Differentiated Threat and the Genesis of Prejudice: Group-Specific Antecedents of Homonegativity, Islamophobia, Anti-Semitism, and Anti-Immigrant Attitudes

Differentiated Threat and the Genesis of Prejudice: Group-Specific Antecedents of Homonegativity,... Abstract In this article, we argue that an exclusive focus on the generalized aspect of prejudice limits understanding of the structure and genesis of prejudice towards particular outgroups. In order to conceptualize the specific nature of particular prejudices, we propose the differentiated threat approach. This framework postulates that different outgroups challenge diverse realistic and symbolic interests, and that these outgroup specific threats affect various socioeconomic strata and cultural groups differentially. The differentiated threat approach is applied to analyse majority-group Belgians’ attitudes towards immigrants, Muslims, Jews, and homosexuals. The results show that a common denominator of prejudice can be distinguished, but that the prejudices towards the various outgroups contain substantively relevant unique components that are influenced by socio-demographic and attitudinal predictors in diverging ways. Gender traditionalism is found to reinforce Homonegativity and temper Islamophobia at the same time. Feelings of relative deprivation are more strongly related to Islamophobia than to other forms of prejudice, and are unrelated to homonegativity. Religious involvement plays a more decisive role in the formation of anti-Semitism and Homonegativity than it does in the other forms of prejudice. Anti-immigration attitudes show a class gradient that is absent in attitudes towards other outgroups. Our results evidence that the concrete realization of attitudes towards a specific outgroup cannot be understood without paying attention to structural and contextual factors, such as social positions, the nature of intergroup relations, power balances, and elite discourses. generalized prejudice, anti-Semitism, homonegativity, Islamophobia, differentiated threat approach In early research on intergroup relations, scholars put forward the idea that negative feelings tend to generalize across various outgroups (Allport 1958; Adorno et al. 1950). Numerous studies have evidenced how this so-called “generalized prejudice” (Akrami, Ekehammar, and Bergh 2011; Bäckström, and Björklund 2007; Duckitt and Sibley 2007) or “group-focused enmity” (Zick et al. 2008) is related to personality traits and ideological orientations such as right-wing authoritarianism and social dominance orientation (Asbrock, Sibley, and Duckitt 2010; Ekehammar et al. 2004). The discovery of a common denominator in negative attitudes towards various outgroups has contributed to our knowledge of intergroup relations. In this article, however, we argue that the exclusive focus on the common denominator of prejudices tends to obscure the particular causes of prejudice towards specific outgroups. Our argument is that—while the formation of general negative outgroup attitudes takes place via more or less universal social-psychological mechanisms—the concrete realization of attitudes towards particular outgroups is crucially dependent on both structural and contextual factors (see Akrami et al. 2011). Social positions of ingroup individuals and their sense of group position, intergroup relations characterized by differential social contact, socioeconomic competition and power balances, as well as available media and elite discourses might explain why prejudice is triggered toward one outgroup but not towards others (Hagendoorn 1995). Although this is acknowledged in the literature (Agnew, Thompson, and Gaines 2000; Ray and Lovejoy 1986; Zick et al. 2008), a systematic theoretical elaboration and empirical test of the specificity of prejudices towards various outgroups is still lacking. In this article, we develop a theoretical framework—the differentiated threat approach—that conceptualizes how different outgroups may be prejudiced against for different reasons by different strata of the ingroup. Subsequently, we provide an empirical application of this approach by engaging with two generalizability questions: (1) To what extent do negative attitudes generalize across different outgroups? Can we ascertain generalized prejudice or do individuals hold differentiated views regarding various social outgroups? (2) Can a single general causal model explain negative attitudes towards various outgroups or are specific prejudices triggered by particular sets of predictors, as the differentiated threat approach predicts? These questions are answered by articulating them in the concrete intergroup context of Belgium, where attitudes of Belgian majority-group members towards four outgroups—namely immigrants, Muslims, Jews and sexual minorities—are scrutinized. THEORY AND HYPOTHESES Generalized Prejudice: Personality and the Common Denominator of Prejudices A dominant paradigm in social-psychological studies of intergroup relations stresses that the inclination to hold prejudices generalizes across outgroups. This idea of “generalized prejudice” was already present in Gordon Allport’s (1958) seminal study, where he noted that “people who reject one outgroup will tend to reject other outgroups. If a person is anti-Jewish, he is likely to be anti-Catholic, anti-Negro, anti any outgroup” (p. 66). More recently, Andreas Zick and colleagues (2008) introduced the concept of “group-focused enmity” (GFE) to refer to a generalized devaluation of outgroups with an ideology of inequality as its common core. Hostile attitudes toward different outgroups, who are considered to be unequal, socially threatening, and/or culturally deviant, are understood as the components of a single syndrome, namely GFE (see also Zick et al. 2008). Over the years, numerous studies have accumulated empirical evidence for this hypothesis. Prejudices towards very diverse social groups—including various ethnic and religious groups, but also women, sexual minorities, elderly people, disabled persons, and homeless people—are indeed found to correlate very strongly (Agnew et al. 2000; Altemeyer 1998; Bierly 1985; Bratt 2005; Ekehammar et al. 2004; Zick et al. 2008). The GFE approach not only implies a stable structure of interrelated prejudices, but also argues that the various components are triggered by a single set of antecedents. Explanations why individuals are so consistent in their acceptance or rejection of various outgroups are often sought in person-based variables such as personality traits, ideological dispositions, and cognitive styles (Hodson and Dhont 2015). Theodor Adorno and colleagues (1950), for example, maintained that general outgroup negativity is rooted in an authoritarian personality structure. Building on this theoretical claim, contemporary scholars argue that individuals with high levels of right-wing authoritarianism tend to reject outgroups they consider as a threat to the social-cultural order—namely ingroup values, norms, stability, and security. Individuals with strong social dominance orientations, on the other hand, are more inclined to devalue groups of perceived lower social status (Duckitt and Sibley 2007). Similarly, generalized prejudice has been linked to Big Five personality traits like agreeableness and openness to experience (Ekehammar et al. 2004). Personality is not the only factor put forward to explain generalized prejudice, however. Low educated, unemployed, and non-manual workers, men, and elderly persons are generally more prone to hostility towards outgroups (Altemeyer 1998; Schiefer 2013; Zick et al. 2011). Also, religious involvement is claimed to play a—albeit somewhat paradoxical—role in the genesis of generalized prejudice. Christian fundamentalism is positively related to feelings of GFE (Koopmans 2015), but other measures of Christian religiosity, such as church attendance, are only weakly associated with general outgroup hostility (Altemeyer and Hunsberger 1992). In summary, the generalized prejudice approach (1) focuses primarily on the common component of generalized prejudice rather than on the particularity of the specific components; and (2) seeks the origins of generalized prejudice in characteristics of the individual, such as cognitive capacities, personality structures, and socio-demographics, rather than in the social context of specific intergroup relations and features of the target group. The Differentiated Threat Approach: Social Context and the Outgroup Specificity of Prejudice An exclusive focus on the generalized character of prejudice limits our understanding of prejudice. Which outgroups become subject to derogation and negative imagery depends crucially on “the options a specific society offers” (Zick et al. 2008:367). Although early (Allport 1958) and more recent (Fiske et al. 2002) social-psychological studies have recognised the importance of situational factors for understanding prejudice (but see also Choma and Hodson 2008), several sociological theories stressed the importance of more insight in the structural and contextual factors of prejudice—including patterns of social contact and competition, economic and power relations, as well as media and elite discourses positions—in order to understand why negative attitudes towards specific (and not all) outgroups are triggered (see, for example, Agnew et al. 2000; Akrami et al. 2011; Ray and Lovejoy 1986). Herbert Blumer’s (1958) concept of “sense of