This paper contributes to the literature on the relationship between adolescent depression and educational attainment in several ways. First, while cross-sectional data are normally used, this paper uses longitudinal data in order to defend against the potential of reverse causality. Second, this is the first paper in the literature to control for sibling-fixed effects in examining the relationship between adolescent depressive symptoms and human capital accumulation. Importantly, this eliminates omitted factors such as family and neighborhood characteristics common to siblings that affect both depressive symptoms and educational attainments (e.g. neighborhood crime, and family resources). Third, this paper examines the effects of both an indicator and scale of depressive symptoms and finds important associations with these depressive symptoms and human capital accumulation. Though the results cannot be given a causal interpretation, the findings show a negative relationship between depressive symptoms and years of schooling. The relationship appears to be driven primarily through increasing the chances of dropping out but may have small impacts on the likelihood of college attendance (conditional on high school graduation). In particular, preferred estimates suggest that a standard deviation increase in depressive symptoms is associated with a 25–30% increase in the likelihood of dropping out. Copyright © 2009 John Wiley & Sons, Ltd.