Journal of Personality and Social Psychology Cherish Yourself: Longitudinal Patterns and Conditions of Self-Esteem Change in the Transition to Young Adulthood Jenny Wagner, Oliver Lüdtke, Kathrin Jonkmann, and Ulrich Trautwein Online First Publication, August 13, 2012. doi: 10.1037/a0029680

CITATION Wagner, J., Lüdtke, O., Jonkmann, K., & Trautwein, U. (2012, August 13). Cherish Yourself: Longitudinal Patterns and Conditions of Self-Esteem Change in the Transition to Young Adulthood. Journal of Personality and Social Psychology. Advance online publication. doi: 10.1037/a0029680

Journal of Personality and Social Psychology 2012, Vol. ●●, No. ●, 000 – 000

© 2012 American Psychological Association 0022-3514/12/$12.00 DOI: 10.1037/a0029680

Cherish Yourself: Longitudinal Patterns and Conditions of Self-Esteem Change in the Transition to Young Adulthood Jenny Wagner and Oliver Lu¨dtke

Kathrin Jonkmann and Ulrich Trautwein

Humboldt-University of Berlin

University of Tu¨bingen

Several recent studies have illustrated a general increase in self-esteem from after adolescence until midlife. However, the specific pattern and possible conditions of self-esteem development from the important transition out of high school into young adulthood are still not well understood. In a longitudinal study (Transformation of the Secondary School System and Academic Careers; TOSCA), German students were interviewed 4 times beginning with their senior high school year (at Time 1 [T1]: N ⫽ 4,532; age: M ⫽ 19.6 years, SD ⫽ 0.9; 55% female). Conditional latent change models were applied and established 3 main findings. First, self-esteem showed a gradual increase across the transition, with both the self-esteem intercept and slope indicating substantial interindividual variability in the transition to young adulthood. Second, structural (having a partnership) as well as personality (Big Five) characteristics were substantially related to self-esteem development in emerging adulthood. Third, there were gender-specific associations between self-esteem and partnership status as well as between self-esteem and neuroticism and agreeableness. Findings point to a general upward development of self-esteem yet show interdependencies with the accomplishment of age-specific challenges in the transition to young adulthood. Keywords: self-esteem development, transition to young adulthood, longitudinal data, time-varying covariates

an accumulation of life challenges. The transition from late adolescence into young adulthood can be regarded as a particularly demanding period with several demographic and subjective challenges (Arnett, 2000). Concentrating on the transition from secondary school into young adulthood, the aim of our study was threefold: First, we examined developmental trajectories of selfesteem across and beyond the transition out of high school. Second, we addressed effects of structural characteristics, such as moving out of the parental home, and personality characteristics, such as extraversion, that are known to be indicative of the specific challenges of late adolescence and early adulthood. Third, we were interested in gender-specific patterns of self-esteem development and the relation with covariates while young women and men are transitioning out of school.

The role of self-esteem— defined as the evaluation and appraisal of or attitude toward the self (e.g., James, 1890/1950; Leary & Baumeister, 2000)—in adolescence and young adulthood has been of enduring interest to personality and developmental psychologists. Previous research has found evidence for both beneficial effects of high self-esteem (Greenberg et al., 1992; Leary, Tambor, Terdal, & Downs, 1995; Robins, Hendin, & Trzesniewski, 2001) and detrimental outcomes of low self-esteem (Crocker & Luhtanen, 2003; Orth, Robins, Trzesniewski, Maes, & Schmitt, 2009; Trzesniewski et al., 2006). At the same time, there is an ongoing debate about the degree to which the pursuit of selfesteem is adaptive or maladaptive (Baumeister, Campbell, Krueger, & Vohs, 2003; Crocker & Park, 2004). Despite this debate, self-esteem is regarded as a fundamental psychological construct with long-term effects for maturation. Just recently, the availability of longitudinal data sets and the latest methodological advancements have drawn more attention to the specific patterns of self-esteem development across the life span (cf. Orth, Robins, & Widaman, 2012; Trzesniewski, Donnellan, & Robins, 2003). Yet there is still a need to better understand change trajectories and possible conditions of development, particularly in life stages with

Self-Esteem Change During the Transition to Young Adulthood Young kids start out with reasonably high levels of self-esteem. However, the biological, social, and cognitive changes of late childhood and early adolescence appear to take a toll on positive views. Self-esteem shows steady decline and steep drops in early adolescence (Robins, Hendin, & Trzesniewski, 2001; Trzesniewski et al., 2003). Increasingly realistic self-views as well as an enhanced attention to negative feedback appear to be plausible reasons. Recent longitudinal analyses have shown that beginning in late adolescence and early adulthood, mean levels of self-esteem begin to increase until well into middle adulthood (Erol & Orth, 2011; Orth, Trzesniewski, & Robins, 2010; Shaw, Liang, & Krause, 2010). Simultaneously, self-esteem becomes more stable (Trzesniewski et al., 2003).

Jenny Wagner and Oliver Lu¨ dtke, Department of Psychology, Humboldt-University of Berlin, Berlin, Germany; Kathrin Jonkmann and Ulrich Trautwein, Center for Educational Science and Psychology, University of Tu¨bingen, Tu¨bingen, Germany. Correspondence concerning this article should be addressed to Jenny Wagner, Department of Psychology, Humboldt-University of Berlin, Unter den Linden 6, 10099 Berlin, Germany. E-mail: [email protected] 1

2

¨ DTKE, JONKMANN, AND TRAUTWEIN WAGNER, LU

Theoretical accounts that help us understand normative selfesteem development can be found in life-span as well as personality psychology. Life-span psychology suggests that normative developmental trajectories in self-esteem are most likely related to changing resource characteristics, such as increased independence by late adolescence or declining health conditions in late adulthood and old age (Baltes, 1987). In young adulthood, developmental gains and an increasingly positive ratio of available resources (Heckhausen, Dixon, & Baltes, 1989; Heckhausen & Schulz, 1995) could lead individuals to experience the self in a more positive way. At the same time, the transition into young adulthood has been recognized as a time of vast changes, of relative independence from social roles, and of (self-) exploration (Arnett, 2000). Thus both positive effects and strenuous challenges are conditions of normative self-esteem development in young adulthood. By contrast, the social investment principle (Roberts & Wood, 2006) emphasizes the interrelatedness of individuals and their social environment for personality development. Age-related social challenges and the pursuit of specific social roles have been shown to drive personality maturation across diverse age groups (e.g., Neyer & Lehnart, 2007; Roberts, Wood, & Smith, 2005). With respect to the Big Five personality characteristics, research clearly supports the finding that investment in adult roles is related to personality maturation, that is, higher emotional stability, agreeableness, and conscientiousness (Roberts, Walton, & Viechtbauer, 2006). With respect to self-esteem, classified as a surface rather than a core characteristic of personality (e.g., Asendorpf & van Aken, 2003), a comparable (or even stronger) interrelatedness with environmental changes can be expected (McCrae et al., 2000). Similar to life-span psychology, this line of reasoning suggests the importance of developmental adjustments but focuses on the driving power of social environments and roles. Importantly, the transition from secondary school into young adulthood may be particularly well-suited as a time in which to explore stability and change in self-esteem as well as relations with possible correlates. Arnett (2000) detailed the distinct characteristics of this life period demographically, subjectively, and with respect to identity exploration. Coming from very similar starting points (e.g., most adolescents live with their parents, go to school, and are unmarried and childless) the years between finishing secondary school and the mid-20s represent numerous decisions individuals must make regarding the self, including the lives they want to live, the relationships they want to have, and the careers they want to pursue. Thus, emerging adulthood (Arnett, 2000) is regarded as high in exploratory and experimental qualities. Furthermore, developmental decisions are known to translate into the timing and direction of subsequent events (Freund & Baltes, 2005). For example, delayed lifetime transitions such as postponing leaving the parental house have been associated with childlessness in midlife (Hagestad & Call, 2007; Mulder, 2003). Given that the process of self-evaluation is essential for initiating selfregulative processes and behavioral changes (Tesser, Crepaz, Collins, Cornell, & Beach, 2000), we submit that self-esteem maturation could be one central process that indicates whether developmental challenges such as those found in emerging adulthood are accomplished successfully (Brandtstädter & Greve, 1994; Freund & Smith, 1999). Comparing one’s current status to an ideal status or with the status of surrounding peers is expected to affect

one’s view of the self. In the following, we describe such expectations in more detail.

Structural and Personality Effects on Self-Esteem Change Research has repeatedly illustrated that there are substantial interindividual differences in levels and rates of change in selfesteem (Erol & Orth, 2011; Orth et al., 2010; Pullmann, Allik, & Realo, 2009; Shaw et al., 2010). Therefore, the investigation of correlates is an important topic for better understanding those between-person differences. Two broad research directions can be considered as possible conditions of change: One major line of research emphasizes the role of contextual characteristics such as the occurrence of life events or the entry into a new social role (e.g., Lehnart, Neyer, & Eccles, 2010; Reitzes & Mutran, 1994). A second line of research focuses on the role of individual characteristics such as the impact of subjective health or personality (e.g., Orth et al., 2010; Shaw et al., 2010). Our study considers both of these research traditions and examines relevant context characteristics as well as personality characteristics as possible correlates of self-esteem change in young adulthood.

Structural Characteristics Human development is affected by social-contextual conditions (Roberts et al., 2006; Ryan & Deci, 2000). Young adulthood is a life period in which a person encounters several tasks regarding social-structural changes that will shape the immediate and the future life course, such as getting a good education. We investigated three structural characteristics that are salient tasks of this life period (Havighurst, 1953). First, one of the major structural challenges of late adolescence and young adulthood is to become more independent of one’s parents. For example, by the age of 18, more than 90% of all Germans still live at home, but by age 25, this percentage has decreased to about 30% (Bundesamt, 2010). Previous research on leaving the parental home has focused on timing, the role of relationship quality, and the support of parents (Graber & Dubas, 1996). Self-esteem has not been directly assessed in this context; however, a cross-sectional study with firstyear college students showed that homesickness was related to lower self-esteem (Tognoli, 2003). Thus, moving out of the parental home has to be regarded as an important developmental step that, on average, should be related to higher self-esteem in young adulthood. A second task of this life period is to establish a long-term romantic relationship. Previous research has illustrated that stable singlehood in young adulthood is related to lower self-esteem, particularly in men (Lehnart et al., 2010). In addition, having a first romantic relationship during this age period has been shown to relate to decreases in neuroticism (Lehnart et al., 2010; Neyer & Lehnart, 2007). Therefore, we expect that partnership status will have an impact on self-esteem development above and beyond the normative changes in young adulthood. The third structural task with high priority in this life period is to pave a career path. Finishing school opens up the possibility of diverse educational pathways such as deciding to invest in further education (e.g., college, university). The life of a student may be less regulated (classes are often optional) compared to people who

SELF-ESTEEM IN YOUNG ADULTHOOD

start jobs right out of school (with a strict 8-hr workday). In this context, research on personality development has indicated specific changes in the contexts of career decisions such that higher initial levels of openness have been found for those enrolling in further college education (Lu¨dtke, Roberts, Trautwein, & Nagy, 2011). We propose that the mere decision about whether or not to continue one’s education (i.e., going to college/university or not continuing with the academic path) could also affect self-esteem development. However, due to unavailable previous research, no clear direction of association can be hypothesized. Specifically, attending a university could be regarded as challenging because it involves orienting oneself to a new city and to a less structured educational environment. Thus, the socially and scientifically demanding situation of attending a university is expected to be a challenge for the evaluation of the self.

