Journal of Vocational Behavior 70 (2007) 463477 www.elsevier.

com/locate/jvb

Self-esteem during university studies predicts career characteristics 10 years later q

Katariina Salmela-Aro *, Jari-Erik Nurmi
Department of Psychology, Agora, P.O. Box 35, University of Jyva skyla , 40014 Jyva skyla , Finland Received 26 April 2006 Available online 6 February 2007

Abstract To examine how self-esteem measured during university studies would impact on the characteristics of the work career 10 years later, 297 university students completed the Rosenbergs self-esteem inventory four times while at university and various career-related questionnaires 10 years later. Latent Growth Curve Modeling showed that a high overall level of self-esteem predicted being in permanent employment 10 years later, having a high salary, and reporting a high level of work engagement, and job satisfaction, and a low level of burnout. By contrast, low self-esteem predicted unemployment, feelings of exhaustion, cynicism and reduced accomplishment at work, and low levels of work engagement and job satisfaction. 2007 Elsevier Inc. All rights reserved.
Keywords: Self-esteem; Job engagement; Burnout; Salary; Unemployment; Longitudinal

1. Introduction Self-esteem is generally regarded as an evaluation in which people express approval or disapproval of themselves and make judgements about their personal worth (Rosenberg, Schooler, Schoenbach, & Rosenberg, 1995; Suls, 1989). Although a large body of research has been carried out on self-esteem (For a review, see Baumeister, Campbell, Krueger, & Vohs, 2003), little is known about how self-esteem predicts later life-trajectories, such as the early phase of the career and working life. It might be assumed that high self-esteem,
q *

This study was funded by grant from the Finnish Academy (121 0319). Corresponding author. Fax: +358 14 260 2841. E-mail address: katariina.salmela-aro@psyka.jyu. (K. Salmela-Aro).

0001-8791/$ - see front matter 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jvb.2007.01.006

464

K. Salmela-Aro, J.-E. Nurmi / Journal of Vocational Behavior 70 (2007) 463477

which is a favourable global evaluation of the self, provides a foundation for success in dealing with challenges at work and in the career. In turn, low self-esteem, an unfavourable evaluation of the self, might lead to diculties in making the transition to work and the early vocational career. The present 15-year, ve-wave longitudinal study examined changes in self-esteem among young people during their university studies and to what extent self-esteem predicted their subsequent work experiences, in terms of both the working life situation (a permanent job, unemployment, and level of salary) and work-related attitudes (work engagement, job satisfaction, and burnout), 10 years later. Self-esteem is dened as the value people place on themselves (Rosenberg et al., 1995). It is the evaluative component of self-knowledge. High self-esteem refers to a highly favourable and low self-esteem to an unfavourable global evaluation of the self (Baumeister, 1993). High self-esteem demonstrates a sense of self-worth with the implication that one will be accepted rather than rejected by others, and that one is not a failure in ones life. Previous cross-sectional research has shown that, after a period of decreasing selfesteem during adolescence (Trzesniewski, Donnellan, & Robins, 2003), self-esteem typically increases during the 20s and the 30s (Kling, Hyde, Showers, & Buswell, 1999; Robins, Trzesniewski, Tracy, Gosling, & Potter, 2002). However, few longitudinal studies have examined changes in self-esteem during young adulthood. In one study, Galambos, Barker, and Krahn (2006) found that self-esteem increased during this period. In another study, Schulenberg, OMalley, Bachman, and Johnston (2005) showed that well-being (a composite measure consisting of self-esteem, self-ecacy, and social support) increased in a longitudinal sample followed from age 18 to 24. However, although self-esteem might improve in young adulthood, such trajectories may also show substantial inter-individual variation. Individuals face many more transitions and life-decisions during young adulthood than at any other stage of life (Caspi, 2002; Grob, Krings, & Bangerter, 2001). Such transitions and decisions include taking responsibility of oneself, achieving nancial independence from parents, building intimate relationships and starting a family (Arnett, 2000; Erikson, 1968; Lefkowitz, 2005; Masten et al., 1999; Schulenberg, Maggs, & OMalley, 2003; Shanahan, 2000; Shiner & Masten, 2002). One transition that the vast majority of individuals face during young adulthood is the transition from nishing their education to entry into working life. Achieving a certain level of education lays the foundation for entry into a particular occupation. Resolution of work-related challenges then provides the basis for a successful transition to adulthood, whereas failure in such resolution may lead to the risk of later developmental disadvantage and problems (Schulenberg et al., 2003). It might be also assumed that individuals self-esteem during young adulthood provides the basis for their success in dealing with the major developmental tasks and challenges they face. The transition into working life is a good example. Entrance into working life, however, is a complex phenomenon involving many processes and outcomes. Not surprisingly, various concepts have been used in describing it, such as occupational success (Judge & Bono, 2001), unemployment (e.g., Kokko & Pulkkinen, 1998; Wineeld, Tiggemann, & Wineeld, 1992), and re-employment (Caplan, Vinokur, Price, & Van Ryn, 1989; Waters & Moore, 2002; Wineeld & Tiggemann, 1985). Entry into working life is not only reected in career characteristics but also in several evaluative components of work, such as work-related satisfaction and engagement on the one hand, and burnout (Schaufeli, Maslach, & Marek, 1993) and workaholism (Burke, 2001; Taris, Schaufeli, & Verhoeven, 2005) on the other.

