Terrorist threat, dehumanization, and right-wing authoritarianism as predictors of discrimination.

This study analyzed the relationship between terrorist threat and discrimination, operationalized by support for retributive justice against Islamic groups suspect of terrorist crimes. Two experimental studies were performed. Study 1 (N = 215) showed that the terrorist threat against the ingroup raises the support for the retributive procedures through the dehumanization of the outgroup. Study 2 (N = 304) analyzed how the mediating role of dehumanization in the relationship between terrorist threat and support for retributive justice is moderated by right‐wing authoritarianism (RWA). In addition, the study aimed to verify if the dehumanization of outgroup and RWA could explain the relationship between terrorist threat and discrimination of Muslim immigrants. The results indicated that adherence to RWA favors dehumanization of the outgroup and, consecutively, the discrimination, operationalized as support for the use of retributive justice. The adherence to RWA has been identified as the mechanism that explains the discrimination against Muslim immigrants.

Keywords: Political psychology; radicalization; extremism; terrorism

Coexistence between native and foreign‐born communities is one of the most relevant social issues of our time. Several studies indicate that this relationship is marked by the expression of negative attitudes and discriminatory behaviors against immigrant groups (e.g., Álvaro et al., [5]; Cea D'Ancona & Valles, [15]). Particularly in recent years, explicit manifestations of hostility against immigrants from Muslim‐majority countries have been increasing (Ekman, [30]; Ward, Szabo & Stuart, [75]). Data from the European Social Survey (Heath & Richards, [45]) indicate that the number of people who perceive those immigrants as responsible for the increase in criminality and insecurity and who are in favor of banning their entry into Europe increased from 2002 to 2014.

One of the factors associated with the escalation of blatant discriminatory behavior against social minorities in Europe is terrorist threat perception. According to the Pew Research Center ([58]), the percentage of people who express concern with attacks motivated by Islamic terrorism has been increasing since 2011, especially in countries with a significant proportion of Muslim immigrants, such as France, Belgium, Spain, and the United Kingdom. Surveys have repeatedly identified that the occurrence of Islamic terrorist attacks is followed by expressions of hostility and discriminatory behaviors against immigrants from Muslim‐majority countries (e.g., Doosje, Zimmermann, Küpper, Zick & Meertens, [23]; Echebarria‐Echabe & Fernández‐Guede, [29]; Heath & Richards, [45]; Imhoff & Recker, [46]; Jost, Napier, Thorisdottir, Gosling, Palfai & Ostafin, [48]).

From the perspective of intergroup relations, discrimination is the negative treatment of a given group or its members (Dovidio, Hewstone, Glick & Mess, [24]). Discrimination may occur at the individual or the institutional level. The former is observed both in avoidance behavior and in explicit hostility against members of a given group (Allport, [2]). While discrimination at the institutional level is characterized by the support for norms, procedures, and actions of institutions that result in unfavorable treatment against certain groups (Sidanius & Pratto, [63]). The effects of threat perception on discrimination have been extensively studied in social psychology (e.g., Sherif, [62]; Stephan, Renfro, Esses, Stephan & Martin, [65]; Stephan & Stephan, [66]). Perceiving an outgroup as a possible threat to the physical integrity of ingroup members is associated with increased expression of negative attitudes and discriminatory behaviors (Esses, Jackson & Bennett‐AbuAyyash, [31]; Stephan et al., [65]). The perception that the West is under the constant threat of radical Islamism has supported the strengthening of extreme right political parties, such as the Front National in France, the Freiheitliche Partei Österrecihs in Austria, and the Partij voor de Vrijheid (PVV) in Holland, which share nationalistic and openly xenophobic views (Ekman, [30]; Greven, [37]).

In this context, wherein terrorist threat is associated with increased hostility and blatantly discriminatory practices against Muslim immigrants, the overall objective of the present research is to analyze discrimination against Islamic terrorist suspects operationalized as the support for retributive justice strategies to the detriment of restorative strategies. The former is characterized by emphasis on punishing those who break social rules, based on the principle that people should receive a punishment commensurate with the gravity of the crime they committed (Gromet & Darley, [38]; Tyler, Boeckmann, Smith & Hou, [69]), whereas the restorative strategy is focused on repairing the damage caused by the offense, involving both the transgressor and those affected by the crime (victims and community members) in combined efforts to resolve the transgression. In contrast to retributive justice, restorative justice operates through the offender completing reparative sanctions (Braithwaite, [13]; Gromet & Darley, [38]). The choice of retributive justice as a measure of discrimination stems from the fact that previous investigations have verified that support for punitive mechanisms in the face of a transgression can be explained by intergroup factors. Research based on Social Identity Theory (Tajfel & Turner, [67]) has determined that support for the use of both retributive and restorative measures varies based on whether social identity is shared with the offender. The use of retributive justice (for example, punishment of the aggressor) is preferred over restorative justice (for example, rehabilitation and social reintegration of the transgressor) when the criminal action is committed by an outgroup member (Sommers & Ellsworth, [64]; Vidmar, [71]). In intergroup conflict scenarios, the preference for retributive mechanisms against outgroup members has practical and symbolic implications. As pointed out by Wenzel, Okimoto, Feather and Platow ([76]), retributive justice, since it consists of a unilateral imposition of punishment, favors the maintenance of differences of status and power between outgroup and ingroup.

