Development of Personality in Early and Middle Adulthood: Set Like Plaster or Persistent Change?

Abstract

Different theories make different predictions about how mean levels of personality traits change in adulthood. The biological view of the Five-factor theory proposes the plaster hypothesis: All personality traits stop changing by age 30. In contrast, contextualist perspectives propose that changes should be more varied and should persist throughout adulthood. This study compared these perspectives in a large (N = 132,515) sample of adults aged 21-60 who completed a Big Five personality measure on the Internet. Conscientiousness and Agreeableness increased throughout early and middle adulthood at varying rates; Neuroticism declined among women but did not change among men. The variety in patterns of change suggests that the Big Five traits are complex phenomena subject to a variety of developmental influences.

Full text

Development of Personality in Early and Middle Adulthood:
Set Like Plaster or Persistent Change?
Sanjay Srivastava and Oliver P. John
University of California, Berkeley Samuel D. Gosling
University of Texas at Austin
Jeff Potter
Cambridge, Massachusetts
Different theories make different predictions about how mean levels of personality traits change in
adulthood. The biological view of the Five-factor theory proposes the plaster hypothesis: All personality
traits stop changing by age 30. In contrast, contextualist perspectives propose that changes should be
more varied and should persist throughout adulthood. This study compared these perspectives in a large
(N⫽132,515) sample of adults aged 21–60 who completed a Big Five personality measure on the
Internet. Conscientiousness and Agreeableness increased throughout early and middle adulthood at
varying rates; Neuroticism declined among women but did not change among men. The variety in
patterns of change suggests that the Big Five traits are complex phenomena subject to a variety of
developmental influences.
How does personality change during adulthood? Psychologists
since William James (1890/1950) have struggled with the question
of whether various aspects of personality, including personality
traits, change in meaningful ways during adulthood, and when
those changes take place. Contemporary hypotheses about the
development of personality traits stem from theories about what
personality traits are. McCrae and Costa’s (1996) five-factor the-
ory asserts that personality traits arise exclusively from biological
causes (i.e., genes) and that they reach full maturity in early
adulthood; thus, this theory predicts little or no change on any
personality dimension after early adulthood. By contrast, contex-
tualist perspectives argue that traits are multiply determined, and
that one important influence on traits is the individual’s social
environment (Haan, Millsap, & Hartka, 1986; Helson, Jones, &
Kwan, 2002). Contextualist perspectives thus predict plasticity:
Change is complex and ongoing, owing to the many factors that
can affect personality traits.
In this study, we set out to understand how personality traits
change in early and middle adulthood by examining the Big Five
personality trait dimensions (Goldberg, 1992; John & Srivastava,
1999; McCrae & Costa, 1999). We used a cross-sectional design to
study how mean levels of personality traits differ by age and
whether those age effects are moderated by gender.
1
We were
particularly interested in examining whether change on all of the
Big Five dimensions stops or slows in middle adulthood, as
predicted by the five-factor theory, or whether change is ongoing
and differentiated, as predicted by contextualist theories.
Past Research on Mean-Level Change on the Big Five
During Adulthood
A recent literature review summarized previous studies of
mean-level change on the Big Five (Roberts, Robins, Caspi, &
Trzesniewski, in press). In this review, Roberts et al. (in press)
rationally categorized a wide variety of personality measures into
the Big Five domains and summarized patterns of mean-level
change that were consistent across studies. They concluded that, in
general, Conscientiousness and Agreeableness tend to go up dur-
ing adulthood, Neuroticism tends to go down, Openness shows
mixed results across studies, and Extraversion shows no general
pattern of change at the factor level. This basic pattern of findings
1
“Change” is a broad concept that can be defined in a variety of other
ways, such as rank-order change (whether people change in their ordering
relative to age mates) and individual differences in change (whether
different individuals change at different rates over time). These other ways
of examining change address somewhat different substantive issues, and it
is possible to obtain conceptually compatible but different results with the
different approaches. (For a fuller discussion of different kinds of change,
see Caspi & Roberts, 1999.)
Sanjay Srivastava and Oliver P. John, Institute of Personality and Social
Research, University of California, Berkeley; Samuel D. Gosling, Depart-
ment of Psychology, University of Texas at Austin; Jeff Potter, Cambridge,
Massachusetts.
Sanjay Srivastava was supported by a National Science Foundation
Graduate Research Fellowship; Sanjay Srivastava and Oliver P. John
were supported by National Institute of Mental Health Grant MH-43948.
We thank Ravenna Helson, Robert R. McCrae, and Frank J. Sulloway for
their helpful comments on drafts of this article. We also thank Frank J.
Sulloway for developing the algorithm to remove repeat responders from
the database.
Correspondence concerning this article should be addressed to Sanjay
Srivastava, who is now at the Department of Psychology, Stanford Uni-
versity, Jordan Hall, Building 420, Stanford, California 94305. E-mail:
sanjay@psych.stanford.edu
Journal of Personality and Social Psychology Copyright 2003 by the American Psychological Association, Inc.
2003, Vol. 84, No. 5, 1041–1053 0022-3514/03/$12.00 DOI: 10.1037/0022-3514.84.5.1041
1041

has been reported in specific studies by researchers who argue that
personality traits are affected by context (e.g., Helson et al., 2002;
Helson & Kwan, 2000) as well as those who favor a strictly
biological interpretation of traits (e.g., McCrae et al., 1999, 2000).
Although Roberts et al.’s (in press) conclusion seems to repre-
sent some common ground among researchers, there is still con-
siderable disagreement: The biological and contextual perspectives
disagree sharply over the timing of changes within the life course
and over whether there are any differences between men’s and
women’s development.
Set Like Plaster: The Five-Factor Theory
According to the five-factor theory, personality traits are “insu-
lated from the direct effects of the environment”(McCrae & Costa,
1999, p. 144) and are exclusively biological in origin. Change is
addressed by Postulate 1c of the five-factor theory: “Traits develop
through childhood and reach mature form in adulthood; thereafter
they are stable in cognitively intact individuals”(McCrae & Costa,
1999, p. 145). More specifically, traits are said to reach maturity
by age 30 (e.g., Costa & McCrae, 1994; McCrae & Costa, 1999;
McCrae et al., 2000). The predicted stability is expected to last
throughout middle age, though in old age personality could change
again, being disrupted by cognitive decline. A commonly used
metaphor for this pattern of change, based on a passage from
William James (1890/1950), is that personality becomes “set like
plaster”by age 30 (see Costa & McCrae, 1994); thus, we refer to
Postulate 1c, in its general form, as the plaster hypothesis.
In its original formulation, the plaster hypothesis stated that
changes in Big Five traits after age 30 were nonexistent or trivial
(Costa & McCrae, 1994; McCrae & Costa, 1990, 1996). More
recently, the authors of the five-factor theory have indicated that
the plaster hypothesis is “ripe for minor revision”(McCrae &
Costa, 1999, p. 145), as studies have shown changes in mean levels
of personality traits after age 30 (e.g., McCrae et al., 1999, 2000;
see also Roberts et al., in press). They interpret such changes as
stemming from intrinsic biological maturation rather than social
influences, and they still regard the plaster hypothesis as basically
true: “From age 18 to age 30 there are declines in Neuroticism,
Extraversion, and Openness to Experience, and increases in Agree-
ableness and Conscientiousness; after age 30 the same trends are
found, although the rate of change seems to decrease”(McCrae et
al., 2000, p. 183).
Despite this conclusion, no study that we are aware of has
directly tested whether mean levels of the Big Five traits do in fact
change less after age 30 than before. This may be in part because
past research on adult development has compared discrete age
groups, rather than treating age as a continuous variable. For
example, McCrae et al.’s (1999, 2000) two recent cross-sectional
studies reported means for groups of 22- to 29-year-olds and
means for groups of 30- to 49-year-olds, but the studies do not
report the amount of change within those critical age ranges.
We thus set out to test the plaster hypothesis by directly com-
paring rates of change during the relevant age periods. In trans-
lating the plaster hypothesis into formal predictions about rates of
change, we specified two versions of it. We call the original
formulation (as described in Costa & McCrae, 1994) the hard
plaster hypothesis: Age effects after age 30 should not be reliably
different from zero, and this should hold for each of the Big Five
dimensions. We call the more recent “minor revision”(McCrae &
Costa, 1999) the soft plaster hypothesis, because here personality
is like plaster that has not fully hardened but is becoming more and
more viscous: Personality traits change more slowly after age 30
than before age 30.
Contextual Perspectives on Personality and Change
In contrast to the plaster hypothesis, contextual theories predict
that personality changes throughout adulthood (e.g., Haan et al.,
1986; Helson, Mitchell, & Moane, 1984; Neugarten, 1972). By
their very definition, contextual theories are necessarily more
varied than the five-factor theory, but viewed together they predict
different changes in personality during different life periods and, in
some formulations, different changes for men and women (Helson,
Pals, & Solomon, 1997; Wink & Helson, 1993).
Personality Changes as Person–Environment
Transactions
Social roles, life events, and social environments change during
the life course, and such factors have been suggested as important
influences on basic personality traits (Haan et al., 1986; Hogan,
1996). A number of researchers have focused on the transactions
between individuals’personalities and experiences. In the trans-
actional view, individuals are seen as active agents who play an
important role in selecting and shaping their environments, and
these environments in turn affect their personalities. Often these
transactions serve to amplify or strengthen earlier dispositions
(Caspi & Moffitt, 1993). For example, personality traits like Open-
ness and ambition predicted women’s level of involvement in the
women’s movement in the 1960s and 1970s; involvement in the
women’s movement, in turn, led to subsequent increases in Open-
ness and ambition (Agronick & Duncan, 1998).
Research on transactional person–environment processes gen-
erally addresses individual differences in change, but the transac-
tional perspective can be applied to understanding mean-level
change as well. Just as individual differences in personality lead
individuals toward different experiences that subsequently affect
their personalities, normative changes in personality help prepare
people for normative adult roles, which in turn can support further
personality changes. Thus, a transactional perspective on mean-
level change in personality would focus on normative role transi-
tions—that is, transitions experienced by large numbers of people.
Probably the three most important social role domains that
undergo changes in early and middle adulthood are work, marriage
or partnership, and parenting. These three role domains correspond
to the major tasks of adulthood identified by Erikson’s (1950)
theory of adult development: work is involved in the adult task of
consolidating an identity; marriage/partnership in the task of inti-
macy; and parenting children in generativity. Although individuals
differ in the exact timing of when they take on work responsibil-
ities, form committed partnerships, and nurture children, there are
normative age ranges for these roles, suggesting that they may be
linked to typical mean-level personality changes.
Which personality factors are related to these role domains?
Conscientiousness has been linked to working, work performance,
and work commitments (Barrick & Mount, 1991; Barrick, Mount,
& Judge, 2001; Roberts, 1997; Vandewater & Stewart, 1998), and
1042 SRIVASTAVA, JOHN, GOSLING, AND POTTER

