The impact of peer influence on academic performance: A three-stage co-evolution model

Estimating peer influence on academic performance is important to both social scientists and policy makers. The fundamental challenges in prior studies that assess peer influence using observational data are about (1) separating the effect of peer influence from homophily and (2) capturing network dynamics. In this paper, we build on related work and propose a three-stage co-evolution model to investigate the impact of peer influence on academic performance by controlling for multiple sources of homophily and network dynamics. Our preliminary analysis provides support for the existence of homophily and peer influence. In addition, we find evidences of asymmetric peer influence, i.e. influence is more significantly positive among higher ability students, than lower ability ones. Early findings from this study shed light on the possible underlying mechanisms of how networks evolve, and how students' behaviors are affected by their peers' performance and decisions, and vice versa. Our findings offer useful insights about the prevalence and role of peer influence to policy makers, educational researchers and teachers.