The group membership of the offender, therefore, influences the support for different forms of justice, and there is greater support for the use of retributive procedures when the offender belongs to a minority group (Álvaro et al., [4]; da Costa Silva, Torres, Álvaro, Garrido, & Linhares, [20]). Nevertheless, the present research adds another psychosocial factor: the dehumanization of the transgressor. The concept of dehumanization may be defined as the denial of qualities or traits that people perceive as solely human or central to human nature (Haslam, [40]). From the perspective of intergroup relations, studies converge in characterizing dehumanization as a perceptive process in which qualities considered essentially human are attributed to a lesser degree to outgroup members than to ingroup members (Betancor, Rodríguez‐Pérez, Delgado & Ariño, [12]; Leyens, Paladino, Rodriguez‐Torres et al., [50]). Previous investigations point out that dehumanization is the mechanism underlying the relationship between social group and support for justice procedures, legitimizing the use of violent measures against social minorities (e.g., Goff, Jackson, DiLeone, Cullota & DiTomasso, [35]; Kteily, Bruneau, Waytz, & Cotterrill, [49]).

Although the term dehumanization contradicts the precept, widely shared in democratic societies, that all humans are equal (e.g., Universal Declaration of Human Rights, [70]), various groups have at times been assessed as "less human," which shows that the perceptions of humanity in social relations vary widely. Nazi propaganda described the Jews as "pests" (Goff, Eberhardt, Williams & Jackson, [34]). Soccer fans verbally assaulted a black player by associating him with a monkey (Ferreira, Leite, Muniz, Batista, Torres & Álvaro, [32]). Then‐presidential candidate Donald Trump labeled refugees from Muslim‐majority countries "animals" (Rawstory, [59]).

Studies on dehumanization derive from research initiated by Leyens et al. ([50]) on a specific type of denial of essentially human attributes, termed infrahumanization. The key aspect of this line of research is that secondary emotions (e.g., shame, hope), which are attributes unique to humans, are more easily associated with ingroup members than with outgroup members, whereas this difference is not observed in the attribution of primary emotions (e.g., anger, joy), which are common to animals and humans. Although denying secondary emotions is a subtle process of dehumanization, it has been systematically assessed as a characteristic of relations between social groups (Delgado, Betancor, Rodríguez‐Perez & Ariño, [21]; Rodríguez‐Pérez, Betancor & Delgado, [61]).

Dehumanization has numerous negative effects on intergroup relations. Denying secondary emotions to outgroup members is related to the perception of outgroup threat (Kteily et al., [49]), to failure to help in emergency situations (Cuddy, Rock & Norton, [18]), and to discrimination against immigrants (Pereira, Vala & Leyens, [56]).

Several studies have extended the dehumanization effect beyond the attribution of secondary emotions, focusing on uniquely human characteristics in general (Haslam & Loughnam, [42]). Haslam ([40]) identifies two types of dehumanization: animalistic dehumanization, which is the denial of uniquely human attributes (e.g., refinement and moral sensibility), and mechanistic dehumanization, which is the denial of human nature (e.g., interpersonal warmth and cognitive openness). Animalistic dehumanization resembles the attribution of more uniquely human emotions (secondary emotions) to the ingroup versus the outgroup (Leyens et al., [50]).

Haslam's dualistic model stimulated the development of measurement instruments where dehumanization is verified through the attribution of uniquely human qualities and qualities of human nature to particular social groups (Bastian & Haslam, [10]; Haslam & Bain, [41]; Viki, Winchester, Titshall, Chisango, Pina & Russell, [74]). When taken together, contemporary research on dehumanization suggests that privileging the "humanity" of one's own group is a common occurrence. Furthermore, dehumanization as denial of uniquely human attributes may explain the support for using retributive strategies (e.g., severe punishment, violence) over restorative strategies (e.g., offender rehabilitation) in the case of an offense committed by an outgroup member. Denying uniquely human attributes is related to supporting their social exclusion and the use of violence (Bastian, Denson & Haslam, [9]; Goff et al., [35]; Viki, Fullerton, Raggett, Tait & Wiltshire, [72]), agreeing with the use of torture techniques (Viki, Osgood & Phillips, [73]), and supporting the use of military interventions (Kteily et al., [49]).

In general, dehumanization favors the use of retributive strategies because the dehumanized individual is perceived as incapable of assimilating the moral principles shared by the community and therefore unworthy of civilized treatment (Opotow, [55]). In societies in which using aggression is prohibited or unacceptable, dehumanization is a mechanism that allows justifying the use of violence against those who practice it as legitimate conduct, without morally condemning such conduct (Bandura, [8]; Harris & Fiske, [39]; Jackson & Gaertner, [47]).

In summary, these studies show dehumanization as a recurrent element in the relationship between dominant groups and minority groups. As a phenomenon of intergroup relations, perceiving outgroup members as lacking essentially human attributes is not a stable and uniform phenomenon. For example, in highly threatening situations, such as terrorist crimes, the attribution of exclusively human emotions to outgroup members is minimal (Goldenberg, Heflick, Vaes, Motyl & Greenberg, [36]), especially, when they are described as perpetrators of these crimes (Betancor et al., [12]; Delgado, Rodríguez‐Pérez, Vaes, Leyens, & Betancor, [22]).

The investigations carried out to date demonstrate that members of minority groups are recurrently perceived as devoid of essentially human characteristics and that such a perception legitimizes the use of violent behavior against them. In addition, it is known that dehumanization can be accentuated by associating outgroup members with responsibility for situations of threat and insecurity, for instance terrorist attacks.

