Copyright 2006 by the American Psychological Association
2006, Vol. 132, No. 1, 29 –32
Personality Traits Change in Adulthood: Reply to Costa and
Brent W. Roberts and Kate E. Walton
University of Illinois at Urbana–Champaign
University of Maastricht
In a response to P. T. Costa, Jr., and R. R. McCrae (2006), the authors show that Costa and McCrae’s
writings on personality suggest a belief in immutability of personality traits. The authors agree with Costa
and McCrae that new personality trait models that provide an accurate lower order structure of
personality traits are needed and explain why the Revised NEO Personality Inventory is not the correct
model for that purpose. The authors provide direct evidence refuting the hypothesis that personality traits
change only because of biologically based intrinsic maturation. The authors present arguments supporting
the contention that meta-analyses should be preferred to single longitudinal studies when drawing
inferences about general patterns of personality development. Finally, the authors point out why the
differences between their position and Costa and McCrae’s are important.
Keywords: personality traits, personality change, immutable, NEO–PI–R, meta-analysis
We are pleased to be given the opportunity to address several
pear to be essentially fixed after age 30” (Costa, McCrae, &
issues raised by Costa and McCrae (2006) in response to our
Siegler, 1999, p. 130).
meta-analysis (Roberts, Walton, & Viechtbauer, 2006) of mean-
As we noted in the introduction to our meta-analysis (Roberts et
level change in personality. In a collegial fashion, Costa and
al., 2006), Costa and McCrae have shown a subtle shift in their
McCrae have brought up a number of serious issues that deserve to
description of personality development in the last few years (Costa
be discussed in greater detail.
& McCrae, 2002). For example, they have begun to admit that
personality changes in adulthood, typically with statements such as
that change continues “at a very modest pace throughout adult-
Are Costa and McCrae a Foil?
hood” (McCrae & Costa, 1999, p. 145). Despite this, they continue
to emphasize stability of personality. For example, just recently
Although Costa and McCrae have periodically admitted that
they wrote, “Taken together these results are consistent with the
personality traits change in adulthood (Costa & McCrae, 1989,
broad prediction of Costa and McCrae (2002) that age changes
2002), they have minimized the change in their writing. For
after age 30 are very limited” (Weiss et al., 2005, p. 184). As we
example, consider the following statements: “It appears from sev-
noted in the discussion of our article, this is one place where the
eral longitudinal studies that most personality traits show little or
meta-analysis is authoritative. There is now strong evidence that
no change in mean levels after age 30” (Costa & McCrae, 1992, p.
personality traits change in adulthood past the age of 30.
89); “Individual differences in personality traits, which show at
least some continuity from early childhood on, are also essentially
How Should We Organize Personality Traits Below the
fixed by age 30” (McCrae & Costa, 1994, p. 173); or more
forcefully, “Despite wide differences in measures, subjects, and
periods of the life span studied, all these studies concurred in
We agree with Costa and McCrae’s argument that future meta-
finding relatively little change in the average level of personality
analyses should organize data according to a replicable lower order
traits and surprisingly high stability of individual differences.
structure of personality traits. The key question is whether there is
Barring interventions or catastrophic events, personality traits ap-
a justifiable lower order structure already available and whether
the Revised NEO Personality Inventory (NEO–PI–R) is represen-
tative of that structure. Our answer is no. The key ingredient for a
system to provide an adequate lower order structure of the Big
Brent W. Roberts and Kate E. Walton, Department of Psychology,
Five is some empirical foundation to selecting lower order traits
University of Illinois at Urbana–Champaign; Wolfgang Viechtbauer,
that goes beyond typical personality inventory construction meth-
Department of Psychology, University of Maastricht, Maastricht, The
ods. Unfortunately, in this respect, the NEO–PI–R is like most
other Big Five measures. The items of the NEO–PI–R were gen-
Preparation of this article was supported by a grant from the National
erated from a top-down model of the Big Five, and as the authors
Institute on Aging (R01 AG21178). We thank Tim Bogg, Jennifer Smith,
themselves have stated, theoretical insight and intuition were the
Chris Fraley, Dustin Wood, and Peter Harms for helpful comments on
guiding influence over the generation of items for each domain and
drafts of this article.
the resulting lower order facets for each of the Big Five (Costa &
Correspondence concerning this article should be addressed to Brent W.