Personality Characteristics The way people think and feel about themselves is affected by personality dimensions (Robins, Norem, & Cheek, 1999). Previous research suggests that personality characteristics are associated with the mean-level and development of self-esteem (Trzesniewski, Robins, Roberts, Caspi, & Paul, 2004). Cross-sectional correlations emphasize that people with a more mature personality profile of being more emotional stable, extraverted, and conscientious tend to show higher self-esteem (Robins, Hendin, & Trzesniewski, 2001; Robins, Tracy, Trzesniewski, Potter, & Gosling, 2001; Watson, Suls, & Haig, 2002). However, the longitudinal interrelatedness has not been considered until recently. Erol and Orth (2011) found that young adults with more mature personality profiles showed higher initial levels of self-esteem but that no differences were observed in change trajectories of self-esteem. One downside of this otherwise original study is that personality characteristics were assessed only once. Because research has clearly illustrated substantial changes in personality development across the life span (Helson, Jones, & Kwan, 2002; Lucas & Donnellan, 2011; Roberts & DelVecchio, 2000; Srivastava, John, Gosling, & Potter, 2003), and specifically in times of transition (Hopwood et al., 2011; Lu¨dtke et al., 2011), we propose that self-esteem development should be regarded under the consideration of changing personality characteristics. We expected that personality characteristics supporting social inclusion such as extraversion or agreeableness might be particularly important and positive for one’s relational value in this life period. For example, sociometer theory posits effects of perceived relational value on self-esteem (Leary & Baumeister, 2000). Similarly, relationship research has illustrated that young adulthood is a time of information acquisition (Carstensen, 1995) and increasing social network size (Wrzus, Köckeritz, Wagner, & Neyer, in press).

Gender and Self-Esteem Change in the Transition to Young Adulthood Gender identity becomes increasingly important for the evaluation of the self in early adolescence. In this period, girls show substantially larger drops in self-esteem compared to boys (McMullin & Cairney, 2004; Robins, Trzesniewski, Tracy, Gosling, & Potter, 2002; Trzesniewski et al., 2004). This pattern of lower self-esteem in girls seems to continue well into young adulthood

3

(Twenge & Campbell, 2001). However, previous studies have shown that gender differences are often small (Kling, Hyde, Showers, & Buswell, 1999; Orth et al., 2010; Robins, Hendin, & Trzesniewski, 2001) or even nonexistent (Erol & Orth, 2011). In a similar vein, findings on personality characteristics have suggested gender-specific variability in the interrelatedness between the social environment and personality development. For example, van Aken and colleagues (van Aken, Denissen, Branje, Dubas, & Goossens, 2006) found that lower work-related stress was associated with increases in extraversion and emotional stability for men but not for women in midlife. Also, negative relations with singlehood and self-esteem were found for men but not for women in young adulthood (Lehnart et al., 2010). Results emphasize that personality development can be diverse across genders irrespective of similar life conditions. Two possible reasons have been considered for differential gender effects during this important life stage. First, a person’s biological sex may activate specific social categories, leading to different expectations for young men and women (Bosak, Sczesny, & Eagly, 2008; Eagly & Steffen, 1984). Second, young men and women have been shown to differ with respect to life goals (Roberts & Robins, 2000): Young men appear to be more interested in “getting ahead,” whereas young women emphasize goals of “getting along.” However, in well-educated samples, such gender differences are sometimes small to nonexistent (Abele, 2002, 2005).

The Present Study In the present study we examine three research questions: First, what are general longitudinal patterns of self-esteem change in the transition to young adulthood? We hypothesized that self-esteem would follow a nonlinear change pattern because of specific challenges directly after the transition, but steady increases in the early to mid-20s. Second, which structural conditions and personality characteristics are related to self-esteem above and beyond the normative change patterns? We hypothesized that self-esteem would be positively related to the on-time accomplishment of age-specific developmental challenges. Third, we explored whether young men and women would differ in self-esteem trajectories and contextual associations. To examine our research questions we analyzed data from a longitudinal study, which began when individuals were in their final year of secondary school and followed them across their transition into young adulthood. Furthermore, self-esteem as well as structural and personality characteristics were assessed at each wave, allowing the current study to extend previous research in several ways: First, we were able to explore self-esteem change in a relatively age-homogeneous sample that—after the transition out of school—was expected to diverge in structural but also personality characteristics. Second, structural as well as personality characteristics were tested as time-varying covariates, a particularly important condition because these characteristics are known to change in this explorative life period. Third, the size and nature of the longitudinal sample allowed us to test for structural differences across time using appropriate methods to explore intra- and interindividual change patterns.

¨ DTKE, JONKMANN, AND TRAUTWEIN WAGNER, LU

4 Method Participants

Our sample consisted of 4,532 participants of an ongoing German longitudinal study (Transformation of the Secondary School System and Academic Careers; TOSCA), which began in 2002 and is now hosted at the University of Tu¨bingen (see Trautwein, Neumann, Nagy, Lu¨dtke, & Maaz, 2010). The study focuses on educational and psychological conditions, adaptations, and outcomes during the transition from school to work or university. Participants were recruited from 149 randomly selected upper secondary schools (99% participation rate for schools) from a single German state; these schools are representative of the German “Gymnasium” school type. Students in schools were also randomly selected and participated at a satisfactory rate of 80%. The first assessment, taking place between February and May 2002, was conducted in schools with two trained interviewers administering all materials to the students. At Time 1 (T1), all students were in their senior high school year and had a mean age of 19.60 (SD ⫽ 0.85) years. About half of the sample was female (n ⫽ 2,495; 55%), and the majority (n ⫽ 3,558; 79%) lived in traditional family settings, that is, with both biological parents (vs. nontraditional family contexts with single or adoptive parents, etc.). After the first assessment, three additional waves were conducted in 2-year intervals: Time 2 (T2) took place between February and May 2004, about two years after graduation from school; Time 3 (T3) between February and May 2006; and Time 4 (T4) between February and May 2008. At T2, only those students who had signed a consent form for further participation at T1 were recontacted (more than 60%). Follow-up assessments were conducted via extensive questionnaires that took about 2 hr to complete. Participants were rewarded with €10 to €15 (about U.S. $12 to $18) each time they participated. The self-esteem scale was completed by 4,471 students at T1, 2,310 at T2, 1,908 at T3, and 1,869 at T4. For attrition analyses, we compared students who continued their participation across all four waves with those students who dropped out of the study. Students who continued their participation were more likely to be female, ␹2(1) ⫽ 90.74, p ⬍ .001, performed better on a test on reasoning abilities (M ⫽ 102.30 vs. M ⫽ 99.20, d ⫽ 0.15), and were more likely to live in a traditional family setting, ␹2(1) ⫽ 8.27, p ⬍ .01. Importantly, continuers and those students who dropped out of the study early on were similar regarding their mean self-esteem at T1 (d ⫽ 0.04). Thus, the few existing differences between groups were marginal in effect size and appear indicative of only small selectivity effects. Analyses were based on all available responses, that is, even participants with only one assessment of self-esteem were included in the analyses. Recent methodological research on missing values has provided powerful algorithms that are able to deal with missing values and still produce accurate and valid parameter estimates (Schafer & Graham, 2002). Even in cases when missing data are not missing at random, these algorithms appear to produce less biased estimates compared to ad hoc procedures such as listwise deletion. To deal with missing values, we used the full information maximum likelihood (FIML) estimator implemented in Mplus (L. K. Muthe´n & Muthe´n, 1998 –2010). Irrespective of missing

observations on some variables, this estimator uses all available information but does not impute missing values.

Measures Self-esteem. Self-esteem was assessed with four items from the German version of the Self-Descriptive Questionnaire (SDQ; Marsh, 1992; (a) “Overall, I have pretty positive feelings about myself”; (b) “All things considered, I really like myself”; (c) “Overall, I don’t have much respect for myself”; (d) “Overall, I am not very accepting of myself”) using a 4-point response scale from 1 (applies not at all) to 4 (applies totally). Across all four assessments, internal consistency was good, with alphas ranging from .75 to .83 (German version: Schwanzer, Trautwein, Lu¨dtke, & Sydow, 2005). Big Five personality. Big Five personality traits were measured using the German version of the NEO–FFI personality inventory (Costa & McCrae, 1992). Neuroticism, extraversion, openness to experience, agreeableness, and conscientiousness were all rated on a 4-point response scale from 1 (applies not at all) to 4 (applies totally). Extensive previous work has shown the reliability, validity, and comparability of the German NEO–FFI inventory (Borkenau & Ostendorf, 1991; Lu¨dtke, Trautwein, Nagy, & Köller, 2004). Across all four waves, alpha reliabilities were satisfactory (Neuroticism: ␣s between .81 and .89; Extraversion: ␣s between .77 and .81; Openness to Experience: ␣s between .73 and .74; Agreeableness: ␣s between .73 and .76; and Conscientiousness: ␣s between .83 and .85). Structural characteristics. From the second assessment onward, participants reported information about their living situation (0 ⫽ in parental home, 1 ⫽ out of parental home), partnership status (0 ⫽ no partner, 1 ⫽ with partner), and current educational path1 (0 ⫽ not a student, 1 ⫽ student). Demographics. At the first assessment, participants reported information on age, gender, and family situation (traditional: living with both biological parents vs. nontraditional: living with single or adoptive parents, etc.). Furthermore, two subtests of a cognitive test battery assessing reasoning abilities were administered (KFT; Heller & Perleth, 2000) and transformed to an IQ-like metric.