group position” explicitly stresses the paramount importance of the concrete contexts and social structures in which intergroup relations are shaped: “[race] prejudice exists basically in a sense of group position rather than in a set of feelings which members of one racial group have toward the members of another racial group. This different way of viewing race prejudice shifts study and analysis from a preoccupation with feelings as lodged in individuals to a concern with the relationship of [racial] groups” (p. 3). This is to say, prejudice is a way to describe, justify, and understand one’s own structural and group position in relation to that of outgroup members. The development of prejudices, according to this logic, is fundamentally a social and collective process originating from dominant group members’ perception that their proprietary claims over certain resources and privileges as well as their shared beliefs and social hierarchies are being challenged by a subordinate group. In a similar vein, scholars like Lawrence Bobo (1999; Bobo and Hutchings 1996) and Pierre Bourdieu (1985, 1990) characterize social space as a universe of ongoing struggle between individuals and groups, making clear that it is meaningless to study particular expressions of prejudice without reference to the specific context of relative group positions and intergroup struggles within different social fields. Such a relational perspective posits prejudices as structured by the ingroup’s relative positioning vis-à-vis outgroups in the power configuration and their ability to mobilize their capital (social, cultural, economic, and/or symbolic) to sustain or improve status (Bobo 1999; Bourdieu 1984). As such, the particular characteristics of the relations between two groups facilitate certain prejudices and inhibit others. Similarly, pioneers of group conflict theory (GCT), such as Valdimer Key (1949) and Hubert Blalock (1967), observe that anti-Black prejudice and discrimination were more prevalent in U.S. regions where blacks made up considerable proportions of the resident population. This led to the idea that prejudices towards outgroups are essentially rooted in the perception that they threaten prerogatives of the own social group (Meuleman, Davidov, and Billiet 2009). In this logic, prejudice is seen as a defensive reaction to a sense of threat caused by conflicts between two or more social groups. This approach emphasizes the role of social comparisons—essentially not egoistic, but fraternal or group relative deprivation—to explain the genesis of prejudice. Individuals who perceive that their ingroup is worse off or disadvantaged in comparison to a relevant outgroup are more likely to express prejudice toward that group (Merton [1957] 1968; Runciman 1966). In order to investigate the outgroup specific antecedents of prejudice, the general idea that outgroup threats cause prejudice among the dominant population needs to be qualified in two ways—we call this the differentiated threat approach: (1) the type of perceived threat is idiosyncratic to specific intergroup relations; and (2) outgroup threats do not affect a society as a whole, but disproportionally affect specific socioeconomic strata and cultural segments of the ingroup. Consequently, the differentiated threat approach—as a theoretical lens—can only be understood and articulated with explicit reference to the specific social contexts. The first component of the differentiated threat approach acknowledges that the nature of the perceived conflict/threat can be very diverse. Outgroups can challenge the collective interest, identity, and power status of the ingroup, resulting in different types of threat (Riek, Mania, and Gaertner 2006). On the one hand, realistic or socioeconomic threat is induced by the (subjective) experience of relative deprivation, caused by competition for material goods such as well-paid jobs, affordable housing, or resources of the welfare state (Olzak 1994). Symbolic or cultural threat, on the other hand, originates in intergroup conflict over the established social order, cultural traditions, and shared beliefs, norms, and values (Stephan et al. 1998). Importantly, distinct outgroups can be perceived either as a realistic threat, a symbolic threat, or a combination of both (Hjerm and Nagayoshi 2011). Based on the nature of perceived threat (realistic versus symbolic) posed by outgroups, various types of “others” can be distinguished (see Figure 1): (1) deviant groups (low realistic threat, high symbolic threat): non-subordinated outgroups who are violating the social norms and threatening the established social-cultural order and societal stability. These groups are perceived as socially/culturally deviant, but not struggling for collective resources and not necessarily of lower social status (e.g., LGBT, Ku Klux Klan members, or atheists in the United States); (2) competing groups (high realistic threat, low symbolic threat): socially subordinated outgroups with low status and power who compete actively for scarce resources and challenge social inequalities, but not culturally deviant. These groups can be classified as disadvantaged norm-conformists since they are not challenging the values of the ingroup, but the distribution of collective resources (e.g., unemployed or homeless people); and (3) dissident groups (high realistic threat, high symbolic threat): outgroups seen as competing with but also threatening the ingroup since they tend to struggle actively for scarce resources and are challenging the ingroup’s values and norms at the same time. These groups could be described as usurpers trying to seize status and power by claiming scarce resources and shifting established cultural order (e.g., immigrants). Figure 1. View largeDownload slide Typology of Outgroups According to the Nature of Threat Posed Figure 1. View largeDownload slide Typology of Outgroups According to the Nature of Threat Posed This categorization of groups bears resemblance to Anne Schneider and Helen Ingram’s (1993) theory on the social construction of target groups that distinguishes groups on the basis of their power resources and the valence of the social constructions (positive versus negative). In line with the dual-process motivational model (Duckitt 2006) and the stereotype content model (Fiske et al. 2002), we argue that different types of outgroups may be prejudiced against for distinct reasons, according to different group-specific mechanisms such as the nature of the perceived threat and the content of ascribed stereotypes. Prejudices against outgroups that are symbolically threatening (e.g., deviant groups) are predominantly driven by authoritarian dispositions. Feelings of relative deprivation, on the other hand, are expected to induce prejudices towards socially subordinate groups competing for scarce resources (e.g., competing groups) (Asbrock et al. 2010). Similarly, Susan Fiske and colleagues (2002) show that prejudices cluster along two dimensions, namely “perceived competence” defined by the relative socioeconomic status of groups (low versus high status) and “perceived warmth” defined by the type of interdependence between groups (competitive versus cooperative). As a consequence, groups that are perceived to be similar on stereotype content are targeted for similar reasons. The second component of the differentiated threat approach is its focus on the social stratification and cultural diversity of the dominant society. The dominant group cannot be reduced to a homogenous bloc, but is instead segmented into different socioeconomic classes (with distinct risks and resources) and cultural groups (with distinct value patterns and ideological dispositions). Therefore, intergroup competition for realistic and symbolic goods does not affect all groups in society to the same degree. Realistic threats emanating from specific outgroups will only affect certain socioeconomic segments of the dominant group, while symbolic threats posed by distinct outgroups will only appeal to some cultural sections of dominant society. Conversely, prejudiced attitudes are not necessarily directed towards all possible outgroups, but are rather focused on those outgroups that are socially constructed as a challenge to realistic and/or symbolic interests (Schneider and Ingram 1993). The differentiated threat approach can thus be translated into a two general propositions: (1) socioeconomic segment A of the dominant society will develop prejudice towards outgroup B if outgroup B threatens the realistic interests of segment A; (2) cultural segment C of the dominant society will develop prejudice towards outgroup D if outgroup D threatens the symbolic interests of segment C. Below, these two propositions are translated into testable hypotheses regarding prejudices towards four outgroups in Belgium. Contextualizing Specific Prejudices: The Case of Belgium The quintessence of the differentiated threat approach is that prejudices cannot be understood in abstracto, but instead they need to be situated in social space (cf. Bourdieu 1985), i.e., the specific structural positions, competitive contexts, and cultural-discursive contexts in which intergroup relations are embedded. The general propositions of the differentiated threat approach cannot be tested directly but provide an operational lens by which prejudice can be studied in a particular social space. This contribution illustrates how this can be done by applying