Putting the numbers to work: Implications for violence prevention

By analysing international crisis behaviour from 1918 to 1994, this study seeks a better understanding of the factors that make interstate and intrastate crises endure and escalate into violence. Observations are crisis actors drawn from the International Crisis Behavior (ICB) dataset. The analysis indicates that an actor is more likely to use violence in a crisis if there is social unrest in that country; if it is contiguous to its main adversary in the crisis; if there is a violent trigger to the crisis; if there is an ethnic dimension to the crisis; and if the crisis is a long one. If the principal adversaries in a dyad are democratic or the actor is democratic, the likelihood of the use of violence is significantly reduced. Crises tend to be longer when there is an ethnic component and if there is violence. Taken together, the two analyses link crisis duration and outbreak of violence. Ethnicity and unrest have both direct and indirect effects on violence. The findings have implications for conflict prevention in two ways. If the factors found to be predictors of an escalation to violence are present during a crisis, then this provides an important early warning for agencies that the outbreak of violence is likely. In particular, early and decisive intervention, assuming it is successful, may be an effective method of preventing escalation to violence. Even in the absence of a crisis, the existence of such factors can be addressed through structural prevention, particularly through the development of democratic and other institutions.