .816 .632 .297 ⫺.212 .960 ⫺.050 ⫺.245 .799 ⫺.073 1.698 .808 .519

ⴱ

Female ⫻ Years Unemployed at T1 (squared) .015 .093 .026 ⫺.171 .130 ⫺.210 .078 .198 .062 .027 .164 .028 ⫺.348 .166 ⫺.361

ⴱ

Female ⫻ Reemployed at T1 .028 .074 .007 .076 .104 .014 .039 .158 .005 .003 .131 .000 .131 .133 .021

Slope ⫺.089 .011

ⴱⴱ

⫺.094 0.014

ⴱⴱ

⫺.131 0.018

ⴱⴱ

⫺.082 .016

ⴱⴱ

⫺.113 .016

ⴱⴱ

Unemployed at T1 ⫺.538 .277 ⫺.250 ⫺.208 .386 ⫺.061 .687 .497 .159 ⫺.326 .446 ⫺.094 ⫺1.098 .440 ⫺.318

ⴱ

Years unemployed at T1 .784 .296 .792

ⴱⴱ

.204 .412 .129 ⫺.557 .530 ⫺.280 .356 .476 .223 1.267 .469 .797

ⴱⴱ

Years unemployed at T1 (squared) ⫺.171 .062 ⫺.574

ⴱⴱ

⫺.074 .086 ⫺.156 .089 .110 .148 ⫺.075 .099 ⫺.157 ⫺.276 .098 ⫺.579

ⴱⴱ

Reemployed at T1 .042 .052 .021 .010 .072 .003 ⫺.077 .093 ⫺.019 ⫺.073 .084 ⫺.023 ⫺.062 .082 ⫺.020

Female .017 .015 .023 .074 .021 .062

ⴱⴱ

⫺.012 .027 ⫺.008 ⫺.033 .024 ⫺.027 .007 .024 .006

Female ⫻ Unemployed at T1 .748 .433 .231 1.394 .604 .270

ⴱ

⫺.462 .778 ⫺.071 .293 .698 .056 1.586 .689 .306

ⴱ

Female ⫻ Years Unemployed at T1 ⫺.965 .469 ⫺.647

ⴱ

⫺1.414 .654 ⫺.596

ⴱ

.419 .842 .140 ⫺.002 .756 ⫺.001 ⫺2.045 .746 ⫺.855

ⴱⴱ

Female ⫻ Years Unemployed at T1 (squared) .185 .097 .422 .300 .135 .429

ⴱⴱ

⫺.068 .173 ⫺.076 ⫺.026 .155 ⫺.037 .428 .153 .607

ⴱⴱ

Female ⫻ Reemployed at T1 ⫺.039 .077 ⫺.013 ⫺.160 .107 ⫺.035 .046 .138 .008 .142 .124 .030 ⫺.027 .122 ⫺.006

␹

(degrees of freedom) 255.03 (85) 433.6 (85) 312.0 (85) 636.2 (85) 1094.3 (85)

CFI .998 .997 .998 .995 .990

RMSEA .017 .025 .020 .031 .042

Intercept R

.039 .038 .037 .072 .032

Intercept variance b ⫽ .236, SE ⫽ .015

ⴱⴱ

b ⫽ .454, SE ⫽ .017

ⴱⴱ

b ⫽ 1.086, SE ⫽ .037

ⴱⴱ

b ⫽ .655, SE ⫽ .032

ⴱⴱ

b ⫽ .648, SE ⫽ .029

ⴱⴱ

Slope R

.016 .023 .005 .013 .016

Slope variance b ⫽ .136, SE ⫽ .011

ⴱⴱ

b ⫽ .344, SE ⫽ .017

ⴱⴱ

b ⫽ .558, SE ⫽ .025

ⴱⴱ

b ⫽ .355, SE ⫽ .022

ⴱⴱ

b ⫽ .352, SE ⫽ .021

ⴱⴱ

Note. In each model, all coefficients are made relative to those employed across all years of the study, where Unemployed at T1 was coded 0 ⫽ not unemployed at T1, 1 ⫽ unemployed at T1; Years

unemployed at T1 was coded from 0 to 4; Years unemployed at T1 (squared) was coded from 0 to 16; Reemployed at T1 was coded 0 ⫽ not reemployed at T1, 1 ⫽ reemployed at T1; Female was

coded 0 ⫽ male, 1⫽ female. In all models (N ⫽ 6,769), additional controls for age and years of education were included, as well as dummy variables to control for changes to marital status, changes

to disability status, and changes to parental status. The full model results, including the coefficients from the covariates, are available upon request. All coefficients (b) indicate standardized effects

(SD ⫽ 1) on personality.

ⴱ

p ⬍ .05.

ⴱⴱ

p ⬍ .01.

1002

BOYCE, WOOD, DALY, AND SEDIKIDES

increase, such that those unemployed for 4 years decrease by only

0.30 relative to those who never experienced unemployment.

Reemployment’s influence on FFM traits. To test whether

personality endures once an individual regains employment (H6),

we examined the slope effect of the reemployed dummy variable.

Significance on the slope of this variable would indicate that the

unemployed who were reemployed by T1 experienced mean-level

changes to their personality relative to those who remained in

employment for all 4 years. An analysis of slope values on this

variable suggest no evidence that reemployed individuals experi-

Figure 2. Predicted mean-level change in personality across men and women across a 4-year unemployment

spell, based the results in agreeableness, conscientiousness, and openness from Table 5. The point estimate for

0 years of unemployment represents the average initial preunemployment value of those that became unem-

ployed across the study. Each subsequent point estimate is derived from the mean-level change in personality

for the group of individuals experiencing the associated number of years of unemployment in the study. The

overall pattern therefore reflects an implied trajectory of personality change during a 4-year unemployment spell.

Dashed lines, for men and women, respectively, represent the background change in personality of those that

remain employed throughout the study.

1003

UNEMPLOYMENT AND PERSONALITY

enced mean-level personality change relative to the employed

group. Although unemployment is likely to have precipitated per-

sonality change, the results suggest that personality rebounds upon

reemployment.

Discussion

We theorized that the experience of unemployment, a major

non-normative life event, would precipitate changes in personality

by giving rise to different ways of thinking, feeling, and behaving.

We aimed to examine whether, relative to individuals who remain

in employment, (a) unemployment advances change in basic per-

sonality traits, (b) the influence of unemployment on personality

depends on its duration, (c) this influence differs by gender, and

(d) personality change is sustained following reemployment. Pre-

vious work found no effect of unemployment on personality

(Specht et al., 2011), but this work did not address some of the

previously stated nuances in the unemployment experience.

We showed, in our study, that agreeableness, conscientiousness,

and openness changed during unemployment relative to employ-

ment, with the influence contingent upon the year of unemploy-

ment, gender, and reemployment. Across some of the FFM per-

sonality traits, we found mean-level changes of approximately .40

(and sometimes much higher) relative to individuals who remained

employed. Studies that used comparable methodologies to test the

influence of other major life events on personality change reported

much weaker effects. For example, across many life events, Specht

et al. (2011) found mean-level changes of approximately .10 to

.20, with their largest effect on divorce (.25). We also note that

unemployment, which has one of the strongest impacts on well-

being,

has an effect size of approximately d ⫽⫺.38 (McKee-

Ryan et al., 2005). When we consider the standardized effect sizes,

we obtain effects that range up to a full standard deviation change

in personality (e.g., d ⫽ 0.97 in conscientiousness for men unem-

ployed for 4 years; d ⬎ 0.70 in openness and agreeableness for

women unemployed for 3 years

). Although we offer some caution

as to drawing large inferences from small subsamples, the effect

sizes of unemployment on personality found here are compara-

tively large. We also note that we observed some changes in in the

group that remained employed. However, these effects were com-

paratively small and are likely to reflect the process of personality

development that typically takes place across the life span (Lucas

& Donnellan, 2011; Roberts et al., 2006a), or may have arisen

because of socioeconomic factors likely to influence everyone,

such as the global financial crisis. Importantly, the effects of

unemployment are observed in addition to these changes.

