Modeling Terrorist Attacks: Assessing Statistical Models to Evaluate Domestic and Ideologically International Attacks

Abstract

Many prior studies have analyzed how country characteristics affect the rate of terrorist violence and there is a growing literature on how group traits influence terrorist violence. The current study expands upon this literature by using multilevel modeling to assess both these units of analysis on the rate of domestic attacks and the rate of attacks against foreign targets. Using data from the Big Allied and Dangerous (BAAD) and the Global Terrorism Database (GTD), a cross-national sample of 224 terrorist groups are modeled in relation to their countries of origin to assess rates of domestic attacks. In this cross-sectional study many of these terrorist groups target multiple foreign countries. Multiple membership random effects modeling (MMREM) is used to assess the impact of multiple countries targeted by a group. The results of the study indicate that multilevel modeling provides an improved statistical fit and the MMREM model provides an improved measurement for analyzing attacks targeting foreign countries.