We empirically study the relationship between the gender composition and the performance of teams. In our setting, teams of students are incentivized to develop business plans as part of a compulsory course in a business administration program. At first sight, and in line with earlier empirical literature, the gender composition seems to matter. Our article differs in that we have access to a measure of task-specific individual ability. When controlling for ability, the relation between gender and team performance vanishes, highlighting the importance of controlling for other characteristics in empirical and experimental studies of gender effects. Objective While various studies document a relationship between the gender composition and the performance of teams, we ask whether such findings may depend on a lack of controlling for other dimensions of heterogeneity among team members. Methods In our setting, teams of students are incentivized to develop business plans as part of a compulsory course in a business administration program. Our article differs from the literature in that we have access to a measure of task-specific individual ability. Using ordinary least squares regressions, controlling for year and team size fixed effects, and conducting various robustness checks, we study the association between team performance and team characteristics. Results At first sight, and in line with earlier empirical literature, the gender composition seems to matter for team performance. However, when controlling for the individual abilities of team members, the relation between the gender composition and team performance vanishes. Conclusion Our study highlights the importance of controlling for other characteristics in empirical and experimental studies of gender effects, which in our context leads them to disappear.