K. Salmela-Aro, J.-E. Nurmi / Journal of Vocational Behavior 70 (2007) 463477

465

A few studies have been carried out on the role of self-esteem in peoples subsequent work career and its evaluative components. Some of them have shown that high selfesteem is linked to a low likelihood of unemployment (e.g., Kokko & Pulkkinen, 1998; Wineeld et al., 1992) and a high likelihood of re-employment (Kasl et al., 1992). Other research has shown that self-esteem is associated with the evaluative components of work (Burke, 2001; Schaufeli et al., 1993; Taris et al., 2005). It has been shown, for example, that self-esteem is related to burnout (Janssen, Schaufeli, & Houkes, 1999). Similarly, Ma kikangas, Kinnunen, and Feldt (2004) found recently that low self-esteem was related to all three components of burnout (i.e., exhaustion, cynicism and reduced accomplishment). By contrast, associations have been found between high self-esteem and work-related engagement (i.e., feelings of vigor, dedication and absorption in ones working life later nez, Marques Pinto, Salanova, & Bakker, 2002). on) (Schaufeli, Mart Although considerable amount of cross-sectional research has been carried out on the associations between self-esteem and individuals transition to working life and career characteristics, few longitudinal studies have examined this topic. Consequently, little is known about whether it is self-esteem that contributes to work and career outcomes, or whether it is rather work characteristics that predict changes in self-esteem. For example, Baumeister et al. (2003) suggested recently that, although there is some empirical evidence that high self-esteem goes together with better job performance, it is likely that occupational success buers self-esteem rather than vice versa. The present study sought to investigate to what extent young adults self-esteem, and changes in it, during university studies would predict the outcomes of their transition to working life (unemployment, permanent job, salary, managerial position, number of subordinates, working hours) and their attitudes towards their work (work engagement, satisfaction, and burnout) 10 years later. There are several mechanisms by which self-esteem may impact on the transition to working life and work-related attitudes during young adulthood. One possibility is that young adults with high self-esteem, as compared to those with low self-esteem, are active in dealing with career-related issues, such as nishing their education successfully, managing their budgetary resources, and nding an appropriate job, which then contributes to their success in the transition to work and the early career. Another possible mechanism is that high self-esteem increases career-related self-ecacy beliefs which then lead to persistence in the face of obstacles (Baumeister et al., 2003) and high quality of performance in the career domain (see Betz, 2004; Lent, Hackett, & Brown, 1999). With respect to the evaluative aspects of work and career, a third possible mechanism is that individuals with low self-esteem tend to be emotionally more vulnerable to negative attitudes towards their work. They may also interact less adequately with other people in their work. These characteristics may in particular increase their susceptibility to burnout and to feelings of reduced accomplishments. The fourth possibility is that strong self-esteem is a force that strengthens the workers belief in his or her own ability and signicance, and thus supports active coping in stressful situations. Consequently, high self-esteem may lead to workrelated engagement, i.e., feelings of vigor, dedication and absorption in ones working life later on (Schaufeli et al., 2002). In the present study, we investigated to what extent young adults self-esteem during university studies would predict their career characteristics and evaluative components of work 10 years later when they were facing the transition to working life as new employees. We assumed that self-esteem would be relatively stable during young adulthood (Trzesniewski et al., 2003), and that self-esteem during a challenging life period, such as

466

K. Salmela-Aro, J.-E. Nurmi / Journal of Vocational Behavior 70 (2007) 463477

university studies, would be predictive of individuals success in dealing with another demanding stage of life, in our case the transition to work. Because self-esteem during the university years might be assumed to be inuenced by academic achievement (Baumeister et al., 2003), which may also contribute to success in the transition to work, the level of academic achievement during those years was controlled for in our analyses when the role of self-esteem was examined. 2. Aims The present study had two aims. Our rst aim was to examine to what extent young adults self-esteem would change over a six-year period of university studies. Our second aim was to investigate to what extent young adults self-esteem, and changes in it, during their university studies, would predict their career characteristics (unemployment, permanent job, salary, managerial position, number of subordinates, working hours) and workrelated satisfaction and attitudes (work engagement, satisfaction, and burnout) 10 years later. 3. Methods 3.1. Participants and procedure This study is part of the ongoing Helsinki Longitudinal Student Study (HELS study). The participants were 297 (78 men, 219 women) 18- to 25-year-old (M = 20.61, SD = 1.80) undergraduates studying various subjects (Biology, Geography, Economics, English, Finnish, French, History, Psychology, and Sociology) at the University of Helsinki at the time of the rst measurement. Because we were interested in young adulthood, seven participants in the original sample (N = 304) who were aged 30 or older were excluded from the analyses. The sample was reasonably representative of the university student population in Finland in terms of ethnicity and socio-economic status. Women were over-represented in the study, since females comprise the majority of students in the humanities and social sciences, from where most of the participants were drawn. The participants were measured at ve time-points. (1) A total of 297 (78 men, 219 women) undergraduates were rst investigated during their rst autumn term (rst study year). All the students present during the day of the study were furnished with preliminary information about the study and asked if they would consent to participate. None of them refused. (2) Two years after Measurement Time 1 (third study year), 263 students (66 men, 197 women) were examined (retention rate 89%). (3) Four years after Measurement Time 1 (fth study year), 237 students (60 men, 177 women) answered the questionnaire (retention rate 80%). (4) Six years after Measurement Time 1 (the seventh year after starting university studies), 243 students (62 men, 181 women) returned the questionnaire (retention rate 82%). On the rst four measurement occasions, the participants lled in a self-esteem scale (Rosenberg et al., 1995), and gave some background information. (5) Ten years after Measurement Time 4 (15 years after Measurement Time 1), the participants (M age = 36.60, SD = 1.80) in the original sample were sent questionnaires measuring their work-related performance and experience (unemployment, temporary or permanent job, salary, managerial position, subordinates, work satisfaction, engagement, burnout, workaholism, working hours). Two-hundred and two students returned the questionnaires (retention

