A latent class analysis of intimate partner violence perpetrators: exploring the role of executive functions

Purpose This study aims to identify subtypes of male perpetrators of intimate partner violence against women (IPVAW) through latent class analysis (LCA), and to examine whether executive functioning (EF) contributes to differentiating these subtypes. Design/methodology/approach The sample comprised 996 men convicted of IPVAW in Spain. First, LCA was conducted using psychological and violence-related variables proposed by Holtzworth-Munroe and Stuart (1994). Subsequently, latent class regression (LCR) analyses were performed to assess the associations between EF variables and class membership. Findings Four perpetrator classes were identified: moderate antisocial (MA), low-risk family-only (LRFO), severely violent antisocial and a novel situational family (SF) group. LCR results indicated that EF variables were not strong predictors of class membership. The only exception was cognitive flexibility: lower attentional cost – reflecting better flexibility – was associated with a higher likelihood of belonging to the LRFO group compared to the MA group. Practical implications These findings highlight the heterogeneity of male IPVAW perpetrators and introduce a new profile, the SF subtype. Although most EF variables were not differentially associated with class membership, the observed link with cognitive flexibility suggests that neuropsychological assessment may provide added value for tailoring intervention strategies to perpetrators’ individual cognitive profiles. Originality/value This study contributes novel insights by assessing whether executive functioning (EF) – working memory, inhibition, decision-making and cognitive flexibility – predicts class membership among IPVAW perpetrators within a LCR framework, based on the Holtzworth-Munroe typology. While previous research relied on psychological and violence-related indicators, this is the first to evaluate EF as a predictor. Although EF variables did not consistently predict class membership, cognitive flexibility showed a small but significant association. Moreover, replicating the original typology revealed a previously unreported class – SF – highlighting overlooked perpetrator profiles and supporting the relevance of integrating cognitive dimensions into typological models.