It is expected that women who run for office will receive significantly more support from women relative to male voters. However, evidence for this gender affinity has been mixed. In fact, recent experimental studies, and elections, have not found it to exist for female candidates. To test the presence of gender affinity, we examine the case where voters have the opportunity to elect the first female mayor in a major US city. Using exit poll data, we find that women (and men) who cited gender as an important quality in choosing a candidate were significantly more likely to support the female candidate, but only a small portion of voters felt gender was the most important candidate quality. For the majority of women, we find that there were confounding factors, such as uncertainty of the female candidate’s leadership skills and issue positions, which prevented her from gaining the majority she needed to win.