K. Salmela-Aro, J.-E. Nurmi / Journal of Vocational Behavior 70 (2007) 463477

467

rate 69%). Of the dropouts, three had died, ve refused to participate, 30 had an unknown address and 12 had moved abroad. At Measurement Time 5, 9% were unemployed, 87% were employed and the rest were studying, or taking care of children at home. Among those who were working, 63% had a permanent and 37% had a temporary job. We compared the participants who dropped out of the study at a particular measurement point to those who did not drop out. Those who had dropped out (M = 21.22, SD = 10.68) before Measurement Time 2 showed less progress in their studies during their rst two university years than those who continued to participate (M = 26.42, SD = 11.59; t (297) = 3.04, p < .01). Those who had dropped out before Measurement Time 3 (M = 16.30, SD = 12.20) showed less progress during their third and fourth university years than those who continued to participate (M = 22.85, SD = 14.11; t (297) = 3.35, p < .001). However, the results showed no dierences in self-esteem between those who dropped out and to those who continued to participate. 3.2. Measures Self-esteem. Self-esteem was measured by Rosenberg et al. (1995) 10-item scale. This scale is a self-report measure of generalized feelings about the self. The self-esteem items (e.g., I feel I have a number of good qualities; At times, I think I am no good at all) were rated on a four-point Likert scale (1, totally disagree; 4, totally agree). The Cronbach alpha coecients for the scale were between 0.87 and 0.89. The correlation between the self-esteem measure and the depressive symptoms measure (Beck, Ward, Mendelsohn, Mosck, & Erlaugh, 1961) at Time 1 was .72 (p < .001) and at Time 3 .65 (p < .001), and the UCLA Loneliness scale (Russell, Peplau, & Cutrona, 1980) at Time 2 was .42, p < .001 (Salmela-Aro, 2005). Unemployment. Unemployment was elicited by the question Are you unemployed at the moment?. Nine percent of the participants were unemployed at Time 5. Those who were working were asked about the following: Permanent job. The question was Do you have a permanent or temporary position?. 37% were in a temporary and 63% in a permanent job. Managerial position. The question was What kind of position you have? with 4 response alternatives (1 Worker, 2 Supervisor, 3 Middle manager, 4 Top manager). Subordinates. The question was How many subordinates do you have at the moment? (Table 1). Working hours. The question was How many hours per week do you work? (Table 1).
Table 1 Means, standard deviations for the outcome variables M Number of subordinates Working hours Job satisfaction Exhaustion Cynicism Reduced sense of accomplishment Workaholism Work engagement 7.17 38.03 3.37 2.31 1.88 4.54 3.50 4.13 SD 33.31 10.90 0.68 1.33 1.37 1.02 1.33 1.06

468

K. Salmela-Aro, J.-E. Nurmi / Journal of Vocational Behavior 70 (2007) 463477

Salary. The participants reported their monthly salary using seven response alternatives (e.g., : 1, under 350; and 7, over 3300). Job satisfaction. The question was Are you satised with your work? and was answered on a four-point Likert scale (Table 1). The variable correlated .51 (p < .001) with job engagement (Schaufeli et al., 2002) and .32 (p < .001) with life satisfaction (Diener, Emmons, Larsen, & Grin, 1985; Salmela-Aro, 2005). Although job satisfaction consists of one item only, a similar scale has been frequently used in previous research (Heslop, Smith, Metcalfe, Macleod, & Hart, 2002). Burnout. Burnout was measured on the revised Maslach Burnout Inventory-General Survey (MBI-GS, Schaufeli, Leiter, Maslach, & Jackson, 1996). This is a 9-item questionnaire measured on a Likert scale from 0 never to 6 daily, which produces three scores: exhaustion, cynicism and professional ecacy (three questions for each sum). The Cronbach alpha reliabilities for the scales were .82, .86, .88 (Table 1). The Burnout scale has been shown to correlate .39 (p < .001) with depressive symptoms (Beck et al., 1961), and .23 (p < .01) with the Utrecht Work Engagement scale (Schaufeli et al., 2002; Salmela-Aro, 2005). Engagement. Work engagement was assessed with the revised version of the Utrecht Work Engagement scale (UWES; Schaufeli et al., 2002). It includes questions related to vigor (e.g., In my work, I feel that Im bursting with energy), dedication (I nd my work full of meaning and purpose), and absorbtion to be answered on a Likert scale from 0 never to 6 daily. The questions formed one factor rather than three separate factors and consequently a sum score was used. The Cronbach alpha reliability for the whole scale was .92 (Table 1). The Engagement scale has been shown to correlate .23 (p < .01) with the burnout, .35 (p < .001) with depressive symptoms, .34 (p < .001) with life satisfaction (Diener et al., 1985) and .51 (p < .001) with job satisfaction (Salmela-Aro, 2005). Workaholism was assessed with four questions (e.g., I am in a hurry all the time and I compete with time, I feel guilty if I am not working on something) to be answered on a Likert scale from 0 never to 6 daily based on Robinson (1999) Compulsive tendencies factor of Work Addiction scale (Flowers & Robinson, 2002; see also Taris et al., 2005). The Cronbach alpha reliability for the scale was .79 (Table 1). It has been shown to correlate .40 (p < .001) with the burnout, .27 (p < .01) with depressive symptoms, .17 (p < .05) with life satisfaction and .16 (p < .05) with job satisfaction (Salmela-Aro, 2005). Register data. Academic Achievement was examined according to the progress the participants had shown in their studies (the total amount of course work passed within the rst four years of university studies) and the grades received (the mean of grades received during the rst four years of university studies, ranging from 0 to 3). This information was gathered from the received university archives. 3.3. Analytical strategy n, The research questions were examined using Latent Growth Curve Modeling (Muthe n & Muthe n, 2000) analyses. The statistical analyses were performed 2001a, 2001b; Muthe in two steps. First, Latent Growth Curve modeling was used to investigate the growth rate of self-esteem during the six-year period of university studies (Time 1 to Time 4). In this analysis, the mean level of self-esteem (the intercept), the average rate of growth (Linear and Quadratic trends), and individual variation in level and in linear and quadratic growth were estimated. Second, career characteristics (including unemployment, salary, manage-