to commitment to a stable partner relationship (Neyer & Asen-
dorpf, 2001). Agreeableness should be most closely linked to
parenting and similar generativity-relevant tasks, as exemplified in
nurturing and prosocial behaviors (Graziano & Eisenberg, 1997;
John & Srivastava, 1999). Role transitions in work, partnership,
and childrearing take place throughout early and middle adult-
hood: Normatively, most people enter new jobs in their early 20s
and begin advancing in their careers thereafter (U.S. Census Bu-
reau, 2000a), marry in their mid to late 20s (U.S. Census Bureau,
2000b), and raise children in their 30s (U.S. Census Bureau,
2000c). If the timing of personality changes is linked to the timing
of role transitions, there should be important changes in Consci-
entiousness and Agreeableness, and these changes should be ap-
parent well into the 30s.
Aside from these normative social role changes, other theories
suggest possible changes in personality traits after age 30. People
get better at emotion regulation as they grow older and thus tend
to have fewer negative emotional experiences (Gross et al., 1997);
this could translate into persistently declining levels of Neuroti-
cism with age. Socioemotional selectivity theory (Carstensen,
Isaacowitz, & Charles, 1999) predicts that as adults progress into
middle and later adulthood, they are less and less interested in
gathering new information and in meeting new people, implying
declining Openness and Extraversion, and more interested in re-
lationships with close others, implying increasing Agreeableness.
Gender Roles and Personality Change
Do men and women differ in their development on any of the
Big Five dimensions? Women and men may develop differently
because of gender-based social experiences (Helson et al., 1997;
Stewart & Ostrove, 1998). In particular, there may be develop-
mental differences on Neuroticism. Adolescent girls show higher
levels of Neuroticism than boys (del Barrio, Moreno-Rosset,
Lopez-Martinez, & Olmedo, 1997; Gullone & Moore, 2000; Mar-
galit & Eysenck, 1990). Yet, studies of subsequent development
during middle adulthood indicate that women’s self-confidence
and coping skills improve with age (Helson & Moane, 1987;
Helson et al., 1997), suggesting decreasing levels of Neuroticism
primarily in women.
Few studies of both men and women have directly compared
changes in adult men’s and women’s Neuroticism.
2
However, a
large study of Finnish twins aged 18–59 followed members of
multiple cohorts longitudinally and found that in both longitudinal
and cross-sectional analyses, women decreased in Neuroticism
with age whereas men did not change (Viken, Rose, Kaprio, &
Koskenvuo, 1994). Similarly, a longitudinal study by Wink and
Helson (1993) found that women became less emotionally depen-
dent and more competent with age; in contrast, men started adult-
hood less dependent and more competent than women but then
remained relatively stable on these traits. Thus, we expected that
the gender difference in Neuroticism found in late adolescence and
college-age samples would narrow with age: women should de-
crease in Neuroticism during adulthood, whereas men should not
change much.
In summary, contextual perspectives diverge from the five-
factor theory’s assertion that all of the Big Five follow just one
principle—no change—starting at age 30. Rather, a variety of
developmental processes may affect each Big Five dimension
differently during particular life periods, possibly in different ways
for men and women. Contextual perspectives, viewed together,
offer a metatheoretical counterpoint to the five-factor theory:
Change on the Big Five is complex and multiply determined, and
remains a fact of life well beyond early adulthood. The nature of
change may be different during different periods of adulthood,
resulting in curvilinear age effects (Helson et al., 2002), and men
and women may change in different ways, resulting in Age ⫻
Gender interactions. Thus, we decided to examine the Big Five
with regression models that would test for such differences.
Design of the Present Study
Our interest in testing hypotheses about different age effects
during different developmental periods, and about different age
effects for men and women, raised the issue of statistical power.
Testing the hard and soft versions of the plaster hypothesis re-
quires obtaining slope estimates for limited age ranges, and these
estimates would be unreliable in small samples. Furthermore, tests
of interactions and curvilinear effects have considerably less
power than tests of main effects and linear trends (Chaplin, 1997;
McClelland & Judd, 1993). In short, we needed a large sample to
test our hypotheses. This concern led us to a medium for data
collection that has been available for only a few years, but offers
access to large numbers of willing participants: the Internet. The
Web revolution of the mid-1990s resulted in the massive intercon-
nection of American society to the Internet, making it possible to
reach large numbers of participants. Research on Internet users
indicates that, although they are not perfectly representative of the
general population, they are quite diverse (Lebo, 2000; Lenhart,
2000), probably at least as much as more traditional samples of
undergraduate psychology students or research volunteers re-
cruited through newspaper advertisements or word of mouth.
Thus, we used an Internet sample to examine two issues central
to adult personality development. First, in a sample sufficiently
large to use new analyses well-suited to this question, we directly
tested the hard and soft versions of the plaster hypothesis to see
whether personality does indeed become “set like plaster”after
age 30. Second, to map out the patterning of change in more detail,
we used regression models with curvilinear and interactive effects
to test for different changes during different life periods and for
gender-specific development.
Method
Participants
Participants were part of the Gosling–Potter Internet Personality Project,
a personality study of volunteers recruited and assessed over the World
Wide Web. Personality and demographic data were available for 132,515
participants (54% female) between the ages of 21 and 60; the mean age of
our participants was 31 years (SD ⫽9 years). All selected participants
lived in the United States or Canada (the latter represented 9.2% of the
sample).
2
In a meta-analysis, Feingold (1994) compared the size of the gender
difference between studies that used high school, college, and adult sam-
ples. Because all adult samples were grouped into a single category,
however, this analysis was not sensitive to any changes in the gender gap
after the age of 21.
1043
PERSONALITY DEVELOPMENT IN ADULTHOOD

