Date of Degree


Document Type


Degree Name



Criminal Justice


Joshua D. Freilich

Committee Members

Kevin T. Wolff

Brian Lawton

Steven M. Chermak

Subject Categories



risk factors, comparative analysis, violent extremists, non-violent extremists, jihadists, far-right


Problem Statement: This study investigated risk factors commonly highlighted by prior studies and risk assessment tools (e.g., Violent Extremism Risk Assessment Version 2 (VERA-2) and Extremism Risk Guidance 22+ (ERG22+)). It compared the risk factors across and within two groups – violent extremists (i.e., jihadists and far-right extremists) and non-violent extremists (i.e., jihadists and far-right financial/material support crime extremists) in the U.S. Jihadists and far-rightists are the two greatest security threats facing the U.S., and the findings can aid counterterrorism efforts on assessment development, identifying effective risk factors across varying groups, and assessing violence risk. Research Questions/Objectives: This study aimed to achieve two central objectives: (1) compare these risk factors across and within two groups: violent extremists (i.e., jihadists and right-wing extremists) and non-violent extremists (i.e., jihadists and right-wing financial/material support crime extremist offenders) and (2) explore underlying associations between risk factors and theoretical domains. Theoretical framework: Although criminological theories are applied to better inform radicalization and violent extremism pathways, psychological theories have dominated the risk assessment and factor development with little influence from criminology. Both tools under study here did not utilize criminology theories to inform their development of these risk factors. To develop the study’s hypotheses about the extremist violence correlates, I applied and integrated two individual-level criminological theories. Both rational choice and social network theories inform risk factor domains and conceptualizes these factors within a nested model, similar to ecological systems theory (Bronfenbrenner & Morris, 2007). Rational choice is an individual-situational theory focusing on individuals’ behaviours resulting from their decisions based on situational characteristics, with the assumption of an individual’s agency. The rational choice theory has links to social network theory. Social network is an individual-group theory outlining individual-group interactions and emphasizes relationship significance and promoting specific actions (e.g., violence). Research design and subjects: This mixed-method study used 1990 – 2018 offender-level data from the open-source U.S. Extremist Crime Database (ECDB) and randomly selected a sample of 420 offenders, including 210 violent extremists (i.e., 105 jihadists and 105 far-rightists) and 210 non-violent financial/material support crime extremists (i.e., 105 jihadists and 105 far-rightists). The ECDB is an open-source relational database and includes many offender-level attributes, both incident-level and individual-level characteristics. Importantly, this study enhanced the ECDB and innovatively collected new variables to fully operationalize these risk factors. The first research question was examined using a series of binary logistic regression models to predict and compare for significant risk factors. The second research question was investigated using tetrachoric correlation coefficients to run an exploratory factor analysis model to determine the associations between the dichotomous variables. The case study constructions across both groups selected three cases each from groups and sampled for heterogeneity (n = 6). The six case studies were developed to explore contextual nuances within the risk factors (i.e., unique behavioral manifestations of risk) at the individual level. Findings: Generally, the risk factors do not necessarily differ across and within both the violent and non-violent groups with several exceptions. Specific risk factors that merit possible logical explanations for their statistical significance are provided within the dissertation. Additionally, the exploratory factor analysis findings concluded almost all the various motivations examined can be grouped into the individual-situational and individual-group domains instead of the individual domain. The qualitative study provided six case studies involving one jihadist and two far-right extremists for each violent and non-violent group. The study provided several examples of each risk factor extracted from the case studies to provide variation. Potential avenues of future research were suggested to examine certain aspects of the project. Points for consideration: First, it should be noted VERA-2 risk factors can be applicable to the non-violent extremist groups as demonstrated from both the quantitative and qualitative evidence. Second, several variables were highly correlated across and within domains and may result in multiple risk factors appearing in an extremist’s profile, inaccurately raising their risk level. Third, the risk factors were harder to map onto the non-violent far-right extremist group extremists (e.g., sovereign citizens, tax defiers and/or protestors) for two reasons. First, open-source data concerning non-violent extremists are limited compared to violent extremists’ (and non-violent jihadists), impacting risk factor variable coding. Second, it is likely the risk assessments were developed with jihadists (both violent and non-violent) as the main focus. The ERG 22+ was developed based on U.K cases, specifically jihadists. As such, more research is still needed to examine whether these risk factors are truly applicable to the non-violent extremist group.

Included in

Criminology Commons