Evaluation of Support for Personality

Change Hypotheses

We made a number of hypotheses as to how unemployment

might be expected to precipitate personality change. Specifically

we hypothesized that the experience of unemployment (relative to

employment) would result in mean-level reductions in both con-

scientiousness (H1) and neuroticism (H2), as well as mean level

changes in agreeableness (H3a), extraversion (H3b), and openness

(H3c). We observed little evidence for any direct effects, with

openness being the only personality trait to respond directly as a

result of unemployment (H3c), thus only partially supporting H3.

However, there was evidence that the time spent unemployed

influenced mean levels of agreeableness (H4a), as well as open-

ness (H4e), partially supporting H4 that personality resulting from

unemployment is largely dependent upon the number of years that

an individual was unemployed. Further, we obtained partial sup-

port for the hypothesis that personality change following unem-

ployment differed by gender. There were mean-level change dif-

ferences between men and women in both conscientiousness and

openness thus offering partial support for H5a. In addition, there

were nonlinear differences between unemployment duration and

agreeableness, conscientiousness, and openness, giving partial

support for H5b. In men, for example, agreeableness increased for

those experiencing 2 or 3 years of unemployment, but decreased

for longer-term unemployment. In women, however, agreeable-

ness decreased at all stages of unemployment. These findings

highlight a critical role for both gender and unemployment dura-

tion in personality change following unemployment. This is con-

ducive to the idea that unemployment will generally create both

threats and opportunities, which will be more or less salient at

various stages of unemployment (short term vs. long term), and

will differ according to gender. In early unemployment stages,

there may be incentives for individuals to behave agreeably in an

effort to secure another job or placate those around them, but in

later years when the situation becomes endemic, such incentives

may weaken. Such tendencies may differ by gender according to

traditional work roles (Forret et al., 2010). Similarly, openness

reduced overall for both men and women, but the degree of

reduction by gender varied according to years spent unemployed,

perhaps reflecting differences in coping strategies (Leana & Feld-

man, 1991).

Although there was no significant effect of unemployment on

conscientiousness for men, Figure 2 suggests that a strong linear

trend may have been present. We therefore reestimated the con-

scientiousness model excluding the quadratic of the years unem-

ployed at T1 variable and obtained a significant effect on the linear

years unemployed at T1 variable. This suggests that the analysis

carried out for conscientiousness in Table 5 represents an overpa-

rameterization with respect to discerning a simple linear trend in

men, and thus lends support for the hypothesis that the longer an

individual is unemployed, the larger the reductions in conscien-

tiousness (H4b). Conscientiousness is important for success at

work (Barrick et al., 1993; Judge & Ilies, 2002; Mueller & Plug,

2006; Nyhus & Pons, 2005), but our results also seem to suggest

that work is important for high levels of conscientiousness. Un-

employed individuals may experience situational pressures to

gradually reduce their level of conscientiousness, as this practice

may constitute an adaptive way of coping with unemployment. For

example, unemployed individuals who are conscientious endure

the largest decreases in life satisfaction following unemployment

In a further analysis, we included changes in life satisfaction as an

additional control to determine whether changes in personality could be

viewed simply as a proxy for changes in life satisfaction. The results

remained significant.

To calculate these estimates, we divided the standardized coefficients

in Table 5 by the standard deviations of the appropriate unemployment

variables (unemployed dummy SD ⫽ 0.173, years unemployed SD ⫽

0.376, years unemployed squared SD ⫽ 1.253, Female ⫻ Unemployed

dummy SD ⫽ 0.115, Female ⫻ Years Unemployed SD ⫽ 0.250, Female ⫻

Years Unemployed squared SD ⫽ 0.848).

1004

BOYCE, WOOD, DALY, AND SEDIKIDES

(Boyce et al., 2010), and conscientiousness is related to enjoyment

of one’s own income which the unemployed lack (Ameriks et al.,

2003; Boyce & Wood, 2011a). These findings may be interpreted

as indicating that, in some ways, it is preferable to be less consci-

entious. However, conscientiousness is related to job search be-

havior and is therefore helpful in finding employment (Kanfer et

al., 2001), hinting to potentially conflicting situational pressures

that may perversely result in prolonged unemployment periods.

Our results additionally show mean-level differences in conscien-

tiousness change by gender in that, whereas both men and women

endured decreases, women regained some of their lost conscien-

tiousness levels in later years of unemployment. These regains

may reflect a greater ease to pursue non-work-related activities

congruent with traditional gender roles (e.g., caregiver; Forret et

al., 2010).

Because unemployment is likely to entail unsettling and stress-

promoting situations (Dooley et al., 2000; Frost & Clayson, 1991)

that contribute to loneliness and low self-esteem (Heinrich &

Gullone, 2006; Waters & Moore, 2002), it is surprising that we

observed no changes in neuroticism. Neuroticism is the personality

trait most strongly linked to well-being (Boyce et al., 2013; De-

Neve & Cooper, 1998; Steel, Schmidt, & Shultz, 2008), and so it

is likely that at least temporary changes in neuroticism would have

taken place. Such a finding, rather than being inconsistent, helps to

illustrate differences between relatively stable predispositions and

temporary shifts to well-being. Yet an alternative explanation is

that unemployment alleviates certain difficulties associated with

the workplace. As an aside, we note that neuroticism did not

“perform” well in the test for measurement invariance, raising

some doubts about the adequacy of the neuroticism scale that we

used.

Our results generally highlight the importance of unemployment

duration, which was overlooked in previous work (Specht et al.,

2011). Examining personality change exclusively in terms of

whether someone is unemployed or not may conceal the possibility

that some periods of unemployment are associated with increases,

whereas other periods are associated with decreases, in certain

traits. Our results call attention to unfair stigmatization as a con-

sequence of unemployment (Karren & Sherman, 2012). Stigma

can be attached to the unemployed by attributing to them certain

negative personality dispositions (McGarty, 2002). Our findings

indicate that, alternatively, the experience of unemployment itself

may create the personality types which would subsequently be

unfairly stigmatized against.

We further hypothesized that becoming reemployed would cul-

minate in additional mean-level changes in personality (H6).

examined this hypothesis by focusing on the group of individuals

who experienced some unemployment during the study, but be-

came reemployed by the second time point in which personality

was measured. However, an examination of the reemployed group

revealed no evidence of mean-level personality change across the

study relative to the employed group, and as such, we cannot reject

the null hypothesis. The implication here is that individuals expe-

riencing unemployment recover their preunemployment levels of

personality. This may be so because the reemployed individuals

did not experience personality change in the first place, but this

interpretation is implausible. Our finding is therefore in line with

suggestions that environmental factors will only influence person-

ality in the long run, provided they are consistent and persistent

(McGue, Bacon, & Lykken, 1993). An alternative explanation

states that individuals who maintained their preunemployment

level of personality traits were more likely to find reemployment.

