Presented at Polnet 2017
Abstract: Existing literature has recognized that actions of repression and dissent are endogenous and strategic, but how to study and model this interdependent dynamic remains an empirical challenge for researchers. This task becomes even more challenging when the measure of interactions is highly aggregated. Compressing actions into a binary violent/nonviolent index misses the variety of tactics being used by both parties and also overlooks the mechanism and process through which tactic interactions lead to conflict escalation or de-escalation. To study this dynamic, I propose using network analysis as a new approach to explore the interdependence. Three tasks are accomplished in this article. First, by developing a network of tactics and strategies focused on actions, rather than actors, I can map out tactic interactions in conflictual events and explore clusters of strategy substitutions between the state and the dissidents. I then employ this new measure to explain electoral violence and mutual conflict escalation, showing that it allows us to examine not only actions at more micro level but also the interplay between repression and dissent which were often analyzed separately due to the limitation of empirical models and the complexity of the interaction. Finally, I show that this network approach also significantly improves our ability to forecast different types of repressive state actions at more micro tactic and event level, adding the important policy value to this approach.