We posit a female advantage in the relationship between self-reliance and leadership evaluations. We test this prediction in four studies. First, using multi-rater evaluations of young managers, we find that self-reliance relates positively to leadership evaluations for women, but not for men. Next, in each of three experiments, we manipulate the gender of a leader and the agentic trait he or she displays (e.g., self-reliance, dominance, no discrete agentic trait). We find that self-reliant female leaders are evaluated as better leaders than self-reliant male leaders. In contrast, we find a male advantage or no gender advantage for dominant leaders or leaders who are described positively, but not in terms of any discrete agentic trait. Consistent with expectancy violation theory, the female advantage in the relationship between self-reliance and leadership evaluations emerges because self-reliant female leaders are seen as similarly competent, but more communal, than self-reliant male leaders. We discuss the implications of these findings for understanding the effects of self-reliance, gender stereotypes, and stereotype violations on leadership evaluations.