334 STEIGER.ALLEMAND.ROBINS.AND FEND self-report method variance has already been statistically removed both self-esteem domains indeed play a critical o for life prior level ver,even during ado First.depn Indeed.one should consid ating domains o self-c rt to build un stable amilia on.and y(190 rather than a onal skills and ir the ating that self- ctor of depr vice versa (e.g.,Orth folow-up survey.we were to seven items indic uch as ne n the pred res,as they were not mea red in the present study tudies to include clinical me ent tools in order to p or depressive disoners as a consequence of selt-esteem de eloo that has ally overapping with low it would be valuable to have the same m e of studies misht extend For 16 d with be that an BDI:BDI-V) d it whether a dis es have more complex ass eem and of the first I elo psy a time frame of ove 20 years.We could reveal that not only al ml in the p s still two de later.The and not to elf-esteem is especially malleable 35m References L.Y..Metalsky.G.I..Alloy.L B.(1989).Hope s in their perso t to do In D.L. -1851.Hs Y:V B.Reese.H.w otogrlnmirodicianoreearchmChad.Moecy brins m CA:Brook eister,R. to be accepted in order to find a good attitude toward onesef.Asweself-report method variance has already been statistically removed by controlling for prior levels of depressive symptoms. Another constraint concerns the measurement of depressive symp￾toms during adolescence. First, depressive symptoms were only col￾lected at one measurement wave at age 16. This design did not allow us to test for effects of depressive symptoms on self-esteem as suggested by Rohde, Lewinsohn, and Seeley (1990). Therefore, we only tested the vulnerability model, which proposes that low self￾esteem is a cause rather than a consequence of depression (e.g., Beck, 1987). However, the causal relationship between self-esteem and depression has been tested in a number of longitudinal studies, all indicating that self-esteem functions as a predictor of depression, and not vice versa (e.g., Orth & Robins, 2012). Second, the adult measure of depressive symptoms in this study lacks diagnostic clinical value. In the follow-up survey, we were restricted to seven items indicating depressive symptoms. Thus, we can only relate self-esteem to a continuum of symptoms of depression without being able to clinically diagnose major depressive disorder, because this scale does not define a cutoff value to diagnose depression. Hence, it may be necessary in future studies to include clinical measurement tools in order to predict major depressive disorders as a consequence of self-esteem develop￾ment. Furthermore, it would be valuable to have the same measure of depression in adolescence and adulthood. Depression was assessed more statelike at age 16 (original BDI) compared with a more traitlike measure at age 35 (simplified version of BDI; BDI-V). Note, how￾ever, that the BDI-V scale is developed on the basis of the original BDI. Furthermore, the two scales have been compared extensively by Schmitt et al. (2003), clearly showing high convergent validity and a similar pattern of external correlates, confirming that they have a similar nomological network. Ideally, however, the two scales would be assessed and compared over time to see whether they behave identically. Future studies might address this issue. Related to this, the question remains whether depression should be defined and assessed as a statelike or traitlike construct. Clinical psychologists usually define depression as a statelike construct, whereas personality psy￾chologists conceptualize depression having both state and trait com￾ponents. As shown in a study using longitudinal modeling of depres￾sion scores across adolescence and adulthood, most of the variance in depression is due to a trait and not to a state component (Cole & Martin, 2005). One further possible critique is that using a traitlike measure of depression at age 35 might cause an increase of the association between adolescent self-esteem and depression. Note, however, that the correlation spans two decades from adolescence when individuals experience considerable changes in their personality and/or affect, to adulthood. A further issue should be mentioned with respect to domain￾specific self-esteem. We used the two domains self-esteem of appear￾ance and academic competence because they are central to adolescent experiences. How adolescents see themselves in terms of academic competence can be crucial when applying for a job. In Germany, at the age of 16, many students finish school and apply for apprentice￾ships. We believe that these experiences are important challenges and represent some of the major developmental tasks during adolescence. The self-esteem domain of perceived physical appearance was inte￾grated in our analysis because adolescence brings along a number of physical changes such as changes in body weight and height or hormonal changes and body maturation. Changes in body image need to be accepted in order to find a good attitude toward oneself. As we have seen, both self-esteem domains indeed play a critical role for life outcomes. However, even though the chosen self-esteem domains are highly important, future studies could additionally test other domains of self-esteem. Indeed, one should consider integrating domains of self-esteem such as perceived interpersonal skills. It is during adoles￾cence that individuals typically start to build up stable outerfamilial relationships such as peer friendships or first romantic relationships. Therefore, these social changes in the environment might affect how adolescents see themselves with regard to interpersonal skills and, in turn, could be predictive of long-term life outcomes. Having suc￾ceeded in building up and maintaining social relationships during adolescence might be another important root of adaptive adjustment later in life. Future studies might also consider to include classical personality traits such as neuroticism as a third variable, as neuroticism might explain some of the variance in the prediction of depressive symptoms in adulthood. Unfortunately, we were unable to include personality measures, as they were not measured in the present study. However, we controlled for prior levels of depression in all our models, which is a central component of the neuroticism/negative affectivity con￾struct that has been discussed as empirically overlapping with low self-esteem and depression (cf. Watson, Suls, & Haig, 2002). Finally, future studies might extend our research by including moderator variables. For example, it could be that an individual’s school grades differ from his or her perception of academic compe￾tence, and it would be interesting to see whether a discrepancy between the subjective perception and the objective evidence leads to maladaptive outcomes. By including moderators, we might be able to uncover even more complex associations between self-esteem and mental health. Overall, this study provides one of the first long-term findings of effects of adaptive self-esteem development during adolescence on long-term health outcomes in adulthood. To the best of our knowl￾edge, no other study tested for effects of adolescent development within a time frame of over 20 years. We could reveal that not only adolescent self-esteem level but also self-esteem change plays a crucial role in the prevention of depressive symptoms still two de￾cades later. These results provide a basis for interventions in enhanc￾ing self-esteem during adolescence—thus, during an age period when self-esteem is especially malleable. References Abramson, L. Y., Metalsky, G. I., & Alloy, L. B. (1989). Hopelessness depression: A theory-based subtype of depression. Psychological Re￾view, 96, 358 –372. doi:10.1037/0033-295X.96.2.358 Alwin, D. F. (1994). Aging, personality, and social change: The stability of individual differences over the adult life span. In D. L. Featherman, R. M. Lerner, & M. Perlmutter (Eds.), Life-span development and behaviour (pp. 135–185). Hillsdale, NJ: Erlbaum. Baltes, P. B., Lindenberger, U., & Staudinger, U. M. (2006). Life span theory in developmental psychology. In W. Damon & R. M. Lerner (Eds.), Handbook of child psychology: Vol. 1, Theoretical models of human development (pp. 569 – 664). New York, NY: Wiley. Baltes, P. B., Reese, H. W., & Nesselroade, J. R. (1977). Life-span developmental psychology: Introduction to research methods. Monterey, CA: Brook & Cole. Baumeister, R. F., Campbell, J. D., Krueger, J. I., & Vohs, K. D. (2005). Exploding the self-esteem myth. Scientific American, 292, 84 –91. doi: 10.1038/scientificamerican0105-84 This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly. 334 STEIGER, ALLEMAND, ROBINS, AND FEND