Classroom Peer Effects, Effort, and Race

Abstract

This dissertation develops a theoretical model of educational peer effects and then empirically tests whether or not they exist. In the theoretical model, each student selects an effort level to maximize utility; this effort choice depends on his peer group's effort and race. The students' equilibrium effort expression results in hypotheses that can be directly investigated empirically, a definition of the social multiplier, and conditions under which a social multiplier exists. The empirical model uses student-level data with observations on complete classrooms and two measures of effort, self-assessed effort and time spent studying, to investigate whether or not peer effects exist. The estimation results of the empirical model, interpreted using a simulation-based technique, find a positive relationship between the amount of time a student spends studying and time spent studying by peers who share his race; for self-assessed effort, the results are ambiguous. Simulations of policy experiments show that effort is higher in more racially homogeneous classrooms and that a social multiplier exists for both a reduction in the time a student spends working at a part time job and an increase in the student's socioeconomic status.