Estimating the effects of sports and physical exercise on bullying

Using three nationally representative high school samples, we estimate the effect of physical activity in general, and sports specifically, on in-school student victimization (i.e., bullying), overall and by gender. Due to non-random selection, we instrument physical activity in a two-stage model using state-level sports participation rates and physical education requirements. We find that physical exercise may decrease, if not prevent, victimization, though competitive sports increase the likelihood of bullying for those on the margin, especially for male students. These results are relevant for administrators and policymakers, who can increase access to physical activities while simultaneously increasing oversight on and off-the-field.