Do students want to succeed? Peer group choice, social influence, and undergraduate performance

Students, like people in other contexts, routinely make choices about various aspects of their environ-ment. One such choice for students is deciding whom to partner with in studying or discussing course material. Study groups may exert peer influence on their members, but that influence, as well as other determinants of grades, may differ depending upon student-level preferences regarding optimal classroom context. If students are able to cater to their preferences, a theory of rational choice in the presence of heterogeneous peer effects would imply that the social influence on their learning that results should be different than if they randomly encounter a learning context. This study searches for evidence of such student-level heterogeneity by exploiting a unique classroom experiment in study group assignment within a University of South Australia introductory economics course. I observe multiple partnerings and performance outcomes for two sets of students: one set is randomly assigned to small study groups, and the other may choose their own groups. Drawing a theoretical framework from the treatment effects literature, which has not to date been brought to bear on educational peer effects estimation, I decompose the peer effect and measure the strength and nature of selection versus other determinants of peer group composition and peer group influence. Implications of the work for peer effects estimation, educational policy, and the treatment effects literature are discussed. ∗I thank seminar participants in the economics group of the Research School of Social Sciences of the Australian National University for helpful comments, as well as UniSA’s Division of Business and Centre for Regulation and Market Analysis for support over the course of this project. All viewpoints and errors contained within are solely my own. 1