Either way, our results pattern highlights the importance of under-

standing personality change in relation to unemployment. The

patter of our results also offers another explanation as to why

previous studies, which did not analyze data from reemployed

individuals separately, failed to find personality changes precipi-

tated by unemployment (Specht et al., 2011).

Implications, Limitations and Future Research

Our study established that personality change takes place in

response to unemployment, a major non-normative life event. In

doing so, our study not only blazes new territory for occupational

research and practice but also carries broader implications for the

conceptualization of personality stability and change. Although

recent evidence points to the malleability of personality (Lucas &

Donnellan, 2011; Roberts et al., 2006a), there has been consider-

able debate on whether change arises because of natural matura-

tion influenced by biological factors (McCrae & Costa, 2008;

McCrae et al., 2000) or to variation in social and occupational

contexts (Haan, Millsap, & Hartka, 1986; Hogan, 1996; Kogan,

1990). Our results side with the contextual perspective. Personality

change is associated with normative events that occur at relatively

predictable life intervals, such as leaving home, beginning a first

job, moving in with a partner, marriage, having a child, retirement,

death of a parent, and death of a spouse (Specht et al., 2011).

Unemployment, however, is unique in that the experience can

happen throughout most of the life cycle and is largely unantici-

pated. Hence, documenting that changes in personality occur fol-

lowing a non-normative event, like unemployment, is crucial for

the contextual perspective. Our predictions for personality change

arising from unemployment were based on limited existing theory,

but we hope that our findings will open up opportunities for

scholars to focus greater theoretical attention on unemployment

and personality change specifically.

The demonstration that personality changes, in conjunction with

occupational context, is particularly crucial, given that personality

predicts a good deal of applied and behavioral outcomes. For

example, FFM traits predict wage earning (Fletcher, 2013; Groves,

2005; Heineck, 2011; Mueller & Plug, 2006; Nyhus & Pons, 2005;

Semykina & Linz, 2007), knowledge sharing with colleagues

(Matzler, Renzl, Muller, Herting, & Mooradian, 2008), job satis-

faction (Winkelmann & Winkelmann, 2008), wealth accumulation

(Ameriks et al., 2003; Ameriks, Caplin, Leahy, & Tyler, 2007),

entrepreneurial behavior (Zhao, Seibert, & Lumpkin, 2010), and

well-being following occupationally related events (Boyce &

Wood, 2011a, 2011b; Boyce et al., 2010; Pai & Carr, 2010). In

parts of the literature, personality has been conceptualized as if it

were unchanging (Boyce, 2010; Ferrer-i-Carbonell & Frijters,

2004) or, even if somewhat changing, as having narrow applied

potential (Cobb-Clark & Schurer, 2012). According to our find-

ings, the assumption that personality change is nonexistent or

unimportant is a mistake. Therefore, our demonstration of substan-

tively large personality change in the context of a commonly

There was also no evidence to support any gender differences for those

becoming reemployed.

1005

UNEMPLOYMENT AND PERSONALITY

occurring labor market event has the potential to shape the way

researchers think about personality development processes in or-

ganizations, while highlighting the need for additional empirical

foci. Can personality change explain essential labor market out-

comes? Could the environment assist the development of person-

ality types that are most useful for occupational success in indi-

viduals, organizations, and society? What are the ethical

implications of enacting policies that, while obtaining some other

core objective, simultaneously have the potential to shape person-

ality?

It is crucial to understand both cognitive skills and personality

traits when assessing early-year educational investment on later

labor-market outcomes (Heckman, Moon, Pinto, Savelyev, &

Yavitz, 2010). However, cognitive skills reach stability relatively

early in life (Borghans et al., 2008), whereas personality traits

continue to change throughout life, and, as demonstrated here, can

do so in response to unemployment. Greater exploration into

personality change may then offer impactful later-life intervention

strategies that will help mitigate possible harmful effects of vari-

ous labor market events and promote adaptive coping in occupa-

tional settings (and beyond). If specific traits are found to be

conducive to a functional workplace, then it will be critical to

initiate accompanying policies (e.g., encouraging fairer employ-

ment practices, expanding access to mental health care; Benach et

al., 2010; Blustein, 2008; Layard, 2006) that seek to foster them.

That conditions promoting personality development may be as-

sisted though policy intervention raises the possibility of concep-

tualizing personality traits as quality-of-life indicators (Boyce et

al., 2013). Such indicators—typically including health, crime, eco-

nomic, and subjective well-being outcomes—change across time

and may therefore provide clues on how individuals and societies

progress (Diener & Suh, 1997). Personality also changes across

time. As such, monitoring personality at the national level may

indicate the achievement of desirable outcomes that raise the

quality of life in society. For example, personality change has been

linked to improved health outcomes (Magee, Heaven, & Miller,

2013; Turiano et al., 2012). Neuroticism, in particular, predicts

various mental and physical health disorders, and hence mental

health will likely track reductions in neuroticism (Lahey, 2009).

Extraversion and agreeableness may offer clues on the extent of

community engagement, whereas conscientiousness may reflect a

society that is well-structured, organized, and conducive to mem-

bers’ pursuit of meaningful goals. Openness to experiences is a

correlate of intelligence (Furnham, Swami, Arteche, & Chamorro-

Premuzic, 2008) and creativity (Chamorro-Premuzic, Reimers,

Hsu, & Ahmetoglu, 2009); this trait, then, may reflect societal

proclivities for inventiveness and flexible responses to new situa-

tions. Taken together, personality may be suitable as a quality of

life indicator.

Exploration of personality change would more generally be

assisted by a wider inclusion of personality measures at multiple

time points in routine large-scale data collection. Personality mea-

sures increasingly appear in large and nationally representative

longitudinal data sets. In our study, such measures were available

at only two time points and not every year. Having only two time

points presents a limitation for longitudinal studies (for a discus-

sion see Ployhart & Vandenberg, 2010). First, this practice restricts

the ability to discern precise forms of change—for example,

whether the change was steady or delayed. Second, this practice

increases the probability of confounding true change with mea-

surement error (Singer & Willett, 2003). However, we examined

the extent to which the impact of unemployment on personality

varied with regard to length of time that each individual had spent

in unemployment at the second time point. This helped overcome

some of the issues regarding the form of change, but we were still

only able to implicate the trajectory of personality changes across

the years of unemployment. Ideally, we would track within-person

changes in personality at each stage of the unemployment process

(e.g., at each year of unemployment, and then once reemployment

was established). Our results might therefore be explained through

an increased tendency for certain individuals both to become

unemployed in the first place and to remain unemployed for longer

time periods. As an example, certain traits (e.g., the proactive

personality) are useful in the job search process (Brown, Cober,

Kane, Levy, & Shalhoop, 2006), and this may have been one

reason for the relation between personality and unemployment

duration (Uysal & Pohlmeier, 2011; Wanberg, 2012). One concern

is whether our observed personality change reflects measurement

error or true change. The evidence points toward true change,

given our tests for measurement invariance. Yet, future empirical

efforts could separate more successfully true change from mea-

surement error by engaging in greater temporal frequency of

personality assessment.

A further limitation in our research, which resulted from the

restricted availability of personality variables, was a somewhat

restricted sample size, especially in regard to those individuals

who experienced unemployment for the longest time periods. This

limitation raises the issue of whether our analysis lacked adequate

power. However, our primary goal was testing whether personality

changed in some way following unemployment—a goal that we

accomplished conclusively in the current sample. We offer caution

against drawing large inferences at the population level, but nev-

ertheless our results are indicative.

