Computational Model of Public Support for Insurgency and Terrorism: A Prototype for More-General Social-Science Modeling.

摘要

This report is part of a longer-term research agenda that began with a 2008-2009 RAND study reviewing and integrating social-science knowledge related to counterterrorism. That study used qualitative conceptual causal models called 'factor trees' to show compactly all the different factors contributing to various aspects of terrorism at a slice in time and how the factors relate to each other qualitatively. That work reflected a considerable base of social-science knowledge. In the 2009-2011 period, the factor trees were subsequently tested and refined in empirical studies; they were also used in a variety of applications. This report describes a next step. Going beyond the conceptual and qualitative, we have specified and implemented a prototype computational model (PSOT) for just one of the earlier factor-tree models, that addressing public support of terrorism and insurgency. The earlier model is extensively discussed in Davis, Larson, et al., 2012. The factor tree itself is shown in Figure S.1. Our prototype seeks to describe, as a function of contributing factors, the extent to which a nation's public supports an ongoing insurgency and its use of terrorism. The support may involve, for example, sympathy, approval, material contributions, sheltering from the government, or direct participation in insurgent operations. Public support is a complicated aggregation because the public is often heterogeneous with numerous disputing factions. Moreover, public support is not a purely 'rational' behavior predictable with a stable utility function. We undertook building a computational model with trepidation because characterizing the mathematics of how factors combine required going well beyond the established science base.