Analytic Strategy Our research questions involved three distinct steps of data analysis: First, we were interested in change trajectories of selfesteem across time and thus applied latent curve models (LCMs; see Bollen & Curran, 2006). Second, to analyze relations between structural and personality characteristics with self-esteem, we estimated conditional latent curve models including time-invariant covariates (TICs) and time-varying covariates (TVCs). Third, we conducted gender-specific analyses by applying multiple-group models. All models were fit to the data using Mplus Version 6.1 (L. K. Muthe´n & Muthe´n, 1998 –2010). Model fit evaluation was based on the full information maximum likelihood estimator (FIML) and used conventional criteria. Because ␹2 test statistics are known to be problematic with increasing sample size such as 1 The educational path differentiated between individuals who entered into a higher educational institution such as a university or college (including the German Fachhochschule) versus all others.

SELF-ESTEEM IN YOUNG ADULTHOOD

in this context, two additional fit indices were included. Rootmean-square errors of approximation (RMSEA) below .08 and .05 as well as comparative fit indices (CFI) greater than .90 and .95 are suggested as guidelines for acceptable and excellent fit to the data, respectively (Marsh, Hau, & Grayson, 2005). For the genderspecific multiple-group models the comparison of two nested models was based on the ␹2 difference test. Latent curve models. Because self-esteem was assessed with four items at each assessment, we were able to apply latent curve models with multiple indicators. Using structural equation modeling techniques, interindividual differences in intraindividual change patterns in self-esteem were analyzed. Latent curve models or so-called second-order latent growth models (Sayer & Cumsille, 2001) control for measurement error at the level of indicators and, thus, have the advantage of being able to discriminate between structural relations and measurement error (Bollen & Curran, 2006). Figure 1 shows the latent curve model for analyzing individual change in self-esteem. Observed self-esteem indicators were used to measure time-specific latent self-esteem. These specific latent factors were used to define two latent factors of self-esteem growth: the latent intercept and the latent slope factor. By fixing all loadings to unity, the intercept factor represents the mean level of self-esteem at the first assessment (i.e., before leaving school). The intercept variance illustrates interindividual variability in self-esteem mean levels at T1. The latent slope factor, by contrast, reflects the amount of individual change that occurred between T1 and T4. The variance component of the slope factor represents interindividual variability in change trajectories across time. In order to provide the best model fit, we tested a number of different forms of change, namely, linear, quadratic,

5

and free slope estimation, where only two loadings are fixed and all other loadings are able to indicate nonlinear (free) patterns of change. This empirical procedure is based on previous studies that have suggested nonlinear change patterns of self-esteem after adolescence (Erol & Orth, 2011). Measurement invariance across time is one precondition for applying LCMs. Our first step was thus to build a self-esteem measurement model with four correlated latent factors, one for each measurement. Testing for measurement invariance guarantees that observed change in manifest variables is real change and is not due to changing relations between the latent variables or their indicators (Bollen & Curran, 2006). Measurement invariance, which is one important strength of second-order latent growth models, was tested in five steps (Table 1; cf. Bollen & Curran, 2006). Results showed that even the most restrictive model (Model 5) still had good model fit. Therefore, all of the following models are based on a measurement model of strict factorial invariance, that is, with invariant factor loadings, item intercepts, and error variances. We also allowed for correlated residuals across time because they account for the effects of the specific indicators (cf. Figure 1) and research suggests that ignoring these dependencies may lead to distorted results in the relations of the factors of interest (Bollen & Curran, 2006; Marsh & Hau, 1996). Conditional latent curve models. We extended the univariate latent curve model by estimating a number of conditional models with covariates (a) that are invariant across time, such as gender, or have been measured only once (TICs), or (b) that change or vary across time, such as personality characteristics (TVCs). The inclusion of TICs is performed by analyzing the impact of the covariate on the mean level (i.e., intercept of self-

Figure 1. Latent curve model illustrating interindividual differences in intraindividual change patterns in self-esteem. In the figure, latent variables are represented by circles or ellipses and observed variables by squares. Straight single-headed arrows illustrate effects on the variable at the head of the arrow, and two-headed arrows represent the linear covariance between connected variables (Bollen & Curran, 2006). T ⫽ time.

¨ DTKE, JONKMANN, AND TRAUTWEIN WAGNER, LU

6

Table 1 Fit Indices for Measurement Invariance Tests for Self-Esteem Model Model Model Model Model Model

1: 2: 3: 4: 5:

Unconstrained model Weak invariance Strong invariance Strict invariance Factor variances invariant across time

␹2

df

CFI

RMSEA

290.31 380.53 695.16 686.25 694.93

74 83 95 104 107

.987 .983 .965 .966 .966

.025 .028 .037 .035 .035

Note. N ⫽ 4,532. Models allow for correlated residuals. CFI ⫽ comparative fit index; RMSEA ⫽ root-meansquare error of approximation.

esteem) and on the change trajectory (i.e., the slope of selfesteem). The consideration of TVCs was accomplished in two different types of analyses. The first type of model was guided by the time-varying-covariates approach presented by B. O. Muthe´n (1993). This approach includes TVCs directly within the LCM to use them as exogenous predictors of the outcome (Figure 2). In this context, self-esteem is modeled as the major developmental process, and the TVCs are modeled as a kind of secondary developmental process (Grimm, 2007). Resulting ␤ parameters can be understood as occasion- or time-specific effects of the covariate on the outcome or as the ability of the covariate to predict specific deviations in the outcome (Preacher, Wichman, MacCallum, & Briggs, 2008). The strength of this approach of analysis is that both self-esteem and TVCs may change across time and, in addition, effects of the TVCs on the outcome may vary across time (but not individuals). Therefore, constraining the ␤ parameters across time without a substantial decrease in model fit illustrates a timeconstant effect of the TVCs on the outcome. Considerations of effects of the change trajectory on the outcome variable thus show the amount of change while controlling for occasion-specific effects of the considered TVCs. Time-specific personality variables were modeled as latent factors. In the second type of model, TVCs were added as additional latent slope factors that refer to the individual’s mean effect of the

TVC on the outcome variable (for a discussion of different approaches, see Curran & Bauer, 2011; Grimm, 2007; Preacher et al., 2008). This alternative approach includes TVCs as so-called definition variables (Mehta & West, 2000; Raudenbush & Chan, 1993; Singer & Willet, 2003) and is similar to how TVCs are addressed in multilevel models of longitudinal data (Figure 3). That is, a TVC is added as an additional latent slope factor that refers to the individual’s mean effect of the TVC on the outcome variable. In this context, personality variables were centered at the individual mean across time (i.e., comparable to group-mean centering in MLM). Unfortunately, models with definition variables do not report the conventional model fit criteria (see Wu, West, & Taylor, 2009). These models are conducted only with respect to personality characteristics and not the dummy-coded structural variables. This type of analysis has the strength of being able to estimate the variability of the effect of the TVC between individuals (but not within individuals across time). In this context, fixed effects represent the average within-subject correspondence between personality characteristics and self-esteem across time, whereas the random part illustrates interindividual variability of the within-subject correspondence across persons. In our models, we analyzed one TVC at a time. Multigroup models. Finally, in order to analyze genderspecific change trajectories of self-esteem, we applied multigroup models to all previous analyses. Personality TVCs were again

Figure 2. Latent curve model including a time-varying covariate (see Preacher et al., 2008, p. 47ff). Measurement models for self-esteem and extraversion are not depicted. T ⫽ time.

SELF-ESTEEM IN YOUNG ADULTHOOD

7

Figure 3. Latent curve model including a time-varying covariate (adapted from Mehta & West, 2000; see also Preacher et al., 2008). T ⫽ time.

modeled as latent variables and set to be invariant across time (Lu¨dtke et al., 2011). Imposing cross-group equality constraints on specific parameters, we were able to test for time-specific gender differences by analyzing changes in the ␹2 difference test.

means and standard deviations, showed substantial mean-level change with respect to all but one comparison between T1 and T2 (d ⫽ 0.05, p ⬎ .05). Consecutive assessments between T2 and T3 (d ⫽ 0.17, p ⬍ .05) as well as between T3 and T4 (d ⫽ 0.12, p ⬍ .05) demonstrated significant but small mean-level differences. As expected, effect sizes increased with increasing distance between assessment points (ds ⱖ 0.22) and thus revealed ongoing selfesteem change across time. Table 2 also shows latent factor retest stabilities. Self-esteem stability decreased across assessments that were further apart but remained above r ⫽ .50 in all instances. Systematic testing of consecutive retest stabilities illustrated statistically significant differences such that stability increased in the beginning of the study and decreased slightly at the end (r12 ⬍ r23 and r23 ⬎ r34, all ps ⬍ .05). Generally, differences were very small and, as expected, consecutive retest stabilities decreased for assessment points that were further apart (i.e., longer time intervals between assessments). Latent curve models. We estimated three different latent curve models with multiple indicators to analyze the change tra-

Results Our Results section is structured as follows: First, we present latent means, latent correlations, and change patterns of selfesteem across the transition into young adulthood. Following this, we focus on LCMs with structural and psychological TVCs. Finally, we examine whether young men and women differ with respect to self-esteem development and their possible effects of TVCs across time.

Self-Esteem Continuity and Change in the Transition to Young Adulthood Descriptive statistics of latent self-esteem variables are presented in Table 2. Standardized mean differences, based on latent

Table 2 Latent Mean Differences and Correlations of Self-Esteem Change Across Four Time Points Mean-level change Self-esteem

M

SE

Var

d1x

d2x

Time 1 Time 2 Time 3 Time 4

3.18 3.20 3.28 3.34

0.01 0.01 0.01 0.01

0.21 0.22 0.23 0.24

— .05 .22 .34

— .17 .29

d3x

— .12

Stability d4x

r1x

r2x

r3x

r4x

—

— .69 .61 .53

— .76 .64

— .74

—

Note. d ⫽ (mean of Tt ⫺ mean of Tt-1)/pooled standard deviation; with pooled SD ⫽ 0.47. Bold mean-level differences are substantial at p ⬍ .01. All correlations are significant at p ⬍ .001.

