Date of Award

Spring 2010

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Graduate Group

Criminology

First Advisor

Lawrence W. Sherman

Second Advisor

Randall Collins

Third Advisor

John M. MacDonald

Abstract

This dissertation, in a 3-paper format, uses three data sets to study three aspects of terrorism in Pakistan. In the first paper, using data from the GTD, I describe empirically the temporal and spatial patterns of terrorism incidents in Pakistan from 1974 to 2007. In addition, I also describe the patterns in target types, weapon types and terrorist types and the patterns prior to and following the U.S.-led invasions of Afghanistan and Iraq. Analysis methods include the univariate time series, descriptive statistics and the GIS. The study offers new insights on the measurement of terrorism, the cyclical nature of terrorism, the role of conflict, the choice of weapons, the sponsorship of terrorism, the selection of targets and the reactionary nature of terrorism. The second paper analyzes personal, socioeconomic and demographic characteristics of arrested terrorists in Pakistan from 1990 to 2009. I analyze police data on 2,344 terrorists using the GIS, univariate and bivariate analyses. Significant group differences, as well as differences based on geography lead to the conclusion that a generic profile of terrorists is almost impossible. One factor universally present in the circumstances of almost all the terrorists is that they belong to some area of conflict. In the third paper, I test Sherman’s theory of defiance (1993) to discern how arrests of different terrorist types and terrorist arrest types result in different types of reactions—defiance, deterrence, or irrelevance. Terrorist types can be divided into hardcore terrorists and peripheral terrorists, and arrest types can be divided into ordinary arrests or arrests by killings. I use 20 years of data from eight regions of the Punjab. Using fixed-effects cross-sectional time series (long panel), instrumental variable approach and Poisson distribution, I conclude that: aggregated arrests, ordinary arrests, and arrests of hardcore terrorists, in the current period, are associated with higher expected incidence and seriousness of terrorism in the same six month period. Further, that as compared to peripheral arrests, hardcore arrests generate more defiance. Lags of arrests and ordinary arrests decrease the expected incidence and seriousness of terrorism, suggesting a possible decay in defiance after the first six months.