Measurement and Conceptual Approaches to Religious Violence: The Use of Natural Language Processing to Generate Religious Violence Event-Data

Abstract

AbstractHow do we measure religious violence? This study is focused on utilizing new methodological approaches and data sources to measure religiously motivated violence. Previous attempts to measure religious violence concentrated on coding U.S. State Department International Religious Freedom reports or utilizing existing datasets on armed conflict/civil wars. These previous attempts provided state-level data of the levels of religiously motivated violence, but due to data limitations cannot provide more fine-grained measures of specific acts of violence tied to religious motivation. In particular, accounting for varying levels of intensity especially in regards to non-lethal acts of religiously motivated violence is missing. This study builds upon previous attempts focusing on the creation of more fine-grained measures and accounting for its variation at the sub-national level utilizing natural language processing. The data generated are used to examine incidences of reported religious violence in India from 2000 to 2015.