Investigating the role of group-based morality in extreme behavioral expressions of prejudice

Throughout history, identity-based hatred and violence have caused immense suffering. Despite arguments suggesting long-term decreases in global violence and prejudice, recent years have witnessed concerning increases in hate crimes and the proliferation of hate groups, particularly in the US and Europe. This resurgence highlights the critical need to understand the underlying mechanisms driving acts of hate. Existing research often emphasizes the role of intergroup threat, suggesting that perceived threats (realistic or symbolic) from outgroups fuel prejudice and extreme behaviors like hate crimes and hate speech. However, this perspective leaves unanswered the crucial question: Why do some perceived threats, and some individuals perceiving them, lead to acts of hate? This study proposes that the moralization of perceived threats is central to understanding acts of hate, a hypothesis referred to as the moralized threat hypothesis. This hypothesis draws upon a significant body of research connecting violence and extreme behavior to moral values, perceptions of moral violations, and feelings of moral obligation. It posits that acts of hate are often motivated by the belief that an outgroup has committed a moral wrong, and that the likelihood of perceiving such violations is influenced by an individual's own moral values. Given recent increases in EBEPs associated with right-wing ideologies and the concern over hate speech's role in violent crimes, the focus here is primarily on EBEPs aligned with right-wing ideologies, anticipating an association with moral values centered on group preservation, which are linked to conservatism and right-wing ideologies in US contexts. Moral Foundations Theory (MFT) provides a framework to operationalize these values, proposing a hierarchical model with individualizing values (care/harm, fairness/cheating) and binding values (loyalty/betrayal, authority/subversion, purity/degradation). The moralized threat hypothesis predicts a correlation between binding values and EBEPs against groups marginalized by the ideological right. This prediction is tested across five studies, starting with a geospatial analysis of US county-level moral values and hate group prevalence, followed by survey-based studies examining the perceived justification of EBEPs.

The literature review establishes the background for this study by reviewing previous research on hate crimes and prejudice. It summarizes findings on the detrimental effects of hate crimes on victims and communities, showing increased rates of depression and anxiety among survivors. It also acknowledges the ongoing debate about long-term trends in global violence, while pointing out recent counter-trends such as increased hate crimes in the US and Europe, often targeting specific minority groups, emphasizing the need for further understanding of these phenomena. The authors discuss prior research focusing on the role of intergroup threat as a mechanism in hate crimes and hate speech, highlighting how perceptions of threat (both realistic and symbolic) can lead to increased prejudice. However, the existing literature is found to lack a comprehensive explanation for why some perceived threats lead to acts of hate while others do not. The study bridges this gap by proposing the moralized threat hypothesis, integrating existing research on the link between violence, moral values, and perceptions of moral violations. The framework offered provides a novel perspective on the motivations behind extreme prejudice by focusing on moral factors.

The study employs a mixed-methods approach, combining geospatial analysis with experimental and survey methodologies across five studies. **Study 1:** This study uses geospatial analysis to investigate the relationship between county-level moral values and the prevalence of hate groups in the US. County-level moral values (individualizing and binding) were estimated using data from YourMorals.org, applying Multilevel Regression and Synthetic Poststratification (MrsP) to account for non-representative sampling. Data on hate group prevalence were obtained from the Southern Poverty Law Center (SPLC). A negative-binomial regression model with Bayesian Spatial Filtering was used to analyze the data, adjusting for various county-level covariates (ethnic composition, educational attainment, poverty level, 2016 Democratic Presidential vote share, rural vs. urban). The analysis also considered models with and without state-fixed effects to explore potential state-level confounding variables. **Studies 2-5:** These studies employ survey and experimental methods to examine the relationship between moral values, perceptions of outgroup moral violations, and perceived justification of EBEPs. Participants were recruited through various online platforms (Qualtrics Panels, Amazon Mechanical Turk). Moral values were measured using the Moral Foundations Questionnaire (MFQ). Perceived moral wrongness of outgroup actions was assessed using specific questions. Perceived justification of EBEPs was measured using items describing different acts of hate, scored on scales ranging from 'not at all justified' to 'extremely justified'. Studies 2 and 3 used observational designs to examine these relationships with different outgroups (Muslims and Mexican immigrants), while Studies 4 and 5 employed experimental manipulations to control for perceived moral threat in fictional outgroup scenarios, focusing on the effects of binding and individualizing values violations. Bayesian hierarchical models were used for data analysis in studies 2-5 to account for varying levels of justification among participants and across different EBEP items. Mediation analysis was performed to determine the role of perceived moral wrongness in the relationship between moral values and EBEP justification. Study 5 further explored moderated mediation to examine whether the effect of perceived moral wrongness was moderated by binding and individualizing values. This study also incorporated the use of vignettes depicting graphic content in order to elicit stronger responses. Across all studies, robust analyses were used to account for potential methodological limitations.

