The majority of contemporary research on terrorism and extremism has focused on jihadi extremism, while other ideologies have received far less attention. In response to this gap, the current study aims to compare the similarities and differences between jihadi, right-wing and left-wing extremists at the individual level. Using the Profiles of Individual Radicalization in the United States (PIRUS) dataset, a multinomial logistic regression model was used to compare individual-level characteristics across ideologies. Additionally, a two-step cluster analysis was conducted to determine whether these similarities and differences can provide additional insights into differentiating types of extremists generally based on their personal characteristics, regardless of their ideological adherence. The results of the multinomial logistic regression illustrate that there are notable differences across extremist ideologies, but also many similarities. Further exploring ways to better highlight individual-level differences, a cluster analysis revealed five distinct groups of extremists based on their personal characteristics, and demonstrate the utility of a typology of individual characteristics that is empirically derived and validated, and is not dependent on the a priori identification or specification of ideological motivation.
This is the abstract of a new Studies in Conflict and Terrorism article by Sara Doering, Garth Davies, and Raymond Corrado. The use of multinomial logistic regression as a data analysis methodology in this paper is interesting. START’s Profiles of Individual Radicalization in the United States dataset is used in the study.