Research on civil violence and political instability has yielded important baseline information about the conditions that have the potential to increase risks for mass violence, political instability, or state failure. Limits on data collection, analysis, and interpretation immediately before and during outbreaks of conflict, however, constrain analysts from identifying which conflict-prone country will descend into political instability or violence in time for a targeted intervention or effective response. This paper presents a conceptual framework for analyzing the heterogeneous and dynamic character of local conflict. This work is anchored in the need to describe conflict dynamics as they occur, to understand in real-time the political, economic, and social drivers and to gather high-resolution (e.g. local, disaggregated) data to analyze social instability. The paper demonstrates the effectiveness of such a framework and applies it to three case studies: the Kenyan presidential election of 2007, the Georgia-South Ossetian war in 2008, and the Mexico drug wars in 2010. The case-study results suggest that the analysis of high-resolution event data immediately prior to two of the conflicts could have enabled early detection and warning of the potential for large-scale civil violence. The third case provides retrospective analytical insight into local conflict dynamics. This paper argues that in an era of non-state actors, emergent conflict, and natural resource pressures, a new conceptual approach to event data collection and analytical process can provide low-cost, near real-time monitoring and evaluation of ongoing and potential conflicts in multiple languages and regions.
Author(s): Davin O'Regan