However, it has not yet been analyzed whether dehumanization of the outgroup and the consequent support for the use of retributive measures may vary depending on whether the ingroup is highlighted as the focus of this threat.

Thus, Study 1 was carried out to verify if, in a terrorist threat scenario where a member of the outgroup (Islamic terrorist) is presented as a suspect in the criminal action, the level of threat against the ingroup (Spanish) could influence the dehumanization of the outgroup and, consequently, discrimination against it, operationalized by support for the use of retributive justice procedures rather than restorative ones. In this study, the model of dehumanization based on the denial of secondary emotions was adopted, since this phenomenon is recurrently verified in intergroup conflicts, in general, and specifically in contexts of terrorist threat (Betancor et al., [12], Delgado et al., [22]; Goldenberg et al., [36]), the scenario adopted in the present research.

Because the perception of threat to ingroup physical integrity increases the expression of discriminatory behaviors (Esses et al., [31]; Stephan et al., [65]), in Study 1 it was expected that in the scenario where the ingroup is the focus of terrorist threats (high threat) by the outgroup, greater support for retributive justice would be observed, with this relationship being explained by the dehumanization of the outgroup (Hypothesis 1a). On the other hand, it was assumed that in a scenario in which the ingroup (Spanish) is not represented as the focus of the terrorist threat (low threat), there would be less support for the use of retributive measures, due to the lower degree of dehumanization of the outgroup in that condition (Hypothesis 1b). Figure represents the theoretical model adopted in this study.

This study included a total of 215 Spanish subjects (53.95% women), members of the general population, recruited by "snowball" sampling with ages ranging from 18 to 58 years (M = 36.97; SD = 13.4). The recruitment of the participants by Sociology students of a Public University in Madrid, who were previously trained on the research objectives and the APA standards regarding research with human beings. The recruitment had as its sole criterion not to select university students, since this type of sample is characterized by low levels of explicit discrimination (Álvaro et al., [4]), making it necessary to investigate in other population segments.

All participants were randomly distributed in a 2 (terrorist threat: high threat against ingroup vs. low threat against ingroup) × 2 (type of emotions attributed to terrorists: primary emotions vs. secondary emotions) design. The first variable relied on inter‐subject design, whereas the second was based on intra‐subject design.

Online questionnaires were administered using the SurveyMonkey software. This software enables the online administration of questionnaires and the randomization of participants in different experimental conditions as well. Each participant answered the questionnaire individually after signing the informed consent form for research.

Experimental manipulation: terrorist threat against ingroup: high threat vs. low threat) was experimentally manipulated by presenting a real news story, previously published on an Internet portal and adapted for the purpose of this research. The news item reported on terrorist actions of Islamic State (ISIS) in Syria, which stood out for its political influence in the Middle East and for claiming responsibility for several terrorist acts during the data collection period (December 2016). The report also noted that ISIS is considered by specialists to be one of the most dangerous groups in the world and is responsible for kidnappings, civilian murders, and torture.

The terrorist threat was presented on two levels. In the condition in which the ingroup was the focus of terrorist threat (high threat), information that Spain was under the threat of radical Islamism and that a suspect linked to ISIS propaganda and indoctrination had been detained in Spain was included in the news item. It reported that this suspect was a member of the terrorist network responsible for the 11‐M attacks, which caused approximately 200 deaths and hundreds of wounded in Madrid (2004). In turn, the condition in which the terrorist threat was not directed toward the ingroup (low threat) excluded this information, focusing only on the terrorist actions of ISIS in Syria.

Dehumanization: a measurement tool based on the instrument developed by Rodríguez, Rodríguez‐Pérez, Vaes, Rodríguez and Leyens ([60]) and consisting of 16 emotions, including eight secondary or exclusively human emotions (4 positive secondary emotions: love, hope, optimism and fascination; and 4 negative secondary emotions: bitterness, unhappiness, worry and shame) and eight primary emotions or emotions shared by humans and non‐human animals (4 positive primary emotions: joy, surprise, trust, anticipation; and 4 negative primary emotions: anger, disgust, fear, sadness) was used. The participants were asked to indicate how much they thought people who belong to terrorist groups such as ISIS experience each of those emotions using a Likert‐type scale (1 = not at all to 5 = fully).

Retributive Justice: participants answered a 10‐item, single‐factor scale regarding their attitudes toward the rehabilitation of suspects of Islamic terrorism (The use of police violence against suspects to prepare jihadist terrorism actions is, in certain cases, justified; The government should invest in programs of rehabilitation and social reintegration of Islamic terrorists (inverse); The punishment of terrorist actions is more urgent than its prevention). This measurement tool was adapted from Cullen, Lutz, Link and Wolfe ([19]). The participants answered the questionnaire using a seven‐point Likert scale (from 1 = totally disagree to 7 = totally agree). Higher scores on the scale indicated greater support for Islamic terrorist rehabilitation (α = 0.872).

To analyze how the dehumanization of terrorists varies as a function of terrorist threat against ingroup, 2 (high threat vs. low threat) × 2 (type of emotions: primary vs. secondary emotions) repeated measures analysis of variance (ANOVA) was performed; threat perception was an inter‐subject variable, and the emotions attributed to terrorists were intra‐subject variables.