K. Salmela-Aro, J.-E. Nurmi / Journal of Vocational Behavior 70 (2007) 463477

469

rial position, subordinates, temporary or permanent job, working hours, workaholism, job satisfaction, burnout, and work engagement) were predicted by individual variation in the level and growth components of self-esteem during university studies. n & The analyses were performed using the Mplus statistical package (version 3; Muthe n, 19982004). The missing-data method enabled all the observations in the data set Muthe to be used in estimating the parameters of the models. This method does not assign values for those that are missing, but uses all the data that are available to estimate the model using the full information maximum likelihood. Each parameter was estimated directly without rst lling in missing values for each individual. Because the self-esteem variables were skewed, the parameters of the models were estimated using the MLR procedure. MLR produces standard errors and v2 test statistics for missing data with non-normal outn & comes by means of sandwich estimators and YuanBentler T 2 test statistics (Muthe n, 19982004). The goodness-of-t of the estimated LGC models was evaluated Muthe by four indicators: the v2 test, Bentler (1990) comparative t index (CFI), Root Mean Square Error of Approximation (RMSEA), and Standardized Root Mean Square Error of Approximation (SRMR). 4. Results 4.1. Self-esteem during university studies First, a latent growth curve model for self-esteem was produced in order to investigate the mean level and changes in it across the four measurement times, and whether there was individual variation in Level, Linear trend and Quadratic trend. The model testing started with three growth components: Initial level, Linear trend and Quadratic trend. It was constructed by setting the loadings of the observed self-esteem variable at Time 1, Time 2, Time 3, and Time 4 1 on Level (intercept factor). The loadings of the observed variables were set to 0, 1, 2, and 3 in the Linear trend and to 0, 1, 4, and 9 in the Quadratic trend. The residual variances of the observed variables were allowed to be freely estimated. The nal model tted the data well: v2 = 13.60, df = 4, CFI = .97, RMSEA = .08, SRMR = .06. The results of a latent growth curve model for self-esteem during the rst six years, from Time 1 to Time 4, showed that, at the mean level, self-esteem increased during university studies. In addition, the quadratic trend was negative and signicant (Fig. 1). This result suggests that at the end of the period of university studies the increase in self-esteem leveled out. The results showed further that there was statistically signicant individual variation at the Initial level, and in the Linear and Quadratic trends of selfesteem. The means and variances for the levels and linear trends (Standard Errors in Parentheses) in the Unstandardized Solution were: for means, intercept 3.149 (.029) p < .001, Linear trend 0.103 (.029), p < .01, and Quadratic trend 0.019 (.009), p < .05; and for variances, intercept 0.201 (.020), p < .001, Linear trend 0.085 (.027), p < .01, and Quadratic trend 0.007 (.003), p < .05. Moreover, the covariance between the intercept and Linear trend was negative, suggesting that the lower the initial level of self-esteem, the greater the increase in it across the period of university studies. In addition, the covariance between the Linear and Quadratic trend was negative, suggesting that the greater the linear increase in self-esteem, the less quadratic change in it across university studies.

470

K. Salmela-Aro, J.-E. Nurmi / Journal of Vocational Behavior 70 (2007) 463477

3,35

3,3

3,25

3,2

3,15

3,1

3,05 21 23 25 27

Age
Fig. 1. Self-esteem during young adulthood.