the theoretical framework of the differentiated threat approach to the particular context of Belgium. We focus on prejudices towards four outgroups that figure prominently in current debates in Belgium: immigrants, Muslims, homosexuals, and Jews. Below, we contextualize the societal positions of the four groups and derive hypotheses on the group-specific antecedents of prejudice that indirectly allow us to assess the basic premises of the differentiated threat approach. In the Belgian context the notion of “immigrants” is predominantly associated with Turkish and Moroccan communities that have their roots in the post-war labor migration. The popular use of the term immigrants often also includes the second and third generation descendants of these labor migrants. Persons of Turkish and Moroccan descent generally occupy disadvantaged socioeconomic positions and are often perceived as a threat for low-skilled jobs and social welfare provisions of the native Belgians (Abts and Kochuyt 2013). As a culturally visible and distinct outgroup, immigrants of Turkish and Moroccan origin are perceived as a threat to the established social-cultural order as well (Swyngedouw 1995). In this sense, this target group can be described as “dissident”: both competing with low-skilled natives for resources and status, and threatening the ingroup’s norms and values. Although many persons of Turkish and Moroccan descent are Muslim, it is of crucial importance to distinguish anti-immigration attitudes from a second form of prejudice, namely Islamophobia. Islamophobia refers to indiscriminate negative attitudes or hostility toward Islam and Muslims as a religious community (Bleich 2011). In Western Europe, since 9/11 the supposed historical incompatibility of European and Islamic values is central to the rise of Islamophobia (Strabac and Listhaug 2008). Anti-Islam discourse depicts Islam as a monolithic, inherently violent, and uniquely sexist religion, whose followers are seen as the ultimate cultural “other” that will never be able to cope with democratic and liberal values of Western society (Kumar 2012; Taras 2013). Islam as a denomination and Muslims as a religious-cultural group are in the eyes of prejudiced individuals a cultural rather than an economic threat, and are therefore considered as a “deviant” group. Third, homonegativity refers to the degree of prejudicial biases towards gay, lesbian, bisexual, and transsexual (LGBT) people (Hudson and Ricketts 1980). Although Belgium is known as a frontrunner in terms of LGBT rights and the social climate has become more positive in the last decade (Aerts et al. 2014), homonegativity is still present in Belgian society. The core of homonegative discourse is that homosexuality poses a threat to the social-cultural order: Homosexuals are a “deviant group,” threatening the Christian value system, violating traditional morality, and challenging conventional gender role patterns. Fourth, anti-Semitism refers not to hostility toward Israel, but to the cultivation of resentments against an imagined and generalized “collective Jew” who is a cultural other that does not make any effort to assimilate, and simultaneously is a powerful other with a strong grip on the economy and competing with domestic actors—thus “dissident” (Meer 2013; Schiffer and Wagner 2011). Yet, the nature of dissidence ascribed to Jews is quite specific. Most dissident groups are precisely blamed for not contributing enough to the economy and being in competition with low-skilled natives for scarce resources (e.g., the Turkish and Moroccan communities). The prototypical ultra-orthodox Haredi Jews (who are strongly concentrated in Antwerp and dominate its diamond industry) are contrarily blamed for trying to control domestic economy and world finances. As such, whereas immigrant communities are dissidents of low social status, Jews are imagined as dissidents of high status (Fiske et al. 2002). Note that our selection of outgroups does not cover all theoretically possible types of others (see Figure 1) since our study comprises two deviant (homosexuals and Muslims) and two dissident (immigrants and Jews) groups. Although it does not include an example of competing others, our study contains sufficient variety in intergroup contexts to illustrate the differentiated threat approach empirically. This contextualization, combined with the propositions of the differentiated threat approach, leads us to expect that the different forms of prejudice are triggered—at least partially—by group-specific antecedents. The first proposition of the differentiated threat approach relates to the axis of realistic threat and stipulates that socioeconomic segments of dominant society will develop prejudice towards outgroups that threaten their specific realistic interests. Since immigrants as well as Jews are seen as an economic threat—albeit each in their distinct way—people with strong feelings of relative deprivation are especially expected to develop prejudice towards these groups. We anticipate relative deprivation to be a more relevant antecedent for anti-immigrant attitudes and anti-Semitism than for the other forms of prejudice like homonegativity or Islamophobia, which are predominantly rooted in perceptions of cultural threat. However, the socioeconomic status and hence nature of the economic competition posed by immigrants and Jews is radically different. As a result of the specific form of ethno-stratification in Belgium, low-skilled natives experience persons of Turkish and Moroccan descent as competitors for scarce economic goods (especially jobs, housing, and welfare provisions). As a result, anti-immigration attitudes are expected to crystallize along class lines, with low-skilled manual workers and the unemployed showing higher levels of anti-immigrant attitudes (Kunovich 2004). In the case of anti-Semitism, such class gradient is not expected. The second proposition of the differentiated threat approach relates to the symbolic axis, and states that the dominant cultural group will develop prejudice towards outgroups that threaten their specific symbolic interests. Islamophobia and homonegativity are both embedded in cultural rather than in economic threat perceptions. Muslims as well as sexual minorities are perceived as a threat to social order and dominant value schemes, but the specific sets of values these outgroups challenge are far from identical. As a result, Islamophobia and homonegativity are triggered according to group-specific mechanisms, albeit in quite different social groups characterized by different value patterns. The relationship between traditional gender role attitudes (Thornton, Alwin, and Camburn 1983) and the formation of prejudices exemplifies this crucial point. Regarding sexual minorities, one of the most persistent stereotypes about homosexuality relates to the violation of (patriarchal) gender norms (Aerts et al. 2014). Consequently, homonegativity is expected to be more prevalent among individuals identifying strongly with traditional gender roles. Similarly, males generally hold stronger homonegative beliefs than women (Herek 1988). This gender dimension of homonegativity is not just an effect of sex per se, but is closely related to the gender role expectations within society, especially the hegemonic (traditional) conceptions of masculinity (Connell 2005; Plummer 1999). Islamophobia is linked to threats to gender role expectations as well, but with different (if not opposite) underlying values and mechanisms. The idea that Islam is an inherently sexist religion is one of the constituting myths of Islamophobic discourse (Kumar 2012). Because the critique that Islam oppresses women resonates among persons who endorse gender equality, negative attitudes against Muslims are especially strong among persons rejecting traditional gender norms, which imply male domination and patriarchy. The paradoxical situation that gender traditionalism reinforces homonegativity and at the same time tempers Islamophobia is also apparent from discourse of political elites. Extreme right-wing political actors regularly mobilize the frame of the “homonegative Muslim community” in order to exhort anti-Islam sentiments (Bracke 2012). Besides gender equality, Islam is also perceived as a threat for certain value sets—such as separation of church and state, individual rights/freedom, democracy, and tolerance—that are strongly endorsed by the higher educated. In the eyes of this presumably tolerant group, Islam skepticism becomes legitimate as a defense of tolerance, liberal democracy, and Enlightenment (Bilsky 2009; Spruyt and Elchardus 2012). In other words, the “Muslim other” is transformed into a “suitable enemy,” i.e., a more legitimate object of prejudice (Fekete 2009). Hence, highly educated citizens might be more likely to hold negative attitudes towards Muslims than towards other outgroups, like immigrants, Jews, or homosexuals. The impact of religiosity on group-specific prejudices can be understood from the framework of cultural threats, but is highly complex. Allport (1958:444) already asserted that religion can both make and unmake prejudice. Although most religions preach tolerance toward outgroups, they permit some forms of prejudice against people who are perceived to violate the religion’s value system. A literature review supports the distinction between proscribed prejudices (race/ethnicity) and permitted prejudices (homosexuality) (Hunsberger