The findings across the five studies consistently support the moralized threat hypothesis. **Study 1:** Revealed a positive association between county-level binding moral values and the prevalence of hate groups, even after adjusting for several covariates. However, this relationship was attenuated when state-fixed effects were included, although this may be due to the reduced statistical power of the fixed-effects model. **Studies 2 and 3:** Demonstrated that individuals with higher binding values were more likely to perceive EBEPs as justified, and that this effect was partially mediated by perceived moral wrongness of the outgroup's actions, even when controlling for political ideology. **Study 4:** Experimentally manipulated perceived moral wrongness of a fictional outgroup, finding that higher perceived moral wrongness led to a greater perceived justification of EBEPs. The effect of the experimental manipulation was fully mediated by perceived moral wrongness. **Study 5:** Experimentally manipulated perceived moral wrongness through the use of graphic vignettes. This study found that higher binding values significantly moderated the mediating role of perceived moral wrongness on EBEP justification. The effect was far more potent when violations of binding values were concerned. In other words, individuals with higher binding values were considerably more likely to view EBEPs as justified when they perceived a moral violation of binding values compared to individualizing values, even when controlling for religiosity and political ideology. In the context of violations of individualizing values, the mediating effects were far smaller.

The findings provide strong support for the moralized threat hypothesis, showing a clear link between group-based moral values, perceptions of outgroup moral violations, and the perceived justification of EBEPs. The results highlight the importance of considering moral factors alongside threat perceptions in understanding extreme prejudice. The significant role of binding values suggests that individuals who prioritize loyalty, authority, and purity may be particularly susceptible to justifying EBEPs when they perceive an outgroup as violating these values. This insight has significant implications for interventions aimed at reducing prejudice and preventing violence, as it points to the need for approaches that address underlying moral beliefs and perceptions. The results, also, are in line with observations that the digital environment can be leveraged by extremists to unite individuals with similar ideologies. By focusing on moral outrage, such strategies avoid the need to explicitly address material threats.

This research advances our understanding of the motivations behind extreme behavioral expressions of prejudice by demonstrating the crucial role of moral values and perceived moral violations. The strong association between binding moral values and the perceived justification of EBEPs, especially when violations of binding values are concerned, suggests that interventions need to address these values explicitly. Future research could investigate the generalizability of these findings to different cultural contexts and explore the potential of interventions that target moral foundations to reduce prejudice and violence. Furthermore, future research should explore how to best design interventions to address these underlying moral frameworks and to prevent them from escalating into extreme behaviors.

While the study employed a robust methodology, some limitations should be considered. The reliance on self-reported data in the surveys introduces the potential for response bias and social desirability effects. The samples, although carefully stratified, may not fully represent the diversity of the US population, potentially limiting the generalizability of some findings. The SPLC data used in Study 1, while extensive, may not be entirely complete or unbiased, although the use of MrsP reduces the effects of this limitation. The use of state-fixed effects in Study 1 caused the statistical models to not converge and thus limited the ability of this study to draw stronger conclusions.