This research examines the role of friendship contingent self-esteem (FCSE), or self-esteem that is dependent on the quality of one's friendships, in predicting depressive symptoms. In Study 1, the authors developed a measure of FCSE. Both FCSE and others' approval correlated with self-esteem and depressive symptoms, but when entered simultaneously in a regression equation, only FCSE significantly predicted self-esteem and depressive symptoms. Study 2 showed that dependency and close friendship competence predicted depressive symptoms only for those high in FCSE. In Study 3, a diary study, FCSE predicted self-esteem instability. Self-esteem instability, in turn, predicted depressive symptoms. Furthermore, a three-way interaction of rumination, FCSE, and the valence of the event predicted momentary self-esteem. Findings are discussed with regard to the importance of considering FCSE when investigating interpersonal risk for depression.

The quality of interpersonal relationships may be a key factor in the etiology of depression. An interpersonal depressive vulnerability has been proposed to interact with interpersonal stressors to predict depressive symptoms. Research examining gender differences in this link has been inconclusive. We propose that research should focus on whether one's self-esteem is based on the quality of interpersonal relationships. We propose a model of gender differences in depression that examines the role of interpersonal contingent self-esteem (ICSE) in the development of depression. Specifically, we propose that ICSE interacts with gender in the face of interpersonal stress to promote decreases in momentary self-esteem. Decreases in momentary self-esteem, in turn, lead to rumination, maladaptive interpersonal behaviors, and depressive symptoms.

No abstract

In our reply to the commentaries on Cambron, Acitelli, & Pettit's paper (this issue), we address the issue discussed in both commentaries: the need to take a developmental approach to the study of gender differences in depression. In response to Mezulis and Funasaki (this issue), we attempt to specify where the Cambron et al. paper supports their point that vulnerability, stress, and vulnerability-stress interaction may vary across domains. We also enumerate several hypotheses that are derived from our model of gender differences in depression. In response to Burwell and Shirk (this issue), we agree with the need to investigate the origin of self-esteem contingencies, and that the measurement of self-esteem contingencies is fraught with problems. We also remind readers that whether examining interpersonal relationship or physical appearance risk factors for depression, we must focus on the extent to which one's self-esteem is based on that domain.