Roberts, Department of Psychology, University of Illinois, 603 East Daniel
McCrae, 1998). The limits of this rational approach to identifying
Street, Champaign, IL 61820. E-mail: firstname.lastname@example.org
the lower order structure of the Big Five were highlighted in a
ROBERTS, WALTON, AND VIECHTBAUER
recent effort to establish an empirically justifiable lower order
relationship is associated with increases in traits from the family of
structure of conscientiousness. Roberts, Chernyshenko, Stark, and
conscientiousness (Roberts & Bogg, 2004; Robins, Caspi, & Mof-
Goldberg (2005) found that no personality inventory in existence,
fitt, 2002). Relationship dissatisfaction and insecurity are associ-
including the NEO–PI–R, provided a structure that covers the
ated with increases in neuroticism (Neyer & Asendorpf, 2001;
Roberts & Chapman, 2000; Robins et al., 2002; Scollon, 2004).
And finally, drug consumption is associated with decreases in
Interpreting the Data
conscientiousness-related traits (Roberts & Bogg, 2004; Stein,
Newcomb, & Bentler, 1987). Given the challenge of running one
Researchers do not always interpret data in the same way, even
longitudinal study, finding any sort of replication across longitu-
when they agree on the numbers. For example, Costa and McCrae
dinal studies is remarkable.
(1988) anticipated our findings when they estimated that the upper
limit of personality trait change across the life course would be
Limits of Meta-Analysis
approximately one standard deviation (p. 860). Costa and McCrae
used this estimate to point out that there was not much change in
We agree with Costa and McCrae that meta-analyses are not
personality traits, whereas we concluded that this was an impres-
definitive. Nonetheless, we believe that they provide a more reli-
sive amount of change. We based our conclusion on the burgeon-
able estimate of personality continuity and change than any single
ing evidence that most effect sizes in psychology are between one
longitudinal study. We say this not as outsiders looking in on
quarter and one half of a standard deviation (e.g., small- to
longitudinal work but as researchers enthusiastic about both meta-
medium-sized effects; Meyer et al., 2001) and that effect sizes of
analysis and longitudinal studies, as we are involved in the assess-
one standard deviation are considered quite large (Cohen, 1992).
ment, maintenance, and evaluation of seven different longitudinal
According to these standards, the individual decade-to-decade
studies in addition to our meta-analytic work.
effect sizes we report in our meta-analysis (Roberts et al., 2006)
One of the primary weaknesses of individual longitudinal stud-
are consistent with most effect sizes in psychology, and the cu-
ies is their idiosyncratic nature. Many studies focus on elite sam-
mulative effect sizes are quite large indeed.
ples, on men or women, or on specific cohorts. Furthermore,
Beyond the numbers is the more interesting question of why we
individual longitudinal studies do not afford us the opportunity to
might find these patterns of personality trait development. This is
test whether many factors can and do affect patterns of personality
where we find ourselves disagreeing most with the five-factor
development. For example, several longitudinal studies are neces-
theory approach to personality development. The five-factor the-
sary to infer cohort effects (e.g., Helson, Jones, & Kwan, 2004),
ory argues that personality trait change is purely biological and is
and the number of studies provided in a meta-analysis offers even
governed solely by genetic factors (McCrae et al., 2000). The
stronger tests of this and similar ideas. In this way, meta-analyses
evidence supporting the position that personality change results
are more definitive than any one longitudinal study. This is not to
from intrinsic maturational processes (i.e., genes) has been refuted
say that they are so reliable as to rule out future longitudinal
by numerous studies (for reviews, see Fraley & Roberts, 2005;
research. Our impression is that, like other studies, meta-analyses
Roberts, Wood, & Smith, 2005). For example, several recent
often invite more questions than they answer and thus generate the
studies have shown that there are impressive levels of individual
need for continued longitudinal investigations of personality de-
differences in trait change in old age (Mroczek & Spiro, 2003;
velopment and other topics.
