Previous studies have drawn inconsistent conclusions regarding the relationship between measures of self-concept and adolescent alcohol use. The prevalence of alcohol use and experimentation among adolescents is strikingly high, thus it is imperative to discover precipitating and predictive factors for adolescent alcohol use. The goal of the current study was to test whether efficacy and self-derogation, both measures of self-concept, predicted subsequent alcohol use differently for girls and boys. It was hypothesized that adolescents who report low efficacy should subsequently report high self-derogation. High self-derogation may in turn make adolescents more likely to drink, especially girls. Participants were students in 22 middle and junior high schools in New York City. The study employed a three-wave longitudinal design in which participants completed questionnaires at baseline, 1-year follow-up, and 2-year follow-up. Questionnaires included measures of efficacy, self-derogation, and alcohol use. Results of longitudinal structural equation modeling revealed that reports of lower self-efficacy at baseline were associated with greater self-derogation at the 1-year follow-up. No differences were observed for gender at this stage. However, increased self-derogation predicted higher levels of alcohol use for girls, but not for boys. These findings have implications for alcohol prevention and treatment programs for girls. Prevention and treatment approaches for girls should focus on building positive self-images. Table, figure, references