We were also unable to examine the extent to which personality

change following unemployment predicted unemployment length.

An alternative explanation of our results is therefore that changes

following personality may have occurred immediately, and it

seems plausible that those who experienced the largest initial

reductions in, for example, conscientiousness following unem-

ployment may have been more likely to undergo longer periods of

unemployment. Alternatively, those who decreased (vs. increased)

in agreeableness may have been more likely to be unemployed

several years later, thus explaining why individuals in longer (vs.

shorter) unemployment periods manifest decreases in agreeable-

ness. These limitations could have been easily overcome had

personality data been available at more than two time points.

Moreover, it is possible that individuals high on such traits as

openness or agreeableness were more likely to become unem-

ployed (Specht et al., 2011). There was some evidence of a

selection effect, as indicated by the zero-order correlations in

Table 2, which illustrates that individuals higher in neuroticism

and lower in openness at T0 were more likely to experience

unemployment. Yet we did account for selection effects by fully

controlling for them in the latent change models.

Our work relied on small item-scales for each of the personality

traits. Although we demonstrated personality change, this is a

limitation that arises from resource constraints in large nationally

representative surveys. More extensive personality scales would

1006

BOYCE, WOOD, DALY, AND SEDIKIDES

have assisted our understanding of how specific facets of each of

the FFM personality dimensions are susceptible to change. As well

as using more detailed self-report personality scales, future work

into personality change may benefit from the use of neurological

assessments of personality (such as eyeblink measures of extra-

version; Blumenthal, 2001). Although the validity of self-report

measures of personality is established (Goldberg, 1993; McCrae &

Costa, 1987, 1997), with perceptions of the self often motivating

behavior and being strongly linked to biological functioning

(O’Cleirigh, Ironson, Weiss, & Costa, 2007; Ryff et al., 2006;

Sedikides, 2012), it would be useful to examine whether the same

patterns emerge when personality is measured with overt behavior

(e.g., agreeableness in a laboratory setting).

Conclusion

Unemployment is an event that can be inflicted upon most

persons. As we have demonstrated, this event can influence an

individual’s core personality—a finding that challenges the notion

of personality being fixed. This challenge will hopefully contribute

to a wider conceptualization of personality in disciplines outside of

psychology (e.g., management, economics, social sciences), while

suggesting that public policy can play a key role in enabling

psychological growth. Increases in national unemployment rates

may have pivotal implications for personality. As such, to the

extent that personality precipitates desirable social and economic

behavior—for example, higher savings rates (Ameriks et al., 2003;

Ameriks et al., 2007), prosocial activities (Binder & Freytag,

2013), or better health behaviors (Lahey, 2009; O’Connor, Conner,

Jones, McMillan, & Ferguson, 2009)—unemployment may pose

additional difficulties beyond the simple economic. Policies de-

signed to curb unemployment preserve not only psychological

health but also, critically, the basic personality traits that charac-

terize personhood.

References

Allemand, M., Zimprich, D., & Hertzog, C. (2007). Cross-sectional age

differences and longitudinal age changes of personality in middle adult-

hood and old age. Journal of Personality, 75, 323–358. http://dx.doi.org/

10.1111/j.1467-6494.2006.00441.x

Ameriks, J., Caplin, A., & Leahy, J. (2003). Wealth accumulation and the

propensity to plan. The Quarterly Journal of Economics, 118, 1007–

1047. http://dx.doi.org/10.1162/00335530360698487

Ameriks, J., Caplin, A., Leahy, J., & Tyler, T. (2007). Measuring self-

control problems. The American Economic Review, 97, 966–972. http://

dx.doi.org/10.1257/aer.97.3.966

Atkinson, T., Liem, R., & Liem, J. H. (1986). The social costs of unem-

ployment: Implications for social support. Journal of Health and Social

Behavior, 27, 317–331. http://dx.doi.org/10.2307/2136947

Barrick, M. R., Mount, M. K., & Strauss, J. P. (1993). Conscientiousness

and performance of sales representatives: Test of the mediating effects

of goal-setting. Journal of Applied Psychology, 78, 715–722. http://dx

.doi.org/10.1037/0021-9010.78.5.715

Bashkov, B. M., & Finney, S. J. (2013). Applying Longitudinal Mean and

Covariance Structures (LMACS) analysis to assess construct stability

over two time points: An example using psychological entitlement.

Measurement and Evaluation in Counseling and Development, 46, 289–

314. http://dx.doi.org/10.1177/0748175613497038

Benach, J., Muntaner, C., Chung, H., Solar, O., Santana, V., Friel, S.,...

Marmot, M. (2010). The importance of government policies in reducing

employment related health inequalities. British Medical Journal, 340,

c2154. http://dx.doi.org/10.1136/bmj.c2154

Benet-Martínez, V., & John, O. P. (1998). Los Cinco Grandes across

cultures and ethnic groups: Multitrait multimethod analyses of the Big

Five in Spanish and English. Journal of Personality and Social Psychol-

ogy, 75, 729–750. http://dx.doi.org/10.1037/0022-3514.75.3.729

Bettencourt, B. A., & Sheldon, K. (2001). Social roles as mechanisms for

psychological need satisfaction within social groups. Journal of Person-

ality and Social Psychology, 81, 1131–1143. http://dx.doi.org/10.1037/

0022-3514.81.6.1131

Binder, M., & Freytag, A. (2013). Volunteering, subjective well-being and

public policy. Journal of Economic Psychology, 34, 97–119. http://dx

.doi.org/10.1016/j.joep.2012.11.008

Bleidorn, W., Kandler, C., Hülsheger, U. R., Riemann, R., Angleitner, A.,

& Spinath, F. M. (2010). Nature and nurture of the interplay between

personality traits and major life goals. Journal of Personality and Social

Psychology, 99, 366–379. http://dx.doi.org/10.1037/a0019982

Bleidorn, W., Kandler, C., Riemann, R., Spinath, F. M., & Angleitner, A.

(2009). Patterns and sources of adult personality development: Growth

curve analyses of the NEO PI-R scales in a longitudinal twin study.

Journal of Personality and Social Psychology, 97, 142–155. http://dx

.doi.org/10.1037/a0015434

Blumenthal, T. D. (2001). Extraversion, attention, and startle response

reactivity. Personality and Individual Differences, 31, 495–503. http://

dx.doi.org/10.1016/S0191-8869(00)00153-7

Blustein, D. L. (2008). The role of work in psychological health and

well-being: A conceptual, historical, and public policy perspective.

American Psychologist, 63, 228–240. http://dx.doi.org/10.1037/0003-

066X.63.4.228

Borghans, L., Duckworth, A. L., Heckman, J. J., & ter Weel, B. (2008).

The economics and psychology of personality traits. The Journal of

Human Resources, 43, 972–1059. http://dx.doi.org/10.1353/jhr.2008

.0017

Boyce, C. J. (2010). Understanding fixed effects in human well-being.

Journal of Economic Psychology, 31, 1–16. http://dx.doi.org/10.1016/j

.joep.2009.08.009

Boyce, C. J., & Wood, A. M. (2011a). Personality and the marginal utility

of income: Personality interacts with increases in household income to

determine life satisfaction. Journal of Economic Behavior & Organiza-

tion, 78, 183–191. http://dx.doi.org/10.1016/j.jebo.2011.01.004

Boyce, C. J., & Wood, A. M. (2011b). Personality prior to disability

determines adaptation: Agreeable individuals recover lost life satisfac-

tion faster and more completely. Psychological Science, 22, 1397–1402.

http://dx.doi.org/10.1177/0956797611421790

Boyce, C. J., Wood, A. M., & Brown, G. D. A. (2010). The dark side of

conscientiousness: Conscientious people experience greater drops in life

satisfaction following unemployment. Journal of Research in Personal-

ity, 44, 535–539. http://dx.doi.org/10.1016/j.jrp.2010.05.001

Boyce, C. J., Wood, A. M., & Powdthavee, N. (2013). Is personality fixed?