The hypothesis that the terrorist threat against ingroup (X) influences support for the use of retributive justice against the outgroup (Y) through the dehumanization of the Islamic terrorist suspect (M) was tested using the PROCESS software for SPSS (Hayes, [43]). The bootstrap CI analysis method (5,000 samples) was adopted, as recommended by Hayes and Scharkow ([44]), because it is a test that presents better reliability than the rest.

The analysis of variance indicated a main effect of the variable terrorist threat against the ingroup (high threat × low threat) on the attribution of primary and secondary emotions to the Islamic terrorism suspect, F(1, 213) = 8.757, p = 0.003, ŋ² = 0.039. As can be seen in Fig. , fewer secondary emotions were significantly attributed to suspects in the context in which the ingroup was perceived as the target of terrorist threat (high threat; M = 2.75, SD = .526) than those attributed when the terrorist threat was perceived as unrelated to the ingroup (low threat; M = 3.11, SD = 0.712), t (204.63) = − 5.102, p < 0.001. However, no significant differences in attribution of primary emotions to suspects as a function of terrorist threat against the ingroup were found, t (213) = −1.751, p = 0.081.

Initially, the terrorist threat against the ingroup was dummy coded as 0 = low threat and 1 = high threat. The analyses indicate that the terrorist threat against the ingroup can predict support for the use of retributive justice against the suspected Islamic terrorist, b = 0.23, t (194) = 2.49, p = 0.01, 95%, CI = 0.05, 0.42. As can be seen in Fig. , the high threat condition, in contrast to the low threat condition, significantly decreases the attribution of secondary emotions to the suspected Islamic terrorist, b = −.27, t (190) = − 6.4, p < 0.001, 95% CI = −3.64, −0.19. In turn, the attribution of secondary emotions to the suspected Islamic terrorist predicts support for the use of retributive justice in a negative way, b = −60, t (194) = −3.99, p < 0.001, 95% CI = −0.90, −0.30.

Upon analyzing the direct and indirect effects of the perceived terrorist threat with respect to retributive justice, it was found that the terrorist threat against the ingroup did not have a direct effect on support for retributive justice, b = 0.07, t (194) = 0.64, p = 0.51, 95% CI: lower = −0.14, upper = 0.28. However, the indirect effect, through the attribution of secondary emotions to the suspected Islamic terrorist, was 0.16, being significant (95% CI: lower = 0.08, upper = 0.27).

In this study, we sought to verify if, in a terrorist threat scenario in which a member of the outgroup (Islamic terrorist) is presented as a suspect in the criminal action, the level of threat against the ingroup (Spanish) could influence the dehumanization of the outgroup and, therefore, discrimination against them, operationalized by the support for the use of retributive justice procedures rather than restorative ones.

The results indicate that terrorist threat against ingroup influences the expression of discriminatory behavior, here assessed by the support for retributive justice strategies. In the condition in which the ingroup was perceived as a possible target of terrorist actions (high threat), participants showed increased support for retributive justice.

The relationship between terrorist threat against ingroup and discrimination was mediated by dehumanization of suspects of Islamic terrorism. When the ingroup was perceived as a possible target of terrorist actions (high threat), attribution of secondary emotions to suspects decreased, which, in turn, implied increased support for retributive justice against this group, thus confirming Hypothesis 1a. Conversely, when the ingroup (Spanish) was not placed as the focus of the terrorist threat (low threat), there was less support for retributive justice, since this condition implied a greater attribution of secondary emotions to the outgroup member, confirming Hypothesis 1b.

In summary, Study 1 concurs with previous investigations indicating that the association of the outgroup with the terrorist threat raises the discrimination against them (Doosje et al., [23]; Echebarria‐Echabe & Fernández‐Guede, [29]; Jost et al., [48]). At the same time, the study corroborates the fact that dehumanization of the outgroup is the mechanism that explains the support for discriminatory behavior (Bastian et al., [9]; Kteily et al., [49]). The present investigation, however, offers an additional contribution by identifying that dehumanization of the outgroup and the consequent support for retributive justice mechanisms are accentuated by the ingroup being the focus of the threat perpetrated by the outgroup.

Although the present study has identified that the context in which the ingroup is the focus of a threat favors the expression of discriminatory behavior, previous investigations are consistent in pointing out that reactions to a threatening intergroup context are not linear processes, but moderated by different psychosocial variables (Álvaro et al., [4]; Baysu, Phalet & Brown, [11]; Matthews & Levin, [53]). Among these variables, we highlight right‐wing authoritarianism (RWA), which interacts with the perception of threat in the expression of negative behaviors against the outgroup (Cohrs & Asbrock, [16]; da Costa Silva et al., [20]; Duckitt, [25]). Consequently, Study 2 was conducted to identify whether adherence to RWA moderates the effect of perceived threat against the ingroup on the support for retributive justice, through dehumanization of the outgroup.

RWA has been repeatedly identified as an important predictor of social attitudes and behaviors (Duckitt & Sibley, [27]). Although it was initially defined as a personality measure, derived from authoritarian personality studies (Adorno, Frenkel‐Brunswick, Levinson & Sanford, [1]), the RWA concept is closer to a measure of social values and attitudes than a personality trait (Duckitt, Bizumic, Krauss & Heled, [26]). Such as defined by Altemeyer ([3]), RWA is a set of social attitudes that encompasses adherence to ingroup norms and social conventions, submission to authority, and support for the use of aggression when sanctioned by authorities. Its relevance in psychosocial investigation is explained by the ability to connect different levels of explanation. At the societal level, studies demonstrate that adherence to authoritarian attitudes is prevalent during periods of heightened political and economic instability (Asbrock & Fritsche, [6]; Peterson & Gerstein, [57]).

