Affirmative action policies promote women and do not harm efficiency in the laboratory.

Gender differences in choosing to enter competitions are one source of unequal labor market outcomes concerning wages and promotions. Given that studying the effects of policy interventions to support women is difficult with field data because of measurement problems and potential lack of control, we evaluated, in a set of controlled laboratory experiments, four interventions: quotas, where one of two winners of a competition must be female; two variants of preferential treatment, where a fixed increment is added to women's performance; and repetition of the competition, where a second competition takes place if no woman is among the winners. Compared with no intervention, all interventions encourage women to enter competitions more often, and performance is at least equally good, both during and after the competition.