Personality changes as much as “variable” economic factors and more

strongly predicts changes to life satisfaction. Social Indicators Research,

111, 287–305. http://dx.doi.org/10.1007/s11205-012-0006-z

Brown, D. J., Cober, R. T., Kane, K., Levy, P. E., & Shalhoop, J. (2006).

Proactive personality and the successful job search: A field investigation

with college graduates. Journal

of Applied Psychology, 91, 717–726.

http://dx.doi.org/10.1037/0021-9010.91.3.717

Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model

fit. In K. A. Bollen & J. S. Long (Eds.), Testing structural equation

models (pp. 136–162). Newbury Park, CA: Sage.

Chamorro-Premuzic, T., Reimers, S., Hsu, A., & Ahmetoglu, G. (2009).

Who art thou? Personality predictors of artistic preferences in a large UK

sample: The importance of openness. British Journal of Psychology,

100, 501–516. http://dx.doi.org/10.1348/000712608X366867

1007

UNEMPLOYMENT AND PERSONALITY

Cheung, G. W., & Rensvold, R. B. (2002). Evaluating goodness-of-fit

indexes for testing measurement invariance. Structural Equation Mod-

eling, 9, 233–255. http://dx.doi.org/10.1207/S15328007SEM0902_5

Chungkham, H. S., Ingre, M., Karasek, R., Westerlund, H., & Theorell, T.

(2013). Factor structure and longitudinal measurement invariance of the

demand control support model: An evidence from the Swedish Longi-

tudinal Occupational Survey of Health (SLOSH). PLoS ONE, 8, e70541.

http://dx.doi.org/10.1371/journal.pone.0070541

Clark, A. E., Diener, E., Georgellis, Y., & Lucas, R. E. (2008). Lags and

leads in life satisfaction: A test of the baseline hypothesis. The Economic

Journal, 118, F222–F243. http://dx.doi.org/10.1111/j.1468-0297.2008

.02150.x

Clark, A. E., Georgellis, Y., & Sanfey, P. (2001). Scarring: The psycho-

logical impact of past unemployment. Economica, 68, 221–241. http://

dx.doi.org/10.1111/1468-0335.00243

Clark, A. E., & Oswald, A. J. (2002). A simple statistical method for

measuring how life events affect happiness. International Journal of

Epidemiology, 31, 1139–1144. http://dx.doi.org/10.1093/ije/31.6.1139

Cobb-Clark, D., & Schurer, S. (2012). The stability of the big-five per-

sonality traits. Economics Letters, 115, 11–15. http://dx.doi.org/10.1016/

j.econlet.2011.11.015

Cohen, S., Gottlieb, B., & Underwood, L. (2000). Social relationships and

health. In S. Cohen, L. Underwood, & B. Gottlieb (Eds.), Measuring and

intervening in social support (pp. 3–25). New York, NY: Oxford Uni-

versity Press.

Costa, P. T., Jr., & McCrae, R. R. (1994). Set like plaster: Evidence for the

stability of adult personality. In T. F. Heatherton & J. L. Weinberger

(Eds.), Can personality change? (pp. 21–40). Washington, DC: Amer-

ican Psychological Association. http://dx.doi.org/10.1037/10143-002

Costa, P. T., Jr., & McCrae, R. R. (2006). Age changes in personality and

their origins: Comment on Roberts, Walton, and Viechtbauer (2006).

Psychological Bulletin, 132, 26–28. http://dx.doi.org/10.1037/0033-

2909.132.1.26

Daly, M., & Delaney, L. (2013). The scarring effect of unemployment

throughout adulthood on psychological distress at age 50: Estimates

controlling for early adulthood distress and childhood psychological

factors. Social Science & Medicine, 80, 19–23. http://dx.doi.org/

10.1016/j.socscimed.2012.12.008

DeNeve, K. M., & Cooper, H. (1998). The happy personality: A meta-

analysis of 137 personality traits and subjective well-being. Psycholog-

ical Bulletin, 124, 197–229. http://dx.doi.org/10.1037/0033-2909.124.2

.197

Diener, E., & Suh, E. (1997). Measuring quality of life: Economic, social,

and subjective indicators. Social Indicators Research, 40, 189–216.

http://dx.doi.org/10.1023/A:1006859511756

Donnellan, M. B., & Lucas, R. E. (2008). Age differences in the Big Five

across the life span: Evidence from two national samples. Psychology

and Aging, 23, 558–566. http://dx.doi.org/10.1037/a0012897

Dooley, D., Prause, J., & Ham-Rowbottom, K. A. (2000). Underemploy-

ment and depression: Longitudinal relationships. Journal of Health and

Social Behavior, 41, 421–436. http://dx.doi.org/10.2307/2676295

Eysenck, H. J. (Ed.). (1981). A model for personality. Berlin, Germany:

Springer-Verlag. http://dx.doi.org/10.1007/978-3-642-67783-0

Ferrer-i-Carbonell, A., & Frijters, P. (2004). How important is methodol-

ogy for the estimates of the determinants of happiness? The Economic

Journal, 114, 641–659. http://dx.doi.org/10.1111/j.1468-0297.2004

.00235.x

Fleeson, W. (2001). Toward a structure- and process-integrated view of

personality: Traits as density distribution of states. Journal of Person-

ality and Social Psychology, 80, 1011–1027. http://dx.doi.org/10.1037/

0022-3514.80.6.1011

Fleeson, W. (2004). Moving personality beyond the person-situation de-

bate: The challenge and the opportunity of within-person variability.

Current Directions in Psychological Science, 13, 83–87. http://dx.doi

.org/10.1111/j.0963-7214.2004.00280.x

Fletcher, J. M. (2013). The effects of personality traits on adult labor

market outcomes: Evidence from siblings. Journal of Economic Behav-

ior & Organization, 89, 122–135. http://dx.doi.org/10.1016/j.jebo.2013

.02.004

Forret, M. L., Sullivan, S. E., & Mainiero, L. A. (2010). Gender role

differences in reactions to unemployment: Exploring psychological mo-

bility and boundaryless careers. Journal

of Organizational Behavior, 31,

647–666. http://dx.doi.org/10.1002/job.703

Frost, T. F., & Clayson, D. E. (1991). The measurement of self-esteem,

stress-related life events, and locus of control among unemployed and

employed blue-collar workers. Journal of Applied Social Psychology,

21, 1402–1417. http://dx.doi.org/10.1111/j.1559-1816.1991.tb00478.x

Furnham, A., Swami, V., Arteche, A., & Chamorro-Premuzic, T. (2008).

Cognitive ability, learning approaches and personality correlates of

general knowledge. Educational Psychology, 28, 427–437. http://dx.doi

.org/10.1080/01443410701727376

Gerlitz, J., & Schupp, J. (2005). Zur erhebung der Big-Fivebasierten

personlichkeitsmerkmale im SOEP [Collecting the Big-Five Personality

in SOEP]. Berlin, Germany: DIW Berlin.

Goldberg, L. R. (1993). The structure of phenotypic personality traits.