RWA expresses a desire to maintain order, stability, and social cohesion (Duckitt & Sibley, [27]). At the intergroup level, it is seen that a higher level of adherence to RWA influences the expression of prejudice and aggressive behaviors against groups perceived as a threat to ingroup security and its core values (Asbrock, Sibley & Duckitt, [7]; Cohrs & Ibler, [17]; Thomsen, Green & Sidanius, [68]). In a context of heightened conflict and intergroup threat, such as the crime of terrorism, RWA was identified as an important predictor of attitudes favoring torture to defend ingroup members (Lindén, Björklund & Bäckström, [51], [52]). Adherence to RWA is associated not only with the perception that minority groups represent a threat, but also with the perception that they are devoid of essentially human characteristics (Jackson & Gaertner, [47]). Because RWA is based on the defense of key ingroup values and on a view of the world as a threatening place, adherence to its attitudinal components leads to the perception of groups whose norms and traditions differ from those of the ingroup, and who are identified as a threat to security and social order, as not only different but also as less human (Kteily et al., [49]).

The investigations carried out thus far present consistent evidence that RWA favors the expression of discriminatory behaviors, especially against culturally distinct groups of the outgroup, identifying them as threatening. At the same time, adherence to RWA is known to be associated with the dehumanization of minority groups. However, it has not yet been analyzed how RWA and the threat against the ingroup interact in the perception that the minority group is devoid of essentially human characteristics.

Considering that Study 1 identified that, in a terrorist threat scenario, the high threat against the ingroup (Spanish) increases the support for retributive justice through dehumanization of the outgroup (suspected of Islamic terrorism), Study 2 sought to analyze how this relationship is qualified by RWA. More specifically, this study replicated the previous one, adding RWA as a moderating variable, with the objective of identifying whether RWA adherence conditions the effect of perceived threat against the ingroup on support for retributive justice, through dehumanization of the outgroup. Figure  represents the theoretical model adopted in this study.

Since adherence to RWA supports the perception that the outgroup poses a threat (Cohrs & Ibler, [17]) and RWA is associated with dehumanization of the outgroup (Jackson & Gaertner, [47]; Kteily et al., [49]) it was expected that participants presenting strong adherence to RWA (+1SD), would dehumanize the outgroup (suspected of Islamic terrorism), both in the condition where the ingroup (Spanish) is the focus of terrorist threats (high threat) and in the condition where it is not identified as the target of these crimes (low threat). Consequently, among these participants no differences in support for retributive justice would be observed when comparing the high threat and low threat conditions (Hypothesis 1a).

On the other hand, it was expected that among the participants with low adherence to RWA (–1SD), dehumanization of the outgroup (suspected of Islamic terrorism) would be greater in the condition where the ingroup (Spanish) is the focus of the terrorist threat (high threat), than in the condition where the ingroup is not mentioned as a target (low threat). Therefore, the high threat condition would result in greater support for retributive justice than the low threat condition, replicating the result found in Study 1 (Hypothesis 1b).

Previous studies identified that the terrorist threat is also associated with the expression of discriminatory conduct against Muslim immigrants (Doosje et al., [23]; Echebarria‐Echabe & Fernández‐Guede, [29]; Imhoff & Recker, [46]; Jost et al., [48]). Therefore, the present study sought to broaden the contributions provided in the literature in analyzing whether the theoretical model adopted to explain support for retributive justice procedures against suspects of Islamic terrorism would explain the relationship between terrorist threat and discriminatory conduct against Muslim immigrants. More precisely, this study had an additional objective to verify whether the terrorist threat against the ingroup would influence institutional discrimination against Muslim immigrants, operationalized through support for increased police surveillance of this specific group of immigrants as a whole. Concomitantly, this study examined whether the dehumanization of the suspected Islamic terrorist would be the underlying mechanism in the relationship between terrorist threat against ingroup and discrimination, as well as whether this relationship is also qualified by RWA.

This study included 304 Spanish subjects (50.98% women), members of the general population, recruited by "snowball" sampling, following the same procedures and criteria adopted in Study 1, with ages ranging from 18 to 60 years (M = 36.19; SD = 14.6).

Study 2 followed the procedures and instruments adopted in Study 1, in addition to the following measures:

Discrimination against Muslim immigrants: using a seven‐point Likert scale (from 1 = totally agree to 7 = totally disagree), the participants were asked to indicate their degree of agreement with the following statement: police should maintain increased surveillance of immigrants from Muslim‐majority countries.

Right‐wing authoritarianism (RWA): a single‐factor scale based on the instrument developed and validated by Cárdenas and Parra ([14]) was adopted. The scale consists of 10 items evaluated based on a seven‐point Likert scale (from 1 = totally disagree to 7 = totally agree), representing beliefs compatible with RWA (our society needs strong leaders who can eradicate the prevailing extremism; the best way to live is to follow traditions and traditional values; society must stop the many radicals who try to spoil our coexistence). The scale had a satisfactory reliability index (α = 0.738). High scores on this scale indicate high adherence to RWA.

