A Bayesian Integrated Theoretical-empirical Model of Conflict in Africa

• In Event: Forecasting Conflict

Abstract

Mass dissent and political violence are not consequences of unitary populations deciding as one to act. Rather, they are the result of a process in which individuals choose to mobilize, often at great risk to themselves. Without a clear understanding of mobilization dynamics at both the individual and the aggregate level, one cannot truly understand anti-state activity, but that understanding requires a theory that bridges the gap between the individual and the aggregate. We propose such a theory, characterized by a reduced form at the individual level that allows for the important aggregation step without loss of population heterogeneity. We employ a theoretically-informed Bayesian Item Response Theory model developed for the purpose of testing our model to directly measure the components of our theory for the test case of anti-state attitudes within Africa. We find that those components are associated with civil unrest cross-nationally in the manner our theory predicts.

Authors

• So Jin Lee, MIT

• Kaitlyn Webster, Duke University

• Margaret J. Foster, University of North Carolina, Chapel Hill

• Marco Morucci, New York University

• Elizabeth C. Carlson, NDSU

• Will H. Moore, Arizona State University

• David A. Siegel, Duke University