Invisible Influence: Bridging Ties Suppress Skew as a Signal of Contagion

Contagious distributions can be recognized by skew more pronounced than that seen in a Poisson distribution. In social contexts, we often interpret excess skew as evidence of social contagion. This article uses a simulation to demonstrate that interpreting Poisson distributions as evidence of independent action rests on the implicit but dubious assumption that social networks are closely approximated by administrative units or spatial areas. I use the case of recreational drug use at college to illustrate that the pattern holds in empirical data. To the extent that networks are decoupled from administrative units or space, skew is suppressed, and we will fail to recognize social contagion.