Gender composition, peer effects and human capital formation: Evidence from rural China

A large body of literature shows that classes with a higher share of girls tend to achieve better academic outcomes, but whether the gender peer effects are universal or context dependent remains unclear. This study uses two rounds of data from more than 4,000 students in randomly assigned classes across 127 rural primary schools in northwestern China, and examines how the proportion of boys relates to academic performance and noncognitive skills. We find that a higher proportion of boys in class significantly improves students’ mathematics achievement and key noncognitive skills including agreeableness, conscientiousness, and openness. The effects are particularly pronounced for boys, which we attribute to the rural context where limited family educational capital, remote school locations, small class sizes, and scarce teacher resources make students more reliant on peer relationships and teacher support than on family resources. Specifically, we identify two complementary mechanisms: a peer substitution mechanism operating among left-behind boys whose fathers work away from home, and a teacher response mechanism arising from classroom management demands in high boy-share classes. By documenting that gender peer effects vary across rural and other contexts, this study complements the existing literature with evidence from an undeveloped setting and enriches our understanding of the mechanisms through peer and teacher channels that shape human capital formation.