Forecasting civil war and political violence
This chapter describes efforts to predict civil wars and related forms of political violence. The chapter first describes the evolution of a conflict prediction paradigm, which is characterized by a commitment to the scientific method, and a consensus on that and how the performance of predictions is to be evaluated. After explaining the standard procedure employed to forecast the onset of war, the chapter then outlines the considerable variation between projects on the outcome that is predicted; the spatial and temporal units for which a prediction is made; the type of predictors that are used; and the computational method that links these predictors with the outcome. This variation is reflective of a young field in which rapid methodological development is in full progress. Finally, the chapter reviews debates on the possibility and desirability of conflict prediction, and the open question of how academic civil war prediction can and should influence policy-making.