Respondents reported their ethnicity as one of six categories: 5,710
(4.5%) respondents were Asian, 3,893 (3%) were Black, 2,414 (2%) were
Latino, 2,094 (2%) were Middle Eastern, 110,004 (86%) were White,
and 3,569 (3%) indicated Other; 4% of respondents declined to report their
ethnicity. We re-ran all of the regression analyses using dummy coded
variables for ethnicity; these controls had very little impact on the findings,
and we report the analyses without these control variables.
We added a question about social class during the survey period, so this
information was available for a subset of the sample. Of these 42,578
participants, 405 (1%) reported “poor,”7,614 (18%) “working
class,”23,024 (54%) “middle class,”10,718 (25%) “upper-middle class,”
and 817 (2%) “upper class.”This social-class distribution (19% below
middle class, 27% above) suggests that we were including participants
from a broad range of backgrounds. As with ethnicity, controlling for
social class did not substantially change any of the effects reported in this
article.
Procedure
The data presented here come from a noncommercial, advertisement-
free Web site (www.outofservice.com) that contains personality measures
as well as several games, quizzes, and questionnaires for entertainment
purposes, and was publicized in a number of ways. Potential respondents
could find out about the site through several channels: it could be found
with major search engines under key words like personality tests;itwas
listed on portal sites, such as Yahoo!, under their directories of personality
tests, and at one point was selected as a Yahoo! “Pick of the Week”; and
individuals who had previously visited the Web site and signed up for its
mailing list received notification when the Big Five survey was added. As
is common on the Internet, news of the site apparently also spread quite
widely through informal channels, such as e-mails or unsolicited links on
other Web sites.
Computerized administration means that data entry and scoring are
automated; thus, it is possible to recruit participants by appealing to their
motivation to receive individualized personality feedback, for the purpose
of self-insight or entertainment. To attract as broad and diverse a sample as
possible, and to examine the effects of different recruiting approaches, we
used two distinct Web pages. One was entitled “All About You—A Guide
to Your Personality”and was said to measure “what many personality
psychologists consider to be the fundamental dimensions of personality.”
The second Web page was entitled “Find Your Star Wars Twin,”with
feedback provided about “the characters from Star Wars with whom you
are most similar (based on the Big Five personality test).”Instructions for
both versions were the same and reminded participants to answer honestly
to get accurate feedback. We consider later in the Results section how these
two Web pages differed, both in who they attracted and in the substantive
findings.
Measurement
The Big Five Inventory (BFI). To measure the Big Five personality
dimensions, we used the BFI (John, Donahue, & Kentle, 1991). The 44 BFI
items consist of short and easy-to-understand phrases to assess the proto-
typical traits defining each of the Big Five dimensions, making it ideal for
a large survey where we could expect respondents to devote a limited
amount of time. The BFI scales have shown substantial internal consis-
tency, retest reliability, and clear factor structure, as well as considerable
convergent and discriminant validity with longer Big Five measures
(Benet-Martinez & John, 1998; John & Srivastava, 1999). The scales have
also shown substantial agreement between self- and peer-reports (John &
Paulhus, 2003; Rammstedt & John, 2003). BFI items are rated on a 5-point
scale ranging from 1 ⫽disagree strongly to5⫽agree strongly (the full
BFI is reprinted in John & Srivastava, 1999).
For this report, we scored the BFI in an intuitive metric known as
percentage of maximum possible (POMP) scores (P. Cohen, Cohen, Aiken,
& West, 1999). A POMP score is a linear transformation of any raw metric
intoa0to100scale, where 0 represents the minimum possible score and
100 represents the maximum possible score; P. Cohen et al. (1999) rec-
ommended POMP scores as a universal metric that is more intuitive than
scale scores with idiosyncratic ranges. In the case of the BFI, we trans-
formed the 1-to-5 BFI metric into POMP scores by subtracting 1 and
multiplying by 25. Sample means and standard deviations, in POMP units,
were as follows: Conscientiousness M⫽63.8, SD ⫽18.3; Agreeableness
M⫽66.4, SD ⫽18.0; Neuroticism M⫽51.0, SD ⫽21.9; Openness
M⫽74.5, SD ⫽16.4; Extraversion M⫽54.6, SD ⫽22.6.
3
Reliabilities, scale intercorrelations, and structural invariance across
age groups. Reliabilities and scale intercorrelations were of special in-
terest in this Internet sample. If there were problems with administering the
BFI on the Internet, such as many random or otherwise unreliable re-
sponses, the coefficient alpha reliabilities of the five scales should be
considerably lower. In contrast, attempts to self-enhance for the sake of
receiving positive feedback should result in higher intercorrelations among
the scales. Our results showed that neither was the case. First, the alpha
reliabilities were very similar to earlier data (see John & Srivastava, 1999):
␣
⫽.82 in the Internet sample (compared with .82) for Conscientiousness;
.79 (.79) for Agreeableness; .84 (.84) for Neuroticism; .80 (.81) for Open-
ness; and .86 (.88) for Extraversion. The scale intercorrelations were also
similar to previous research; the mean of the absolute discriminant corre-
lations among the BFI scales was .16 as compared with .20 reported by
John and Srivastava (1999), and the highest correlation between any two
scales was only .29, as compared with .33.
Another concern was whether the BFI structure was invariant across
ages; if the pattern of factor loadings was different at different ages, that
would complicate the task of comparing scale scores. To test for this, we
split the sample into four age groups spanning a range of 10 years each,
then conducted factor analyses within each age group, extracting five
factors in each analysis. The Big Five factors clearly replicated within each
age group. We then computed factor congruence coefficients between age
groups; the average congruence coefficient for the same conceptual factor
across age groups was .99, reflecting both a high degree of structural
invariance and the unusually low sampling error in such a large sample.
Furthermore, we also computed the scale reliabilities separately by age and
found that they did not vary with age.
Results
How did scores on the Big Five personality dimensions change
with age? We report our analyses in two sections. First, we
examined the two versions of the plaster hypothesis by (a) testing
whether age slopes after age 30 are different from zero (hard
plaster) and (b) comparing age slopes after age 30 to those before
age 30 (soft plaster). Second, we tested models of the data that
allow curvilinear age effects and gender differences in the magni-
tude of age effects, using regressions with polynomial age effects
and gender interaction terms. We also estimated these models
separately for the two Web page formats, to see whether effects
generalized across recruitment strategies.
Is Personality Fixed After Age 30? Testing the Plaster
Hypothesis
The hard plaster hypothesis asserts that there should be no age
effects on any Big Five dimension after age 30; the soft plaster
hypothesis asserts that age effects after age 30 should be weaker
3
Means and standard deviations broken down by age, gender, and Web
page are available by request from the authors.
1044 SRIVASTAVA, JOHN, GOSLING, AND POTTER

than age effects before age 30. To test these hypotheses, we
computed age slopes (i.e., regression coefficients from the Big
Five dimensions regressed on age) within the two theoretically
important age ranges, 21–30 and 31–60; such slopes indicate how
much the predicted Big Five score increased or decreased per year.
The slopes and their 95% confidence intervals are presented in
Table 1.
All slopes were computed separately for men and for women.
Because the plaster hypothesis makes the same predictions for men
and for women, results ought to replicate across gender. Statistical
tests of both versions of the plaster hypothesis are reported in
Table 1. The test for the hard plaster hypothesis is the standard t
test of whether the slope for ages 31–60 was different from zero.
To test the soft plaster hypothesis, we used a ztest (Equa-
tion 3.6.11 in J. Cohen & Cohen, 1983) to compare whether the
slope for ages 21–30 was stronger than the slope for ages 31–60.
4
Conscientiousness. As can be seen in Table 1, the raw Con-
scientiousness slope for men during ages 31–60 was B⫽.31, and
the slope for women was B⫽.26. Both of these slopes were
significantly and substantially different from zero, a clear rejection
of the hard plaster hypothesis. The soft plaster hypothesis predicts
that the slopes from age 31 to 60 should not be as strong as the
slopes from age 21 to 30. For both men and women, the earlier
slope (B⫽.46 for women and B⫽.48 for men) was stronger. The
ztests indicated that the differences between the slopes were
significant, supporting the soft plaster hypothesis for Conscien-
tiousness. In short, this pattern indicates that people changed less
in Conscientiousness after age 30 than before age 30, but they
clearly did not stop changing.
Agreeableness. The results for Agreeableness disagreed
sharply with both versions of the plaster hypothesis, as is evident
in Table 1. Agreeableness increased significantly from age 31
to 60 for both men and women, contradicting the hard plaster
hypothesis. Moreover, the age slopes for ages 31–60 were sub-
stantially greater than the age slopes for ages 21–30; that is, the
data not only failed to support the soft plaster hypothesis, but they
went significantly in the opposite direction.
Neuroticism. Neuroticism yielded different results for men
and for women. For men, the slope for ages 31–60 was not
significantly different from zero, consistent with the hard plaster
hypothesis; in fact, men did not show a significant age effect in
either age period. Women, however, declined consistently in Neu-
roticism. The slope for ages 31–60 was significantly different from
zero, contradicting the hard plaster hypothesis. The age 21–30
slope for women was not significantly weaker than the age 31–60
slope, a failure to support the soft plaster hypothesis.
4
The hard plaster hypothesis predicts a zero effect, that is, a null
hypothesis; thus, a significant result of this test would reject the hard
plaster hypothesis. In contrast, the soft plaster hypothesis predicts a direc-
tional effect, that is, an alternative to a null hypothesis; thus, a significant
result of this test would support the soft plaster hypothesis. The different
logic of these two tests means that statistical power makes them sensitive
in opposite ways. Because of the large sample size, even a very weak
change after age 30 would lead to a rejection of the hard plaster hypothesis;
but the large sample size also means that even a very weak soft-plaster
effect would lead to a confirmation of the soft plaster hypothesis.
Table 1
Linear Slopes (and 95% Confidence Intervals) for the Age Effect on Each of the Big Five Factors, Computed Separately by Gender
and Age Period
Big Five factor Age 21–30 Age 31–60 ttest to reject
hard plaster ztest to confirm
soft plaster Implication of tests
Conscientiousness
Women .48 (⫾.06) .26 (⫾.03) 18.4*** 6.6*** Change slows but does not stop after age 30
Men .46 (⫾.06) .31 (⫾.04) 17.2*** 4.2*** Change slows but does not stop after age 30
Agreeableness
Women .10 (⫾.06) .28 (⫾.03) 21.1*** ⫺5.4*** Change increases after age 30
Men ⫺.01 (⫾.07) .20 (⫾.04) 11.1*** ⫺5.6*** Change increases after age 30
Neuroticism
Women ⫺.25 (⫾.07) ⫺.25 (⫾.03) ⫺14.5*** ⫺0.2 Change is similar before and after age 30
Men ⫺.06 (⫾.08) ⫺.03 (⫾.04) ⫺1.4 ⫺0.7 Little change before or after age 30
Openness
Women .04 (⫾.06) ⫺.04 (⫾.03) ⫺3.1** ⫺0.2 Change is of similar strength but in opposite
directions before and after age 30
Men .04 (⫾.05) ⫺.15 (⫾.03) ⫺9.4*** ⫺3.3*** Change increases after age 30
Extraversion
Women .09 (⫾.08) ⫺.07 (⫾.04) ⫺4.0*** 0.5 Change is of similar strength but in opposite
directions before and after age 30
Men .14 (⫾.08) .05 (⫾.04) 2.1* 2.0* Change slows but does not stop after age 30
Note. Sample size for age 21–30: women, n⫽41,840; men, n⫽40,831. Sample size for age 31–60: women, n⫽30,027; men, n⫽19,817. Numbers
in parentheses are 95% confidence intervals. The metric for the slopes is the percentage of maximum possible units per year. Negative values of the ztest
indicate that the direction of the effect was contrary to that predicted by the soft version of the plaster hypothesis.
*p⬍.05. ** p⬍.01. *** p⬍.001.
1045
PERSONALITY DEVELOPMENT IN ADULTHOOD