4.2. Eects of self-esteem during university studies on career characteristics The second aim was to examine the extent to which self-esteem during the period of university studies predicted the characteristics and experiences of young adults careers 10 years later. The results showed rst that when all the working-life-related measures were entered as outcomes of self-esteem in the same model, the model tted the data well (v2 = 68.82, df = 36, CFI = .97, RMSEA = .05, SRMR = .04). The results showed (see Table 2 and Fig. 2) that the level of self-esteem during university studies predicted employment status 10 years later: the lower the level of the students self-esteem during university studies, the more likely they were to be unemployed 10 years later. Moreover, the level of self-esteem also predicted the participants work situation in terms of whether they either had permanent or temporary work, and their salary levels: the higher the level of self-esteem during university studies, the more likely they were to have a permanent rather than temporary job, and the higher their salary was 10 years later. However, self-esteem during university studies was not related to the number of working hours, workaholic behavior, having subordinates or being in a managerial position 10 years later. Moreover, the results showed that the level of self-esteem contributed to the participants job satisfaction: the higher the level of their self-esteem during university studies, the more satised they were with their job 10 years later (Fig. 3). Moreover, both the level and the Linear trend of self-esteem predicted engagement in work: the higher the level of self-esteem and the greater the increase in it during university studies, the higher the level of engagement in work 10 years later (Fig. 3). Finally, the level of self-esteem during university studies was related negatively to the level of all three burnout components, exhaustion, cynicism and reduced accomplishment: the higher the level of self-esteem during their university studies, the less exhaustion and cynicism, and the greater the sense of accomplishment the participants reported in relation to their work 10 years later (Fig. 3). Finally, the Linear trend of self-esteem during university studies predicted exhaustion and reduced

K. Salmela-Aro, J.-E. Nurmi / Journal of Vocational Behavior 70 (2007) 463477 Table 2 The Level and Linear trend of self-esteem during university studies and working life 10 years later Intercept of self-esteem std. Unemployment Temporary/permanent Subordinates Managerial position Salary Workaholism Work hours Burnout Exhaustion Cynicism Reduced accomplishment Work engagement Job satisfaction
* ** ***

471

Linear trend of self-esteem std. .26 .18 .11 .04 .03 .12 .08 .32*** .20 .32* .30* .20

.27* .33*** .17 .11 .25** .13 .04 .38*** .27** .35*** .26*** .24**

p < .05. p < .01. p < .001.

-.09* Unemployment

Self-esteem Intercept

.16* Permanent job .41**

-.35*

Salary

Self-esteem Linear Trend

Time 1-4

Ten years later

Fig. 2. Self-esteem and changes in it during young adulthood and working life 10 years later. Note: *p < .05, **p < .01, ***p < .001.

accomplishment: the greater the increase in self-esteem during university studies, the less exhaustion and the greater the sense of accomplishment in their working life the participants reported 10 years later (Fig. 3). The results showed, however, that neither the char-

472

K. Salmela-Aro, J.-E. Nurmi / Journal of Vocational Behavior 70 (2007) 463477

-.50*** Self-esteem Intercept Exhaustion -.38*** -.46*** Cynicism -.36*** -.35* -.33** .28* Self-esteem Linear Trend .32** Work engagement .16* Job satisfaction Time 1-4 Ten Years later Reduced sense of accomplishment

Fig. 3. Self-esteem and changes in it during young adulthood and burnout, work engagement and job satisfaction 10 years later. Note: *p < .05, **p < .01, ***p < .001.

acteristics nor experiences of working life 10 years later were predicted by the Quadratic Trend of self-esteem. In order to ensure that the level of the participants academic achievement is not responsible for the paths found from the Level and Linear trend of self-esteem to the variables measuring the characteristics and experiences in working life, we carried out some additional analyses in which the paths from academic achievement to the level and slope of self-esteem was added, as well as the paths from academic achievement to the workinglife outcome variables. Academic achievement was measured either as the mean of academic achievement across the rst four years of university studies or as the sum total of progress in studies across the same years. Both variables were entered in separate analyses. The results showed that neither progress in studies nor academic achievement predicted any of the work-related variables. Although progress in studies predicted an increase (Linear trend) in self-esteem (stand. est. = 2.34, p < .05), controlling for progress in studies or academic achievement did not reduce the impact of the Level of self-esteem on the characteristics and experiences of working life 10 years after. However, after entering progress in studies into the model the impact of the Linear trend of self-esteem on reduced sense of accomplishment and work engagement was no longer statistically signicant. The fact that the previous analyses were carried out using the missing-data method might have inuenced some of the results concerning the mean level changes in self-esteem across time. Consequently, analogous analyses to the previous ones were carried out by

