Expanding the Coverage of Conflict Event Datasets: Three Proofs of Concept

Abstract

ABSTRACT Many contemporary studies on political violence/social unrest rely on conflict event datasets derived primarily from major international/national news reports. Yet, a large body of research identifies systematic patterns of ‘missingness’ in these data, calling into question statistical results drawn from them. In this project, we explore three specific opportunities for additional data collection to help recover systematically excluded events and to potentially assist in addressing resulting bias. We find that all three approaches result in additional and often systematically different material than that reported in news-based datasets, and we reflect on the advantages and drawbacks of these approaches.