American Psychologist, 48, 26–34. http://dx.doi.org/10.1037/0003-

066X.48.1.26

Gramzow, R. H., Sedikides, C., Panter, A. T., Sathy, V., Harris, J., &

Insko, C. A. (2004). Patterns of self-regulation and the Big Five. Euro-

pean Journal of Personality, 18, 367–385. http://dx.doi.org/10.1002/per

.513

Groves, M. O. (2005). How important is your personality? Labor market

returns to personality for women in the US and UK. Journal of Eco-

nomic Psychology, 26, 827–841. http://dx.doi.org/10.1016/j.joep.2005

.03.001

Haan, N., Millsap, R., & Hartka, E. (1986). As time goes by: Change and

stability in personality over fifty years. Psychology and Aging, 1, 220–

232. http://dx.doi.org/10.1037/0882-7974.1.3.220

Hafdahl, A. R., Panter, A. T., Gramzow, R. H., Sedikides, C., & Insko,

C. A. (2000). Free-response self-discrepancies across, among, and

within FFM personality dimensions. Journal of Personality, 68, 111–

151. http://dx.doi.org/10.1111/1467-6494.00093

Hayduk, L. A. (1988). Structural equation modeling with LISREL: Essen-

tials and advances. Baltimore, MD: Johns Hopkins University Press.

Heckman, J. J., Moon, S. H., Pinto, R., Savelyev, P. A., & Yavitz, A.

(2010). The rate of return to the High/Scope Perry Preschool Program.

Journal of Public Economics, 94, 114–128. http://dx.doi.org/10.1016/j

.jpubeco.2009.11.001

Heineck, G. (2011). Does it pay to be nice? Personality and earnings in the

UK. Industrial & Labor Relations Review, 64, 1020–1038.

Heinrich, L. M., & Gullone, E. (2006). The clinical significance of lone-

liness: A literature review. Clinical Psychology Review, 26, 695–718.

http://dx.doi.org/10.1016/j.cpr.2006.04.002

Hogan, R. (1996). A socioanalytic perspective on the five-factor model. In

J. S. Wiggins (Ed.), The five-factor model of personality: Theoretical

perspectives (pp. 180–207). New York, NY: Guilford Press.

Hooper, D., Coughlan, J., & Mullen, M. R. (2008). Structural equation

modelling: Guidelines for determining model fit. Electronic Journal of

Business Research Methods, 6, 53–60.

Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance

structure analysis: Conventional criteria versus new alternatives. Struc-

tural Equation Modeling, 6, 1–55. http://dx.doi.org/10.1080/

10705519909540118

Judge, T. A., & Ilies, R. (2002). Relationship of personality to performance

motivation: A meta-analytic review. Journal of Applied Psychology, 87,

797–807. http://dx.doi.org/10.1037/0021-9010.87.4.797

1008

BOYCE, WOOD, DALY, AND SEDIKIDES

Kandler, C. (2012). Nature and nurture in personality development. Cur-

rent Directions in Psychological Science, 21, 290–296. http://dx.doi

.org/10.1177/0963721412452557

Kanfer, R., Wanberg, C. R., & Kantrowitz, T. M. (2001). Job search and

employment: A personality-motivational analysis and meta-analytic re-

view. Journal of Applied Psychology, 86, 837–855. http://dx.doi.org/

10.1037/0021-9010.86.5.837

Karren, R., & Sherman, K. (2012). Layoffs and unemployment discrimi-

nation: A new stigma. Journal of Managerial Psychology, 27, 848–863.

http://dx.doi.org/10.1108/02683941211280193

Kogan, N. (1990). Personality and aging. In J. E. Birren & S. W. Schaie

(Eds.), Handbook of the psychology of aging (3rd ed., pp. 330–346). San

Diego, CA: Academic Press. http://dx.doi.org/10.1016/B978-0-12-

101280-9.50026-7

Lahey, B. B. (2009). Public health significance of neuroticism. American

Psychologist, 64, 241–256. http://dx.doi.org/10.1037/a0015309

Lang, F. R. (2005). Erfassung des kognitiven leistungspotenzials und der

“Big Five” mit computer-assisted-personal-interviewing (CAPI): Zur

reliabilität und validität zweier ultrakurzer tests und des BFI-S [Assess-

ment of cognitive capabilities and the Big Five with computer-assisted

personal interviewing (CAPI): Reliability and validity]. Berlin, Ger-

many: DIW Berlin.

Lang, F. R., John, D., Lüdtke, O., Schupp, J., & Wagner, G. G. (2011).

Short assessment of the Big Five: Robust across survey methods except

telephone interviewing. Behavior Research Methods, 43, 548–567.

http://dx.doi.org/10.3758/s13428-011-0066-z

Layard, R. (2006). The case for psychological treatment centres. British

Medical Journal, 332, 1030–1032. http://dx.doi.org/10.1136/bmj.332

.7548.1030

Leana, C. R., & Feldman, D. C. (1991). Gender differences in responses to

unemployment. Journal of Vocational Behavior, 38, 65–77. http://dx.doi

.org/10.1016/0001-8791(91)90018-H

Lenton, A. P., Bruder, M., Slabu, L., & Sedikides, C. (2013). How does

“being real” feel? The experience of state authenticity. Journal of

Personality, 81, 276–289. http://dx.doi.org/10.1111/j.1467-6494.2012

.00805.x

Lewin, K. (1951). Field theory in social science. New York, NY: Harper.

Löckenhoff, C. E., Terracciano, A., Patriciu, N. S., Eaton, W. W., & Costa,

P. T., Jr. (2009). Self-reported extremely adverse life events and longi-

tudinal changes in five-factor model personality traits in an urban

sample. Journal of Traumatic Stress, 22, 53–59. http://dx.doi.org/

10.1002/jts.20385

Lucas, R. E. (2007). Long-term disability is associated with lasting changes

in subjective well-being: Evidence from two nationally representative

longitudinal studies. Journal of Personality and Social Psychology, 92,

717–730. http://dx.doi.org/10.1037/0022-3514.92.4.717

Lucas, R. E., & Donnellan, M. B. (2011). Personality development across

the life span: Longitudinal analyses with a national sample from Ger-

many. Journal of Personality and Social Psychology, 101, 847–861.

http://dx.doi.org/10.1037/a0024298

MacLean, K. A., Johnson, M. W., & Griffiths, R. R. (2011). Mystical

experiences occasioned by the hallucinogen psilocybin lead to increases

in the personality domain of openness. Journal of Psychopharmacology,

25, 1453–1461. http://dx.doi.org/10.1177/0269881111420188

Magee, C. A., Heaven, P. C. L., & Miller, L. M. (2013). Personality change

predicts self-reported mental and physical health. Journal of Personality,

81, 324–334. http://dx.doi.org/10.1111/j.1467-6494.2012.00802.x

Magnusson, D., & Endler, N. S. (1977). Interactional psychology: Present

status and future prospects. In D. Magnusson & N. S. Endler (Eds.),

Personality at the crossroads: Current issues in interactional psychol-

ogy (pp. 3–31). New York, NY: Erlbaum.

Matzler, K., Renzl, B., Muller, J., Herting, S., & Mooradian, T. A. (2008).

