Reconsidering Peer Effects from Schoolmates Exposed to Violence at Home

Abstract

Prior research suggests that students exposed to domestic violence impose substantial negative externalities on peers’ learning and labor market outcomes. We test whether these findings generalize using nine-years of student data from a large urban school district linked to geocoded police records of at-home violence. In contrast to prior work, we find negligible peer effects. Adding one exposed peer to a classroom of 20 students reduces English and math scores by a statistically insignificant 0.006 standard deviations and increases disciplinary incidents by 3.2%. We reject achievement reductions greater than 0.017 standard deviations and disciplinary increases larger than 8.3%.