K. Salmela-Aro, J.-E. Nurmi / Journal of Vocational Behavior 70 (2007) 463477

473

using the list-wise option method. The results were, however, similar to those reported above. 5. Discussion This 15-year longitudinal study examined whether self-esteem during university studies would predict the characteristics and evaluation of individuals working career 10 years later. The results showed that self-esteem during university studies predicted how individuals working life turned out 10 years later. For example, a high level of self-esteem during university studies predicted being in a permanent job and having a better salary in the beginning of working life. In turn, a low level of self-esteem predicted being unemployed 10 years later. Moreover, a high level of self-esteem during university studies predicted experiencing higher work satisfaction and work engagement later on. In contrast, low self-esteem predicted high levels of burnout at work. The rst aim of the present study was to examine changes in self-esteem over a six-year period of university studies. The results of a latent growth curve model for self-esteem during the six years showed that, at the mean level, self-esteem increased during university studies. In addition, at the end of this period the increase in self-esteem leveled o. The results also showed that the lower the initial level of self-esteem was, the more it increased across the period of university studies. This result, showing that self-esteem increased during university studies, is a notable nding, given the striking gap in the literature on changes in self-esteem during young adulthood between ages 20 and 27. Overall, the ndings of the present study are consistent with trends reported in other recent longitudinal studies on self-esteem in young adults (Galambos et al., 2006; Schulenberg et al., 2005) as well as the results of cross-sectional studies (e.g., Wade & Cairney, 1997). Although the 20s may be characterized by a prolonged period in which there is a diversity of life choices, this seems to be a time during which young people are coming to grips with themselves and their lives (Arnett, 2000). The second aim of the present study was to examine the extent to which self-esteem during university studies would predict individuals career characteristics (unemployment, permanent job, salary, managerial position, subordinates, working hours) 10 years later. The results showed, rst, that a low level of self-esteem during university studies predicted being unemployed 10 years later, whereas a high level of self-esteem predicted having permanent rather than temporary work. Level of self-esteem also contributed to salary 10 years later: the higher the self-esteem, the higher the salary. What was particularly notable in the latter result was that the magnitude of the eect was relatively high. The present ndings suggesting that young adults self-esteem during their university studies predicts the characteristics of their careers 10 years later are important, because no such longitudinal studies have been carried out previously. The results are, however, consistent with the studies by Wineeld and Tiggemann (1985), Caplan et al. (1989), Vinokur and Schul (1997) who found that high self-esteem assisted unemployed individuals to gain employment. The present nding that showed that self-esteem was a strong predictor of the level of salary may explain previous results suggesting that socioeconomic status is associated with self-esteem (Twenge & Cambell, 2002). The results concerning the role of self-esteem in individuals work-related satisfaction and attitudes showed, rst, that the higher the level of self-esteem participants reported during their university studies, the more satised they were with their work 10 years later.

474

K. Salmela-Aro, J.-E. Nurmi / Journal of Vocational Behavior 70 (2007) 463477

This is along with the meta-analysis by Judge and Bono (2001) who suggest that there is a positive, although relatively weak, correlation between self-esteem and job satisfaction. Self-esteem is likely to unfold its eect on job satisfaction through at least two types of processes. First, self-esteem may inuence the kinds of work environments people seek and whether they successfully attain such environments (i.e., type of job). This then leads to specic experiences at work, which may impact on the level of job satisfaction. Second, those with high self-esteem may be more likely to appraise critical events at work as a challenge and experience less fear of failure compared to those with low self-esteem (Locke, McClear, & Knight, 1996). The results showed further that the level of self-esteem during university studies predicted all three burnout components: exhaustion, cynicism and reduced sense of accomplishments. The higher the level of self-esteem young adults reported during university studies, the less work-related exhaustion and cynicism, and the more work-related accomplishments they showed 10 years later. These are important results, because they suggest that positive self-esteem is a force that buers against burnout, possibly because it strengthens the individuals belief in his or her own capability and signicance, and supports active coping in stressful situations. The ndings of the present study are also in accordance with some earlier results suggesting that self-esteem is negatively associated with overall burnout (Kalimo, Pahkin, Mutanen, & Toppinen-Tanner, 2003) and emotional exhaustion, and positively with personal accomplishment (Janssen et al., 1999). The results showed further that an increase in self-esteem during university studies was related to exhaustion and reduced sense of accomplishment. However, their impact disappeared after controlling for academic achievement during university studies. This result suggests that the impact of an increase in self-esteem during university studies on burnout 10 years later is due to the positive eect academic achievement has on individuals selfesteem during their university years. Recently, there has been a shift in burnout research toward its opposite; work engagement has been introduced as the other pole of burnout: vigor, dedication and absorption in ones work (Schaufeli et al., 2002). The results of the present study showed that high self-esteem predicted high work engagement 10 years later, and that this impact was of moderate size. This is an important nding, because it shows that high self-esteem contributes to work-related engagement, i.e., to vigor, dedication and absorption in ones working life later on (Schaufeli et al., 2002). In their review on the consequences of self-esteem, Baumeister et al. (2003) suggested recently that, although there is some empirical evidence that high self-esteem goes together with better job performance, it is likely that occupational success buers self-esteem rather than vice versa. Although the setting of the present study does not allow to make causal inferences to be made along this relationship, it nevertheless shows that self-esteem measured during young adults university years not only predicted their subjective evaluations of their work-situation (job satisfaction, work engagement and burnout) 10 years later, but that self-esteem also contributed to their more objective career characteristics, such as employment type and unemployment, as well as their salary. These results were true even after controlling for level of academic achievement. Overall, these results suggest that individuals self-esteem plays a role in their success in dealing with the entry into working life, even after a 10-year period. There are several possible explanations for the discrepancy between the results of the present study and the conclusions of Baumeister et al. (2003). First, the majority of the studies in the review by Baumeister et al. (2003) were either

