Abstract:

This paper uses micro data on friendship networks in high schools from the National Longitudinal Study of Adolescents Health to estimate the effects of network’s age diversity on academic performance. The investigation addresses concerns about self selection into networks and unobserved school level differences by using within-school variation and instrumental variable method. The findings provide evidence that having an age diversified friendship network results in significantly lower grade point average in the final year of high school. Furthermore, individuals in mixed-age friendship networks are found to be less likely to attend college. Contrary to the previous research, no evidence is found that having a best friend of a different age, or a group of friends of average age that differs from an individual’s age is associated with worse outcomes in education. The results remain robust after the sample is limited to students with no criminal background and those that are in the expected grade for their given age. Possible explanations for these findings include higher rates of conflict, less communication and cooperation in mixed age friendship networks.

Keywords:

Age diversity, peer effects, academic performance, social networks.