The effect of language on performance: do gendered languages fail women in maths?

Research suggests that gendered languages are associated with gender inequality. However, as languages are embedded in cultures, evidence for causal effects are harder to provide. We contribute to this ongoing debate by exploring the relationship between gendered languages and the gender gap in mathematics achievements. We provide evidence for causality by exploiting the prominent (but not exclusive) practice in gendered languages of using masculine generics to address women. In an experiment on a large representative sample of the Hebrew-speaking adult population in Israel, we show that addressing women in the feminine, compared to addressing them in the masculine, reduces the gender gap in mathematics achievements by a third. These effects are stronger among participants who acquired the Hebrew language early in childhood rather than later in life, suggesting that it is the extent of language proficiency that generates one's sensitivity to being addressed in the masculine or in the feminine. Moreover, when women are addressed in the masculine, their efforts (in terms of time spent on the maths test) decrease and they report feeling that "science is for men" more than when addressed in the feminine. We supplement the analysis with two experiments that explore the roles of general and task-specific stereotypes in generating these effects.