Openness. Openness slopes for both men and women were
significantly negative after age 30, contradicting the hard plaster
hypothesis. However, age effects for Openness did not differ
significantly from zero for the age 21–30 decade. Because both
men and women marginally increased in Openness up to age 30
and then decreased in Openness after age 30, we tested the differ-
ence of the absolute slopes to see if the magnitude was weaker
after age 30. For men, the decline after age 30 was stronger than
the increase up to age 30, a significant rejection of the soft plaster
hypothesis. For women, the magnitude of the decline after age 30
was not significantly different from the magnitude of the increase
up to age 30, a failure to support the soft plaster hypothesis.
Extraversion. Extraversion decreased significantly from age 31
to 60 for women; the increase for men from age 31 to 60 was weak
and barely significant (p⫽.04) by conventional criteria. These
findings contradicted the hard plaster hypothesis, though given the
statistical power of the analysis, the weak result for men was not
a resounding rejection. For men, the soft plaster hypothesis was
confirmed for Extraversion. For women, though, the absolute
strength of the increase from age 21 to 30 was not different from
the absolute strength of the decrease from age 31 to 60, failing to
support the soft plaster hypothesis.
Overall, then, we did not find widespread support for either
version of the plaster hypothesis. Out of 10 tests of each hypoth-
esis—five dimensions tested separately for men and women—
only one fit the hard version (two if the result for men’s Extra-
version is counted) and only four fit the soft version.
The Pattern of Big Five Development: Modeling Age and
Gender Effects
Figure 1 shows the mean scores on each of the Big Five
dimensions for each age, separately for men and women. We used
regression models to map out the relations of age and gender to
personality. In fitting these regression models, the large sample
allowed for very sensitive tests of polynomial (curvilinear) and
interactive effects. However, a model selection process in which
one continues to add polynomial terms as long as they meet
conventional significance criteria was likely to produce unrepli-
cable, uninterpretable, and unnecessarily complex models in such
a large sample. Examining previous studies of curvilinear age
effects on personality (e.g., Haan et al., 1986; Helson et al., 2002),
we concluded that cubic (i.e., third-order) models with gender
interactions seemed like the most complex models that could make
a substantive contribution, so we set this as a practical limit on
model complexity. Thus, for each Big Five factor, we considered
three possible models:
Linear: B5 ⫽b0⫹b1共AGE兲⫹b2共GEN兲⫹b3共AGE*GEN兲.
Quadratic: B5 ⫽b0⫹b1共AGE兲⫹b2共GEN兲⫹b3共AGE*GEN兲
⫹b4共AGE兲2⫹b5共AGE2*GEN兲.
Cubic: B5 ⫽b0⫹b1共AGE兲⫹b2共GEN兲⫹b3共AGE*GEN兲
⫹b4共AGE兲2⫹b5共AGE2*GEN兲⫹b6共AGE兲3⫹b7共AGE3*GEN兲.
(In the above equations, B5 stands for the Big Five dimension
being modeled, AGE represents age centered around its mean, and
GEN is a contrast code for gender.) To select from among these
three models, we set the criterion that a more complex model
would be retained only if it improved fit at F⬎25 ( p⬍10
–5
) over
a simpler model. We used this stringent cutoff to select models that
could realistically be replicated and examined further in future
studies with smaller samples.
Conscientiousness. Patterns of change in work and partnership
suggested that the most pronounced increases in Conscientious-
ness might occur during the 20s, followed by continuing growth at
a slower rate. In statistical terms, this implied that the age effect on
Conscientiousness would follow a decelerating function, with a
steeply increasing slope in early adulthood that becomes flatter at
later ages. Our decision criteria supported a quadratic model for
Conscientiousness (see Table 2). A positive linear age term indi-
cated that people were increasing in Conscientiousness at all ages,
but a negative quadratic age term indicated that the rate of increase
was greater at younger ages than at older ages. None of the age
terms interacted with gender, indicating that men and women did
not change in Conscientiousness at different rates. In Figure 1A,
the quadratic functions for men and women are plotted along with
the observed Conscientiousness means at each age. The close
correspondence between the function and the means from the raw
data suggests that the quadratic function captures the normative
trend quite well.
Agreeableness. If Agreeableness is linked to nurturing and
raising children, we would expect that it would increase the most
during the late 20s and 30s. Because such an increase would occur
in the middle of the age range being analyzed, a cubic model
would be necessary to fit such an effect. The data provided clear
support for a cubic model (see Table 2) under the decision criteria;
the sample means and the cubic fit line are plotted in Figure 1B.
The increase in Agreeableness accelerated in the late 20s, and
Agreeableness continued to increase rapidly through the 30s be-
fore slowing down (but continuing to increase) in the 40s; thus, the
period of most rapid increase coincided with the ages at which
people are typically giving birth and nurturing their dependent
children. There was a significant interaction between gender and
the first-order age term, but no gender interactions with the qua-
dratic or cubic term (see Table 2); in other words, women in-
creased overall in Agreeableness more than men did, but the
degree of curvature in the functions did not differ substantially
between men and women. Though we did not predict it, this would
be consistent with women, on average, being involved in nurturing
roles to a greater extent than men.
Neuroticism. Neuroticism was the strongest candidate for a
Gender ⫻Age interaction. The decision criteria indicated that
Neuroticism was well described by the linear model. This model,
along with the observed Neuroticism means, is graphed in Figure
1C. The significant Age ⫻Gender interaction (see Table 2)
revealed that women declined substantially in Neuroticism
throughout adulthood; men declined quite modestly. By late adult-
hood, the earlier gender difference had nearly disappeared.
Openness. For Openness, our model-selection criteria resulted
in a linear model (see Table 2; a quadratic model would have
improved the fit but only at F[2, 132509] ⫽10.7). The linear
model was consistent with previous developmental findings—both
men and women declined in Openness with age, but only slightly.
A small, unpredicted interaction suggested that men began adult-
hood slightly higher in Openness but then declined at a faster rate
(see Figure 1D).
1046 SRIVASTAVA, JOHN, GOSLING, AND POTTER

Figure 1. Mean Big Five scores broken down by age and gender, with fit curves from the regression models
(see Table 2). Vertical lines at age 30 indicate when the plaster hypothesis predicts that personality stops
changing. POMP ⫽percentage of maximum possible.
1047
PERSONALITY DEVELOPMENT IN ADULTHOOD

