Violence against civilians: a disaggregated analysis

Disaggregated approaches to conflict research have led to new insights into the patterns and processes of political violence in developing countries. This article uses the most comprehensive subnational political violence data (ACLED) to observe where and when violence against civilians occurs within civil wars. Several new conclusions are evident from an event-based analysis of civilian violence: retribution or collateral damage are poor explanations for attacks on the unarmed. Instead, civilians are targeted because they are accessible; rebel groups kill more civilians, often in an attempt to create new frontlines for conflict. However, governments are also responsible for high rates of civilian death, yet they often "contract" this violence out to militias. This analysis confirms that there are multiple violent groups within civil war spaces, and small opposition groups commit higher levels of violence against civilians in local spaces. The strength of a violent group compared to its competition shapes how much civilian violence it commits. The results suggest that theories that emphasize civilian support and retribution as a basis for violence against civilians have overlooked the importance of how multiple violent opposition groups compete within civil wars, and how civilians suffer as a result.