The dehumanization of terrorists was analyzed by 2 (terrorist threat against ingroup: ingroup perceived as the target of terrorist threat – high threat vs. terrorist threat unrelated to the ingroup – low threat) × 2 (Type of emotions attributed to suspects of terrorism: primary emotions vs. secondary emotions) repeated measures analysis of variance (ANOVA); terrorist threat against ingroup was an inter‐subject variable, and emotions attributed to suspects of terrorism were intra‐subject variables.

To test the hypotheses, a moderated mediation model was tested, taking the terrorist threat against the ingroup (high threat × low threat) as a predictive variable (X), the attribution of secondary emotions to the Islamic terrorism suspect as a mediating variable (M), support for retributive justice and discrimination against Muslim immigrants as dependent variables (Y), and finally, RWA as a moderator (W) of the relationship between terrorist threat against the ingroup and attribution of feelings to the suspected Islamic terrorist. For this, the PROCESS software for SPSS (Hayes, [43]) was used. As in Study 1, the bootstrap CI method (5,000 samples) was adopted to evaluate the direct and indirect effects (via mediator).

The main effect of the variable, terrorist threat against the ingroup, on attribution of primary and secondary emotions to suspects of Islamic terrorism was assessed, F (1, 302) = 18.89, p < 0.001, ŋ² = 0.062. Fewer secondary emotions were attributed to suspects of terrorism in the context in which the ingroup was perceived as the target of a terrorist threat (high threat) (M = 2.86, SD = 0.65) than in the context in which the ingroup was not the target (low threat) (M = 3.28, SD = .67), t (302) = −5.44, p < 0.001. As can be seen in Fig. , no significant differences in the attribution of primary emotions to terrorists as a function of terrorist threat against the ingroup were identified, t (285.7) = −1.69, p = 0.09.

Pearson product‐moment correlation analysis was used to determine the relationships between variables (Table ).

Descriptive statistics and Pearson product moment correlation coefficient values for the study variables

• 2 Note
• 3 p < .001.

The values outlined in Table  indicate the occurrence of a significant and negative correlation between adherence to RWA and the attribution of secondary emotions to terrorism suspects. Significant and negative relationships were observed between attribution of secondary emotions, support for retributive justice, and discrimination against Muslim immigrants. However, RWA showed significant and positive relationships with support for retributive justice and with discrimination against Muslim immigrants.

In line with the initial predictions, support for retributive justice against the outgroup member was significantly higher in the high threat condition ( $\widehat{\mathrm{Y}}$ c = 3.97) than in the low threat condition ( $\widehat{\mathrm{Y}}$ c = 3.41), b = .27, t (298) = 2.06, p = 0.039, 95% CI = 0.01, 0.54. Support for retributive justice was also explained by the lower attribution of secondary emotions to the suspected Islamic terrorist, b = −0.21, t (298) = −2.11, p = 0.03, 95% CI = −0.40, −0.01. In addition, attribution of secondary emotions to the suspected Islamic terrorist was significantly influenced by the terrorist threat against the ingroup, such that the high threat condition resulted in a lower attribution of secondary emotions to the suspected Islamic terrorist ( $\widehat{\mathrm{Y}}$ a = 2.82) compared to the low threat condition ( $\widehat{\mathrm{Y}}$ a = 3.28), b = −0.40, t (298) = −6.08, p < 0.001, 95% CI = −0.54, −0.27.

More relevant to our hypotheses, the effect of the saliency of the terrorist threat against the ingroup on the attribution of secondary emotions to the suspected Islamic terrorist was moderated by adherence to RWA, b = 0.19, t (198) = 2.28, p = 0.02, 95% CI = 0.02, 0.36, resulting in a significant indirect conditional effect only in low adherence to RWA (–1SD, 95% CI = 0.011, 0.30), since the indirect effect did not occur in high adherence to RWA (+1SD, 95% CI = −0.01, 0.14). The indirect conditional effect of RWA is explained by the fact that in low adherence to RWA (–1SD), the result was a lower attribution of exclusively human emotions to the Islamic terrorism suspect in the high threat condition ( $\widehat{\mathrm{Y}}$ a = 2.94) compared to the low threat condition ( $\widehat{\mathrm{Y}}$ a = 3.51), b = 0.09, t (300) = 2.28, p < 0.001, 95% CI = 0.01, 0.18, confirming Hypothesis 1b. These differences were also observed in the condition of high adherence to RWA (+1SD), but to a less significant degree, since there is a low attribution of secondary emotions in both the high threat condition ( $\widehat{\mathrm{Y}}$ a = 2.73) and the low threat condition ( $\widehat{\mathrm{Y}}$ a = 2.98), b = −1.22, t (300) = –2.7, p = 0.007, 95% CI = −0.21, −0.03, partially confirming Hypothesis 1a.

In addition, it was tested whether RWA also moderated the relationship between attribution of secondary emotions to the Islamic terrorism suspect and support for retributive justice. No significant relationship was found in this regard, b = 0.04, t (298) = 0.45, p = 0.68, 95% CI = −0.16, 0.24. In addition, it was found that RWA also does not moderate the relationship between terrorist threat against the ingroup and support for retributive justice, b = −0.125, t (298) = −0.74, 95% CI = −0.45, 0.20.