Extraversion. Extraversion was best fit with a linear model
(see Table 2). An unpredicted Age ⫻Gender interaction indicated
that men increased slightly in Extraversion with age whereas
women decreased slightly, resulting in a diminishment of gender
differences with age (see Figure 1E).
Replication of the models across Web site formats. The
p-values associated with the tests of the polynomial models are
generally quite small, some so small as to exceed the computa-
tional limits of our data analysis software (SPSS 10.0.7, which ran
out of decimal places at p⬍10
–22
). Small p-values are one
indicator of the reliability of an effect; cross-validation is another.
The study design allowed for an internal replication analysis
between the two versions of the questionnaire.
Were the respondents to the two Web pages different enough to
constitute replication samples? Although we would not want to
overstate the differences, the Web pages arguably appealed to
somewhat different motives of potential participants. As the name
implies, All About You was described as an opportunity to learn
about oneself, whereas Star Wars was designed to appeal primarily
as an intriguing and fun experience. The two sites did, in fact, draw
somewhat different profiles of participants. All About You drew
66% women and 34% men, whereas Star Wars drew 39% women
and 61% men. Respondents to All About You were on average
about 2 years older than respondents to Star Wars (equivalent to an
effect size of r⫽.11). Controlling for gender, the greatest
between-sites difference on the Big Five was for Openness; Star
Wars respondents tended to be slightly more open to experience
(partial r⫽.10).
In spite of these differences, though, the regression models
clearly replicated across the two Web formats; Table 2 presents the
standardized betas from these replication analyses for comparison
purposes. When we analyzed the respondents for the two Web
formats separately, we found that for Conscientiousness, Agree-
ableness, and Neuroticism, all the same terms were significant in
the same direction and general magnitude, and the shape of the
plotted curves was remarkably similar. For Openness, the linear
term representing the predicted decline replicated across both Web
formats; the unpredicted Age ⫻Gender term found in the full
sample was not significant in the Star Wars data. For Extraversion,
the main effect of gender and the Age ⫻Gender interaction
replicated across both Web formats. These replications further
underscored the reliability of the effects.
Discussion
In this article, we examined the relation between age, gender,
and personality traits in adulthood. We took advantage of a large
sample size and continuous age distribution to test hard and soft
Table 2
Regression Models of the Relations of Age and Gender to the Big Five
Regression term BSE
␤
pReplication
␤
s
Conscientiousness (R⫽.18)
Constant 63.984 .066
Age 0.341 .008 .16 ⬍10
⫺22
.19/.13
Gender ⫺1.847 .066 ⫺.10 ⬍10
⫺22
⫺.09/⫺.12
Age ⫻Gender ⫺0.005 .008 ⫺.00 ⬎.50 .00/.00
Age
2
⫺0.004 .001 ⫺.03 ⬍10
⫺10
⫺.03/⫺.02
Age
2
⫻Gender 0.001 .001 .01 ⬎.09 .00/.01
Agreeableness (R⫽.16)
Constant 65.735 .076
Age 0.244 .009 .12 ⬍10
⫺22
.13/.10
Gender ⫺1.829 .076 ⫺.10 ⬍10
⫺22
⫺.10/⫺.10
Age ⫻Gender ⫺0.051 .009 ⫺.03 ⬍10
⫺7
⫺.03/⫺.02
Age
2
0.011 .001 .08 ⬍10
⫺19
.11/.03
Age
2
⫻Gender 0.001 .001 .01 ⬎.39 .02/.01
Age
3
⫺0.0005 .000 ⫺.08 ⬍10
⫺17
⫺.11/⫺.05
Age
3
⫻Gender ⫺0.0000 .000 ⫺.00 ⬎.77 ⫺.01/⫺.01
Neuroticism (R⫽.23)
Constant 50.690 .059
Age ⫺0.125 .007 ⫺.05 ⬍10
⫺22
⫺.06/⫺.05
Gender ⫺4.937 .059 ⫺.22 ⬍10
⫺22
⫺.19/⫺.22
Age ⫻Gender 0.111 .007 .04 ⬍10
⫺22
.05/.03
Openness (R⫽.11)
Constant 74.628 .045
Age ⫺0.082 .005 ⫺.04 ⬍10
⫺22
⫺.05/⫺.02
Gender 1.542 .045 .09 ⬍10
⫺22
.09/.04
Age ⫻Gender ⫺0.021 .005 ⫺.01 ⬍10
⫺4
⫺.02/⫺.00
Extraversion (R⫽.11)
Constant 54.442 .062
Age ⫺0.006 .007 ⫺.00 ⬎.44 .00/⫺.01
Gender ⫺2.410 .062 ⫺.11 ⬍10
⫺22
⫺.10/⫺.11
Age ⫻Gender 0.038 .007 .02 ⬍10
⫺6
.01/.02
Note. N ⫽132,515. Replication
␤
s are from the All About You and Star Wars versions of the questionnaire,
respectively. Age is mean-centered at 30.6. Gender is contrast-coded: female ⫽⫺1, male ⫽1. All terms are
from final models.
1048 SRIVASTAVA, JOHN, GOSLING, AND POTTER

versions of the plaster hypothesis, which assert that personality
change stops (hard plaster) or slows (soft plaster) after age 30. On
no Big Five dimension did we find support for the hard plaster
hypothesis among both men and women, and only Conscientious-
ness demonstrated the soft plaster effect for both genders. To test
for more complex change patterns, we then fitted curvilinear and
interactive models to the data that yielded clear and robust effects.
Conscientiousness increased throughout the age range studied,
most strongly during the 20s; Agreeableness increased the most
during the 30s; and Neuroticism declined with age for women but
not much for men. Openness showed small declines with age, and
Extraversion declined for women but not men.
Addressing Sampling and Cohort Concerns
In a study that uses a cross-sectional design to make inferences
about developmental effects, differential sampling by age and
cohort differences are both potential sources of confounds. Al-
though no cross-sectional design can completely rule out such
factors, they need not be fatal flaws if results are interpreted in
relation to existing theory and research. In this section, we con-
sider several factors that might address concerns about sampling
and cohort confounds. We then compare age trends in the Internet
sample with those found in recently published data to see whether
the Internet results are comparable with those found with more
traditional sampling methods and in other countries.
Sampling considerations. One concern was that Internet re-
cruitment might produce age-confounded sampling biases.
5
Younger Internet users might be a fairly broad group, whereas
older Internet users might be a more select subset of older people
in general; if so, age effects might be artifacts of who was likely
to end up in our study rather than true age effects. In part, our focus
on early and middle adulthood provided some assurance against
such a possibility: Survey research on Internet usage and age
indicates that although there is a sharp change in usage patterns
around retirement age, people of preretirement age (i.e., ages
50–64) have similar access rates to younger people, and preretire-
ment Internet users are representative of the overall Internet pop-
ulation in how they spend their time online (Fox et al., 2001). We
were also reassured by our comparisons of the two Web page
recruiting strategies. Age effects replicated across the self-
learning-oriented “All About You”page and the entertainment-
oriented “Find Your Star Wars Twin”page, suggesting that if there
was a selection bias in our sampling, it would have to have been
identical across recruiting strategies.
We also considered what age-confounded sampling bias would
look like if it had happened. A possible scenario was that the
Internet may be less familiar to older people, and thus older people
may have to be higher in Openness in order to seek out and
participate in an online study. In fact, we found a small drop in
Openness with age, not the increase one would predict if older
adults needed to be especially open in order to participate. If there
was any Openness-related sampling bias, it was sufficiently weak
that it did not overwhelm the age effect.
6
Untangling developmental effects from secular trends and co-
hort effects. Schaie (1977) discussed three kinds of effects that
developmental researchers need to consider in designing studies:
(a) effects of the individual’s development, which are frequently of
primary interest to developmental researchers; (b) secular trends
(i.e., changes in current social climate), which can cause variation
between measurements taken at different historical times; and (c)
cohort effects (i.e., generational differences), which can cause
variation between people born and raised in different historical
periods. Most common designs cannot fully disentangle these
effects: Cross-sectional designs cannot intrinsically differentiate
between developmental and cohort effects; longitudinal designs
cannot separate developmental and historical effects; and time-lag
designs (which compare different samples measured in different
years) cannot separate historical and cohort effects. However,
jointly considering different designs can provide insights. If the
results of a cross-sectional study agree with results from longitu-
dinal studies, they can be interpreted as arising from development,
which is the only common effect between the two designs; and if
cross-sectional results agree with results from time-lag studies,
they can similarly be interpreted as arising from cohort differences.
(The two interpretations are not mutually exclusive.)
In this case, the Internet findings agree with the broad trends
among longitudinal studies reviewed by Roberts et al. (in press):
Conscientiousness and Agreeableness went up, Neuroticism went
down (their review did not differentiate men and women), and
Openness and Extraversion changed little. Time-lag studies of the
Big Five are far rarer than cross-sectional and longitudinal studies;
two important recent meta-analyses offer a comparison source for
Neuroticism (Twenge, 2000) and Extraversion (Twenge, 2001).
These studies examined how, across studies, publication year
correlated with sample means for anxiety and Extraversion. Both
traits increased over time from the middle to late 20th century. If
these results were due to cohort effects (rather than secular trends),
that would show up in the cross-sectional Internet data as older
people being both lower in Neuroticism and lower in Extraversion
than younger people. The former was true for women but not for
men, and the latter did not appear to be the case. Thus, compari-
sons suggest that at the level of general trends, it seems most
appropriate to attribute the Internet results to developmental ef-
fects, though we cannot rule out the possibility that cohort differ-
ences may contribute additionally to explaining the age effects. It
is important to note that both developmental effects and cohort
effects are inconsistent with the five-factor theory; cohort effects
represent an environmental influence that traits are supposed to be
immune to. Thus, the age effects obtained here are inconsistent
5
Another potential sampling confound, not specific to an Internet de-
sign, could be selective mortality. Individuals high in Conscientiousness
tend to die earlier (Friedman et al., 1993); thus, cross-sectional increases in
Conscientiousness could occur because at older ages, fewer low-
Conscientiousness individuals are available as participants. To test for such
effects, we examined the standard deviations computed separately at each
age. If age effects were due to selective mortality, the standard deviations
should get smaller with increasing age, as the less conscientious partici-
pants were selected out. In fact, age was not related to the standard
deviation, suggesting that the Conscientiousness effect was not driven by
selective mortality.
6
Again, the standard deviation of Openness did not increase with age, as
would be predicted by the hypothesis that the sampling procedure favored
participants higher in Openness at older ages.
1049
PERSONALITY DEVELOPMENT IN ADULTHOOD

