According to vulnerability-stress theories of depression, negative cognitive styles moderate the relation between negative life events and depression. Findings from research testing this vulnerability-stress hypothesis in children, however, have been mixed (for a review, see Gibb & Coles, 2005). In one model developed specifically to explain depression risk in youth, Cole (1990, 1991), proposed that self-perceived competence, mediates (rather than moderates) the link between negative events and depression in children. Studies have found support for the mediation model, but there is less support for the vulnerability-stress model in children. Methodological limitations of previous research might account for the null moderation findings. Therefore, the primary goal of the proposed study was to (1) test whether domains of self-perceived competence, alone or interacting with overall negative events, contribute to prospective changes in depressive symptoms and onsets of depressive diagnoses, and (2) test whether specific, domains of self-perceived competence interact with matching domains of negative events to predict residual changes in depressive symptoms. A secondary goal was to examine the specificity of the model by testing the effects of low self-perceived competence, alone, and interacting with negative events, on symptoms and diagnoses of social anxiety disorder. Findings revealed that self-perceived physical appearance contributed significantly to a worsening in symptoms of depression and social anxiety over the follow-up. There was no support, however, for the main effect of the other domains of, competence or the vulnerability stress model.