The proposed theoretical model did not prove useful in explaining discrimination against Muslim immigrants, since the relationship between terrorist threat against the ingroup and discrimination against Muslim immigrants was not mediated by the attribution of secondary emotions to the suspected Islamic terrorist, whether in the condition of low adherence to RWA (–1SD, 95% CI = −0.24, 0.26) or in the condition of high adherence to RWA (+1SD, 95% CI = −0.9, 0.12). However, it was found that RWA moderated the relationship between terrorist threat against the ingroup and discrimination against Muslim immigrants, b = −0.20, t (300) = −2.09, p = 0.03, 95% CI = −0.39, −0.01. This effect is explained by the fact that participants with low adherence to RWA (–1SD) manifested higher discrimination against Muslim immigrants in the high threat condition ( $\widehat{\mathrm{Y}}$  = 3.85) than in the low threat condition ( $\widehat{\mathrm{Y}}$  = 3.14), b = 0.35, t (300) = 2.67, p = 0.007, 95% CI = 0.09, 0.61. While participants with high adherence to RWA (+1SD) manifested discrimination against Muslim immigrants in both the high threat condition ( $\widehat{\mathrm{Y}}$  = 5.0) and the low threat condition ( $\widehat{\mathrm{Y}}$  = 4.9), b = 0.01, t (300) = 0.15, p = 0.87, 95% CI = −0.18, 0.21.

The present study aimed to identify whether adherence to RWA conditions the effect of perceived threat against the ingroup on support for retributive justice, through dehumanization of the outgroup. The results confirm the observations of Study 1. It was verified that it is not the simple association of the minority group with a criminal act that explains the discrimination, analyzed here by the support for retributive justice procedures, but rather the fact that the ingroup is identified as the target of this threat. Again, it has been identified that dehumanization of the outgroup is the mechanism underlying this process. Compared to the condition of low threat against the ingroup, the high threat condition resulted in less attribution of secondary emotions to the suspected Islamic terrorist, which therefore increased support for retributive forms of justice.

Study 2, in addition, broadens the contribution of the former by demonstrating that the relationship between terrorist threat against the ingroup and dehumanization of the outgroup is qualified by adherence to RWA. Low adherence to RWA accentuates the differences between the conditions of high and low threat against the ingroup in the attribution of secondary emotions to the outgroup, since the condition that the ingroup is the focus of the terrorist threat is a necessary for the dehumanization of the outgroup, and the consequent support for the use of retributive procedures against it, confirming Hypothesis 1b. These differences in the attribution of secondary emotions to the outgroup were also observed among participants with high adherence to RWA, although in a less significant way, which partially confirms Hypothesis 1a.

Although previous studies have demonstrated that RWA is associated with the perception that the outgroup poses a threat (Cohrs & Ibler, [17]), this study offers an additional contribution by demonstrating that, in a context of heightened intergroup threat such as the crime of terrorism, adherence to RWA supports the perception that the outgroup is devoid of essentially human characteristics, without the need for the ingroup to be perceived as the focus of the terrorist threat for dehumanization to occur.

The results indicated, however, that RWA did not moderate the relationship between terrorist threat against the ingroup and discrimination, much less the relationship between the attribution of secondary emotions to the Islamic terrorism suspect and discrimination. It is possible that such effects did not occur because discrimination was indirectly measured by means of support for retributive justice processes, whereas RWA is commonly associated with flagrant and ostensible forms of discrimination (Duriez & Van Hiel, [28]; Fiske, [33]), an aspect that can be explored in future investigations.

This study also had the additional objective of verifying whether the proposed model adopted to explain the support for retributive justice procedures against suspected Islamic terrorists would explain the relationship between terrorist threat and discriminatory conduct against Muslim immigrants, operationalized in the support for increased police surveillance of this specific group of immigrants as a whole. Although the dehumanization of the Islamic terrorism suspect did not explain the relationship between terrorist threat against the ingroup and discrimination against Muslim immigrants, it was observed that adherence to RWA moderated the relationship between perceived terrorist threat against the ingroup and discrimination against Muslim immigrants. High adherence to RWA favored discrimination against Muslim immigrants in both the high threat and low threat contexts, demonstrating that adherence to RWA and the scenario of Islamic terrorism translate into a process of homogenization and social exclusion of the outgroup.

The overall objective of the present research was to analyze the discrimination against Islamic terrorist suspects operationalized as the support for retributive justice strategies to the detriment of the restorative strategies. Given the relevance of terrorism based on Islamic fundamentalism to the development of negative attitudes against immigrants (Doosje et al., [23]; Heath & Richards, [45]; Imhoff & Recker, [46]; Jost et al., [48]), this study further aimed to assess whether terror threat against ingroup also influences institutional discrimination against Muslim immigrants.

Because the expression of discriminatory behavior against outgroup members is influenced by the perception that they threaten the security and integrity of ingroup members (Esses et al., [31]; Stephan et al., [65]), this research experimentally manipulated terrorist threat by creating a condition highlighting the ingroup as the target of Islamic terrorist threats (high threat) and by creating another condition in which the ingroup was not mentioned (low threat).

The results of two studies with an experimental design showed that the terrorist threat to the ingroup influenced discrimination against suspects of Islamic terror through dehumanization. Study 1 showed that the relationship between terrorist threat against ingroup and discrimination was mediated by dehumanization of suspects of Islamic terrorism. In the context in which the ingroup is perceived as the target of a terrorist threat, dehumanization of Islamic terrorism suspects increased, which in turn resulted in increased support for retributive justice.

