Brittany Alexander, ISTPP Predoctoral Fellow and doctoral student in statistics, presented her paper “Understanding Terrorism Policy Preferences through Bayesian Model Averaging and Multiple Imputation” at the Joint Statistical Meetings on August 12, 2021. ISTPP Director, Arnold Vedlitz, coauthored this study.
Their research uses a two-wave panel survey conducted by ISTPP in 2016 with funding from the National Science Foundation and a six-wave panel survey conducted by Decision Research. The surveys measured a variety of issues and attitudes, including perceived likelihood of terrorism, different concerns about terrorism, support for federal and local spending, and support for specific policies to prevent terrorism.
A Bayesian model was fit to the data and used both surveys to create estimates of public support for a list of policies to prevent terrorism. The support was averaged across the policies in the survey. The model was able to show that there was no statistically significant change in policy support at the population level across the seven months in the study. This finding is far more conclusive than previous work as the model was capable of detecting change at the level of movement of one unit change on a single item in the seven item scale used to measure policy support.