Linear-in-means models are widely used in different contexts to estimate peer effects. In these models, there are two potential sources of endogeneity: in the interaction network and the individual’s characteristics. This ...