Modeling the Social Amplification of Risk following a Terrorist Strike: Methodological Challenges

Publication Type: 
William J. Burns
Understanding those factors critical to predicting public response is crucial to our ability to model the consequences of a terrorist strike in an urban area. Sixteen hypothetical damage scenarios were systematically varied according to non-terrorism vs. terrorism, explosions vs. infectious disease releases, terrorists’ motives as demands to release prisoners vs. solely to instill fear, non-terrorists’ motives as nonintentional vs. intentional (criminal), victims as government officials vs. tourists, non-terrorist incidents as involving no negligence vs. negligence, terrorist acts as non-suicidal vs. suicidal, and number of casualties (0, 15, 495). The setting was a local theme park. Students at a university in San Diego County were randomly assigned to different scenario conditions. For these scenarios, they were asked to address a number of questions regarding their perceptions and likely behaviors during and following an accident or terrorist strike. Results from regression modeling and Multivariate Analysis of Variance (MANOVA) indicated that terrorism and the mechanism used were most influential followed by the presence of suicide or negligence, motive, and victim. Number of casualties made little difference once these other factors were accounted for. To forecast community response, a system dynamics model was introduced that incorporated the study’s survey findings. This model simulated the immediate and mid-term diffusion of fear in a community for different types of accidental and terrorist events. These findings should prove useful to those wishing to predict public response to a variety of different contingencies involving terrorism.