Causal inference on the engagement effects of athletic participation from within‐student variation

We study how athletic participation relates to two measures of engagement, school attendance and disciplinary suspension, among students in an urban school district. Following one strand of the literature, we study within‐student variation, comparing the same student when playing sports versus not. To this literature, we contribute a microeconomic model to better interpret estimates obtained using such variation, and we propose and employ novel instrumental variables based on lagged season‐specific sports choices and the sports‐specific participation trajectories of other students. Our most rigorous models suggest positive effects of athletics on student attendance, but no significant effects on disciplinary suspension.