K. Salmela-Aro, J.-E. Nurmi / Journal of Vocational Behavior 70 (2007) 463477

475

correlational or focused on relatively narrow outcomes, such as performance quality and persistence in tasks. The reason of this is that few longitudinal studies were available. The second possible explanation for this inconsistency is that young adults facing the transition to working life are currently in a dierent life situation. For example, nding a permanent job in the present Western world is a hard task for most young adults. As was suggested by Baumeister et al., in such a situation, when the prospect of failure is salient, a higher level of self-esteem may benet task performance. This may particularly be the case for young adults entering working life. Overall, there is an evident need to carry out future research on the role of self-esteem in young adults success in dealing with a variety of transitions not only in occupational but also in interpersonal life. 5.1. Limitations The following limitations should be taken into account in any attempt to generalize the ndings of the present study. First, the sample examined consisted of university students and thus was not representative of young adults. Although previous studies have shown no dierences in psychological well-being between non-university and university respondents (Hankin, Abramson, & Mott, 1998), the results should be interpreted as a follow-up of this population only. Young adults in vocational schools, for example, may show a dierent pattern. Second, the sample consisted of Finnish students. Several features of Finnish universities, such as higher age at entry to university, tuition-free studies, and the diculty in gaining admission to a university, may mean that some of the results would have been dierent in countries with a dierent higher education system. Clearly, there is a need to replicate some of the results among other groups. Finally, self-esteem was not measured on the last measurement occasion. Consequently, it would be important to examine the extent to which self-esteem changes during working life. 6. Conclusion The results of this study showed that a high overall level of self-esteem predicted being in permanent work 10 years later, having a high salary, and reporting a high level of work engagement and job satisfaction but a low level of burnout. By contrast, low self-esteem predicted unemployment and feelings of exhaustion, cynicism and reduced accomplishment at work. References
Arnett, J. (2000). Emerging adulthood: a theory of development from the late teens through the twenties. American Psychologist, 55(5), 469480. Baumeister, R. F. (1993). Self-esteem: The puzzle of low self-regard. New York: Plenum Press. 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, 144. Beck, A. T., Ward, C. H., Mendelsohn, M., Mosck, L., & Erlaugh, J. (1961). An inventory of measuring depression. Archives of General Psychiatry, 4, 561571. Bentler, P. M. (1990). Comparative t indexes in structural models. Psychological Bulletin, 107, 238246. Betz, N. E. (2004). Contributions of self-ecacy theory to career counselling: a personal perspective. The Career Development Quarterly, 62, 340353.

476

K. Salmela-Aro, J.-E. Nurmi / Journal of Vocational Behavior 70 (2007) 463477

Burke, R. J. (2001). Workaholism components, job satisfaction, and career progress. Journal of Applied Social Psychology, 31, 23392356. Caplan, R. D., Vinokur, A. D., Price, R. H., & Van Ryn, M. (1989). Job seeking, reemployment, and mental health: a randomized eld experiment in coping with job loss. Journal of Applied Psychology, 74, 759769. Caspi, A. (2002). Social selection, social causation, and developmental pathways: empirical strategies for better understanding how individuals and environments are linked across the life-course. In L. Pulkkinen & A. Caspi (Eds.), Paths to successful development: Personality in the life course (pp. 281301). New York: Cambridge University Press. Diener, E., Emmons, R., Larsen, R., & Grin, S. (1985). The satisfaction with life scale. Journal of Personality Assessment, 49, 7175. Erikson, E. (1968). Identity: Youth and crisis. New York: Norton. Flowers, C., & Robinson, B. R. (2002). A structural and discriminant analysis of the Work Addiction Risk Test. Educational and Psychological Measurement, 62, 517526. Galambos, N. L., Barker, E. T., & Krahn, H. J. (2006). Depression, self-esteem, and anger in emerging adulthood: seven-year trajectories. Developmental Psychology, 42, 350365. Grob, A., Krings, F., & Bangerter, A. (2001). Life markers in biographical narratives of people from three cohorts: a life-span perspective in its historical context. Human Development, 44, 171190. Hankin, B. L., Abramson, L. Y., & Mott, T. E. (1998). Development of depression from preadolescence to young adulthood: emerging gender dierences in a 10-year longitudinal study. Journal of Abnormal Psychology, 107, 128140. Heslop, P., Smith, G. D., Metcalfe, C., Macleod, J., & Hart, C. (2002). Change in job satisfaction and its association with self-reported stress, cardiovascular risk factors and mortality. Social Science & Medicine, 54, 15891599. Janssen, P. P. M., Schaufeli, W. B., & Houkes, I. (1999). Work-related and individual determinants of three burnout dimensions. Work & Stress, 13, 7486. Judge, T. A., & Bono, J. E. (2001). Relationship of core self-evaluation traits in self-esteem, generalized selfecacy, locus of control, and emotional stability with job satisfaction and job performance: a meta-analysis. Journal of Applied Psychology, 86, 8092. Kalimo, R., Pahkin, K., Mutanen, P., & Toppinen-Tanner, S. (2003). Staying well or burning out at work: work characteristics and personal resources as long-term predictors. Work & Stress, 17, 109122. Kasl, S., Bernacki, E., Bromet, E., Curtis, E. C., Eaton, W. W., Fine, L. J., et al. (1992). Surveillance of psychological disorders in the workplace. In K. G. Puryear & S. L. Sauter (Eds.), Work and well-being: An agenda for the 1990s (pp. 7395). Washington, DC, US: American Psychological Association. Kling, K. C., Hyde, J. S., Showers, C. J., & Buswell, B. N. (1999). Gender dierences in self-esteem: a metaanalysis. Psychological Bulletin, 125, 470500. Kokko, K., & Pulkkinen, L. (1998). Unemployment and psychological distress: mediator eects. Journal of Adult Development, 5, 205217. Lefkowitz, E. S. (2005). Things Have Gotten Better: developmental changes among emerging adults after the transition to university. Journal of Adolescent Research, 20, 4063. Lent, R. W., Hackett, G., & Brown, S. D. (1999). A social cognitive view of school-to-work transition. The Career Development Quarterly, 47, 297311. Locke, E. A., McClear, K., & Knight, D. (1996). Self-esteem and work. International Review of Industrial and Organizational Psychology, 11, 132. Ma kikangas, A., Kinnunen, U., & Feldt, T. (2004). Self-esteem, dispositional optimism, and health: evidence from cross-lagged data on employees. Journal of Research in Personality, 38, 556575. Masten, A. S., Hubbard, J., Gest, S. D., Tellegen, A., Garmezy, N., & Ramirez, M. (1999). Adversity, resources and resilience: pathways to competence from childhood to late adolescence. Development and Psychopathology, 11, 143169. n, B. O. (2001a). Latent variable mixture modeling. In G. A. Marcoulides & R. E. Schumacker (Eds.), Muthe Advanced structural equation modeling: New developments and techniques (pp. 133). Mahwah, NJ: Erlbaum. n, B. O. (2001b). Second-generation structural equation modelling with a combination of categorical and Muthe continuous latent variables: new opportunities for latent class/latent growth modelling. In L. M. Collins & A. Sayer (Eds.), New methods for the analysis of change (pp. 291322). Washington, DC: American Psychological Association. n, L. K., & Muthe n, B. O. (19982004). Mplus Users Guide. Los Angeles, CA: Muthe n & Muthen. Muthe

