Low self-esteem has been established as a vulnerability factor for depression. In line with recent research, we suggest that a full understanding of the role of self-esteem in depression requires consideration of contingent self-esteem as well. For most people, competence is an important source of self-esteem. Students in particular link their self-esteem to academic competence. To test the hypothesis that academic contingent self-esteem (aCSE) predicts depressive symptoms (DS), two studies were conducted. Preceding the investigation of our hypothesis, the first purpose of Study 1 was to describe the development of aCSE, self-esteem (SE) level, and DS in adolescence in a sample of German students aged 10–16 (N = 1888) in order to provide a foundation for further analyses. Then, to address the main question, age and gender differences in aCSE, SE level, and DS as well as their relations were investigated. The results show that (1) gender differences emerged after the age of 10/11. Girls scored higher on aCSE and DS and lower on SE level than did boys, and aCSE and DS decreased and SE level increased over time in boys, while the rather disadvantageous pattern in girls remained stable. (2) After controlling for SE level and aCSE, the effects of gender and age × gender interaction on DS disappeared, suggesting an influence of aCSE on DS. (3) aCSE predicted DS over and above SE level. Since the results of Study 1 did not allow for causal conclusions, a longitudinal study (N = 160) was conducted to further investigate the causal role of aCSE. According to the diathesis-stress model, aCSE was expected to serve as a diathesis for developing DS in the face of academic stress (daily hassles) during an academic semester at university. The results of Study 2 revealed that aCSE interacted with corresponding hassles to predict increases in DS. High levels of academic stress led to increases in DS only among students who strongly based their SE on academic competence. Implications for prevention and intervention of depression are discussed.

Researchers have argued that the strategies individuals use for self-esteem regulation are interchangeable. In the present study, we examined whether previous self-affirmation reduces the amount of subsequent claimed self-handicapping. More importantly, we tested potential moderators of these effects. Following negative feedback on an intelligence test, 56 female college students were given the opportunity to affirm themselves either within the threatened intelligence domain or within a domain unrelated to the source of threat (e.g., musicality). Results revealed that subjects handicapped less when they had previously affirmed themselves in a domain which was unrelated to the threatening domain (contextual moderator). However, these effects were moderated by dispositional self-esteem (individual moderator). High self-esteem participants claimed fewer handicaps the more they felt self-affirmed whereas claimed self-handicapping among low self-esteem participants was not affected by previous self-affirmation. Altogether, our findings suggest certain limitations on the substitutability of self-protection processes.