and Jackson 2005). Following the Catholic teaching that identifies homosexuality as opposing religious morality and their value system, the Church enables prejudice against LGBT, as it is consistent with religious beliefs (Whitley 2009). Similarly, the portrayal of the role of Jews in the crucifixion of Jesus Christ or, more generally, their threat to Christianity might promote anti-Semitism among church-involved Christians (Pargament et al. 2007). Hypotheses Based on the generalized prejudice and differentiated threat approaches, we can derive two sets of hypotheses regarding the commonality (H1-H9) and specificities (H10-H20) of prejudice towards immigrants, homosexuals, Muslims, and Jews (see Table 1 for a summary). According to social-psychological theory, a strong common denominator—namely GFE or generalized prejudice—can be distinguished in prejudices towards the four outgroups (H1). The respective prejudices are anticipated to be shaped by the same antecedents. Higher levels of generalized prejudice are expected among men (H2), elderly persons (H3), lower educated individuals (H4), lower social classes (H5), and religious persons (H6). Authoritarianism (H7), feelings of relative deprivation (H8), and gender traditionalism (H9) are hypothesized to reinforce generalized prejudice. Table 1. Overview of Hypotheses Regarding the Common and Specific Determinants of Prejudices General Pattern Deviations from the General Pattern GFE Homonegativity Anti-immigrant Attitudes Islamophobia Anti-Semitism Gender (being male) + (H2) + (H11) Old age + (H3) Low education + (H4) − (H17) Low social class + (H5) + (H15) Strong religious involvement + (H6) + (H12) + (H19) Authoritarianism + (H7) Relative deprivation + (H8) − (H13) + (H16) + (H20) Gender traditionalism + (H9) + (H14) − (H18) General Pattern Deviations from the General Pattern GFE Homonegativity Anti-immigrant Attitudes Islamophobia Anti-Semitism Gender (being male) + (H2) + (H11) Old age + (H3) Low education + (H4) − (H17) Low social class + (H5) + (H15) Strong religious involvement + (H6) + (H12) + (H19) Authoritarianism + (H7) Relative deprivation + (H8) − (H13) + (H16) + (H20) Gender traditionalism + (H9) + (H14) − (H18) Table 1. Overview of Hypotheses Regarding the Common and Specific Determinants of Prejudices General Pattern Deviations from the General Pattern GFE Homonegativity Anti-immigrant Attitudes Islamophobia Anti-Semitism Gender (being male) + (H2) + (H11) Old age + (H3) Low education + (H4) − (H17) Low social class + (H5) + (H15) Strong religious involvement + (H6) + (H12) + (H19) Authoritarianism + (H7) Relative deprivation + (H8) − (H13) + (H16) + (H20) Gender traditionalism + (H9) + (H14) − (H18) General Pattern Deviations from the General Pattern GFE Homonegativity Anti-immigrant Attitudes Islamophobia Anti-Semitism Gender (being male) + (H2) + (H11) Old age + (H3) Low education + (H4) − (H17) Low social class + (H5) + (H15) Strong religious involvement + (H6) + (H12) + (H19) Authoritarianism + (H7) Relative deprivation + (H8) − (H13) + (H16) + (H20) Gender traditionalism + (H9) + (H14) − (H18) The differentiated threat approach, on the other hand, predicts that prejudices towards immigrants, homosexuals, Muslims, and Jews cannot be reduced completely to a single common denominator and expects considerable particularities for the different prejudices (H10). Furthermore, the four prejudices are assumed to originate along specific mechanisms. Note that the hypotheses regarding group-specific antecedents are formulated in terms of deviations from the general pattern (i.e., the explanatory model for GFE). Homonegativity: Gender is expected to affect homonegativity even more strongly than GFE. While men are more prejudiced than women towards outgroups in general (see H2), the gender gap is expected to be more outspoken when homonegativity is concerned (H11). Similarly, religious involvement is anticipated to have a stronger impact on homonegativity (H12), while feelings of relative deprivation are assumed to be less relevant when explaining negative attitudes towards LGBT people (H13). Homonegativity is expected to be more outspoken among gender traditionalists (H14). Anti-immigrant attitudes: Belonging to the lower social classes (H15) and relative deprivation (H16) are expected to stimulate anti-immigrant sentiments more strongly than generalized prejudice. Islamophobia: Whereas a low educational level is supposed to be a strong predictor of general prejudice (see H4), we expect a less outspoken educational gradient for Islamophobia (H17). Furthermore, gender traditionalists are hypothesized to show lower levels of Islamophobia (H18). Anti-Semitism: Religious involvement (H19) as well as relative deprivation (H20) are assumed to have a stronger positive effect/relation with on anti-Semitism than on GFE. Importantly, these hypotheses illustrate that generalized prejudice and differentiated threat approaches are not mutually exclusive, but rather complement each other. Whereas the idea of generalized prejudice provides the general underlying mechanism through which prejudices originate, the differentiated threat approach provides more specific insights into which outgroups are targeted by which socioeconomic and cultural categories. DATA AND METHODS Data Set To test the hypotheses, the Belgian data collection for the fourth round of the European Values Study is used. In 2009, a two-stage probability sample of the Belgian population older than 18 years were approached by means of a computer assisted personal interview (CAPI). The realized sample size is 1,509 (response rate: 50.0 percent). Upon completion of the interview, participants were asked to fill out a drop-off questionnaire and send it back by mail. Seventy-five percent of the participants in the CAPI (i.e., 1,136 respondents) returned a completed questionnaire. Non-response analysis shows that certain groups (such as persons aged 40 to 54 and those with a tertiary education) are overrepresented in the realized sample, while others (persons without secondary education, individuals younger than 25 and those over 70) are underrepresented. To remediate this imbalance, we apply post-stratification weights correcting for the joint distribution of gender, age, and education. Because this study focuses on the structure of prejudices among majority-group members, respondents of immigrant background (operationalized as not having Belgian nationality or being born abroad) are excluded from the analysis, leading to a sample size of 1,007. Indicators Measures of Prejudice Each form of prejudice is measured by four to six Likert-items. These items were included in the drop-off questionnaire with a maximal spread in order to minimize memory effects and contamination. Homonegativity is measured by means of statements referring to the (ab)normality of homosexuality (q7_1; q7_5), homosexuality as a normative threat (q7_2), openness of homosexual identities (q7_3), and non-discrimination of homosexuals (q7_4). Anti-immigrant attitudes are operationalized by items regarding the trustworthiness of immigrants (q8_1), social security threat posed by immigrants (q8_2), and the impact of immigration on cultural life (q8_3; q8_4). The introductory text to this battery states that the term immigrants refers in the first place to persons of Turkish of Moroccan origin living in Belgium. Four statements referring to the presumed influence (q15_2), self-complacency (q15_3), trustworthiness (q15_4), and greediness (q15_5) of Jewish persons indicate anti-Semitism. Finally, Islamophobia is measured by means of items on the role of women in Islam (q24_2), the presumed violent nature of Islam (q24_4), and Islam as a geopolitical (q24_3) or cultural (q24_5) threat to Western society. The scales are partially balanced. Besides statements expressing a negative attitude, positively worded items are also included. To answer these statements, respondents were offered five-point scales, ranging from “agree completely” (1) to “disagree completely” (5). Validity, reliability, and dimensionality of the items were tested and confirmed by means of confirmatory factor analysis (CFA) (see Table A1 in the Appendix). Explanatory Variables Various indicators of social-structural position are included, namely gender, age (in years), and educational level (up to lower secondary degree, higher secondary degree, tertiary degree). In the operationalization of religious involvement, we make a distinction between Christians (mostly Catholics) who attend religious services on a regular basis (i.e., at least once per month); persons who consider themselves as Christians but who do not or only occasionally attend services, and non-believers and atheists. Employment status is based on the Erikson-Goldthorpe-Portocarero (EGP) (Ganzeboom and Treiman 1996) social class scheme, and distinguishes between (1) service class I and the self-employed; (2) non-manual workers (i.e., service class II and routine non-manuals); (3) manual workers (skilled and unskilled); and (4) those who never worked (and for which the EGP category can consequently not be determined). A second set of predictors concerns three attitudinal dimensions, each operationalized by means of multiple indicators (five-point agree-disagree statements). Authoritarianism is measured by means of three items on the importance of obedience and respect (q12_4), strict laws (q12_5), and punishment of immoral persons (q12_3). Three