Beyond binary: Analyzing closed-source data to compare specific roles and behaviors within violent and nonviolent terrorist involvement.

Increasingly, studies compare risk and protective factors for involvement in violent and nonviolent terrorist behaviors. This exploratory study investigates whether this distinction is sufficient, or whether it should be disaggregated further into more granular terrorist roles and behaviors. Using data on 404 referrals to a UK countering violent extremism Prevent hub specializing in mental health and associated needs, we compare violent and nonviolent referrals, and then more specific behaviors (vulnerability, proactive extremism, foreign fighting, and violence planning). Bivariate and multivariate analyses show there is value in disaggregating beyond the binary violence versus nonviolence distinction, as more (and more detailed) relationships emerged when using the disaggregated set of behaviors. While gender did not differentiate violent and nonviolent referrals, women were more likely to be referred for radicalization vulnerability or potential foreign fighting. Extreme right-wing and extreme Islamist referrals were no more or less violent overall, but Islamist referrals were disproportionately referred for both the most and least violent behaviors. Personality and developmental disorders were associated with violence, and disaggregated behaviors provided detailed insight into the drivers of these associations. These exploratory findings, while interesting, likely do not generalize beyond our specific sample. Instead, this study's value lies in demonstrating the utility for both research and, eventually, practice of disaggregating beyond violence and nonviolence. The results demonstrate clear operational implications for threat assessment in the need to include a more refined set of risk factors to aid in assessing risk of more relevant outcomes than terrorist involvement overall.