Cross-Dataset Comparisons and an Integrated Picture of Terrorism
Under Review
Abstract
A large body of research examines why terrorist attacks occur in certain places and times but not others. Despite advances in data collection and empirical methods, this literature produces wildly divergent results while investigating a common set of hypotheses about the economic, political, and social causes of terrorism. For each study that finds a negative relationship between one theorized determinant there is at least another that finds a positive relationship or none at all. It is hard to know what to make of these disagreements as studies adopt different research designs, using disparate datasets at different levels of analysis with alternative measures while covering different countries and historical periods. This article demonstrates how the application of the xSub (cross-national data on sub-national violence) data integration protocols on the most widely used terrorism datasets can improve cross-dataset comparison and integration and allow scholars to isolate explanations for the heterogeneities in findings. Additionally, by processing events onto a common event typology with consistent event categories and units across space and time it facilitates integration and comparison of terrorism datasets with general conflict datasets already available in xSub. This process has the potential to advance our understanding of political violence by enabling researchers to distinguish between determinants that uniquely explain the occurrence of terrorism from those that simply explain violence in general. The ability to confidently integrate terrorism event datasets also improves measurement and analysis of terrorism by providing more comprehensive coverage.