items (q30_1; q30_2; q30_3) referring to the feeling of being disadvantaged as a group by government and institutions indicate group relative deprivation. Finally, gender traditionalism is measured by means of items regarding gender-stereotypical education for boys and girls (q14_1; q14_2) and gender roles on the labor market (q14_3; q14_4). CFA showed that these items are sufficiently valid and reliable measurements for the intended concepts (see Appendix B). Statistical Modelling The statistical analysis is carried out in two major steps. First, in order to study how strong the common and group-specific components of prejudice are, we estimate a second-order confirmatory factor model (see also Bratt 2005; Zick et al. 2008). The point of departure of this approach is that a respondent’s response to a statement regarding a specific outgroup reflects three components: (1) second-order factor GFE, i.e., the common component of prejudices; (2) group-specific prejudice, i.e., prejudice towards the outgroup explicitly mentioned in the items; and (3) random measurement error at the item level.1 1 Unfortunately, our model only assumes random measurement error, and does not take systematic measurement error into account. This has important consequences for the distinction between common and specific prejudices. Insofar as the different items are affected by the same bias—i.e., so-called common method variance (CMV) (Podsakoff et al. 2003)—the common factor will absorb the systematic measurement errors. Thus, the second-order CFA approach overestimates rather than underestimates the commonality of prejudices—a fact that is often neglected in empirical research. The model thus specifies a so-called second-order factor (i.e., GFE) that captures the common variance of the first-order factors (i.e., the group-specific prejudices) (see Figure 2). Figure 2. View largeDownload slide Second-Order Factor Model with Standardized Factor Loadings Note: N = 1,007 Figure 2. View largeDownload slide Second-Order Factor Model with Standardized Factor Loadings Note: N = 1,007 Second, in order to study whether the roots of prejudice are common or group-specific (the second generalizability question), structural and attitudinal variables are added to the second-order CFA model (see Figure 3). This model captures the similarity of antecedents by estimating effects of the predictors on the second-order factor GFE (H2-H9). Group-specific explanations (H11-H20) are incorporated by allowing direct effects from the predictors on the various forms of prejudice (the first-order factors). These direct effects indicate deviations from the common explanatory model, i.e., that the impact of the respective predictor on the specific prejudice is different from that on the GFE factor. For reasons of parsimony, only effects from the predictors to the specific prejudices that lead to a significant improvement are introduced in the final model. Figure 3. View largeDownload slide GFE and Forms of Prejudice Explained by Means of Structural and Attitudinal Predictors – Significant Direct Effects Note: Direct arrows from predictors on first-order prejudices are striped. Figure 3. View largeDownload slide GFE and Forms of Prejudice Explained by Means of Structural and Attitudinal Predictors – Significant Direct Effects Note: Direct arrows from predictors on first-order prejudices are striped. All reported models are estimated using Mplus version 7.1. The ordered categorical nature of the indicators is taken into account by using the robust weighted least squares estimator (WLSMV). All reported parameters below are standardized, apart from the effects of the dummy variables, which are semi-standardized. As a result, the effects of dummy variables refer to the difference with the reference category in terms of standard deviations on the dependent variable. RESULTS The Common Denominator of Prejudices: A Second-Order Factor Model In order to analyse to what extent group-specific prejudices share a common denominator, we estimate a model with four first-order factors (anti-immigrant attitudes, homonegativity, Islamophobia, and anti-Semitism) loading on a second-order factor, capturing the communality of prejudices or GFE (see Figure 2). The anti-immigrant factor is allowed to correlate with Islamophobia, since the immigrant group referred to—people of Turkish and Moroccan origin—is predominantly Muslim. The first-order factor loadings are quite high (mostly >.60; many >.80), indicating that items are sufficiently valid and reliable measurements. This second-order factor model has an acceptable model fit (see Model 1 in Table 2): The root mean squared error of approximation (RMSEA) equals .065, and both the comparative fit index (CFI; .968) and Tucker-Lewis index (TLI; .962) are sufficiently close to 1. The second-order factor model has a substantially better fit than a model in which the four specific prejudices are not related to each other (implying that there is no communality; see Model 2 in Table 2) or a model in which all items load on a single first-order factor (meaning that prejudice has no group-specific components; see Model 3 in Table 2). This confirms our hypotheses: overlap indeed exists between the specific prejudices (H1), but at the same time they contain specific components and can therefore not be reduced to a single GFE dimension (H10). Table 2. Fit Indices for Various Decompositions of the Common and Specific Components of Prejudice Chi-square Df RMSEA CFI TLI Model 1 Specific and common components: second-order factor model 602.48 114 .065 .968 .962 Model 2 Only specific components: four unrelated first-order factors 4669.59 119 .195 .705 .663 Model 3 Only common component: all items load on a single factor 4761.76 119 .197 .699 .656 Chi-square Df RMSEA CFI TLI Model 1 Specific and common components: second-order factor model 602.48 114 .065 .968 .962 Model 2 Only specific components: four unrelated first-order factors 4669.59 119 .195 .705 .663 Model 3 Only common component: all items load on a single factor 4761.76 119 .197 .699 .656 Table 2. Fit Indices for Various Decompositions of the Common and Specific Components of Prejudice Chi-square Df RMSEA CFI TLI Model 1 Specific and common components: second-order factor model 602.48 114 .065 .968 .962 Model 2 Only specific components: four unrelated first-order factors 4669.59 119 .195 .705 .663 Model 3 Only common component: all items load on a single factor 4761.76 119 .197 .699 .656 Chi-square Df RMSEA CFI TLI Model 1 Specific and common components: second-order factor model 602.48 114 .065 .968 .962 Model 2 Only specific components: four unrelated first-order factors 4669.59 119 .195 .705 .663 Model 3 Only common component: all items load on a single factor 4761.76 119 .197 .699 .656 The second-order factor loadings contain information on the relative importance of the specific and common components. All factor loadings are high (.54, .53, .86, and .60; see Table 3), expressing that the specific prejudices are strongly influenced by the common GFE factor. Interestingly, the loading for anti-immigrant attitudes is substantially stronger than for the other three forms of prejudice. A general inclination to prejudice informs attitudes towards all outgroups, but it does even more so for anti-immigrant attitudes. The factor loading approaches 1, indicating that it becomes very hard to distinguish between GFE and anti-immigrant attitudes. In the Belgian public opinion, the category “immigrant” is the master template onto which processes of the formation of generalized prejudice are projected. Immigrants, especially of Turkish and Moroccan descent, have become the Belgian archetypical object of othering. Table 3. Common and Specific Components of the Four Forms of Prejudice Loading on 2nd Order Factor GFE Communality Uniqueness Homonegativity .54 .29 .71 Anti-immigrant attitudes .86 .74 .26 Anti-Semitism .53 .28 .72 Islamophobia .60 .36 .64 Loading on 2nd Order Factor GFE Communality Uniqueness Homonegativity .54 .29 .71 Anti-immigrant attitudes .86 .74 .26 Anti-Semitism .53 .28 .72 Islamophobia .60 .36 .64 Table 3. Common and Specific Components of the Four Forms of Prejudice Loading on 2nd Order Factor GFE Communality Uniqueness Homonegativity .54 .29 .71 Anti-immigrant attitudes .86 .74 .26 Anti-Semitism .53 .28 .72 Islamophobia .60 .36 .64 Loading on 2nd Order Factor GFE Communality Uniqueness Homonegativity .54 .29 .71 Anti-immigrant attitudes .86 .74 .26 Anti-Semitism .53 .28 .72 Islamophobia .60 .36 .64 Based on the second-order factor loadings, the common (communalities) and specific (uniqueness) components of the four forms of prejudice can be calculated. The communalities represent the proportion of variance that the forms of prejudice share with the common GFE factor. Table 3 shows that the different prejudices have substantial unique components. Homonegativity, anti-Semitism, and Islamophobia share roughly one third of their variance with GFE, while two thirds of the variance is outgroup specific. The overlap between anti-immigration attitudes and GFE is considerably larger (namely 74 percent). But even in this case, a non-negligible portion of variation (26 percent) is specific for the outgroup referred to. By consequence, an exclusive focus on this communality—that constitutes only the smaller share of the variance—leads by definition to a limited understanding of the genesis of prejudices. Common or Differential Antecedents: A Full Structural Equation Model To study whether the different antecedents precede the specific forms of prejudice, structural and attitudinal predictors are added to the second-order factor model (see Figure 3). Table 4 displays the direct effects of the predictors on GFE (showing the common impact of the antecedents on prejudice in general) as well as the direct effects on the four specific prejudice-components (i.e., the deviations from the general pattern). Sometimes, these direct effects can be misleading as they represent impacts controlling for all other predictors (also those further in the causal chain). Therefore, Table 5 additionally displays the total effects of the predictors on the prejudices. These total effects are the sum of the direct effects shown in Table 4 and the indirect effects that run through the mediating variables. Detailed insight in the explanatory model requires information on both effects. While the direct effects are especially useful to uncover differential impacts of predictors, the total effects provide insight in the general patterns in the data. Table 4. Direct Effects of Structural and Attitudinal Characteristics on Forms of Prejudice GFE Homonegativity Anti-immigrant attitudes Islamophobia Anti-Semitism Par. Est. p-value Par. Est. p-value Par. Est. p-value Par. Est. p-value Par. Est. p-value Gender (male, ref. cat.)  