A new method for estimating causal effects in social networks suggests that a prior study of peer effects from obesity reported spurious findings.

Menu