Small, Hertzog, Hultsch, & Dixon, 2003). Also, researchers have
Environment interaction effects for traits linked to
Why Should We Care?
low agreeableness and neuroticism (Caspi et al., 2002, 2003), and
these studies have now been replicated several times (Moffitt,
In sum, a sober examination of our differences with Costa and
Caspi, & Rutter, in press). And finally, our meta-analysis (Roberts
McCrae boils down to the fact that we are willing to state clearly
et al., 2006) of mean-level change revealed several cohort effects
that personality traits change after age 30 and that the environment
on personality trait change. If people demonstrate individual dif-
plays a role in that change. Costa and McCrae, to date, have shown
ferences in change that partially contradict normative trends, if
a reticence to describe personality development in this way.
individual differences result from Gene
We think this is a crucial difference for several reasons. First, at
tions, and if cohort standing can affect personality change, then
the moment, we do not fully understand the implications of per-
how could personality development be driven only by genes?
sonality trait change. Personality traits are related to many domains
The acknowledgment that the environment plays a role in per-
of life, ranging from health to marriage to work (Caspi, Roberts, &
sonality development invites the legitimate question of whether
Shiner, 2005). How, then, will change in personality traits affect
there are replicable associations between life experiences and
these relationships? For example, conscientiousness is related to
change in personality traits. We take issue with Costa and Mc-
all of the leading behavioral factors connected to premature mor-
Crae’s (2006) statement that it has been “difficult to demonstrate
tality (Bogg & Roberts, 2004). If people increase in conscientious-
any replicable environmental effects on personality traits” (p. 29).
ness with age, are they literally adding years to their life? There is
Several reviews of this literature have been previously published
already evidence to that effect for neuroticism. Men who increased
and show compelling replications across separate longitudinal
less than one half of a standard deviation in neuroticism in adult-
studies (Roberts, Robins, Caspi, & Trzesniewski, 2003; Roberts,
hood were at a 37% higher risk for dying than their counterparts
Wood, & Smith, 2005). For example, success in paid work is
(Mroczek & Spiro, 2005). In short, personality change may have
associated with increases in measures of dominance (Roberts,
profound implications for very important outcomes, such as health
1997; Roberts, Caspi, & Moffitt, 2003). Remaining in a stable
PERSONALITY TRAITS CHANGE IN ADULTHOOD: REPLY
Second, if we accept that personality traits continue to change
longitudinal samples. Journal of Personality and Social Psychology, 83,
even in old age, then we are forced to ask a series of questions that
receive little attention in five-factor theory, such as why certain
Mayer, J. D. (2005). A tale of two visions: Can a new view of personality
traits stay consistent and why other traits change. What role do
help integrate psychology. American Psychologist, 60, 294 –307.
environmental factors play in maintaining continuity and promot-
McCrae, R. R., & Costa, P. T. (1994). The stability of personality: Obser-
vation and evaluations. Current Directions in Psychological Science, 3,
ing change in personality traits? How do genes interact with
environments to arrive at what we would see as personality traits?