with the five-factor theory regardless of how such effects are
substantively interpreted.
Addressing sampling and cohort effects through empirical com-
parisons. Another way to address concerns about this new
method is through comparisons with data from studies that use
more traditional approaches. If age patterns in the Internet data
matched those found in cross-sectional data collected by other
sampling methods, such cross-method agreement would suggest
that the Internet results were not a product of sampling confounds.
Furthermore, some researchers have argued that cohort effects, if
they exist, should be culture-specific (Yang, McCrae, & Costa,
1998); if so, then similarities to cross-cultural findings should
diminish concerns about cohort confounds.
Data in two recent articles by McCrae et al. (1999, 2000) stood
out as a particularly appropriate source for comparing main effects
of age. These articles, which used the Revised NEO Personality
Inventory (NEO PI-R) and the NEO Five Factor Inventory (NEO-
FFI) measures (Costa & McCrae, 1992), reported Big Five means
for several adult age groups, allowing us to make comparisons of
the direction of change and, to some extent, the approximate shape
of change. Big Five self-report data were reported separately for
nine different countries (i.e., Britain, Croatia, Czechoslovakia,
Germany, Italy, Portugal, South Korea, Spain, and Turkey). These
data had been collected under a wide variety of recruitment and
sampling procedures, including a twin study, a study of parents
and their teenage children, and several other convenience samples.
We performed secondary analyses of the results reported from
these nine samples,
7
aggregating the Tscores across the nine
countries into averages (weighted by sample size) for the three age
groups used by the original authors: 22–29, 30–49, and 50⫹.We
were thus able to derive rough age trends (based on three data
points) across the nine samples; we refer to these aggregated data
as the NEO-International data. Results from data aggregated across
these various cultures and recruiting procedures probably reflect a
conservative number of reasonably robust effects, providing a
stable baseline for making comparisons.
Although the BFI and NEO questionnaires are mostly similar,
the instruments do define two of the five factors, Openness and
Extraversion, somewhat differently (John & Srivastava, 1999).
Thus, we also wanted to compare results with a sample that used
the BFI. To do this, we derived age trends from published Big Five
data collected in Germany (Lang, Luedtke, & Asendorpf, 2001).
Participants from the Berlin, Germany, metropolitan area were
recruited by a gold-standard procedure, stratified random sampling
(stratification was on age and gender). They completed a German
translation of the BFI. Only two age groups (20–40 and 45–65) in
the BFI-German data overlapped with the Internet sample, so we
could examine only one age interval.
Results of our secondary analyses of the NEO-International data
and the BFI-German data are presented in Table 3. On Conscien-
tiousness, both comparison samples showed increases throughout
adulthood that were greater in earlier than in later adulthood,
consistent with the Internet results presented in this report. Agree-
ableness also went up in both comparison samples; the NEO-
International data showed similar-sized increases during both in-
tervals, which is consistent with the cubic trend found in the
Internet data (with only three data points, it is impossible to
differentiate a cubic trend from a linear trend; this can be demon-
strated by placing a straight ruler on Figure 1B such that it crosses
the cubic curve at three points). Neuroticism declined in the
comparison samples, but because the reports on the comparison
samples did not provide results separately for men and women, it
is not possible to know whether the trend was driven primarily by
women, as it was in our sample. Openness and Extraversion
declined in the NEO-International data but did not change sub-
stantially in the BFI-German; the latter is consistent with findings
in the Internet data. Thus, the general direction of effects in the
Internet data was similar to that of studies collected with different
methods.
Nevertheless, none of the above considerations completely rule
out sampling or cohort confounds, and the present study should not
be taken as the last word about Big Five development in early and
middle adulthood. Future longitudinal and sequential studies are
needed to further address the possibility of such confounds. More
importantly, we hope that such studies, which are more time-
intensive and expensive than cross-sectional studies, can use the
present results as a springboard to generate and test focused
hypotheses about the timing and theoretical significance of
changes in personality traits.
Personality Theories and Adult Development
We tested predictions about Big Five development made from
two perspectives: the biologistic Five-factor theory, and ap-
proaches that allow for contextual influences on traits. What do the
results say about personality theories and adult development?
7
A subset of the German data reported in McCrae et al. (1999) was also
included in McCrae et al. (2000). Therefore, only the 1999 data were used
for the comparison baselines.
Table 3
Relative Magnitude of Annual Change During Two Age
Intervals
Big Five dimension Interval 1 Interval 2
Age groups that define interval
NEO-International 22–29 vs. 30–49 30–49 vs. 50⫹
BFI-German 20–40 vs. 45–65
Conscientiousness
NEO-International 11 1
BFI-German 11
Agreeableness
NEO-International 11
BFI-German 1
Neuroticism
NEO-International 222
BFI-German 2
Openness
NEO-International 22
BFI-German 0
Extraversion
NEO-International 22 2
BFI-German 0
Note. The number of arrows indicates the relative rates of change within
each row. NEO-International data come from McCrae et al. (1999, 2000);
BFI-German data come from Lang et al. (2001), NEO ⫽NEO Personality
Inventory, Revised; BFI ⫽Big Five Inventory.
1050 SRIVASTAVA, JOHN, GOSLING, AND POTTER

