Towards Dynamic Modeling of Violence: Cultural Violence and Civilian Violence in Syria

• In Event: Addressing Human Rights, Conflict, and Public Heath

Abstract

Modelling a conflict’s sequence of events is an imperfect science at best, particularly when they are inherently interrelated, like targeted violence towards cultural sites and civilians. Both are “caused” by the broader conflict. Looking at their interrelationship introduces an issue with simultaneity. One solution is to choose one measure as the primary dependent variable and employ spatiotemporal models that treat either time or space as static elements. Yet doing so reduces the event dynamics’ complexity. This study employs dynamic spatiotemporal modeling to characterize the simultaneity in the relationship between the targeted violence towards cultural sites and civilians during the Syrian War from 2014 to 2017. We combine civilian fatalities from the Uppsala Data Conflict Program with novel data on we coded cultural heritage destruction to create a spatiotemporal dataset. We then employ a holistic design, starting with a replication of Braithwaite and Johnson’s (2015) spatiotemporal modeling approach and use those results to develop a dynamic autoregressive distributed lag (D-ARDL) model. Our analysis finds that a more dynamic approach provides important nuance to such microdynamics. We discuss the implications of these findings for dynamic modelling and conflict event sequencing.

Authors

• Michelle Fabiani, University of New Haven

• Gregory A. Wadsworth, University of New Haven

• Fiona Greenland, University of Virginia