Even today, large and economically significant gender gaps in pay and representation persist. Despite high female participation in the workforce, progress towards equal pay has stagnated and few women reach the very top. With the removal of formal barriers to women's advancement, and the fact that in many countries women are better educated than men, we need to find new explanations for the persistent gender gaps in the labor market. This research project aims to study a relatively unexplored mechanism behind gender differences in the labor market: the gender composition of work groups. Are women's performance and well-being negatively affected by male-dominated environments? This is difficult to study in a credible way, partly because of insufficient data, and partly because the gender composition of groups often correlates with other factors that affect gender differences in outcomes. We will use Swedish register data to analyze how gender differences in doctoral students' study results, dropouts and labor market outcomes are affected by the gender composition of doctoral programs. By exploiting the fact that the proportion of men in each doctoral program varies between adjacent cohorts, we can capture the causal effect of the proportion of male doctoral students in a program. The results are expected to contribute to a better understanding of the mechanisms behind persistent gender gaps in the labor market, and thus be of interest to international researchers and policy makers alike.