Study 2 replicated the previous study, adding RWA as a moderating variable in the relationship between terrorist threat against the ingroup and dehumanization of the suspected Islamic terrorist. In addition, the study added discrimination against Muslim immigrants as a dependent variable. The results confirmed those obtained in the previous study and, at the same time, indicated that high adherence to RWA facilitates dehumanization of the outgroup and consequent support for the use of retributive justice, regardless of whether the terrorist threat has the ingroup as the focus.

Furthermore, Study 2 showed that terrorist threat against the ingroup interacts with RWA in the expression of institutional discrimination against Muslim immigrants. In a context of Islamic terrorist threat, adherence to RWA raises the support for increased police surveillance of this group of immigrants regardless of whether the ingroup is identified as the focus of the terrorist threat.

The studies reported here contribute to the current literature by clarifying how dehumanization and RWA act as underlying processes in the relationship between terrorist threat against ingroup and discriminatory behavior. Previous studies based on Social Identity Theory (Tajfel & Turner, [67]) had already identified that the support for the use of retributive procedures increases when the criminal action is committed by an outgroup member (Álvaro et al., [4]; da Costa Silva et al., [20]; Sommers & Ellsworth, [64]; Wenzel et al., [76]). Moreover, these minority groups are perceived as devoid of essentially human characteristics (Betancor et al., [12]; Haslam, [40]; Leyens et al., [50]) and such dehumanization increases the support for severe punishments against these groups (Goff et al., [35]), including violence (Viki et al., [73]). Nevertheless, the present research expands the explanatory scope of previous studies by indicating that outgroup dehumanization and its consequent discrimination varies based on the degree of threat it represents to ingroup members. Therefore, the present study showed that the support for retributive justice strategies is a discriminatory process because it emerges in a context of threat and insecurity where the ingroup is perceived as the target of the criminal actions attributed to the outgroup.

The moderating effect of adherence to RWA in the relationship between perceived terror threat against the ingroup and discrimination against Islamic terrorists, found in Study 2, corroborates findings from studies indicating RWA as a variable associated with negative conduct against outgroups perceived as a threat to ingroup security (Asbrock et al., [7]; Cohrs & Asbrock, [16]; Lindén et al., [51], [52]; Thomsen et al., [68]) and with dehumanization of outgroups (Jackson & Gaertner, [47]; Kteily et al., [49]). However, the study broadens previous contributions about the role of RWA in intergroup relations by demonstrating that adherence to its attitudinal components means that perceiving the ingroup as the target of terrorist threats is not a necessary condition for dehumanizing the outgroup, and consequently, supporting retributive mechanisms against it.

Moreover, Study 2 adds evidence that adherence to RWA is a mechanism through which the occurrence of Islamic terrorism crimes results in discrimination against immigrants from Muslim‐majority countries. Such discriminatory behavior can be assessed in the current socio‐political context in which the strengthening of extreme right‐wing political parties and their anti‐immigrant agendas is accompanied by the growth of xenophobic and openly hostile attitudes towards immigrants from Muslim‐majority countries (e.g., Doosje et al., [23]; Ekman, [30]; Imhoff & Recker, [46]).

These effects can be explained because the conservative and authoritarian attitudes of right‐wing authoritarianism are based on the perception of the world as a dangerous place and express a motivation to defend the security of ingroup members and its core values and norms (Duckitt & Sibley, [27]), so adherence to RWA, in a terrorist threat context, promotes the expression of discriminatory behaviors against groups culturally different from the ingroup.

The present research has some limitations. First, the process of dehumanization analyzed here was infrahumanization, which is a more subtle and indirect form of dehumanization (Haslam & Loughnan, [42]). Later conceptualizations of dehumanization have expanded this model by including flagrant dehumanization measures in which specific outgroup members are explicitly assessed as animal‐like (Haslam, [40]; Viki et al., [72]) or as less evolved and civilized (Kteily et al., [49]). Therefore, the context of terror threat most likely triggers different mechanisms of dehumanization, in addition to that analyzed in the present study. Further research should analyze how subtle and flagrant forms of dehumanization explain different discriminatory and intergroup conflict processes.

Although the present research has examined the moderating role of RWA in the relationship between perceived terrorist threat and outgroup dehumanization, other investigations have pointed to the importance of social dominance orientation (SDO) for discrimination against minority groups (Asbrock et al., [7]; Duckitt & Sibley, [27]). Like RWA, SDO encompasses socio‐political attitudinal components that favor support for torture and other human rights violations in terrorist threat contexts (Lindén et al., 2016, 2018). Thus, future research could address the effect of SDO on the process of dehumanization of minority groups.

The mediating effect of dehumanization on the relationship between threat against ingroup and discrimination against immigrants from Muslim‐majority countries was assessed in the context of terrorist acts. Although this scenario is relevant because it involves the perception that the entire ingroup is a target of potential threats, evidence that other contexts of threat, such as economic (Meuleman, Davidov & Billiet, [54]) and the perception of a symbolic terror threat, are also associated with discrimination against minorities, has been published. Symbolic terror threat perception is characterized by the perception that specific outgroups jeopardize the integrity of the ingroup's core values and beliefs. Future studies should examine how dehumanization and RWA explain the relationship between different types of threat and discrimination.

Finally, considering that the context analyzed here involves a group‐level threat and that the degree of identification with the ingroup is related to a greater motivation to defend its status and to discriminate against the outgroup (Tajfel & Turner, [67]), future research must expand the contributions of this study by analyzing the role of identification with the ingroup in the relationship between terrorist threat and discriminatory behavior.

This research had the financial support of Capes/DGPU.