8

¨ DTKE, JONKMANN, AND TRAUTWEIN WAGNER, LU

jectory of self-esteem across time (Figure 1). As hypothesized, our results supported a nonlinear trend in self-esteem change beyond the linear change model. More precisely, the quadratic LCM showed the best fit to the data, ␹2(108) ⫽ 660.05, RMSEA ⫽ .034, CFI ⫽ .968, with substantial improvement from the linear model, ⌬␹2(4) ⫽ 38.24, p ⬍ .01, as well as from the free change model, ⌬␹2(2) ⫽ 34.85, p ⬍ .01. The latent intercept of the LCM (3.18), representing the initial value of self-esteem before leaving school, corresponds to the latent mean of T1. The variance component, indicating the reliable interindividual variability around that mean, showed substantial differences in self-esteem between students at the first assessment (.16, p ⬍ .001). Both linear (.03, p ⬍ .01) and quadratic (.01, p ⬍ .05) slope factors, representing the rates of change across the 6-year interval, were substantial and supported previous findings as well as our own expectations. Self-esteem increased across the full span of the study in a nonlinear pattern. The rate of change appeared to be small between the first and second assessments but accelerated with respect to the third and fourth times of assessment. Figure 4 shows this change trajectory in the transition from high school into young adulthood. Importantly, the variance components suggest that there were reliable interindividual differences in change trajectories of self-esteem across the transition to young adulthood (.06 and .01, respectively, all ps ⬍ .001). Raudenbush and Bryk (2002, p. 78) suggested indicating the magnitude of variability across the sample by calculating a 95% plausible value range (PVR) of the slope. With respect to the quadratic (linear) slope, results revealed that self-esteem change parameters ranged between –.19 (–.45) and .21 (.51). Thus, the 95% PVR illustrates change in both directions; a substantial number of individuals even experience a decline in self-esteem. The correlation between the intercept and the linear slope was small (r ⫽ –.13, p ⬍ .05) and even nonsignificant with the quadratic slope (r ⫽ –.03, ns), indicating that rates of change were fairly unrelated to the initial value of self-esteem. Time-invariant covariates. The consideration of demographic characteristics of the first assessment revealed some differences in self-esteem mean levels, but none of the characteristics affected rates of change in self-esteem across time. On the basis of previous findings regarding the role of abilities for self-esteem in school settings (Trautwein, Lu¨dtke, Köller, & Baumert, 2006), we tested for general relations between cognitive ability and the self-esteem intercept and slope. Cognitive ability was assessed only at the first measurement occasion and modeled as a TIC. Higher cognitive abilities in reasoning at T1 were related to a higher self-esteem intercept (.07, p ⬍ .05) but unrelated to the linear (.05, p ⬎ .05) and quadratic (–.07, p ⬎ .05) slope factors. Traditional (vs. nontraditional) family contexts did not relate to the level or rates of change in self-esteem (all ps ⬎ .28).

Relations of Structural and Personality Characteristics With Self-Esteem Change In the next part of our analyses, we considered the association of two types of covariates that were expected to change across time (TVCs) with self-esteem: (a) structural characteristics such as moving out of the parental home and (b) personality characteristics such as extraversion.

Structural characteristics. In a first step, the latent growth curve model of self-esteem included a number of structural TVCs that captured specific developmental tasks of emerging adulthood. One developmental task shared by the entire sample is the transition out of the secondary school to a university or to work. This shared developmental transition was already captured in the overall LCM because it is experienced by all participants. Apart from this general task, we identified three challenges of emerging adulthood that are considered to be prominent in the transition to young adulthood: (a) moving out of the parental home, (b) having a stable partnership, and (c) entering an educational path. The distributions of percentages clearly showed the time-varying nature of all three structural characteristics (Table 3). As expected, the number of young adults who moved out of the parental house and those who had stable partnerships increased across assessment points, whereas the number of students decreased. Structural characteristics were dummy coded and entered as uncentered predictors into the LCMs. When controlling for normative change in self-esteem, partnership status was related to deviations in self-esteem (Table 4). Importantly, the model comparison supported time-specific associations of this structural TVC, ⌬␹2(2) ⫽ 15.15, p ⬍ .01. Having a partnership revealed a positive association with self-esteem but only in the second part of the study, that is, on average by the ages of 23 and 25. In the previous assessment at T2, the existence of a partnership was unrelated to self-esteem. Either being a student or moving out of the parental home was only marginally related to self-esteem while controlling for interindividual differences in intraindividual change in self-esteem, and no time-specific relations occurred. Results highlight the time-specific link between the developmental transition of having a partner with self-esteem in the transition to young adulthood above and beyond the normative change pattern of self-esteem. Personality characteristics. We next examined relations between Big Five personality characteristics and self-esteem (Table 4). When controlling for changes in self-esteem between the ages of 19 and 25, neuroticism was negatively related and all other personality factors were positively related to self-esteem. In other words, young people in the transition to young adulthood who tended to be less neurotic but more extraverted, open to experiences, agreeable, and conscientious showed a higher level of self-esteem above and beyond the normative self-esteem increase. Importantly, extraversion illustrated time-specific associations across the span of the study. Irrespective of substantial associations in all four assessments, associations between self-esteem and extraversion appeared to be substantially stronger at the second and third times of assessment. In periods of finding and establishing new educational and social settings, higher extraversion appears to be particularly beneficial for the positive evaluation of the self. Table 3 Descriptive Data on Structural Characteristics Across Time Structural TVCs

Time 2

Time 3

Time 4

Moving out (1 ⫽ yes) Partnership (1 ⫽ yes) Student (1 ⫽ yes)

58 54 76

71 58 76

81 65 42

Note.

Data are percentages. TVCs ⫽ time-varying covariates.

SELF-ESTEEM IN YOUNG ADULTHOOD

9

Table 4 Fit Indices of Conditional Quadratic Latent Curve Models of Self-Esteem in Young Adulthood With Structural and Psychological Time-Varying Covariates (TVCs) and Standardized Regression Coefficients of TVCs Model fit TVC Structural TVCs Moving out (1 ⫽ yes) Partnership (1 ⫽ yes) Student (1 ⫽ yes) Psychological TVCs Neuroticism Extraversion Openness to experience Agreeableness Conscientiousness

␹2

TVCs constrained

Standardized regression coefficients

df

RMSEA

CFI

404.64 349.84 385.74

155 153 155

.033 .032 .032

.978 .979 .979

2,176.26 1,652.00 1,637.59 1,274.16 1,406.15

462 459 462 462 462

.029 .024 .024 .020 .021

.962 .971 .968 .975 .977

␤T1

⫺.64ⴱⴱⴱ .28ⴱ .24ⴱⴱ .36ⴱⴱ .39ⴱⴱⴱ

␤T2

␤T3

␤T4

.03† .01 .02†

.02† .08ⴱⴱⴱ .02†

.02† .11ⴱⴱⴱ .03†

⫺.66ⴱⴱⴱ .61ⴱⴱⴱ .23ⴱⴱ .35ⴱⴱⴱ .38ⴱⴱⴱ

⫺.66ⴱⴱⴱ .38ⴱ .22ⴱⴱ .34ⴱⴱⴱ .37ⴱⴱⴱ

⫺.66ⴱⴱⴱ .58ⴱⴱⴱ .23ⴱⴱ .36ⴱⴱⴱ .38ⴱⴱⴱ

⌬␹2

df

3.57 15.15ⴱⴱ 5.54

2 2 2

2.59 8.77ⴱ 3.78 4.12 0.73

3 3 3 3 3

Note. CFI ⫽ comparative fit index; RMSEA ⫽ root-mean-square error of approximation. p ⬍ .10. ⴱ p ⬍ .05. ⴱⴱ p ⬍ .01. ⴱⴱⴱ p ⬍ .001.

†

In the second type of model, we added TVCs as additional latent slope factors (i.e., definition variable) indicating the individual mean association between TVC and self-esteem. The fixed effects parts of these models, representing the within-subject correspondence between personality characteristics and self-esteem, supported our findings regarding the presented analyses. On average, self-esteem again covaried significantly as a function of all five personality characteristics. Individuals low in neuroticism (␤ ⫽ –.47, p ⬍ .001) and high in extraversion (␤ ⫽ .36, p ⬍ .001), openness to experience (␤ ⫽ .11, p ⬍ .01), agreeableness (␤ ⫽ .25, p ⬍ .001), and conscientiousness (␤ ⫽ .29, p ⬍ .001) appeared to have more favorable self-esteem development across the transition to young adulthood. Importantly, and as expected, variance components of all five random definition variables indicated substantial variability in the within-person correspondence between self-esteem and personality characteristics (neuroticism: .12, p ⬍ .001; extraversion: .09, p ⬍ .01; openness to experience: .15, p ⬍ .05; agreeableness: .09, p ⬍ .01; and conscientiousness: .11, p ⬍ .001). Young participants thus differed with respect to the association between personality maturation and self-esteem development. Together, both structural as well as personality characteristics showed substantial relations with self-esteem across time. Importantly, looking at the average trajectories of conditional models clarified that the consideration of covariates was fairly unrelated to the level of self-esteem at the beginning of the study. However, if young adults did not accomplish structural tasks, the positive developmental trend of self-esteem across the transition to young adulthood was decelerated, as can be seen in the nonsignificant slope parameters for most characteristics. By contrast, conditional self-esteem intercept and slope parameters in the personality models did not differ from the unconditional findings, suggesting that the average trend of personality change was related to the average self-esteem trajectory. Our findings emphasize the relation between the Big Five personality characteristics and self-esteem across time.

Gender and Self-Esteem Change in the Transition to Young Adulthood In the final part of our analyses, we examined possible genderspecific patterns of self-esteem development. First, we examined

whether the normative self-esteem trajectory differed by gender, and we tested this by including a dummy-coded gender variable as a time-constant covariate in the general LCM of self-esteem. Results showed that, on average, girls reported lower self-esteem than boys (–.12, p ⬍ .01) at T1, but gender did not relate to rates of change in self-esteem across time. Figure 4 illustrates that young adults, independent of being male or female, show similar increases in their self-evaluations but start out and remain at different levels. Applying multiple-group models, we tested for gender-specific patterns of TVCs on self-esteem (Table 5). Regarding structural TVCs, neither being a student nor moving out of the parental home indicated gender-specific relations with self-esteem. However, controlling for mean changes in self-esteem, the relation between romantic partnership status and self-esteem was substantially different for young men and women. Fixing male and female effects

Figure 4. Latent mean trajectory on self-esteem across the four assessments for the entire sample and separate latent mean trajectories for males and females.