Female −.143 .082 −.169 .013 Age −.155 .009 .178 .000 .183 .000 Education (lower secondary, ref. cat.)  Higher secondary .036 .707  Tertiary −.039 .718 Social class (manual workers, ref. cat.)  Service class/self-employed .165 .280 −.068 .538  Non-manual workers .169 .117 −.182 .015  Never worked −.018 .904 −.156 .209 Religious involvement (regular church attendance, ref. cat)  Occasional church attendance .070 .527 −.329 .002 −.236 .034  Non-believers and atheists .039 .661 −.181 .015 −.025 .776 Authoritarianism .594 .000 Relative deprivation .297 .000 −.158 .001 .109 .009 Gender traditionalism .130 .007 .364 .000 −.207 .000 Explained variance (%) 58.9 53.9 69.8 50.1 28.3 Chi-square: 1226.29; Df: 506; RMSEA: .038; CFI: .967; TLI: .961 N = 1,007 GFE Homonegativity Anti-immigrant attitudes Islamophobia Anti-Semitism Par. Est. p-value Par. Est. p-value Par. Est. p-value Par. Est. p-value Par. Est. p-value Gender (male, ref. cat.)  Female −.143 .082 −.169 .013 Age −.155 .009 .178 .000 .183 .000 Education (lower secondary, ref. cat.)  Higher secondary .036 .707  Tertiary −.039 .718 Social class (manual workers, ref. cat.)  Service class/self-employed .165 .280 −.068 .538  Non-manual workers .169 .117 −.182 .015  Never worked −.018 .904 −.156 .209 Religious involvement (regular church attendance, ref. cat)  Occasional church attendance .070 .527 −.329 .002 −.236 .034  Non-believers and atheists .039 .661 −.181 .015 −.025 .776 Authoritarianism .594 .000 Relative deprivation .297 .000 −.158 .001 .109 .009 Gender traditionalism .130 .007 .364 .000 −.207 .000 Explained variance (%) 58.9 53.9 69.8 50.1 28.3 Chi-square: 1226.29; Df: 506; RMSEA: .038; CFI: .967; TLI: .961 N = 1,007 Notes: The parameters displayed are semi-standardized when the independent variable is a dummy (thus representing the difference with the reference category in terms of standard deviations of the dependent variable) and fully standardized in all other cases. The analysis is weighted for gender, age, and education. Table 4. Direct Effects of Structural and Attitudinal Characteristics on Forms of Prejudice GFE Homonegativity Anti-immigrant attitudes Islamophobia Anti-Semitism Par. Est. p-value Par. Est. p-value Par. Est. p-value Par. Est. p-value Par. Est. p-value Gender (male, ref. cat.)  Female −.143 .082 −.169 .013 Age −.155 .009 .178 .000 .183 .000 Education (lower secondary, ref. cat.)  Higher secondary .036 .707  Tertiary −.039 .718 Social class (manual workers, ref. cat.)  Service class/self-employed .165 .280 −.068 .538  Non-manual workers .169 .117 −.182 .015  Never worked −.018 .904 −.156 .209 Religious involvement (regular church attendance, ref. cat)  Occasional church attendance .070 .527 −.329 .002 −.236 .034  Non-believers and atheists .039 .661 −.181 .015 −.025 .776 Authoritarianism .594 .000 Relative deprivation .297 .000 −.158 .001 .109 .009 Gender traditionalism .130 .007 .364 .000 −.207 .000 Explained variance (%) 58.9 53.9 69.8 50.1 28.3 Chi-square: 1226.29; Df: 506; RMSEA: .038; CFI: .967; TLI: .961 N = 1,007 GFE Homonegativity Anti-immigrant attitudes Islamophobia Anti-Semitism Par. Est. p-value Par. Est. p-value Par. Est. p-value Par. Est. p-value Par. Est. p-value Gender (male, ref. cat.)  Female −.143 .082 −.169 .013 Age −.155 .009 .178 .000 .183 .000 Education (lower secondary, ref. cat.)  Higher secondary .036 .707  Tertiary −.039 .718 Social class (manual workers, ref. cat.)  Service class/self-employed .165 .280 −.068 .538  Non-manual workers .169 .117 −.182 .015  Never worked −.018 .904 −.156 .209 Religious involvement (regular church attendance, ref. cat)  Occasional church attendance .070 .527 −.329 .002 −.236 .034  Non-believers and atheists .039 .661 −.181 .015 −.025 .776 Authoritarianism .594 .000 Relative deprivation .297 .000 −.158 .001 .109 .009 Gender traditionalism .130 .007 .364 .000 −.207 .000 Explained variance (%) 58.9 53.9 69.8 50.1 28.3 Chi-square: 1226.29; Df: 506; RMSEA: .038; CFI: .967; TLI: .961 N = 1,007 Notes: The parameters displayed are semi-standardized when the independent variable is a dummy (thus representing the difference with the reference category in terms of standard deviations of the dependent variable) and fully standardized in all other cases. The analysis is weighted for gender, age, and education. Table 5. Total Effects of Structural and Attitudinal Characteristics on Forms of Prejudice GFE Homonegativity Anti-immigrant Attitudes Islamophobia Anti-Semitism Par. Est. p-value Par. Est. p-value Par .Est. p-value Par .Est. p-value Par. Est. p-value Gender (male, ref. cat.)  Female −.087 .303 −.344 .000 −.072 .303 .019 .742 −.046 .306 Age .067 .227 .294 .000 .055 .221 .178 .000 .036 .221 Education (lower secondary, ref. cat.)  Higher secondary −.221 .040 −.129 .045 −.182 .037 −.137 .068 −.117 .040  Tertiary −.682 .000 −.350 .000 −.562 .000 −.449 .000 −.363 .000 Social class (manual workers, ref. cat.)  Service class/self-employed −.186 .239 −.112 .223 −.221 .010 −.118 .255 −.099 .235  Non-manual workers .008 .945 −.016 .813 −.176 .029 .011 .888 .004 .945  Never worked −.002 .991 .177 .053 −.157 .224 −.110 .282 −.001 .991 Religious involvement (regular church attendance, ref. cat.)  Occasional church attendance .130 .319 −.330 .002 .107 .316 .124 .166 −.167 .130  Non-believers and atheists −.077 .423 −.328 .000 −.063 .419 .014 .836 −.066 .455 Authoritarianism .594 .000 .288 .000 .489 .000 .386 .000 .316 .000 Relative deprivation .297 .000 −.014 .730 .245 .000 .303 .000 .158 .000 Gender traditionalism .130 .007 .427 .000 .107 .004 −.122 .005 .069 .005 Explained variance (%) 58.9 53.9 69.8 50.1 28.3 Chi-square: 1226.29; Df: 506; RMSEA: .038; CFI: .967; TLI: .961 N = 1,007 GFE Homonegativity Anti-immigrant Attitudes Islamophobia Anti-Semitism Par. Est. p-value Par. Est. p-value Par .Est. p-value Par .Est. p-value Par. Est. p-value Gender (male, ref. cat.)  Female −.087 .303 −.344 .000 −.072 .303 .019 .742 −.046 .306 Age .067 .227 .294 .000 .055 .221 .178 .000 .036 .221 Education (lower secondary, ref. cat.)  Higher secondary −.221 .040 −.129 .045 −.182 .037 −.137 .068 −.117 .040  Tertiary −.682 .000 −.350 .000 −.562 .000 −.449 .000 −.363 .000 Social class (manual workers, ref. cat.)  Service class/self-employed −.186 .239 −.112 .223 −.221 .010 −.118 .255 −.099 .235  Non-manual workers .008 .945 −.016 .813 −.176 .029 .011 .888 .004 .945  Never worked −.002 .991 .177 .053 −.157 .224 −.110 .282 −.001 .991 Religious involvement (regular church attendance, ref. cat.)  Occasional church attendance .130 .319 −.330 .002 .107 .316 .124 .166 −.167 .130  Non-believers and atheists −.077 .423 −.328 .000 −.063 .419 .014 .836 −.066 .455 Authoritarianism .594 .000 .288 .000 .489 .000 .386 .000 .316 .000 Relative deprivation .297 .000 −.014 .730 .245 .000 .303 .000 .158 .000 Gender traditionalism .130 .007 .427 .000 .107 .004 −.122 .005 .069 .005 Explained variance (%) 58.9 53.9 69.8 50.1 28.3 Chi-square: 1226.29; Df: 506; RMSEA: .038; CFI: .967; TLI: .961 N = 1,007 Notes: The parameters displayed are semi-standardized when the independent variable is a dummy (thus representing the difference with the reference category in terms of standard deviations of the dependent variable) and fully standardized in all other cases. The analysis is weighted for gender, age, and education. Table 5. Total Effects of Structural and Attitudinal Characteristics on Forms of Prejudice GFE Homonegativity Anti-immigrant Attitudes Islamophobia Anti-Semitism Par. Est. p-value Par. Est. p-value Par .Est. p-value Par .Est. p-value Par. Est. p-value Gender (male, ref. cat.)  Female −.087 .303 −.344 .000 −.072 .303 .019 .742 −.046 .306 Age .067 .227 .294 .000 .055 .221 .178 .000 .036 .221 Education (lower secondary, ref. cat.)  Higher secondary −.221 .040 −.129 .045 −.182 .037 −.137 .068 −.117 .040  Tertiary −.682 .000 −.350 .000 −.562 .000 −.449 .000 −.363 .000 Social class (manual workers, ref. cat.)  Service class/self-employed −.186 .239 −.112 .223 −.221 .010 −.118 .255 −.099 .235  Non-manual workers .008 .945 −.016 .813 −.176 .029 .011 .888 .004 .945  Never worked −.002 .991 .177 .053 −.157 .224 −.110 .282 −.001 .991 Religious involvement (regular church attendance, ref. cat.)  