Personality traits and knowledge sharing. Journal of Economic Psychol-

ogy, 29, 301–313. http://dx.doi.org/10.1016/j.joep.2007.06.004

McArdle, J. J. (2009). Latent variable modeling of differences and changes

with longitudinal data. Annual Review of Psychology, 60, 577–605.

http://dx.doi.org/10.1146/annurev.psych.60.110707.163612

McCrae, R. R., & Costa, P. T., Jr. (1987). Validation of the five-factor

model of personality across instruments and observers. Journal of Per-

sonality and Social Psychology, 52, 81–90. http://dx.doi.org/10.1037/

0022-3514.52.1.81

McCrae, R. R., & Costa, P. T., Jr. (1997). Personality trait structure as a

human universal. American Psychologist, 52, 509–516. http://dx.doi

.org/10.1037/0003-066X.52.5.509

McCrae, R. R., & Costa, P. T., Jr. (2008). The five-factor theory of

personality.

In O. P. John, R. W. Robins, & L. A. Pervin (Eds.),

Handbook of personality: Theory and research (3rd ed., pp. 159–181).

New York, NY: Guilford Press.

McCrae, R. R., Costa, P. T., Jr., Ostendorf, F., Angleitner, A., Hrebícková,

M., Avia, M. D.,...Smith, P. B. (2000). Nature over nurture: Tem-

perament, personality, and life span development. Journal of Personality

and Social Psychology, 78, 173–186. http://dx.doi.org/10.1037/0022-

3514.78.1.173

McCrae, R. R., Costa, P. T., Jr., Pedroso de Lima, M., Simões, A.,

Ostendorf, F., Angleitner, A.,...Piedmont, R. L. (1999). Age differ-

ences in personality across the adult life span: Parallels in five cultures.

Developmental Psychology, 35, 466–477. http://dx.doi.org/10.1037/

0012-1649.35.2.466

McGarty, C. (2002). Stereotype formation as category formation. In C.

McGarty, V. Y. Yzerbyt, & R. Spears (Eds.), Stereotypes as explana-

tions: The formation of meaningful beliefs about social groups (pp.

16–37). New York, NY: Cambridge University Press. http://dx.doi.org/

10.1017/CBO9780511489877.003

McGue, M., Bacon, S., & Lykken, D. T. (1993). Personality stability and

change in early adulthood: A behavioural genetic analysis. Developmen-

tal Psychology, 29, 96–109. http://dx.doi.org/10.1037/0012-1649.29

.1.96

McKee-Ryan, F., Song, Z., Wanberg, C. R., & Kinicki, A. J. (2005).

Psychological and physical well-being during unemployment: A meta-

analytic study. Journal of Applied Psychology, 90, 53–76. http://dx.doi

.org/10.1037/0021-9010.90.1.53

Mischel, W., & Shoda, Y. (1995). A cognitive-affective system theory of

personality: Reconceptualizing situations, dispositions, dynamics, and

invariance in personality structure. Psychological Review, 102, 246–

268. http://dx.doi.org/10.1037/0033-295X.102.2.246

Mueller, G., & Plug, E. J. S. (2006). Estimating the effect of personality on

male and female earnings. Industrial & Labor Relations Review, 60,

3–22.

Neyer, F. J., & Lehnart, J. (2007). Relationships matter in personality

development: Evidence from an 8-year longitudinal study across young

adulthood. Journal of Personality, 75, 535–568. http://dx.doi.org/

10.1111/j.1467-6494.2007.00448.x

Nyhus, E. K., & Pons, E. (2005). The effects of personality on earnings.

Journal of Economic Psychology, 26, 363–384. http://dx.doi.org/

10.1016/j.joep.2004.07.001

O’Cleirigh, C., Ironson, G., Weiss, A., & Costa, P. T., Jr. (2007). Consci-

entiousness predicts disease progression (CD4 number and viral load) in

people living with HIV. Health Psychology, 26, 473–480. http://dx.doi

.org/10.1037/0278-6133.26.4.473

O’Connor, D. B., Conner, M., Jones, F., McMillan, B., & Ferguson, E.

(2009). Exploring the benefits of conscientiousness: An investigation of

the role of daily stressors and health behaviors. Annals of Behavioral

Medicine, 37, 184–196. http://dx.doi.org/10.1007/s12160-009-9087-6

Osafo Hounkpatin, H., Wood, A. M., Boyce, C. J., & Dunn, G. (in press).

An existential-humanistic view of personality change: Co-occurring

changes with psychological well-being in a ten year cohort study. Social

Indicators Research.

1009

UNEMPLOYMENT AND PERSONALITY

Oswald, A. J., & Powdthavee, N. (2008). Death, happiness, and the

calculation of compensatory damages. The Journal of Legal Studies, 37,

S217–S251. http://dx.doi.org/10.1086/595674

Pai, M., & Carr, D. (2010). Do personality traits moderate the effect of

late-life spousal loss on psychological distress? Journal of Health

and Social Behavior, 51, 183–199. http://dx.doi.org/10.1177/

0022146510368933

Pervin, L. A. (1968). Performance and satisfaction as a function of

individual-environment fit. Psychological Bulletin, 69, 56–68. http://dx

.doi.org/10.1037/h0025271

Piedmont, R. L. (2001). Cracking the plaster cast: Big Five personality

change during intensive outpatient counseling. Journal of Research in

Personality, 35, 500–520. http://dx.doi.org/10.1006/jrpe.2001.2326

Pitariu, A. H., & Ployhart, R. E. (2010). Explaining change: Theorizing and

testing dynamic mediated longitudinal relationships. Journal of Man-

agement, 36, 405–429. http://dx.doi.org/10.1177/0149206308331096

Ployhart, R. E., & Vandenberg, R. J. (2010). Longitudinal research: The

theory, design, and analysis of change. Journal of Management, 36,

94–120. http://dx.doi.org/10.1177/0149206309352110

Powdthavee, N., Boyce, C. J., & Wood, A. M. (2011). Estimating the effect

of unearned income on non-cognitive skills: Evidence from surveys of

lottery winners. Unpublished manuscript, IZA Discussion Paper No.

6131.

Reynolds, K. J., Turner, J. C., Branscombe, N. R., Mavor, K. I., Bizumic,

B., & Subasic, E. (2010). Interactionism in personality and social psy-

chology: An integrated approach to understanding the mind and behav-

iour. European Journal of Personality, 24, 458–482. http://dx.doi.org/

10.1002/per.782

Roberts, B. W. (1997). Plaster or plasticity: Are adult work experiences

associated with personality change in women? Journal of Personality,

65, 205–232. http://dx.doi.org/10.1111/j.1467-6494.1997.tb00953.x

Roberts, B. W., & Bogg, T. (2004). A longitudinal study of the relation-

ships between conscientiousness and the social-environmental factors

and substance-use behaviors that influence health. Journal of Personal-

ity, 72, 325–354. http://dx.doi.org/10.1111/j.0022-3506.2004.00264.x

Roberts, B. W., Caspi, A., & Moffitt, T. E. (2003). Work experiences and

personality development in young adulthood. Journal of Personality and

Social Psychology, 84, 582–593. http://dx.doi.org/10.1037/0022-3514

.84.3.582

Roberts, B. W., Walton, K., Bogg, T., & Caspi, A. (2006). De-investment

in work and non-normative personality trait change in young adulthood.

European Journal of Personality, 20, 461–474. http://dx.doi.org/

10.1002/per.607

Roberts, B. W., Walton, K. E., & Viechtbauer, W. (2006a). 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., Walton, K. E., & Viechtbauer, W. (2006b). Personality traits

change in adulthood: Reply to Costa and McCrae (2006). Psychological

Bulletin, 132, 29–32. http://dx.doi.org/10.1037/0033-2909.132.1.29

Roberts, B. W., Wood, D., & Caspi, A. (2008). The development of

personality traits in adulthood. In O. P. John, R. W. Robins, & L. A.