¨ DTKE, JONKMANN, AND TRAUTWEIN WAGNER, LU

10

Table 5 Fit Indices of Conditional Quadratic Latent Curve Models of Self-Esteem in Young Adulthood With Structural and Psychological Time-Varying Covariates (TVCs) and Gender-Specific Standardized Regression Coefficients of TVCs Model fit TVC Structural TVCs Moving out (1 ⫽ yes) Male Female Partnership (1 ⫽ yes) Male Female Student (1 ⫽ yes) Male Female Psychological TVCs Neuroticism Male Female Extraversion Male Female Openness to experience Male Female Agreeableness Male Female Conscientiousness Male Female

␹2

Gender constrained

Standardized regression coefficients ␤T2

␤T3

␤T4

⌬␹2

df

.973

.03 .03

.03 .02

.02 .02

6.84

5

.036

.972

.06 ⫺.02

.13ⴱⴱⴱ .04

.13ⴱⴱ .10ⴱⴱ

21.34ⴱⴱ

5

322

.035

.974

.02 .02

.02 .02

.03 .02

10.05

5

3,268.26

940

.033

.947

⫺.71ⴱⴱⴱ ⫺.55ⴱⴱⴱ

⫺.76ⴱⴱⴱ ⫺.58ⴱⴱⴱ

⫺.73ⴱⴱⴱ ⫺.65ⴱⴱⴱ

⫺.56ⴱ ⫺.79ⴱⴱⴱ

2,496.61

947

.027

.961

.58ⴱⴱⴱ .60ⴱⴱⴱ

.59ⴱⴱⴱ .58ⴱⴱⴱ

.59ⴱⴱⴱ .56ⴱⴱⴱ

.59ⴱⴱⴱ .55ⴱⴱⴱ

10.50

7

2,953.94

947

.031

.945

.23ⴱⴱ .24ⴱⴱ

.23ⴱⴱ .23ⴱⴱ

.23ⴱⴱ .23ⴱⴱ

.23ⴱⴱ .22ⴱⴱ

9.05

7

2,406.12

940

.026

.954

.45ⴱⴱⴱ .29ⴱⴱ

.22 .51ⴱⴱⴱ

15.32ⴱ

7

2,219.41

947

.024

.969

.39ⴱⴱⴱ .38ⴱⴱⴱ

.38ⴱⴱⴱ .37ⴱⴱⴱ

3.13

7

df

RMSEA

CFI

625.03

322

.036

575.80

317

608.43

␤T1

.27 ⫺.06 .38ⴱⴱⴱ .40ⴱⴱⴱ

⫺.12 .63ⴱ .38ⴱⴱⴱ .36ⴱⴱⴱ

29.46ⴱⴱ

14

Note. CFI ⫽ comparative fit index; RMSEA ⫽ root-mean-square error of approximation. ⴱ p ⬍ .05. ⴱⴱ p ⬍ .01. ⴱⴱⴱ p ⬍ .001.

to be equal decreased the model fit substantially, ⌬␹2(5) ⫽ 21.34, p ⬍ .01, as well as fixing effects to be invariant across time reduced the model fit substantially, ⌬␹2(4) ⫽ 16.24, p ⬍ .01. Thus, effects of having a partnership varied across gender and across time. Being in a partnership was associated with higher self-esteem for men by T3 (␤T3 ⫽ .13, p ⬍ .001; ␤T4 ⫽ .13, p ⬍ .001), hence, beginning approximately by age 23. In contrast, being in a partnership was not associated with women’s selfesteem before the last assessment, by approximately age 25. Regarding personality characteristics, patterns between selfesteem and extraversion, openness to experience, as well as conscientiousness were similar for men and women in the transition to young adulthood. For both men and women, time-varying scores of extraversion, openness, and conscientiousness were positively related to time-varying scores of self-esteem while controlling for normative changes in self-esteem between the ages of 19 and 25 (Table 5). However, results differed with respect to associations with neuroticism and agreeableness. Constraining effects of neuroticism to be equal for men and women decreased the model fit substantially, ⌬␹2(14) ⫽ 29.46, p ⬍ .05.2 Furthermore, testing for gender-specific time-varying versus time-constant associations between neuroticism and self-esteem showed that model fit was best for time-varying effects, ␹2(940) ⫽ 3,268.26, RMSEA ⫽ .033, CFI ⫽ .957, ⌬␹2(13) ⫽ 23.70, p ⬍ .05. Whereas associations between neuroticism and self-esteem decreased at the last assess-

ment for young men, this association increased throughout the periods of assessment for young women. With respect to agreeableness, again, the model with gender-specific and time-varying effects showed the best model fit. Fixing male and female effects to be equal, ⌬␹2(7) ⫽ 15.32, p ⬍ .05, but also specifying genderspecific time-invariant effects of agreeableness decreased the model fit substantially, ⌬␹2(6) ⫽ 15.31, p ⬍ .05. Controlling for normative changes in self-esteem, higher agreeableness in women was increasingly positively associated with self-esteem. In contrast, men showed substantial positive associations between agreeableness and self-esteem at only one assessment.

Discussion The aim of the current study was to examine patterns of selfesteem development across the transition from secondary school into young adulthood. Applying latent growth curve modeling, our main findings indicated that self-esteem showed a small increase following secondary school, followed by an accelerated increase 2 In the male-specific LCM with neuroticism modeled as a TVC, the quadratic slope did not show substantial variability in the context of gender constraints. Hence, in the male model only, the slope itself and all corresponding covariances have been fixed to zero. This leads to different degrees of freedom in this model.

SELF-ESTEEM IN YOUNG ADULTHOOD

beginning at approximately age 21. Both structural and personality characteristics related to self-esteem development above and beyond the general self-esteem trajectory. Most notably, extraversion exerted time-specific associations with self-esteem. Finally, gender differences appeared to be mainly small, but illustrated timespecific relations regarding the existence of a stable romantic partnership, the negative association with neuroticism, and the positive association with agreeableness.

Self-Esteem Increases in the Transition to Young Adulthood Our findings support previous research showing that self-esteem increases during late adolescence and into young adulthood (Erol & Orth, 2011; Huang, 2010; O’Malley & Bachman, 1983; Shaw et al., 2010). The established quadratic change pattern may explain differential findings with respect to self-esteem change patterns in this age group (e.g., self-esteem decrease; McMullin & Cairney, 2004). Reducing the sample to the first two years of assessment would have yielded at least the conclusion of robust self-esteem mean levels in young adulthood. Going beyond most previous studies, we tracked our participants while they were transitioning out of secondary school. Previous findings with this same sample (Lu¨dtke et al., 2011) highlighted that this transition is related to an acceleration of personality maturation (see also Hopwood et al., 2011). However, self-esteem development appears to differ from such patterns. Referring to latent mean-level changes between the first and second assessments, self-esteem showed no substantial change but displayed an increase between the subsequent measurement points. Entering postschool life might lead to sudden personality maturation, but the integration of different roles, goals, and selves may postpone the boost in self-esteem (McAdams & Olson, 2010). Focusing on self-esteem retest stability, our findings only partially support previous results of a steadily increasing stability (Trzesniewski et al., 2003). Rather, our results suggest a relative stagnation (or only small fluctuations) in self-esteem stability between ages 19 and 25. This might support the notion that this period of transition is a time of both exploration and consolidation (Arnett, 2000, 2007). Our patterns of findings for mean-level change and retest stability indicate that self-esteem development may be due to a different mechanism. Self-esteem development is often considered to be relatively similar to the development of personality characteristics in young adulthood (Trzesniewski et al., 2004), but the explorative period of emerging adulthood has the potential to both continuously boost and dampen self-esteem (Arnett, 2000). In general, changing social environments (going abroad while studying, finishing one’s studies and starting a job, changing from dating to a stable partnership) may lead to gradual changes and less stable patterns in self-esteem compared to Big Five personality. Importantly, we found substantial interindividual differences in the starting points and change trajectories of self-esteem. Thus, young adults differed in the ways in which they evaluated themselves in their final year of school but also in the ways in which their self-evaluations changed across the transition and further on (see also Erol & Orth, 2011). Above and beyond the normative positive trend of self-esteem development, there were also people who showed an even more accelerated boost but also those who actually decreased in self-esteem across time. Given this pattern,

11

further analyses examined whether structural conditions and personality characteristics were associated with differential levels and rates of change.

Normative Development Relates to Higher Self-Esteem Two characteristics— cognitive abilities and family of origin— were assessed only in high school and conditional models illustrated that higher cognitive abilities were related to higher selfesteem in high school. Such results concur with previous findings from educational research highlighting the association between abilities and the way adolescents evaluate themselves (Trautwein et al., 2006). However, we did not observe relations between cognitive abilities and self-esteem change across the transition to young adulthood. Furthermore, we found that the structure of the family of origin did not affect the way individuals evaluated themselves in the transition to young adulthood. Such results may support and broaden findings suggesting that the diversity of new family types actually does not affect relationships to others and to the self (Lang, Wagner, Wrzus, & Neyer, 2012). In line with previous research, partnership status was related to self-esteem (Lehnart et al., 2010; Neyer & Lehnart, 2007). We found that self-esteem was positively associated with having a stable romantic partnership above and beyond the normative selfesteem change pattern. Importantly, we tested for time-specific relations of the existence of a romantic partnership and having a stable romantic partnership showed a substantial positive relation with self-esteem at later stages of our study when participants were around the ages of 23 and 25 years. This relation was qualified by an additional gender effect, which we discuss further in the next section. We did not find any effects of moving out of the parental house or the educational path on levels or rates of change in self-esteem. It appears that above and beyond the normative development of self-esteem, these structural changes do not affect the evaluation of the self in young adults. Students are often still financially dependent on their parents, which could be closely associated with whether or not they moved out of the parental home. Also, the diversity in life paths that people take after leaving school might need more fine-grained analyses of such developmental transitions (cf. Jackson, Thoemmes, Jonkmann, Lu¨dtke, & Trautwein, 2012). Personality maturation is another major aspect of development in young adulthood. Our findings clearly support the idea that all five positively evaluated personality characteristics— that is, low neuroticism and high extraversion, openness to experience, agreeableness, and conscientiousness—relate to higher self-esteem across time. Again, these associations occurred above and beyond the normative development of self-esteem. Except for extraversion, relations with self-esteem change were stable across time. Self-esteem showed a stronger relation with extraversion specifically in the second and third assessments. Being talkative, assertive, and enthusiastic appeared to be particularly important around the ages of 21 to 23. Being extraverted may ease a person’s ability to settle into new social environments, find new peers, and adjust to new roles. For example, extraversion has been shown to predict changes in relationships with friends and colleagues (Neyer & Asendorpf, 2001). Thus, the general motivation to gain more information (Carstensen, 1995) and increase one’s social networks (Wrzus et al., in press) could lead to

12

¨ DTKE, JONKMANN, AND TRAUTWEIN WAGNER, LU

the specific role of extraversion in early times of settlement in new social environments such as in university or job contexts.