Occasional church attendance .130 .319 −.330 .002 .107 .316 .124 .166 −.167 .130  Non-believers and atheists −.077 .423 −.328 .000 −.063 .419 .014 .836 −.066 .455 Authoritarianism .594 .000 .288 .000 .489 .000 .386 .000 .316 .000 Relative deprivation .297 .000 −.014 .730 .245 .000 .303 .000 .158 .000 Gender traditionalism .130 .007 .427 .000 .107 .004 −.122 .005 .069 .005 Explained variance (%) 58.9 53.9 69.8 50.1 28.3 Chi-square: 1226.29; Df: 506; RMSEA: .038; CFI: .967; TLI: .961 N = 1,007 GFE Homonegativity Anti-immigrant Attitudes Islamophobia Anti-Semitism Par. Est. p-value Par. Est. p-value Par .Est. p-value Par .Est. p-value Par. Est. p-value Gender (male, ref. cat.)  Female −.087 .303 −.344 .000 −.072 .303 .019 .742 −.046 .306 Age .067 .227 .294 .000 .055 .221 .178 .000 .036 .221 Education (lower secondary, ref. cat.)  Higher secondary −.221 .040 −.129 .045 −.182 .037 −.137 .068 −.117 .040  Tertiary −.682 .000 −.350 .000 −.562 .000 −.449 .000 −.363 .000 Social class (manual workers, ref. cat.)  Service class/self-employed −.186 .239 −.112 .223 −.221 .010 −.118 .255 −.099 .235  Non-manual workers .008 .945 −.016 .813 −.176 .029 .011 .888 .004 .945  Never worked −.002 .991 .177 .053 −.157 .224 −.110 .282 −.001 .991 Religious involvement (regular church attendance, ref. cat.)  Occasional church attendance .130 .319 −.330 .002 .107 .316 .124 .166 −.167 .130  Non-believers and atheists −.077 .423 −.328 .000 −.063 .419 .014 .836 −.066 .455 Authoritarianism .594 .000 .288 .000 .489 .000 .386 .000 .316 .000 Relative deprivation .297 .000 −.014 .730 .245 .000 .303 .000 .158 .000 Gender traditionalism .130 .007 .427 .000 .107 .004 −.122 .005 .069 .005 Explained variance (%) 58.9 53.9 69.8 50.1 28.3 Chi-square: 1226.29; Df: 506; RMSEA: .038; CFI: .967; TLI: .961 N = 1,007 Notes: The parameters displayed are semi-standardized when the independent variable is a dummy (thus representing the difference with the reference category in terms of standard deviations of the dependent variable) and fully standardized in all other cases. The analysis is weighted for gender, age, and education. We start the interpretation with the common impact of antecedents, i.e., the effects on GFE. Each of the three attitudinal variables is significantly related to GFE (see Table 4). As hypothesized, persons with feelings of relative deprivation (H8) and an authoritarian disposition (H7) tend to hold more prejudiced views in general. Especially authoritarianism is strongly linked to GFE: An increase of one standard deviation on the authoritarianism scale goes hand in hand with an increase of .594 standard deviations of the GFE-score. Also the endorsement of traditional gender roles turns out to be positively linked to GFE, although its impact is considerably weaker (.130). Of all structural variables, only education and age are significantly related to GFE. As expected, education has a strong tempering effect on generalized prejudice (H4). Looking at the total effects (Table 5), those with a degree from tertiary education score .682 standard deviations lower on GFE than persons with a lower secondary degree (i.e., the reference category). Also between individuals with a higher secondary and a lower secondary degree a significant difference (of -.221) is found. Education has no significant direct effect on GFE, however (see Table 4), indicating that the educational effect is fully mediated by the three attitudinal variables. The fact that the higher educated report less prejudice is completely accounted for by the fact that this group has a less authoritarian worldview, feels less deprived, and is more critical of traditional gender roles. In contrast to H3, age does not have a significant overall impact on GFE. When the attitudinal characteristics of the older cohorts are controlled for, we detect a negative net age effect (-.155; see Table 4). Gender (H2), social class (H5), and religious involvement (H6) are not significantly related to GFE, which contradicts our hypotheses as well. Taken together, the structural and attitudinal predictors explain almost 60 percent of the total variance of GFE. The primary purpose of this analysis, however, is to find out whether certain predictors have group-specific effects that deviate from this general pattern. The fact that several predictors have a direct effect on the first-order prejudices indicates that such differential effects are indeed present. It was already mentioned that relative deprivation had a moderate, positive effect on GFE. Direct effects of relative deprivation on homonegativity and Islamophobia, however, evidence that attitudes towards these two specific outgroups deviate from this general effect of relative deprivation. In case of homonegativity, the direct effect is negative (−.158 in Table 4). Compared to other forms of prejudice, the impact of relative deprivation on homonegativity is significantly weaker, which is exactly what was anticipated (H13). The total effects in Table 5 show that, overall, relative deprivation is not significantly related to homonegativity. Clearly, the perception of being deprived can translate into prejudiced attitudes, but not towards homosexuals. On the contrary, a positive direct effect of relative deprivation on Islamophobia (.109) illustrates that relative deprivation instigates anti-Muslim sentiments even more that it does so for prejudice towards outgroups. Gender traditionalism has a differentiated impact on homonegativity and Islamophobia. Traditional gender attitudes have a strong positive direct effect on homonegativity (.364). While gender traditionalism has a quite modest positive impact on prejudice in general, attitudes towards homosexuals are much more clearly structured along gender role expectations, thereby confirming H14. Of all predictors of homonegativity, gender traditionalism has the most explanatory power. For those adhering to traditional gender roles, homosexuality clearly presents a strong value threat. The situation is quite different in the case of Islamophobia: As expected (H18) traditional gender attitudes have a negative direct effect (−.207) that counteracts the overall prejudice-inducing effect of gender traditionalism completely. The total effect of gender traditionalism on Islamophobia is significantly negative. Thus, persons stressing the importance of gender traditionalism tend to be more prejudiced in general, but are found to score lower on Islamophobia. This is not surprising, given that the subordinate position of Muslim women is one of the constitutive myths of Islamophobia. As expected, the impact of authoritarianism, finally, is far less differentiated. For each of the specific forms of prejudice, authoritarianism is the single most important predictor, and no prejudice-specific effects are detected. Differentiated effects are also found for several structural variables. While no significant gender differences exist regarding GFE, women do report less homophobic attitudes than men (direct effect: −.169; see Table 4), which confirms H11. Also as expected, social class has a differential impact on anti-immigrant attitudes (H15). In terms of the direct effects, there is a significant gap in anti-immigrant attitudes between non-manual and manual workers. In the total effects it can be seen that anti-immigrant attitude is the only form of prejudice that is structured among class lines: The higher service class/self-employed and the non-manual workers report significantly less anti-immigrant sentiments than the manual workers (who serve as reference category in this analysis). Also religious involvement has differential impacts on homonegativity (H12) and anti-Semitism (H19). As expected, the gap between church-goers and marginal Christians is more outspoken regarding homonegativity and anti-Semitism than for other prejudices. This pattern illustrates the role of homosexuals and Jews as permitted targets of prejudice among church-involved Christians. In addition to the hypothesized effects, we also find a small but significant group-specific impact of age. In the case of homonegativity and Islamophobia, age has a direct positive effect (.178 and .183 respectively; see Table 4). The same pattern can also be seen in the total effects (Table 5): Homonegativity and Islamophobia are the only two forms of prejudice for which a significant age effect is found. Homosexuality and Islam are the only issues for which an opinion gap between younger and older cohorts is observed. Finally, the hypothesis regarding the specific effect of education on Islamophobia (H17) was not confirmed. Although the hypotheses tested here do not provide a direct test of the differentiated threat approach, this analysis indirectly renders clear support for its basic propositions. Our results provide a differentiated and nuanced view on the predictors of various prejudices. Besides a set of common antecedents, such as authoritarianism and educational level, the four forms of prejudice have quite different relationships to several predictors in the model. Furthermore, the pattern of group-specific antecedents largely confirms the key idea that prejudices are embedded in realistic and symbolic intergroup conflicts; prejudices toward a particular outgroup develop among specific socioeconomic (or cultural) segments of society who see their realistic (or symbolic) interest threatened. For homonegativity, this point