A randomized trial was conducted to test the efficacy of three self-directed prevention intervention workbooks for depression. Cognitively at-risk college freshmen were randomly assigned to one of three conditions: traditional cognitive, non-traditional cognitive, and academic skills. Consistent with hypotheses, participants who were high in rumination and experienced stress exhibited significantly greater levels of depressive symptoms after completing the traditional cognitive skills workbook than after completing the other two workbooks. This pattern of results held post-intervention and 4 months later. These findings indicate that rumination may hinder ones ability to identify and dispute negative thoughts (at least without the help of a trained professional). The results underscore the importance of identifying individual difference variables that moderate intervention efficacy. They also raise concerns about the potential benefits of self-help books, an industry that generates billions of dollars each year.