McCrae, R. R., & Costa, P. T., Jr. (1999). A five-factor theory of person-
Developmentally speaking, why do humans remain open systems
ality. In L. A. Pervin & O. P. John (Eds.), Handbook of personality:
well past childhood and adolescence? Moreover, accepting the fact
Theory and research (2nd ed., pp. 139 –153). New York: Guilford Press.
that personality traits change in adulthood highlights the inadequa-
McCrae, R. R., Costa, P. T., Jr., Ostendorf, F., Angleitner, A., Hrebickova,
cies of almost all theoretical positions found in personality psy-
M., Avia, M. D., et al. (2000). Nature over nurture: Temperament,
chology and personality development (Roberts & Caspi, 2003).
personality, and life span development. Journal of Personality and
Addressing the questions brought about by personality change will
Social Psychology, 78, 173–186.
inevitably move the field to a new set of theories and potentially a
Meyer, G. J., Finn, S. E., Eyde, L. D., Kay, G. G., Moreland, K. L., Dies,
new vision of personality psychology that is more dynamic, inclu-
R. R., et al. (2001). Psychological testing and psychological assessment.
sive of both person and environmental variables, and hopefully
American Psychologist, 56, 128 –165.
more accurate (e.g., Mayer, 2005). Though differences between
Moffitt, T. E., Caspi, A., & Rutter, M. (in press). Measured gene–
environment interactions in psychopathology: Concepts, research strat-
our position and Costa and McCrae’s may seem minor, those
egies, and implications for research, intervention, and public understand-
differences have profound implications for how we understand and
ing of genetics. Perspectives on Psychological Sciences.
study personality and its development.
Mroczek, D. K., & Spiro, A., III. (2003). Modeling intraindividual change
in personality traits: Findings from the Normative Aging Study. Journals
of Gerontology: Series B: Psychological Sciences and Social Sciences,
Mroczek, D. K., & Spiro, A., III. (2005, January). Personality change
Bogg, T., & Roberts, B. W. (2004). Conscientiousness and health behav-
influences mortality in older men. Paper presented at the 6th Annual
iors: A meta-analysis of the leading behavioral contributors to mortality.
Meeting of the Association for Research in Personality, New Orleans,
Psychological Bulletin, 130, 887–919.
Caspi, A., McClay, J., Moffitt, T. E., Mill, J., Martin, J., Craig, I. W., et al.
Neyer, F. J., & Asendorpf, J. B. (2001). Personality–relationship transac-
(2002, August 2). Role of genotype in the cycle of violence in maltreated
tion in young adulthood. Journal of Personality and Social Psychology,
children. Science, 297, 851– 854.
81, 1190 –1204.
Caspi, A., Roberts, B. W., & Shiner, R. (2005). Personality development.
Roberts, B. W. (1997). Plaster or plasticity: Are work experiences associ-
Annual Review of Psychology, 56, 453– 484.
ated with personality change in women? Journal of Personality, 65,
Caspi, A., Sugden, K., Moffitt, T. E., Taylor, A., Craig, I. W., Harrington,
H., et al. (2003, July 18). Influence of life stress on depression: Mod-
Roberts, B. W., & Bogg, T. (2004). A 30-year longitudinal study of the
eration by a polymorphism in the 5-HTT gene. Science, 301, 386 –389.
Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155–159.
relationships between conscientiousness-related traits, and the family
Costa, P. T., & McCrae, R. R. (1988). Personality in adulthood: A six-year
structure and health-behavior factors that affect health. Journal of Per-
longitudinal study of self-reports and spouse ratings on the NEO Personality
sonality, 72, 325–354.
Inventory. Journal of Personality and Social Psychology, 54, 853– 863.
Roberts, B. W., & Caspi, A. (2003). The cumulative continuity model of
Costa, P. T., Jr., & McCrae, R. R. (1989). The NEO-PI/NEO-FFI manual
personality development: Striking a balance between continuity and
supplement. Odessa, FL: Psychological Assessment Resources.
change in personality traits across the life course. In R. M. Staudinger &
Costa, P. T., & McCrae, R. R. (1992). Multiple uses for longitudinal
U. Lindenberger (Eds.), Understanding human development: Lifespan
personality data. European Journal of Personality, 6, 85–102.
