We examine how and why gender composition impacts evaluations of women’s leadership ability in groups. Integrating theories of impression formation and impression management, the present research proposes that a woman is rated as higher in leadership ability when she is the only woman in the group (i.e., a solo) compared to being one of only two women in the group (i.e., a minority duo) because people are most concerned with appearing biased when evaluating a solo female target. Across 5 experiments (total N = 2,060; 3 pre-registered) we find that people evaluate a solo woman as having greater leadership ability (Studies 1, 2, and 5) than a duo woman. Further, concerns that the evaluator will be seen as biased are higher when evaluating a solo than a non-solo female (Studies 3 and 4). However, when the evaluator’s moral credentials have been established in advance, there is no difference between solos and non-solos conditions on evaluations of leadership ability (Study 2) or reported concern (Study 3). Finally, bias concerns explain why female solos are evaluated as stronger leaders than non-solos (Study 4). We also explore but do not find support for the idea that a minimum of three women in a group represent a critical mass that reduces stereotyping. We discuss implications for managing diverse teams.