Rejection of the plaster hypothesis and its implications. Over-
all, we found a general lack of support for the plaster hypothesis,
and considerable evidence that directly contradicted it. Mean lev-
els of personality traits changed gradually but systematically
throughout the life span, sometimes more after age 30 than before.
By standardized effect size indices—multiple Rs—the effects of
age and gender on personality can be characterized as small to
moderate (J. Cohen, 1992). So how should these findings be
interpreted? The multiple Rtreats individual differences among
participants of the same age and gender as error variance in the
denominator. Thus, the Ris essentially an index of predictability
(Wiggins, 1973), quantifying the answer to the question, “How
well can I predict an individual’s personality knowing only age
and gender?”Against the backdrop of all possible predictors of
personality—thousands of genes, the influence of peers and fam-
ily, a lifetime of individually constructed experiences—it should
not be surprising that age and gender provide at best a small to
moderate amount of predictive information. In this context, the age
effects reported here are indeed modest, but nevertheless constitute
clear evidence of ongoing change.
The rejection of the plaster hypothesis also involves a rejection
of two of its implications. The plaster hypothesis, as stated by the
five-factor theory, is a general law of adult development: thus, it
implies that all Big Five dimensions develop in the same ways, and
that men’s and women’s development is the same. Both of these
corollaries were contradicted by the evidence in this study, which
showed that the course of development was different for different
traits, and sometimes different for men and women. This suggests
that personality traits cannot be easily explained by a small, tidy
set of principles.
8
Personality theories and traits. Perhaps ironically, personality
theories that place traits at the root of everything else that is
interesting about personality (McCrae & Costa, 1999) and those
that consider them of little interest (Mischel & Shoda, 1995) share
a common assumption: traits are separate from the rest of the
personality system, static and unresponsive to social environments
and life contexts. The personality system has been compared with
a multistory house, with traits at the bottom and other aspects of
personality—motivations, adaptations, personal narratives, and so
on—built above (Little, 1996; see also McAdams, 1996). If we
follow that metaphor, then the results in this article suggest what
might be called a Californian view of the personality system—in
California, even foundations shift a little bit year by year (and
some get shaken up every once in a while).
What does this view imply for theories of personality? Strictly
speaking, the evidence in this article does not by itself rule out
strictly biological accounts of personality, though it does contra-
dict the five-factor theory’s brand of biologism. Could the changes
found here result solely from intrinsic, genetically determined (and
sexually dimorphous) maturation? There is no denying that genes
contribute to individual differences in Big Five trait levels (Loeh-
lin, McCrae, Costa, & John, 1998). However, longitudinal behav-
ior genetic studies, which can examine both typical levels of traits
and patterns of change in traits, have suggested that change in
personality traits is at least as well explained by environments as
by genes, and heritabilities of personality traits diminish from early
to middle adulthood (Loehlin, 1992; McGue, Bacon, & Lykken,
1993; Viken et al., 1994). Baltes (1997) referred to the “incomplete
architecture of human ontogeny,”by which he meant that the
legacy of biological evolution has provided the most structure for
the earlier part of the life span, when people have not yet survived
long enough to pass along their genes; as individuals get older,
culture becomes a more important influence on how well they
adapt to their environments. Along with Baltes, we believe that
adult personality is characterized more by plasticity than by in-
creasing calcification, and the mechanisms of personality change
can be understood best by considering the life contexts that ac-
company change.
In this study, Conscientiousness showed substantial change
throughout early and middle adulthood, with the strongest effects
in earlier adulthood; this is the time of life when adults are entering
and advancing in the work force and forming committed partner-
ships, life tasks that have been linked to changes in Conscientious-
ness (Neyer & Asendorpf, 2001; Roberts, 1997). Agreeableness
showed the largest changes somewhat later, when adults are typ-
ically caring for children. Neuroticism declined only for women,
who probably had faced particularly difficult social environments
earlier in their lives (Orenstein, 1994).
Several processes may link these personality and environmental
changes to one another. In the case of Conscientiousness, for
example, maturational changes may prompt individuals to seek out
roles that fit their newly changed personalities. Conversely, new
work responsibilities and relationship commitments resulting from
adult social roles may require individuals to become more orga-
nized and responsible in order to meet the demands of their new
situations. Or, both mechanisms may take place as part of a
transactional system—personality changes may lead people to
select and shape environments that further reinforce their person-
alities (Caspi & Roberts, 1999). Future research should further
examine these processes and address more precisely what aspects
of social roles and environments are transacting with personality
traits. For example, are aspects of work, per se, important in the
development of Conscientiousness, or is the importance of work
more a matter of work’s place in broader life themes (such as being
self-sufficient and taking over financial responsibilities previously
assumed by parents)? Similarly, is there something specific about
raising children that promotes Agreeableness (in which case the
minority of adults who do not have children might prove an
8
The size of developmental effects has sometimes been indexed by the
dstatistic, expressing the difference between older and younger partici-
pants in standard deviation units. One disadvantage of this approach is that,
like R, it puts all competing sources of variance (including measurement
error) in the denominator; moreover, dstatistics are better suited for truly
categorical variables than for arbitrarily selected age groups. Nevertheless,
for comparison purposes, we computed ds for each of the Big Five
separately for men and women, using the difference between the highest
and lowest point of each fit curve. Averaging absolute ds across all five
dimensions, women changed about .33 standard deviations and men
changed about .24 standard deviations. According to general scientific
usage, these are medium to small effects (J. Cohen, 1992); that is, these
effects do not indicate little or no change, as argued by McCrae and Costa
(1996, 1999). Furthermore, averaging across the Big Five goes against an
important finding for this study, namely that change differs importantly
across the Big Five. In fact, we found a substantial range of effect sizes;
absolute ds ranged from 0.07 (Extraversion) to 0.54 (Conscientiousness)
for women, and from 0.03 (Neuroticism) to 0.59 (Conscientiousness) for
men.
1051
PERSONALITY DEVELOPMENT IN ADULTHOOD

exception to the general finding here), or is there a broader
increase in prosocial concerns in the 30s and 40s? In the case of
Neuroticism, how are women able to find or create more emotion-
ally stable social environments as they grow older?
Transactional perspectives have an important place in theories
of adult development. Erikson (1950) believed that adult develop-
ment was guided by a series of major challenges that grew out of
biological and cultural imperatives. Identity, intimacy, and gener-
ativity—the three tasks that coincide with the age range we stud-
ied—manifest themselves in adult efforts to find work and become
self-sufficient, to find a partner, and to care for others, and they are
tasks that are provoked by biological imperatives but played out in
the realm of adult social life and situated in a specific sociohis-
torical context. Neugarten (1972) theorized that biology and social
environments set broad limits on adult development, within which
the individual construes and creates his or her own experiences. If
so, then a study of mean-level changes like this one suggests
something about those broad limits, describing the typical path
around which individuals vary. Roberts et al. (in press) suggested
that increasing Conscientiousness and Agreeableness and decreas-
ing Neuroticism in adulthood may indicate increasing maturity—
people becoming, on average, better adapted as they get older. The
findings in this article suggest that people continue to mature well
into middle adulthood.
References
Agronick, G. S., & Duncan, L. E. (1998). Personality and social change:
Individual differences, life path, and importance attributed to the wom-
en’s movement. Journal of Personality and Social Psychology, 74,
1545–1555.
Baltes, P. B. (1997). On the incomplete architecture of human ontogeny:
Selection, optimization, and compensation as foundation of develop-
mental theory. American Psychologist, 52, 366–380.
Barrick, M. R., & Mount, M. K. (1991). The Big Five personality dimen-
sions and job performance: A meta-analysis. Personnel Psychology, 44,
1–26.
Barrick, M. R., Mount, M. K., & Judge, T. A. (2001). Personality and
performance at the beginning of the new millennium: What do we know
and where do we go next? International Journal of Selection and
Assessment, 9, 9–30.
Benet-Martinez, 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.
Carstensen, L. L., Isaacowitz, D. M., & Charles, S. T. (1999). Taking time
seriously: A theory of socioemotional selectivity. American Psycholo-
gist, 54, 165–181.
Caspi, A., & Moffitt, T. E. (1993). When do individual differences matter?
A paradoxical theory of personality coherence. Psychological Inquiry, 4,
247–271.
Caspi, A., & Roberts, B. W. (1999). Personality continuity and change
across the life course. In L. A. Pervin & O. P. John (Eds.), Handbook of
personality: Theory and research (2nd ed., pp. 300–326). New York:
Guilford Press.
Chaplin, W. F. (1997). Personality, interactive relations, and applied psy-
chology. In R. Hogan, J. A. Johnson, & S. R. Briggs (Eds.), Handbook
of personality psychology (pp. 873–890). San Diego, CA: Academic
Press.
Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155–159.
Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation
analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum.
Cohen, P., Cohen, J., Aiken, L. S., & West, S. G. (1999). The problem of
units and the circumstances for POMP. Multivariate Behavioral Re-
search, 34, 315–346.
Costa, P. T., Jr., & McCrae, R. R. (1992). NEO PI-R professional manual.
Odessa, FL: Psychological Assessment Resources.
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.
del Barrio, V., Moreno-Rosset, C., Lopez-Martinez, R., & Olmedo, M.
(1997). Anxiety, depression and personality structure. Personality and
Individual Differences, 23, 327–335.
Erikson, E. H. (1950). Childhood and society. New York: Norton.
Feingold, A. (1994). Gender differences in personality: A meta-analysis.
Psychological Bulletin, 116, 429–456.
Fox, S., Rainie, L., Larsen, E., Horrigan, J., Lenhart, A., Spooner, T.,
& Carter, C. (2001). Wired seniors: A fervent few, inspired by family
ties. Washington, DC: Pew Internet & American Life Project. Retrieved
June 28, 2002, from http://www.pewinternet.org/reports/toc.
asp?Report⫽40
Friedman, H. S., Tucker, J. S., Tomlinson-Keasey, C., Schwartz, J. E.,
Wingard, D. L., & Criqui, M. H. (1993). Does childhood personality
predict longevity? Journal of Personality and Social Psychology, 65,
176–185.
Goldberg, L. R. (1992). The development of markers for the Big-Five
factor structure. Psychological Assessment, 4, 26–42.
Graziano, W. G., & Eisenberg, N. (1997). Agreeableness: A dimension of
personality. In R. Hogan, J. A. Johnson, & S. R. Briggs (Eds.), Hand-
book of personality psychology (pp. 795–824). San Diego, CA: Aca-
demic Press.
Gross, J. J., Carstensen, L. L., Pasupathi, M., Tsai, J., Skorpen, C. G., &
Hsu, A. Y. C. (1997). Emotion and aging: Experience, expression, and
control. Psychology and Aging, 12, 590–599.
Gullone, E., & Moore, S. (2000). Adolescent risk-taking and the five-factor
model of personality. Journal of Adolescence, 23, 393–407.
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.
Helson, R., & Kwan, V. S. Y. (2000). Personality development in adult-
hood: The broad picture and processes in one longitudinal sample. In
S. E. Hampson (Ed.), Advances in personality psychology (Vol. 1, pp.
77–106). Philadelphia: Taylor & Francis.
Helson, R., Jones, C., & Kwan, V. S. Y. (2002). Personality change over 40
years of adulthood: Hierarchical linear modeling analyses of two longi-
tudinal samples. Journal of Personality and Social Psychology, 83,
752–766.
Helson, R., Mitchell, V., & Moane, G. (1984). Personality and patterns of
adherence and nonadherence to the social clock. Journal of Personality
and Social Psychology, 46, 1079–1096.
Helson, R., & Moane, G. (1987). Personality change in women from
college to midlife. Journal of Personality and Social Psychology, 53,
176–186.
Helson, R., Pals, J. L., & Solomon, M. (1997). Is there adult development
distinctive to women? In R. Hogan, J. A. Johnson, & S. R. Briggs (Eds.),
Handbook of personality psychology (pp. 291–314). San Diego, CA:
Academic Press.
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: Guilford Press.
James, W. (1950). The principles of psychology. New York: Dover. (Orig-
inal work published 1890)
John, O. P., Donahue, E. M., & Kentle, R. L. (1991). The Big Five
Inventory—Versions 4a and 54. Berkeley: University of California,
Berkeley, Institute of Personality and Social Research.
1052 SRIVASTAVA, JOHN, GOSLING, AND POTTER

