Psychometric evaluation of the Militant Extremist Mindset Scale

Abstract

Due to its wider social impact, there is a growing interest in the study of radicalization and violent extremism. Radicalization and violent extremism can be operationalized through three-dimensional Militant-Extremist Mindset (MEM), including Pro-violence (PV) – acceptance, justification, and advocacy of the use of violence, Divine Power (DP) – beliefs in heaven and God, the role of martyrdom, and afterlife pleasures, and Vile World (VW) – the belief there is something wrong with the world we live in, and that the present-day world is vile and miserable. Despite the growing support of three-dimensional structure of the construct in adults, there is a lack of studies assessing the MEM among the youth, which may be particularly susceptible to radicalization and violent extremism. This study explored the latent structure and psychometric properties of the MEM scale on a sample of 283 high school students from Belgrade and Sandžak aged between 15 and 18 (M = 16.30, SD = 0.71;71.7% females). They completed a 24-item version of the MEM scale. The results of confirmatory factor analysis (Maximum Likelihood) of MEM demonstrated insufficiently good fit [χ2(249) = 543.06, p < .001, CFI = .86, TLI = .83, RMSEA = .06, 90%CI: .06 - .07] for initial three-factor model of correlated dimensions of PV, VW, and DP, with a single significant latent correlation of .32 between PV and VW factors. All three subscales [PV KMO = .96, α = .83, H2 = .89; VW KMO = .95, α = .84, H2 = .90; DP KMO = .92, α = .74, H2 = .84] exhibited high indices of item sample adequacy, internal consistency, and homogeneity. The model was modified by exclusion of three items from PV and two items from DP subscales from the final version due to their poor psychometric properties and low primary or relatively high secondary factor loadings. The resulting set of 19 items demonstrated a satisfactory fit to the three-factor model of MEM [χ2(151) = 302.11, p < .001, CFI = .92, TLI = .90, RMSEA = .06, 90%CI: .05 - .07]. Psychometric properties of the shorter versions of PV [KMO = .97, α = .86, H2 = .95] and VW scales [KMO = .94, α = .82, H2 = .91] proved to be markedly higher after the exclusion of the problematic items. Results on the MEM’s latent structure are in line with studies using adult samples. Results indicate that the adapted, shorter version of MEM can be used for assessing radicalization and violent extremism among youth.