Forecasting Crisis Dynamics with Machine Coded Data: A Model of Power Projection, Influence and Escalation
Abstract
Research Problem: International relations offers two contrasting schools of thought on crisis dynamics. Deterrence theory advocates bold words or actions intended to inhibit escalation; the spiral model discourages these same behaviors, fearing they can produce a security dilemma. Contradictory forecasts are unhelpful for policy; which prescriptions should one follow? Both perspectives have also failed to evolve with new thinking about the nature of conflict. Nations appear to fight for reasons that differ from those assumed by either deterrence theory or the spiral model. Resolving this disconnect is crucial to anticipate crisis dynamics. This project aims to answer the central question of How can one explain and predict actions during an international crisis?Proposed Methods: Phase 1: Data Processing~overcoming the ongoing intellectual impasse by applying ~big data~ techniques to create an empirical model of crisis dynamics. The smaller steps in this process include intercoder aggregation, event taxonomy, and measuring temporal distortion. Phase 2: Data Analysis~led by Drs. Douglass and Scherer, supervised by Professor Gartzke and will be conducted by cPASS~s affiliated graduate and undergraduate students. The smaller steps in this process include data preparation, selecting the appropriate function form, interpreting out of sample accuracy, interpreting residuals, interpreting covariate importance, interpreting covariate effects, and repeat. Phase 3: Theory Refinement~critically assess the existing canonical theories and propose refinements and extensions of theory in light of what is discovered in the analysis conducted in Phase 2.Anticipated Outcomes: The project aims to produce an improved International Crisis Behavior dataset, a probabilistic decision tree, the ability to identify the most appropriate historical analogies, and a theory of crisis dynamics which will allow for the prediction of outcomes from different types of actions in crises in combination with the probabilistic decision tree.Impact on DoD/Broader Implications for National Defense: The anticipated Theory of Crisis Dynamics will help policy makers by resolving the intellectual logjam between deterrence theory and the spiral model. Decision makers will be able to apply the theory to the immediate situations they face, scenarios they plan for, and future evolutions of foreign and defense policy.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 15, 2019
- Source ID
- N000141912491
Entities
People
- Erik Gartzke
Organizations
- Office of Naval Research
- United States Navy
- University of California, San Diego