Leveraging a Multi-Method Approach to Improve Mass Atrocity Forecasting

Forecasting mass atrocities is a central concern for academics, policymakers, and practitioners, but determining where and when mass atrocities will occur is far from straightforward. Over the last few decades, researchers around the world have developed several forecasting models. Some of these models—like the Political Instability Task Force model or the Australia Forecasting Project model—have emphasized quantitative assessments of the risk of mass atrocity. Others—like the UN Framework of Analysis for Atrocity Crimes—have focused on how case-specific factors coalesce to impact the onset of mass atrocity. In this article, we suggest that a multi-methods framework that capitalizes on the strengths of each of these approaches enhances the ability to correctly forecast mass atrocities. Specifically, we rely upon case-based analysis that identifies combinations of factors that are associated with the absence of atrocities as well as two quantitative approaches geared toward predicting the onset of mass atrocity. After integrating results, we assess how well the forecasts fare and discuss the possible uses of our multi-methods approach.