Innovations in Measuring Strategic Interactions During Violent Conflict

Session Submission Type: Full Paper Panel

Session Description

This panel brings together four projects that leverage advances in data collection, measurement, and modeling to improve our understanding of how relationships between adversaries develop over repeated interactions. Recent innovations in these three steps allow the authors to improve the match between theory, data, and methodologies addressing questions of strategic interaction between conflict actors and (in one case) geography. In doing so, the four projects provide dynamic and more fine-grained insights into conflict processes.

The first project uses data on negotiations to measure the trajectory of negotiations in civil war history between conflict parties and show that constructive interactions lead to positive feedback loop that make settlements more likely. Similarly, the second project uses a Bayesian nonparametric approach to model how protests tend to cluster in time and space, allowing empirical models to better match theories of snowballing behavior. The third project uses textual event data to show that violent non-state organizations are more likely to gain concessions from governments when they vary the tactics they employ. The final project explores different approaches for converting spatial data on roads into network data, and illustrates how different decisions affect the inferences reached when analyzing civil war event data on road networks.

Together, the four projects present new theoretical explanations for how patterns of strategic interaction evolve over time between adversaries. Empirically, they offer new ways to model interactions between governments, rebel groups, terrorist organizations, and protest movements.

Sub Unit

• DIVISION 21: Conflict Processes

Cosponsor

• DIVISION 8: Political Methodology

Individual Presentations

• Evaluating the Cumulative Effects of Negotiations in Civil War Resolution - Elizabeth J. Menninga, University of Iowa; Alyssa Prorok

• Clustered Political Violence Events as Dirichlet Process Models - Kaitlyn Webster, Duke University; Margaret J. Foster, Duke University

• Extracting Road Networks from Spatial Data - Juan Fernando Tellez, Duke University; Rob Williams, Washington University in St. Louis

Chair

• Johannes Karreth Ursinus College

Discussants

• Nils W. Metternich University College London

• Karsten Donnay University of Konstanz