Pervin (Eds.), Handbook of personality: Theory and research (3rd ed.,

pp. 375–398). New York, NY: Guilford Press.

Robinson, M. D., & Sedikides, C. (2009). Traits and the self: Toward an

integration. In P. J. Corr & G. Matthews (Eds.), Cambridge handbook of

personality psychology (pp. 457–472). Cambridge, UK: Cambridge

University Press. http://dx.doi.org/10.1017/CBO9780511596544.029

Ryff, C. D., Dienberg Love, G., Urry, H. L., Muller, D., Rosenkranz,

M. A., Friedman, E. M.,...Singer, B. (2006). Psychological well-being

and ill-being: Do they have distinct or mirrored biological correlates?

Psychotherapy and Psychosomatics, 75, 85–95. http://dx.doi.org/

10.1159/000090892

Scollon, C. N., & Diener, E. (2006). Love, work, and changes in extraversion

and neuroticism over time. Journal of Personality and Social Psychology,

91, 1152–1165. http://dx.doi.org/10.1037/0022-3514.91.6.1152

Sedikides, C. (2012). Self-protection. In M. R. Leary & J. P. Tangney

(Eds.), Handbook of self and identity (2nd ed., pp. 327–353). New York,

NY: Guilford Press.

Sedikides, C., & Gregg, A. P. (2003). Portraits of the self. In M. A. Hogg

& J. Cooper (Eds.), Sage handbook of social psychology

(pp. 110–138).

London,

UK: Sage.

Semykina, A., & Linz, S. J. (2007). Gender differences in personality and

earnings: Evidence from Russia. Journal of Economic Psychology, 28,

387–410. http://dx.doi.org/10.1016/j.joep.2006.05.004

Sheldon, K. M., Ryan, R. M., Rawsthorne, L. J., & Ilardi, B. (1997). Trait

self and true self: Cross-role variation in the big-five personality traits

and its relations with psychological authenticity and subjective well-

being. Journal of Personality and Social Psychology, 73, 1380–1393.

http://dx.doi.org/10.1037/0022-3514.73.6.1380

Shoda, Y., & Mischel, W. (1998). Personality as a stable cognitive-

affective activation network: Characteristic patterns of behavior varia-

tion emerge from a stable personality structure. In S. J. Read & L. C.

Miller (Eds.), Connectionist and PDP models of social reasoning and

social behavior (pp. 175–208). Mahwah, NJ: Erlbaum.

Singer, J. D., & Willett, J. B. (2003). Applied longitudinal data analysis.

New York, NY: Oxford University Press. http://dx.doi.org/10.1093/

acprof:oso/9780195152968.001.0001

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

Srivastava, S., John, O. P., Gosling, S. D., & Potter, J. (2003). Develop-

ment of personality in early and middle adulthood: Set like plaster or

persistent change? Journal of Personality and Social Psychology, 84,

1041–1053. http://dx.doi.org/10.1037/0022-3514.84.5.1041

Steel, P., Schmidt, J., & Shultz, J. (2008). Refining the relationship

between personality and subjective well-being. Psychological Bulletin,

134, 138–161. http://dx.doi.org/10.1037/0033-2909.134.1.138

Turiano, N. A., Pitzer, L., Armour, C., Karlamangla, A., Ryff, C. D., &

Mroczek, D. K. (2012). Personality trait level and change as predictors

of health outcomes: Findings from a national study of Americans

(MIDUS). The Journals of Gerontology: Series B: Psychological and

Social Sciences, 67B, 4–12. http://dx.doi.org/10.1093/geronb/gbr072

Uysal, S. D., & Pohlmeier, W. (2011). Unemployment duration and per-

sonality. Journal of Economic Psychology, 32, 980–992. http://dx.doi

.org/10.1016/j.joep.2011.03.008

Vandenberg, R. J., & Lance, C. E. (2000). A review and synthesis of the

measurement invariance literature: Suggestions, practices, and recom-

mendations for organizational research. Organizational Research Meth-

ods, 3, 4–70. http://dx.doi.org/10.1177/109442810031002

Wagner, G. G., Frick, J. R., & Schupp, J. (2007). The German Socio-

Economic Panel Study (SOEP): Scope, evolution and enhancements.

Schmollers Jahrbuch, 127, 139–169.

Wanberg, C. R. (2012). The individual experience of unemployment.

Annual Review of Psychology, 63, 369–396. http://dx.doi.org/10.1146/

annurev-psych-120710-100500

Waters, L. E., & Moore, K. A. (2002). Self-esteem, appraisal and coping:

A comparison of unemployed and re-employed people. Journal of Or-

ganizational Behavior, 23, 593–604. http://dx.doi.org/10.1002/job.156

Watson, D., & Humrichouse, J. (2006). Personality development in emerg-

ing adulthood: Integrating evidence from self-ratings and spouse ratings.

Journal of Personality and Social Psychology, 91, 959–974. http://dx

.doi.org/10.1037/0022-3514.91.5.959

1010

BOYCE, WOOD, DALY, AND SEDIKIDES

Widiger, T. A. (2009). Neuroticism. In M. R. Leary & R. H. Hoyle (Eds.),

Handbook of individual differences in social behavior (pp. 129–146).

New York, NY: Guilford Press.

Winkelmann, L., & Winkelmann, R. (2008). Personality, work, and satis-

faction: Evidence from the German Socio-Economic Panel. The Journal

of Positive Psychology, 3, 266–275. http://dx.doi.org/10.1080/

17439760802399232

Wood, A. M., & Boyce, C. J. (in press). Personality. In A. Michalos

(Ed.), Encyclopaedia of quality of life research. New York, NY: Sprin-

ger.

Zhao, H., Seibert, S. E., & Lumpkin, G. T. (2010). The relationship of

personality to entrepreneurial intentions and performance: A meta-

analytic review. Journal of Management, 36, 381–404. http://dx.doi.org/

10.1177/0149206309335187

Appendix

SOEP Personality Scale

In the questionnaire section of the SOEP entitled “What kind of

personality do you have?” individuals are asked whether they see

themselves as someone who...

1. . . . does a thorough job

2. ...iscommunicative, talkative

3. ...issometimes somewhat rude to others

4. ...isoriginal, comes up with new ideas

5. . . . worries a lot

6. . . . has a forgiving nature

7. . . . tends to be lazy

8. ...isoutgoing, sociable

9. . . . values artistic experiences

10. . . . gets nervous easily

11. . . . does things effectively and efficiently

12. ...isreserved

13. ...isconsiderate and kind to others

14. . . . has an active imagination

15. ...isrelaxed, handles stress well

Individuals are asked whether the statement applies to them on

a 1-to-7 scale, with “1” meaning the statement does not apply to

them at all and “7” that it applies perfectly. Questions 3, 6, and 13

relate to the Agreeableness scale; 1, 7, and 11 relate to the

Conscientiousness scale; 2, 8, and 12 relate to the Extraversion

scale; 5, 10, and 15 relate to the Neuroticism scale; and 4, 9, and

14 relate to the Openness to Experience scale. Scores for each of

the traits are obtained by aggregating across each of the three items

by trait after reverse-coding Questions 3, 7, 12, and 15.

Received June 13, 2012

Revision received November 17, 2014

Accepted December 5, 2014 䡲

1011

UNEMPLOYMENT AND PERSONALITY