Disentangling peer effects from other confounding effects is difficult,and separately identifying endogenous and contextual effects is impossible for the linear-in-means model. This study confronts these problems by using spatial autoregressive models with group fixed effects. The nonlinearity introduced by the variations in the peer measurements provides information to identify both endogenous and contextual effects,thus resolving the "reflection problem." The group fixed effects term captures the confounding effects of the common variables.Applying the model to data sets from the National Longitudinal Studyof Adolescent Health, I find strong evidence for both endogenous and contextual effects in student academic achievement. (c) 2010 by The University of Chicago. Allrights reserved.

This paper extends Brock and Durlauf's (2001a, 2001b) binary choice complete network (or group interaction) model with homogeneous rational expectations to a general network model with heterogeneous rational expectations. In our model, individuals will form expectations regarding peers' behaviors taking into account their characteristics. Endogenous, contextual, and correlated effects are all identifiable. Conditions for unique equilibrium are established. For a complete network with heterogeneous rational expectations, multiple equilibria can be characterized by an aggregate scalar index. The empirical results on adolescents' smoking behaviors show significant endogenous and contextual effects, even after controlling for school-grade random effects and school fixed effects.