John, O. P., & Paulhus, D. L. (2003). Self-enhancement has more than one
dimension: Evidence for distinct agentic and communal biases in self-
evaluation. Unpublished manuscript.
John, O. P., & Srivastava, S. (1999). The Big Five Trait taxonomy: History,
measurement, and theoretical perspectives. In L. A. Pervin & Oliver P.
John (Eds.), Handbook of personality: Theory and research (2nd ed., pp.
102–139). New York: Guilford Press.
Lang, F. R., Luedtke, O., & Asendorpf, J. B. (2001). Validity and psycho-
metric equivalence of the German version of the Big Five Inventory in
young, middle-aged and old adults. Diagnostica, 47, 111–121.
Lebo, H. (2000). The UCLA internet report: Surveying the digital future.
Los Angeles: University of California Regents. Retrieved October 31,
2000, from http://www.ccp.ucla.edu/UCLA-Internet-Report-2000.pdf
Lenhart, A. (2000). Who’s not online: 57% of those without Internet access
say they do not plan to log on. Washington, DC: Pew Internet &
American Life Project. Retrieved October 31, 2000, from http://www.
pewinternet.org/reports/toc.asp?Report⫽21
Little, B. R. (1996). Free traits, personal projects and idio-tapes: Three tiers
for personality psychology. Psychological Inquiry, 7, 340–344.
Loehlin, J. C. (1992). Genes and environment in personality development.
Newbury Park, CA: Sage.
Loehlin, J. C., McCrae, R. R., Costa, P. T., Jr., & John, O. P. (1998).
Heritabilities of common and measure-specific components of the Big
Five personality factors. Journal of Research in Personality, 32, 431–
453.
Margalit, M., & Eysenck, S. (1990). Prediction of coherence in adoles-
cence: Gender differences in social skills, personality, and family cli-
mate. Journal of Research in Personality, 24, 510–521.
McAdams, D. P. (1996). What this framework can and cannot do. Psy-
chological Inquiry, 7, 378–386.
McClelland, G. H., & Judd, C. M. (1993). Statistical difficulties of detect-
ing interactions and moderator effects. Psychological Bulletin, 114,
376–390.
McCrae, R. R., & Costa, P. T., Jr. (1990). Personality in adulthood. New
York: Guilford Press.
McCrae, R. R., & Costa, P. T., Jr. (1996). Toward a new generation of
personality theories: Theoretical contexts for the five-factor model. In
J. S. Wiggins (Ed.), The five-factor model of personality: Theoretical
perspectives (pp. 51–87). New York: Guilford Press.
McCrae, R. R., & Costa, P. T., Jr. (1999). A five-factor theory of person-
ality. In L. A. Pervin & O. P. John (Eds.), Handbook of personality:
Theory and research (2nd ed., pp. 139–153). New York: Guilford Press.
McCrae, R. R., Costa, P. T., de Lima, M. P., Simoes, A., Ostendorf, F.,
Angleitner, A., et al. (1999). Age differences in personality across the
adult life span: Parallels in five cultures. Developmental Psychology, 35,
466–477.
McCrae, R. R., Costa, P. T., Ostendorf, F., Angleitner, A., Hrebickova, M.,
Avia, M. D., et al. (2000). Nature over nurture: Temperament, person-
ality, and life span development. Journal of Personality and Social
Psychology, 78, 173–186.
McGue, M., Bacon, S., & Lykken, D. T. (1993). Personality stability and
change in early adulthood: A behavioral genetic analysis. Developmental
Psychology, 29, 96–109.
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.
Neugarten, B. (1972). Personality and the aging process. Gerontologist, 12,
9–15.
Neyer, F. J., & Asendorpf, J. B. (2001). Personality–relationship transac-
tion in young adulthood. Journal of Personality and Social Psychol-
ogy, 81, 1190–1204.
Orenstein, P. (1994). Schoolgirls: Young women, self-esteem, and the
confidence gap. New York: Doubleday.
Rammstedt, B., & John, O. P. (2003). Die grossen Fuenf: Measurement,
language, and culture effects in Germany and the U.S. Manuscript
submitted for publication.
Roberts, B. W. (1997). Plaster or plasticity: Are adult work experiences
associated with personality change in women? Journal of Personal-
ity, 65, 205–232.
Roberts, B. W., Robins, R. W., Caspi, A., & Trzesniewski, K. H. (in press).
Personality trait development in adulthood. In J. L. Mortimer & M.
Shanahan (Eds.), Handbook of the life course. New York: Plenum Press.
Schaie, K. W. (1977). Quasi-experimental research designs in the psychol-
ogy of aging. In J. E. Birren & K. W. Schaie (Eds.), Handbook of the
psychology of aging (pp. 39–58). New York: Van Nostrand Reinhold.
Stewart, A. J., & Ostrove, J. M. (1998). Women’s personality in middle
age: Gender, history, and midcourse corrections. American Psycholo-
gist, 53, 1185–1194.
Twenge, J. M. (2000). The age of anxiety? The birth cohort change in
anxiety and neuroticism, 1952–1993. Journal of Personality and Social
Psychology, 79, 1007–1021.
Twenge, J. M. (2001). Birth cohort changes in extraversion: A cross-
temporal meta-analysis, 1966–1993. Personality and Individual Differ-
ences, 30, 735–748.
U.S. Census Bureau. (2000a). Labor force, employment, and earnings
(Section 13). Statistical abstracts of the United States. Washington, DC:
U.S. Department of Commerce.
U.S. Census Bureau. (2000b). Population (Section 1). Statistical abstracts
of the United States. Washington, DC: U.S. Department of Commerce.
U.S. Census Bureau. (2000c). Vital statistics (Section 2). Statistical ab-
stracts of the United States. Washington, DC: U.S. Department of
Commerce.
Vandewater, E. A., & Stewart, A. J. (1998). Making commitments, creating
lives: Linking women’s roles and personality at midlife. Psychology of
Women Quarterly, 22, 717–738.
Viken, R. J., Rose, R. J., Kaprio, J., & Koskenvuo, M. (1994). A devel-
opmental genetic analysis of adult personality: Extraversion and neurot-
icism from 18 to 59 years of age. Journal of Personality and Social
Psychology, 66, 722–730.
Wiggins, J. S. (1973). Personality and prediction: Principles of personality
assessment. Reading, MA: Addison-Wesley.
Wink, P., & Helson, R. (1993). Personality change in women and their
partners. Journal of Personality and Social Psychology, 65, 597–606.
Yang, J., McCrae, R. R., & Costa, P. T., Jr. (1998). Adult age differences
in personality traits in the United States and the People’s Republic of
China. Journals of Gerontology: Series B: Psychological Sciences and
Social Sciences, 53, 375–383.
Received February 19, 2002
Revision received August 2, 2002
Accepted August 22, 2002 䡲
1053
PERSONALITY DEVELOPMENT IN ADULTHOOD