Gender and Self-Esteem Change in the Transition to Young Adulthood Our results support earlier findings showing that men typically report higher self-esteem levels than women but no differences in change trajectories (Kling et al., 1999; McMullin & Cairney, 2004; Orth et al., 2010; Robins, Hendin, & Trzesniewski, 2001). Similar to most previous samples, our differences appeared to be relatively small. Such findings have been contrasted by studies without differences or with even contrary results (e.g., Erol & Orth, 2011). In this context, McMullin and Cairney (2004) discussed the possibility that the relations between self-esteem and possible moderators such as age, gender, or socioeconomic status may be more complicated. Furthermore, relations showed different patterns in the context of domain-specific self-esteem considerations (Gentile et al., 2009). Thus, to avoid false consequences (e.g., boosting only girls’ self-esteem and disregarding possible difficulties of young boys), a closer look at possible gender-specific moderators of self-esteem development is necessary. Interestingly, our findings show that young men and women differ in a number of conditional associations with self-esteem. First, having a stable romantic partnership was related to higher self-esteem beginning at age 23 for men, whereas the effect for women was only found at age 25, indicating lagged effects. Potential explanations of the general time-varying relations, beginning only in the mid-20s, might be due to mechanisms of social exclusion and value similarity (Leary & Baumeister, 2000). For instance, based on the human need to belong, Leary and Baumeister (2000) proposed that self-esteem is a gauge (sociometer) of one’s relational value. Supporting such notions, social inclusion by peers and (important) others has been shown to positively affect self-esteem in the short and long term (Denissen, Penke, Schmitt, & van Aken, 2008; Leary et al., 1995). Being single might be “normal” and, thus, ordinary in social encounters in one’s early 20s. Yet, with an increasing number of peers entering into stable partnerships, similarity in behaviors and values might be higher for those living in comparable life contexts. Experiences of belonging and shared value systems with peers could have profitable effects on self-esteem. Addressing our gender differences, we extended previous findings by Lehnart et al. (2010), who illustrated relations between self-esteem and partnership status only for men. We additionally found time-lagged relations for women who entered into a stable partnership and showed a substantial association with self-esteem only at approximately age 25. Thus, achieving the goal of having a romantic partnership appears to boost young men’s self-esteem earlier than would occur for young women. Lehnart et al. suggested that such differential gender effects are rooted in more beneficial consequences of a partnership (or marriage) for men compared to women, particularly with respect to health outcomes as well as increased social support. Likewise, men and women are known to have different mating preferences, with women being particularly choosy about their long-term partners (Buss, 2004; Gebauer, Leary, & Neberich, in press), which makes the selection of a reasonable romantic partner take more time. Finally, associations between personality characteristics and self-esteem appeared to be stable and similar for both genders in

the transition to young adulthood. In all cases, higher extraversion, openness to experience, and conscientiousness were related to higher self-esteem. Differential relations appeared regarding neuroticism and agreeableness that even showed gender and timespecific variability. Supporting previous findings (Roberts & Robins, 2004), being neurotic was negatively associated with self-esteem. Such detrimental associations of neuroticism occurred for both genders but appeared to decrease for male and increase for female participants across time. Although the tendency to be anxious and more vulnerable to stress has been shown to be higher in women (e.g., Costa, Terracciano, & McCrae, 2001; Donnellan & Lucas, 2008), reasons for such differential relations with selfesteem need further consideration. In contrast, young women reported higher self-esteem compared to men when they were more agreeable. This association between self-esteem and agreeableness, above and beyond general self-esteem development, increased across assessments only for women but not men. To be pleasant and accommodating in social situations appears to be particularly self-enhancing for young women. Such differences in behavioral tendencies could mimic the necessity to live up to gender-specific or normative behavior (Eagly, 2007; Eagly & Steffen, 1984). Being pleasant could relate to more positive feedback and thus provide a higher relational value for women, whereas it could be perceived as unmanly for young male adults. Likewise, this pattern might be a result of differential life goals in young students (Ferriman, Lubinski, & Benbow, 2009; Roberts & Robins, 2000): Male students were shown to be predisposed to “getting ahead,” pursuing goals of career and earning money, whereas female students tended to pursue goals under the broad heading of “getting along,” such as maintaining social relationships. With agreeableness being known to serve important functions in social relationship maintenance and communal strivings, women are more likely to foster this personality characteristic and, along with this, to take care of their self-esteem (Crocker, 2011; Gebauer, Wagner, Sedikides, & Neberich, 2012).

Limitations and Outlook Despite the strengths of having a longitudinal design with a large random sample and mostly multiple-item measures that allow for latent variable approaches, we need to consider a number of caveats with regard to our study. First, even with a longitudinal design, causal inferences are complicated by unobserved variables (Foster, 2010), which may affect self-esteem and its association with structural and personality time-varying covariates. In addition to third variables, Foster (2010) recognized that most likely there is no single causal effect. Specifically, young adulthood has been shown to be a time characterized by the coincidence of several developmental tasks. Thus, the relations between covariates and self-esteem across time not only varies across individuals, as shown here, but could also be affected by other constructs and life tasks that we were not able to include. Nevertheless, we maintain that the longitudinal design, the use of latent variable models, and the partial replication of results with different data analytical methods support our conclusions. Second, even though our sample was drawn randomly from schools, interpretations have to be narrowed to attendees of the secondary school system. Arnett (2000) has already drawn attention to the “forgotten half,” that is, people who do not attend college. We included a part of this half

SELF-ESTEEM IN YOUNG ADULTHOOD

because only 76% of our sample actually proceeded with further education. Nevertheless, there are a number of young adults with lower educational resources who should be given further consideration in psychological studies of self-esteem development. Third, all constructs that we used were based on self-report measures only. Although we believe that self-evaluation represents a valid method to capture self-esteem (cf. Donnellan, Trzesniewski, Conger, & Conger, 2007; Robins et al., 1999), it is also important to examine whether our findings would be replicated using other data sources (e.g., behavioral measures or peer reports). Finally, with respect to our structural characteristics we, for example, were not able to specify whether romantic partners were mutually exclusive and/or whether the quality of romantic relations relates to self-esteem development. So far it is an open question if having an exclusive and/or good partner relationship is important or just having one with respect to longitudinal self-esteem development. Regarding the two proposed theoretical accounts, we found support for both consulted approaches of life-span theory and social-investment principle: Supporting the life-span perspective (Baltes, 1987), self-esteem appears to have distinct normative change patterns across the transition into young adulthood. Thus, the gradual increase in self-esteem across time might illustrate both age-related positive (i.e., increased interdependence) as well as challenging (i.e., increased responsibility) conditions. Similarly, we found support for principles of social investment by showing that structural characteristics were related to self-esteem (Roberts & Wood, 2006). We clearly extended previous research by showing time-specific relations of having a partnership and being more extraverted with selfesteem. Such findings might be able to broaden or refine our knowledge of developmental processes of trait-like characteristics. In sum, the transition out of secondary school has to be regarded as an important developmental period with profound effects on the self in general (Arnett, 2000, 2007). In extending previous findings, our results reveal that self-esteem shows an initially slow but accelerating increase until the mid-20s. Young adults, moreover, illustrate substantial interindividual variability in starting points of self-esteem in the last school year and regarding change patterns across time. Both directions of a more positive but also a more negative evaluation of the self occur. To have a stable partnership and the development of a mature personality constitute important contexts for the positive evaluation of the self above and beyond normative changes in self-esteem. Finally, conditions of selfesteem change show time-specific as well as gender-specific relations that should be regarded in upcoming research.

References Abele, A. E. (2002). Ein Modell und empirische Befunde zur beruflichen Laufbahnentwicklung unter besonderer Beru¨cksichtigung des Geschlechtsvergleichs [A model and empirical findings on career development under special considerations of gender comparisons]. Psychologische Rundschau, 53, 109 –118. doi:10.1026//0033-3042.53.3.109 Abele, A. E. (2005). Ziele, Selbstkonzept und Work-life-Balance bei der längerfristigen Lebensgestaltung: Befunde der Erlanger Längsschnittstudie bela-e mit Akademikerinnen und Akademikern [Goals, self-concept, and work–life balance in long-term way of life: Findings of the Erlanger longitudinal study bela-e with academics]. Zeitschrift fu¨r

13

Arbeits- und Organisationspsychologie, 49, 176 –186. doi:10.1026/ 0932-4089.49.4.176 Arnett, J. J. (2000). Emerging adulthood: A theory of development from the late teens through the twenties. American Psychologist, 55, 469 – 480. doi:10.1037/0003-066X.55.5.469 Arnett, J. J. (2007). Emerging adulthood: What is it, and what is it good for? Child Development Perspectives, 1, 68 –73. doi:10.1111/j.17508606.2007.00016.x Asendorpf, J. B., & van Aken, M. A. G. (2003). Personality-relationship transaction in adolescence: Core versus surface personality characteristics. Journal of Personality, 71, 629 – 666. doi:10.1111/14676494.7104005 Baltes, P. B. (1987). Theoretical propositions of life-span developmental psychology: On the dynamics between growth and decline. Developmental Psychology, 23, 611– 626. doi:10.1037/0012-1649.23.5.611 Baumeister, R. F., Campbell, J. D., Krueger, J. I., & Vohs, K. D. (2003). Does high self-esteem cause better performance, interpersonal success, happiness, or healthier lifestyles? Psychological Science in the Public Interest, 4, 1– 44. Bollen, K. A., & Curran, P. J. (2006). Latent curve models: A structural equation perspective. Hoboken, NJ: Wiley. Borkenau, P., & Ostendorf, F. (1991). Ein Fragebogen zur Erfassung fu¨nf robuster Persönlichkeitsfaktoren [A questionnaire for the assessment of five robust personality factors]. Diagnostica, 37, 29 – 41. Bosak, J., Sczesny, S., & Eagly, A. H. (2008). Communion and agency judgments of women and men as a function of role information and response format. European Journal of Social Psychology, 38, 1148 – 1155. doi:10.1002/ejsp.538 Brandtstädter, J., & Greve, W. (1994). The aging self: Stabilizing and protective processes. Developmental Review, 14, 52– 80. doi:10.1006/ drev.1994.1003 Bundesamt, S. (2010). Frauen und Männer in verschiedenen Lebensphasen [Women and men in different life stages]. Wiesbaden, Germany: Statistisches Bundesamt. Buss, D. M. (2004). Evolutionäre Psychologie [Evolutionary psychology]. Mu¨nchen, Germany: Pearson Studium. Carstensen, L. L. (1995). Evidence for a life-span theory of socioemotional selectivity. Current Directions in Psychological Science, 4, 151–156. doi:10.1111/1467-8721.ep11512261 Costa, P. T., Jr., & McCrae, R. R. (1992). Normal personality assessment in clinical practice: The NEO personality inventory. Psychological Assessment, 4, 5–13. doi:10.1037/1040-3590.4.1.5 Costa, P. T., Jr., Terracciano, A., & McCrae, R. R. (2001). Gender differences in personality traits across cultures: Robust and surprising findings. Journal of Personality and Social Psychology, 81, 322–331. doi:10.1037/0022-3514.81.2.322 Crocker, J. (2011). Presidential address: Self-image and compassionate goals and construction of the social self: Implications for social and personality psychology. Personality and Social Psychology Review, 15, 394 – 407. doi:10.1177/1088868311418746 Crocker, J., & Luhtanen, R. K. (2003). Level of self-esteem and contingencies of self-worth: Unique effects on academic, social and financial problems in college students. Personality and Social Psychology Bulletin, 29, 701–712. doi:10.1177/0146167203029006003 Crocker, J., & Park, L. E. (2004). The costly pursuit of self-esteem. Psychological Bulletin, 130, 392– 414. doi:10.1037/00332909.130.3.392 Curran, P. J., & Bauer, D. J. (2011). The disaggregation of within-person and between-person effects in longitudinal models of change. Annual Review of Psychology, 62, 583– 619. doi:10.1146/annurev .psych.093008.100356 Denissen, J. J. A., Penke, L., Schmitt, D. P., & van Aken, M. A. G. (2008). Self-esteem reactions to social interactions: Evidence for sociometer

