Much Ado about Something: Predicting Interstate Crises Onset and Resolution over Contentious Issues

We explore whether dynamic information from event data can improve our ability to forecast the onset and termination of international crises. Studies of dyadic interstate conflicts have demonstrated that taking into account contentious issues between countries can improve on conventional models explaining dyadic violent conflict based on purely structural factors (e.g., Gibler 2012, Gleditsch and Ward 2013). However, data on contentious issues are often limited to particular types such as claims over territory, river, or sea, although states can experience violent conflict over various other issues such as weapons program or regime type that may not be recorded in existing data. Recent work on civil and interstate wars suggests that information from news can improve structural conflict prediction that can help capture changes over time (Chadefaux 2014, Muller & Rauh 2018). We seek to combine measures of contentious issues and levels of conflict and cooperation between states based on event data to forecast the onset and resolution of ICB conflict data.