Susceptibility to violent extremism and cognitive rigidity: Registered replication, corroboration and open questions for criminological research and practice

Abstract

Does cognitive inflexibility predict vulnerability to violent extremism? Schumann, Salman, Clemmow, and Gill (2021) conducted a registered direct replication of Zmigrod, Rentfrow, and Robbins’ (2019, Frontiers in Psychology, 10, 989) studies on the relationship between neuropsychologically assessed cognitive inflexibility and extremist attitudes. Replicating the original study, Schumann et al. (2021) demonstrated that cognitive inflexibility on the Remote Associates Test was related to individuals' willingness to fight and die for their ideological group (Bayes Factor = 58.7). In a further corroboration, the study found that individuals who indicated they would self-sacrifice in a trolley-dilemma paradigm were more psychologically rigid on the Remote Associates Test than those who opted for self-preservation (Bayes Factor = 402.3). A mini meta-analysis with the original study revealed a significant negative correlation between cognitive flexibility and violent extremist behaviour intentions, thereby supporting the conclusions of the original study by Zmigrod and colleagues (2019). Nevertheless, the replication study highlighted moderators of the effect sizes of the reported relationships and revealed subtleties in the theoretical relationships between cognitive style and tendencies towards extremist attitudes. Follow-up analyses with Bayesian statistics demonstrated that—corroborating the original study—cognitive inflexibility on both the Remote Associates Test (Bayes Factor = 7.2) and the Wisconsin Card Sorting Test (Bayes Factor = 2.3) are linked to individuals' willingness to fight for an ideological cause. However, the replication study revealed that these effects are specific to individuals who endorse ideological self-sacrifice. Furthermore, deviating from the original study, Schumann and colleagues found that cognitive flexibility on the Wisconsin Card Sorting Test may in fact positively predict willingness to die for a group under some conditions. This suggests that flexibility may be a double-edged sword when it comes to ideological conversion. These insights indicate that a fine-grained examination of risk sub-profiles and interactions among predictors is essential. The findings of the replication also stimulate a discussion of what constitutes a phenomenologically or practically meaningful effect size for psychological and criminological research and practice—and the implications for cognitive methodologies in risk assessment.