14

¨ DTKE, JONKMANN, AND TRAUTWEIN WAGNER, LU

mechanisms across days, people, and nations. Journal of Personality and Social Psychology, 95, 181–196. doi:10.1037/0022-3514.95.1.181 Donnellan, M. B., & Lucas, R. E. (2008). Age differences in the Big Five across the life span: Evidence from two national samples. Psychology and Aging, 23, 558 –566. doi:10.1037/a0012897 Donnellan, M. B., Trzesniewski, K. H., Conger, K. J., & Conger, R. D. (2007). A three-wave longitudinal study of self-evaluations during young adulthood. Journal of Research in Personality, 41, 453– 472. doi:10.1016/j.jrp.2006.06.004 Eagly, A. H. (2007). Female leadership advantage and disadvantage: Resolving the contradictions. Psychology of Women Quarterly, 31, 1–12. doi:10.1111/j.1471-6402.2007.00326.x Eagly, A. H., & Steffen, V. J. (1984). Gender stereotypes stem from the distribution of women and men into social roles. Journal of Personality and Social Psychology, 46, 735–754. doi:10.1037/0022-3514.46.4.735 Erol, R. Y., & Orth, U. (2011). Self-esteem development from age 14 to 30 years: A longitudinal study. Journal of Personality and Social Psychology, 101, 607– 619. doi:10.1037/a0024299 Ferriman, K., Lubinski, D., & Benbow, C. P. (2009). Work preferences, life values, and personal views of top math/science graduate students and the profoundly gifted: Developmental changes and gender differences during emerging adulthood and parenthood. Journal of Personality and Social Psychology, 97, 517–532. doi:10.1037/a0016030 Foster, E. M. (2010). Causal inference and developmental psychology. Developmental Psychology, 46, 1454 –1480. doi:10.1037/a0020204 Freund, A. M., & Baltes, P. B. (2005). Entwicklungsaufgaben als Organisationsstrukturen von Entwicklung und Entwicklungsoptimierung [Developmental tasks as organizing structures of development]. In S. H. Filipp & U. M. Staudinger (Eds.), Enzyclopädie der psychologie. Göttingen, Germany: Hogrefe. Freund, A. M., & Smith, J. (1999). Content and function of the selfdefinition in old and very old age. The Journals of Gerontology: Series B: Psychological Sciences and Social Sciences, 54(B), P55–P67. doi: 10.1093/geronb/54B.1.P55 Gebauer, J. E., Leary, M. R., & Neberich, W. (in press). Big Two personality and Big Three mate preferences: Similarity attracts, but countrylevel mate preferences crucially matter. Personality and Social Psychology Bulletin. Gebauer, J. E., Wagner, J., Sedikides, C., & Neberich, W. (2012). Agencycommunion and self-esteem relations are moderated by culture, religiosity, age, and sex: Evidence for the “self-centrality breeds selfenhancement” principle. Journal of Personality. Advance online publication. doi:10.1111/j.1467-6494.2012.00807.x Gentile, B., Grabe, S., Dolan-Pascoe, B., Twenge, J. M., Wells, B. E., & Maitino, A. (2009). Gender differences in domain-specific self-esteem: A meta-analysis. Review of General Psychology, 13, 34 – 45. doi: 10.1037/a0013689 Graber, J. A., & Dubas, J. S. (1996). Leaving home: Understanding the transition to adulthood. San Francisco, CA: Jossey-Bass. Greenberg, J., Solomon, S., Pyszczynski, T., Rosenblatt, A., Burling, J., Lyon, D., . . . Pinel, E. (1992). Why do people need self-esteem? Converging evidence that self-esteem serves an anxiety-buffering function. Journal of Personality and Social Psychology, 63, 913–922. doi: 10.1037/0022-3514.63.6.913 Grimm, K. J. (2007). Multivariate longitudinal methods for studying developmental relationships between depression and academic achievement. International Journal of Behavioral Development, 31, 328 –339. doi:10.1177/0165025407077754 Hagestad, G. O., & Call, V. R. A. (2007). Pathways to childlessness: A life course perspective. Journal of Family Issues, 28, 1338 –1361. doi: 10.1177/0192513X07303836 Havighurst, R. (1953). Developmental tasks and education. New York, NY: Longmans, Green. Heckhausen, J., Dixon, R. A., & Baltes, P. B. (1989). Gains and losses in

development throughout adulthood as perceived by different adult age groups. Developmental Psychology, 25, 109 –121. doi:10.1037/00121649.25.1.109 Heckhausen, J., & Schulz, R. (1995). A life-span theory of control. Psychological Review, 102, 284 –304. doi:10.1037/0033-295X.102.2.284 Heller, K. A., & Perleth, C. (2000). Kognitiver Fähigkeitstest fu¨r 4. bis 12. Klassen (KFT 4-12⫹R) [Test of Cognitive Functioning for Grades 4 to 12]. Göttingen, Germany: Beltz-Test. Helson, R., Jones, C., & Kwan, V. S. Y. (2002). Personality change over 40 years of adulthood: Hierarchical linear modeling analyses of two longitudinal samples. Journal of Personality and Social Psychology, 83, 752–766. doi:10.1037/0022-3514.83.3.752 Hopwood, C. J., Donnellan, M. B., Blonigen, D. M., Krueger, R. F., McGue, M., Iacono, W. G., & Burt, S. A. (2011). Genetic and environmental influences on personality trait stability and growth during the transition to adulthood: A three-wave longitudinal study. Journal of Personality and Social Psychology, 100, 545–556. doi: 10.1037/ a0022409 Huang, C. (2010). Mean-level change in self-esteem from childhood through adulthood: Meta-analysis of longitudinal studies. Review of General Psychology, 14, 251–260. doi:10.1037/a0020543 Jackson, J. J., Thoemmes, F., Jonkmann, K., Lu¨dtke, O., & Trautwein, U. (2012). Military training and personality trait development: Does the military make the man or does the man make the military? Psychological Science, 23, 270 –277. doi:10.1177/0956797611423545 James, W. (1950). The principles of psychology. New York, NY: Dover. (Original work published 1890) Kling, K. C., Hyde, J. S., Showers, C. J., & Buswell, B. N. (1999). Gender differences in self-esteem: A meta-analysis. Psychological Bulletin, 125, 470 –500. doi:10.1037/0033-2909.125.4.470 Lang, F. R., Wagner, J., Wrzus, C., & Neyer, F. J. (2012). Personal efforts in regulating social relationships across adulthood. Manuscript submitted for publication. Leary, M. R., & Baumeister, R. F. (2000). The nature and function of self-esteem: Sociometer theory. In M. P. Zanna (Ed.), Advances in experimental social psychology (Vol. 32, pp. 1– 62). San Diego, CA: Academic Press. Leary, M. R., Tambor, E. S., Terdal, S. K., & Downs, D. L. (1995). Self-esteem as an interpersonal monitor: The sociometer hypothesis. Journal of Personality and Social Psychology, 68, 518 –530. doi: 10.1037/0022-3514.68.3.518 Lehnart, J., Neyer, F. J., & Eccles, J. (2010). Long-term effects of social investment: The case of partnering in young adulthood. Journal of Personality, 78, 639 – 670. doi:10.1111/j.1467-6494.2010.00629.x Lucas, R. E., & Donnellan, M. B. (2011). Personality development across the life span: Longitudinal analyses with a national sample from Germany. Journal of Personality and Social Psychology, 101, 847– 861. doi:10.1037/a0024298 Lu¨dtke, O., Roberts, B. W., Trautwein, U., & Nagy, G. (2011). A random walk down university avenue: Life paths, life events, and personality trait change at the transition to university life. Journal of Personality and Social Psychology, 101, 620 – 637. doi:10.1037/a0023743 Lu¨dtke, O., Trautwein, U., Nagy, G., & Köller, O. (2004). Eine Validierungsstudie zum NEO–FFI in einer Stichprobe junger Erwachsener: Effekte des Itemformats, faktorielle Validität und Zusammenhänge mit Schulleistungsindikatoren [A validation of the NEO–FFI in a sample of young adults: Effects of the response format, factorial validity, and relations with indicators of academic achievement]. Diagnostica, 50, 134 –144. Marsh, H. W. (1992). Content specificity of relations between academic achievement and academic self-concept. Journal of Educational Psychology, 84, 35– 42. doi:10.1037/0022-0663.84.1.35 Marsh, H. W., & Hau, K.-T. (1996). Assessing goodness of fit: Is parsi-