K. Salmela-Aro, J.-E. Nurmi / Journal of Vocational Behavior 70 (2007) 463477

477

n, B. O., & Muthe n, L. K. (2000). Integrating person-centered and variable-centered analyses: growth Muthe mixture modeling with latent trajectory classes. Alcoholism: Clinical and Experimental Research, 24, 882891. Robinson, B. E. (1999). The Work Addiction Risk Test: development of a tentative measure of workaholism. Perceptual and Motor Skills, 88, 199210. 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, 423434. Rosenberg, M., Schooler, C., Schoenbach, C., & Rosenberg, F. (1995). Global self-esteem and specic selfesteem: dierent concepts, dierent outcomes. American Sociological Review, 60, 141156. Russell, D., Peplau, L., & Cutrona, C. (1980). The revised UCLA loneliness scale: concurrent and discriminant validity evidence. Journal of Personality and Social Psychology, 39, 472480. Salmela-Aro, K. (2005). Unpublished data. University of Jyva skyla , Finland. Schaufeli, W. B., Leiter, M. P., Maslach, C., & Jackson, S. E. (1996). MBI - general survey. Palo Alto, CA: Consulting Psychologists Press. nez, I. M., Marques Pinto, A., Salanova, M., & Bakker, A. B. (2002). Burnout and Schaufeli, W. B., Mart engagement in university students: a cross-national study. Journal of Cross-Cultural Psychology, 33, 464481. Schaufeli, W. B., Maslach, C., & Marek, T. (1993). Professional burnout: Recent developments in theory and research. Philadelphia, US: Taylor & Francis. Schulenberg, J. E., Maggs, J. L., & OMalley, P. M. (2003). How and why the understanding of developmental continuity and discontinuity is important: the sample case of long-term consequences of adolescent substance abuse. In J. T. Mortimer & M. J. Shanahan (Eds.), Handbook of the life course. New York: Kluwer Academic/ Plenum Publishers. Schulenberg, J. E., OMalley, P. M., Bachman, J. G., & Johnston, L. D. (2005). Early adult transitions and their relation to well-being and substance use. In R. A. Settersten, Jr., F. F. Furstenberg, Jr., & R. G. Rumbaut (Eds.), On the frontier of adulthood: Theory, research, and public policy (pp. 417453). Chicago: University of Chicago Press. Shanahan, M. J. (2000). Pathways to adulthood in changing societies: variability and mechanisms in life course perspective. Annual Review of Sociology, 26, 667692. Shiner, R. L., & Masten, A. S. (2002). Transactional links between personality and adaptation from childhood through adulthood. Journal of Research in Personality, 36, 580588. Suls, J. (1989). Self-awareness and self-identity in adolescence. In J. Worell & F. Danner (Eds.), The adolescent as decision-maker: Applications to development and education (pp. 143179). San Diego, CA: Academic Press. Taris, T. W., Schaufeli, W. B., & Verhoeven, L. C. (2005). Workholism in the Netherlands: measurement and implications for job strain and worknonwork conict. Applied Psychology: An International Review, 54, 3760. 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, 205220. Twenge, J. M., & Cambell, W. K. (2002). Self-esteem and socioeconomic status: a meta-analytic review. Personality and Social Psychology Review, 6, 5971. Vinokur, A. D., & Schul, Y. (1997). Mastery and inoculation against setbacks as active ingredients in the JOBS intervention for the unemployed. Journal of Consulting and Clinical Psychology, 65, 867877. Wade, T. J., & Cairney, J. (1997). Age and depression in a nationally representative sample of Canadians: a preliminary look at the National Population Health Survey. Canadian Journal of Public Health, 88, 297304. Waters, L. E., & Moore, K. A. (2002). Predicting self-esteem during unemployment: the eect of gender nancial deprivation, alternate roles and social support. Journal of Employment Counseling, 39, 171189. Wineeld, A. H., & Tiggemann, M. (1985). Psychological correlates of employment and unemployment: eects, predisposing factors, and sex dierences. Journal of Occupational Psychology, 58, 229242. Wineeld, A. H., Tiggemann, M., & Wineeld, H. R. (1992). Spare time use and psychological well-being in employed and unemployed young people. Journal of Occupational and Organizational Psychology, 65, 307313.