The Primacy of the Local: Identifying Terrorist Hot Spots Using Geographic Information Systems

Abstract

Despite the wide range of studies focused on the causes of terrorism, most use the state as the unit of analysis. Doing so, however, overlooks important variation that occurs within the state. Our research seeks to determine the causes of domestic terrorism through a more refined unit of analysis. We do this by using the PRIO-GRID cell structure spatially merged with a geocoded version of the GTD dataset. We then perform a Getis-Ord Gi* hot spot analysis to uncover those local areas most prone to domestic terrorism. Our results indicate the following attributes increase the likelihood of terrorism: mountainous terrain, close proximity to a state capital, large population, high population density, and poor economic conditions. When testing between regime types, we find that factors such as population, economic conditions, and the number of ethnic groups are significant only in democracies, while distance to capital is significant only in autocracies.