Causal Effects in Social Networks

This paper reviews the current literature on the estimation of causal peer effects. After a discussion of causality in general, I introduce the standard peer effect model in networks and illustrate the reflection problem. I then present approaches to causal inference with observational data before introducing experimental approaches. I review estimation issues arising from measurement and sampling errors, and discuss how they affect causal inference in network and peer effect experiments. The last section of the paper broadens the discussion to encompass dynamic peer effects and link formation and illustrates the different meanings that causality can take in the estimation of peer effects.