SELF-ESTEEM IN YOUNG ADULTHOOD mony always desirable? The Journal of Experimental Education, 64, 364 –390. Marsh, H. W., Hau, K.-T., & Grayson, D. (2005). Goodness of fit in structural equation models. In A. Maydeu-Olivares & J. J. McArdle (Eds.), Contemporary psychometrics: A Festschrift for Roderick P. McDonald (pp. 275–340). Mahwah, NJ: Erlbaum. McAdams, D. P., & Olson, B. D. (2010). Personality development: Continuity and change over the life course. Annual Review of Psychology, 61, 517–542. doi:10.1146/annurev.psych.093008.100507 McCrae, R. R., Costa, P. T., Jr., Ostendorf, F., Angleitner, A., Hrebı´ckova´, M., Avia, M. D., . . . Smith, P. B. (2000). Nature over nurture: Temperament, personality, and life span development. Journal of Personality and Social Psychology, 78, 173–186. doi:10.1037/0022-3514.78.1.173 McMullin, J. A., & Cairney, J. (2004). Self-esteem and the intersection of age, class, and gender. Journal of Aging Studies, 18, 75–90. doi:10.1016/ j.jaging.2003.09.006 Mehta, P. D., & West, S. G. (2000). Putting the individual back into individual growth curves. Psychological Methods, 5, 23– 43. doi: 10.1037/1082-989X.5.1.23 Mulder, C. H. (2003). The effects of singlehood and cohabitation on the transition to parenthood in the Netherlands. Journal of Family Issues, 24, 291–313. doi:10.1177/0192513X02250885 Muthe´n, B. O. (1993). Latent variable modeling of growth with missing data and multilevel data. In C. M. Cuadras & C. R. Rao (Eds.), Multivariate analysis: No. 2. Future directions (pp. 199 –210). Amsterdam, the Netherlands: North Holland. Muthe´n, L. K., & Muthe´n, B. O. (1998 –2010). Mplus user’s guide (6th ed.). Los Angeles, CA: Muthe´n & Muthe´n. Neyer, F. J., & Asendorpf, J. B. (2001). Personality-relationship transaction in young adulthood. Journal of Personality and Social Psychology, 81, 1190 –1204. doi:10.1037/0022-3514.81.6.1190 Neyer, F. J., & Lehnart, J. (2007). Relationships matter in personality development: Evidence from an 8-year longitudinal study across young adulthood. Journal of Personality, 75, 535–568. doi:10.1111/j.14676494.2007.00448.x O’Malley, P. M., & Bachman, J. G. (1983). Self-esteem: Change and stability between ages 13 and 23. Developmental Psychology, 19, 257– 268. doi:10.1037/0012-1649.19.2.257 Orth, U., Robins, R. W., Trzesniewski, K. H., Maes, J. R., & Schmitt, M. (2009). Low self-esteem is a risk factor for depressive symptoms from young adulthood to old age. Journal of Abnormal Psychology, 118, 472– 478. doi:10.1037/a0015922 Orth, U., Robins, R. W., & Widaman, K. F. (2012). Life-span development of self-esteem and its effects on important life outcomes. Journal of Personality and Social Psychology, 102, 1271–1288. doi:10.1037/ a0025558 Orth, U., Trzesniewski, K. H., & Robins, R. W. (2010). Self-esteem development from young adulthood to old age: A cohort-sequential longitudinal study. Journal of Personality and Social Psychology, 98, 645– 658. doi:10.1037/a0018769 Preacher, K. J., Wichman, A. L., MacCallum, R. C., & Briggs, N. E. (2008). Quantitative Applications in the Social Sciences: Vol. 157. Latent growth curve modeling. Los Angeles, CA: Sage. Pullmann, H., Allik, J., & Realo, A. (2009). Global self-esteem across the life span: A cross-sectional comparison between representative and self-selected Internet samples. Experimental Aging Research, 35, 20 – 44. doi:10.1080/03610730802544708 Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models: Applications and data analysis methods. Thousand Oaks, CA: Sage. Raudenbush, S. W., & Chan, W. S. (1993). Application of a hierarchical linear model to the study of adolescent deviance in an overlapping cohort design. Journal of Consulting and Clinical Psychology, 61, 941–951. doi:10.1037/0022-006X.61.6.941 Reitzes, D. C., & Mutran, E. J. (1994). Multiple roles and identities:

15

Factors influencing self-esteem among middle-aged working men and women. Social Psychology Quarterly, 57, 313–325. doi:10.2307/ 2787158 Roberts, B. W., & DelVecchio, W. F. (2000). The rank-order consistency of personality traits from childhood to old age: A quantitative review of longitudinal studies. Psychological Bulletin, 126, 3–25. doi:10.1037/ 0033-2909.126.1.3 Roberts, B. W., & Robins, R. W. (2000). Broad dispositions, broad aspirations: The intersection of personality traits and major life goals. Personality and Social Psychology Bulletin, 26, 1284 –1296. doi: 10.1177/0146167200262009 Roberts, B. W., & Robins, R. W. (2004). Person– environment fit and its implications for personality development: A longitudinal study. Journal of Personality, 72, 89 –110. doi:10.1111/j.0022-3506.2004.00257.x Roberts, B. W., Walton, K. E., & Viechtbauer, W. (2006). Patterns of mean-level change in personality traits across the life course: A metaanalysis of longitudinal studies. Psychological Bulletin, 132, 1–25. doi: 10.1037/0033-2909.132.1.1 Roberts, B. W., & Wood, D. (2006). Personality development in the context of the neo-socioanalytic model of personality. In D. K. Mroczek & T. D. Little (Eds.), Handbook of personality development (pp. 11–39). Mahwah, NJ: Erlbaum. Roberts, B. W., Wood, D., & Smith, J. L. (2005). Evaluating five factor theory and social investment perspectives on personality trait development. Journal of Research in Personality, 39, 166 –184. doi:10.1016/ j.jrp.2004.08.002 Robins, R. W., Hendin, H. M., & Trzesniewski, K. H. (2001). Measuring global self-esteem: Construct validation of a single-item measure and the Rosenberg Self-Esteem Scale. Personality and Social Psychology Bulletin, 27, 151–161. doi:10.1177/0146167201272002 Robins, R. W., Norem, J. K., & Cheek, J. M. (1999). Naturalizing the self. In L. A. Pervin & O. P. John (Eds.), Handbook of personality: Theory and research (2nd ed., pp. 443– 4777). New York, NY: Guilford Press. Robins, R. W., Tracy, J. L., Trzesniewski, K., Potter, J., & Gosling, S. D. (2001). Personality correlates of self-esteem. Journal of Research in Personality, 35, 463– 482. doi:10.1006/jrpe.2001.2324 Robins, R. W., Trzesniewski, K. H., Tracy, J. L., Gosling, S. D., & Potter, J. (2002). Global self-esteem across the life span. Psychology and Aging, 17, 423– 434. doi:10.1037/0882-7974.17.3.423 Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55, 68 –78. doi:10.1037/0003-066X.55.1.68 Sayer, A. G., & Cumsille, P. E. (2001). Second-order latent growth models. In L. M. Collins & A. G. Sayer (Eds.), New methods for the analysis of change (pp. 179 –200). Washington, DC: American Psychological Association. Schafer, J. L., & Graham, J. W. (2002). Missing data: Our view of the state of the art. Psychological Methods, 7, 147–177. doi:10.1037/1082989X.7.2.147 Schwanzer, A. D., Trautwein, U., Lu¨dtke, O., & Sydow, H. (2005). Entwicklung eines Instruments zur Erfassung des Selbstkonzepts junger Erwachsener [Development of a self-concept questionnaire for young adults]. Diagnostica, 51, 183–194. doi:10.1026/0012-1924.51.4.183 Shaw, B. A., Liang, J., & Krause, N. (2010). Age and race differences in the trajectories of self-esteem. Psychology and Aging, 25, 84 –94. doi: 10.1037/a0018242 Singer, J. D., & Willet, J. B. (2003). Applied longitudinal data analysis: Modeling change and event occurrence. New York, NY: Oxford University Press. Srivastava, S., John, O. P., Gosling, S. D., & Potter, J. (2003). Development of personality in early and middle adulthood: Set like plaster or persistent change? Journal of Personality and Social Psychology, 84, 1041–1053. doi:10.1037/0022-3514.84.5.1041 Tesser, A., Crepaz, N., Collins, J. C., Cornell, D., & Beach, S. R. H.

16

¨ DTKE, JONKMANN, AND TRAUTWEIN WAGNER, LU

(2000). Confluence of self-esteem regulation mechanisms: On integrating the self-zoo. Personality and Social Psychology Bulletin, 26, 1476 – 1489. doi:10.1177/01461672002612003 Tognoli, J. (2003). Leaving home: Homesickness, place attachment, and transition among residential college students. Journal of College Student Psychotherapy, 18, 35– 48. doi:10.1300/J035v18n01_04 Trautwein, U., Lu¨dtke, O., Köller, O., & Baumert, J. (2006). Self-esteem, academic self-concept, and achievement: How the learning environment moderates the dynamics of self-concept. Journal of Personality and Social Psychology, 90, 334 –349. doi:10.1037/0022-3514.90.2.334 Trautwein, U., Neumann, M., Nagy, G., Lu¨dtke, O., & Maaz, K. (Eds.). (2010). Schulleistungen von Abiturienten: Die neu geordnete gymnasiale Oberstufe auf dem Pru¨fstand [School performance of high-school graduates: Testing the new ordered upper school]. Wiesbaden, Germany: VS Verlag fu¨r Sozialwissenschaften. Trzesniewski, K. H., Donnellan, M. B., Moffitt, T. E., Robins, R. W., Poulton, R., & Caspi, A. (2006). Low self-esteem during adolescence predicts poor health, criminal behavior, and limited economic prospects during adulthood. Developmental Psychology, 42, 381–390. doi: 10.1037/0012-1649.42.2.381 Trzesniewski, K. H., Donnellan, M. B., & Robins, R. W. (2003). Stability of self-esteem across the life span. Journal of Personality and Social Psychology, 84, 205–220. doi:10.1037/0022-3514.84.1.205 Trzesniewski, K. H., Robins, R. W., Roberts, B. W., Caspi, A., & Paul, C. (2004). Personality and self-esteem development across the life span. In

P. T. Costa Jr. & I. C. Siegler (Eds.), Recent advances in psychology and aging (Vol. 15, pp. 163–185). Amsterdam, the Netherlands: Elsevier Science. Twenge, J. M., & Campbell, W. K. (2001). Age and birth cohort differences in self-esteem: A cross-temporal meta-analysis. Personality and Social Psychology Review, 5, 321–344. doi:10.1207/S15327957PSPR0504_3 van Aken, M. A. G., Denissen, J. J. A., Branje, S. J. T., Dubas, J. S., & Goossens, L. (2006). Midlife concerns and short-term personality change in middle adulthood. European Journal of Personality, 20, 497–513. doi:10.1002/per.603 Watson, D., Suls, J., & Haig, J. (2002). Global self-esteem in relation to structural models of personality and affectivity. Journal of Personality and Social Psychology, 83, 185–197. doi:10.1037/0022-3514.83.1.185 Wrzus, C., Köckeritz, M., Wagner, J., & Neyer, F. J. (in press). Social network changes and life events across the life span: A meta-analysis. Psychological Bulletin. Advance online publication. doi:10.1037/ a0028601 Wu, W., West, S. G., & Taylor, A. B. (2009). Evaluating model fit for growth curve models: Integration of fit indices from SEM and MLM frameworks. Psychological Methods, 14, 183–201. doi:10.1037/ a0015858

Received December 6, 2011 Revision received June 22, 2012 Accepted July 17, 2012 䡲