Level and stability of self-esteem as predictors of depressive symptoms

In a recent study, Kernis, Grannemann and Mathis (Journal of Personality and Social Psychology, 61, 80-84) reported that stability of self-esteem (SE) moderates the relation between level of SE and depression. Specifically, level of SE predicted depression more strongly in persons with temporally stable SE. We attempted to replicate this finding across three independent data sets (total N = 504). Although level of SE was a strong prospective predictor of depressive symptoms in all three studies, in none were significant interactions obtained between level and stability of SE in the form suggested by Kernis and his colleagues. Overall, our results suggest that Kernis et al.'s finding is unreliable. Given recent theory suggesting that instability in SE (Roberts & Monroe, Clinical Psychology Review, 14, 161-181, 1994) and neuroticism (Martin, Personality and Individual Differences, 6, 353-365, 1985) predispose to depression subsequent to life Stressors, these findings are not surprising. Data are presented that demonstrate that stable SE and affect measured 'on-line' through daily assessments are, in fact, associated with lower levels of neuroticism.