psychology in exchange with other disciplines (pp. 183–214). Dordrecht,
Costa, P. T., & McCrae, R. R. (1998). Six approaches to the explication of
The Netherlands: Kluwer Academic.
facet-level traits: Examples from conscientiousness. European Journal
Roberts, B. W., Caspi, A., & Moffitt, T. (2003). Work experiences and
of Personality, 12, 117–134.
personality development in young adulthood. Journal of Personality and
Costa, P. T., Jr., & McCrae, R. R. (2002). Looking backward: Changes in
Social Psychology, 84, 582–593.
the mean levels of personality traits from 80 to 12. In D. Cervone & W.
Roberts, B. W., & Chapman, C. (2000). Change in dispositional well-being
Mischel (Eds.), Advances in personality science (pp. 219 –237). New
and its relation to role quality: A 30-year longitudinal study. Journal of
York: Guilford Press.
Research in Personality, 34, 26 – 41.
Costa, P. T., Jr., & McCrae, R. R. (2006). Age changes in personality and
Roberts, B. W., Chernyshenko, O., Stark, S., & Goldberg, L. (2005). The
their origins: Comment on Roberts, Walton, and Viechtbauer (2006).
structure of conscientiousness: An empirical investigation based on seven
Psychological Bulletin, 132, 28 –30.
major personality questionnaires. Personnel Psychology, 58, 103–139.
Costa, P. T., Jr., McCrae, R. R., & Siegler, I. C. (1999). Continuity and
Roberts, B. W., Robins, R. W., Caspi, A., & Trzesniewski. K. (2003).
change over the adult life cycle: Personality and personality disorders. In
Personality trait development in adulthood. In J. Mortimer & M. Shana-
C. R. Cloninger (Ed.), Personality and psychopathology (pp. 129 –154).
han (Eds.), Handbook of the life course (pp. 579 –598). New York:
Washington, DC: American Psychiatric Press.
Fraley, C., & Roberts, B. W. (2005). Patterns of continuity: A dynamic
Roberts, B. W., Walton, K. E., & Viechtbauer, W. (2006) Patterns of
model for conceptualizing the stability of individual differences in
mean-level change in personality traits across the life course: A meta-
psychological constructs across the life course. Psychological Review,
analysis of longitudinal studies. Psychological Bulletin, 132, 3–27.
112, 60 –74.
Roberts, B. W., Wood, D., & Smith, J. L. (2005). Evaluating five-factor
Helson, R., Jones, C., & Kwan, V. S. Y. (2004). Personality change over
theory and social investment perspectives on personality trait develop-
40 years of adulthood: Hierarchical linear modeling analyses of two
ment. Journal of Research in Personality, 39, 166 –184.
ROBERTS, WALTON, AND VIECHTBAUER
Robins, R. W., Caspi, A., & Moffitt, T. E. (2002). It’s not just who you’re
Stein, J. A., Newcomb, M. D., & Bentler, P. M. (1987). Personality and
with, it’s who you are: Personality and relationship experiences across
drug use: Reciprocal effects across four years. Personality and Individ-
multiple relationships. Journal of Personality, 70, 925–964.
ual Differences, 8, 419 – 430.
Scollon, C. N. (2004). Predictors of intraindividual change in personality
Weiss, A., Costa, P. T., Jr., Karuza, J., Duberstein, P. R., Friedman, B., &
and well-being. Unpublished doctoral dissertation, University of Illinois,
McCrae, R. R. (2005). Cross-sectional age differences in personality
among Medicare patients aged 65 to 100. Psychology and Aging, 20,
Small, B. J., Hertzog, C., Hultsch, D. F., & Dixon, R. A. (2003). Stability
and change in adult personality over 6 years: Findings from the Victoria
Longitudinal Study. Journals of Gerontology: Series B: Psychological
Received July 20, 2005
Sciences and Social Sciences, 58, P166 –P176.
Accepted August 19, 2005