Detecting a Copycat Effect in School Shootings Using Spatio-Temporal Panel Count Models

Abstract

School shootings are often motivated by the perpetrators' desire for media attention and notoriety. Since school shootings receive intense regional and national media coverage, a high likelihood for copycat attacks can be expected. However, it has not yet been shown convincingly that this hypothesized copycat effect exists. We investigate whether a copycat effect can be detected in US state-level school shooting data from 1990 to 2017. We do so by estimating spatio-temporal panel count models and control for socio-demographic characteristics, as well as state and Federal gun control laws. Positive spatial and temporal dependence indicate that the risk for additional school shootings in the same and neighboring states increases after the initial attack.