ABSTRACT

While research on terrorism has grown rapidly in recent years, few studies have applied criminological theories to the analysis of individual-level political extremism. Instead, research on radicalization has drawn primarily from political science and psychology and has overwhelmingly focused on violent extremists, leaving little variation in the dependent variable. Using a newly available dataset, we test whether a set of variables derived from prominent criminological theories are helpful in distinguishing between non-violent and violent extremists. Results show that variables related to social control (lack of stable employment), social learning (radical peers), psychological perspectives (history of mental illness), and prior criminal record all have significant effects on participation in violent political extremism and are robust across multiple techniques for imputing missing data. At the same time, other common indicators of social control (e.g., education, marital status) and social learning perspectives (e.g., radical family members) were not significant in the multivariate models. We argue that terrorism research would benefit from including criminology insights and by considering political radicalization as a dynamic, evolving process, much as life-course criminology treats more common forms of crime.