Stereotypical Selection

Women are still under-represented and struggling to establish careers in traditionally male-dominated fields. Does minority status in and of itself create a barrier to women's success? Experiments suggest that under-representation exacerbates the detrimental effect of the negative stereotypes that often characterize women's ability in these fields. However, in real-world environments, these results might not hold. While lab experiments typically shut down the selection channel altogether, the choice to enter male-dominated fields is endogenous, and may in part be motivated by challenging these stereotypes. This paper assesses how minority status affects performance when selection is endogenous by studying the performance of 14,000 students at an elite university across 16 departments, in a real-world setting that combines a choice with well-defined stereotypes - university major - with exogenous variation in peer identity - quasi-random allocation of students across class groups within the same course. The evidence indicates that those who go against stereotypes (e.g. women in math) do not suffer from being in the minority, but they impose negative externalities on those who select on stereotypes (e.g. men in math). In line with social identity considerations being incorporated into educational choices, the evidence points towards ex-ante "sensitivity" to social norms and preferences to engage with same-gender peers inducing students to select different majors and then reacting to the composition of the environment in a self-fulfilling way.