Research

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. 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, it facilitates integration and comparison of terrorism datasets with general conflict datasets, enabling researchers to distinguish between determinants that uniquely explain the occurrence of terrorism from those that simply explain violence in general.

Why do some rebel groups adopt diverse portfolios of violence, while others appear to specialize in a single form of contention at a time? In this article, I theorize that rebel groups expand violent repertoires in response to changes in pressure levied upon them by the state and rival challengers. Just as competition fosters innovation in commercial firms, vying over a monopoly on violence promotes diversity in rebel group behavior. By varying targets and the use of direct and indirect force, rebels can alleviate external pressure, increase operational tempo and lethality. I find consistent support for these propositions after integrating event data for 779 rebel groups in 42 civil wars in 31 countries between 1997 and 2020. These findings have important implications for research and policy that aims to explain rebel group behavior and evaluate the efficacy of countermeasures.

What explains variation in patterns of violence across time and space? Documenting patterns is critical for the study of political violence and the formation of related policy aimed at reducing its occurrence. However, there is little consistency in how “patterns of violence” is conceptualized or applied. I organize and integrate data from some of the most widely used conflict databases along the following dimensions - target, force, and magnitude and generate a novel patterns of violence (POV) dataset. The dataset catalogs 340,810 unique events for 1,591 actors in 33 civil wars in 26 countries. I demonstrate how this configuration, scaled across different spatial and temporal levels of analysis for different actors, can identify unique trends and heterogeneities in violence.