How social ties affect peer-group effects: a case of university students

Among the key issues of peer effects estimation is the correct identification of relevant peers. In this study, we explore how the individual performance of university students is influenced by characteristics and achievements of peers from individual’s social network. The analysis uses data from two directed networks: a network of friends and a network of study partners for thirdyear students at a top-tier Russian university. Data on network ties in randomly formed student groups enables us to address the endogeneity problem and disentangle the influence of peers’ performance from the effect that a peer’s background has on students. We show that both the GPA of peers and their ability measures are significant in the estimated regression model. A onepoint increase in the average GPA of peers is associated with an increase in an individual student’s own GPA of approximately one fourth. The regression on the data from the network of study partners has slightly greater explanatory power than the analys is based on data from the network of friends. No effect from a student’s classmates is found in the model that assumes group interactions occur between group mates