is most obvious: The impact of 5 out of 8 predictors deviates significantly from the general pattern, and homonegativity is consistently stronger among the elderly, males, churchgoers and gender traditionalists—i.e., groups experiencing homosexuality as a stronger threat to the moral order. Also for Islamophobia, several differential effects are found that interestingly enough often go in to opposite direction as those for homonegativity. The fact that Islamophobia and homonegativity correlate positively (see section above titled “The Common Denominator of Prejudices”) does not exclude that both forms of prejudice have partially different social roots and are shaped by different processes. Anti-immigrant attitudes deviate least from the common explanatory model. This is hardly surprising, given the large overlap between anti-immigrant prejudice and GFE. The particularity of the antecedents’ prejudices is also evidenced by the sizeable differences in explanatory power of the predictors: The model explains 69.8 percent of the variation in anti-immigrant attitudes versus 28.3 percent in the case of anti-Semitism. CONCLUSION In this article, we argue that an exclusive focus on the generalized aspect of prejudices limits our understanding of the structure and genesis of particular prejudices. The realization of attitudes towards specific outgroups is contingent on structural and contextual factors, such as social positions, the nature of intergroup relations, power balances, and elite discourses. In order to conceptualize the outgroup specific nature of prejudices, we develop a theoretical framework—the differentiated threat approach—that hinges on two crucial principles. First, the model acknowledges that the interests that are challenged by outgroups are diverse (realistic versus symbolic threats), and that various types of outgroups—deviant, competing, and dissident—can be distinguished. Second, the differentiated threat approach stresses that societies are structurally and culturally stratified, and that perceived outgroup threats are not equally distributed across society, but disproportionally affect specific socioeconomic strata and cultural groups depending on the nature of conflict. As a result, individuals’ prejudices are not necessarily directed towards all possible outgroups, but are instead focused on outgroups that are perceived to challenge the realistic and/or symbolic interests of the relevant reference group. Paramount to the differentiated threat approach is that outgroup threats are embedded in intergroup relations and have to be interpreted within specific contexts. Hence, in-depth knowledge about the intergroup relations in particular settings are required. As such, the relevance of the differentiated threat approach was tested empirically by analysing the attitudes of Belgian majority-group members towards four outgroups, namely immigrants, Muslims, Jews, and sexual minorities. Our study shows that, although a common denominator of generalized prejudice can be distinguished, negative attitudes towards the various outgroups contain substantively relevant unique components. The degree to which a specific prejudice is linked with GFE varies across outgroups. A very strong overlap exists between GFE and anti-immigration attitudes, which suggests that in Belgian public opinion the category “immigrant” is the master template onto which processes of the formation of generalized prejudice are projected. Our analysis furthermore reveals that certain factors—such as low education, feelings of relative deprivation, and an authoritarian disposition—consistently contribute to generalized prejudice. But we also find that this general explanation model ignores that specific prejudices are partly triggered by particular sets of predictors. Clearly, there are particular reasons and motivations to be prejudiced idiosyncratic to the outgroup. The differential impact of gender traditionalism on Islamophobia and homonegativity illustrates this point perfectly. Gender traditionalism is the strongest factor inducing homonegativity, but tends to reduce anti-Islam feelings at the same time. The reason for this paradoxical effect should be sought in the specific content of prejudices towards homosexuals and Muslims. One of the most persistent stereotypes about homosexuality relates to the violation of traditional gender norms (including heteronormativity and patriarchy), while the critique that Islam is an inherently sexist religion oppressing women is precisely one of the constitutive elements of Islamophobia. Homosexuals as well as Muslims are disliked because they are perceived as a threat to certain values. But the specific value sets concerned and segments of society that feel threated are fundamentally different. Various other predictors where found to have an outgroup specific impact that deviates from the explanation model for generalized prejudice. Feelings of relative deprivation, for example, are more strongly related to Islamophobia than to other forms of prejudice, and are unrelated to homonegative views. Religious involvement plays a more decisive role in the formation of anti-Semitism and homonegativity than it does in the other forms of prejudice. And anti-immigration attitudes show a class gradient that is absent in attitudes towards other outgroups. Notably, the analyses showed that especially the origins of homonegativity deviated from those of GFE. As such, homonegativity can be considered as a rather “exceptional” case of prejudice. Looking at the position of LGBT in the Belgian context, it is not hugely surprising that homosexuality—as opposed to immigration or the increasing visibility of Islam—is more accepted by public opinion and almost a non-issue in political discourse. This finding once more stresses the necessity of a contextualized approach. The structure of prejudices and the classification of outgroups (into dissident, deviant, and competing groups) are not universal givens, but vary across societies and could undergo modifications over time. Zick and colleagues (2011), for instance, found that sexism and homophobia were much less associated with prejudice in Hungary and Poland, compared to more Western countries. Besides, prejudice towards certain outgroups might move away from the GFE component, while attitudes towards others might move into the center or even become the pinnacle of prejudice. This is illustrated by the finding that, due to the increased marketization of Western societies, prejudices toward unprofitable groups (e.g., unemployed and homeless) are claiming a more prominent position in the structure of prejudice (Hövermann et al. 2015). Our results confirm various hypotheses that were derived from the differentiated threat approach, and as such illustrate the necessity of a contextualized approach. These findings do not invalidate the theory of generalized prejudice, however. The generalized prejudice and differentiated threat approaches complement each other in the understanding of prejudice and excluding one from the other would provide a distorted and one-sided picture. As such, future research examining the origins of prejudice should try to differentiate between the group-specific and generalized component of prejudice and emphasize the contextualisation of the group-specific component. Furthermore, because the differentiated threat approach critically highlights the importance of context, future research should evaluate its theoretical underpinnings in different social settings. Doing so, future research could include examples of “competing” groups—a type of others that was not included in our empirical analysis. However, it cannot be overemphasized that studies using the differenitated threat approach need to be mindful that the outgroups that were analyzed in this article might hold different social positions and evoke different types of threat in other societies (for example, immigrants as a predominantly economic threat in Italy versus cultural threat in France; Bergamaschi 2013). Finally, although our approach has a strong focus on threat, it has shown that prejudice is not solely based on uncontrollable personality traits but largely depends on situational group membership and the search for a positive self-identity and personal/community security (Kinder and Kam 2010). Hence, it can easily be translated for the study of group identities, which is related to the study of prejudices. For example, the differentiated threat approach, and its focus on the position of groups vis-à-vis each other, can be used by social identity researchers interested in how the positionality of the different others to the ingroup may invoke differenitated othering mechanisms. Alternatively, social identity researchers can expand our differenitated threat model by adding the layer of social identitification to the intergroup positionality. The authors would like to thank the participants of the Leuven-Tilburg Sociology seminar as well as three anonymous reviewers for their helpful comments. This study was made possible by grants from KU Leuven research council (OT/13/30), the Belgian National Lottery, and the National Science Foundation FWO-Vlaanderen (Grant Number ZKC6622-Asp/13). APPENDIX Table A1 CFA Results for the Specific Prejudice Factors Homonegativity Anti-immigrant Attitudes Anti-Semitism Islamophobia Standardized factor loadings  q7_1 Children should learn that being homosexual is completely normal. −.67  q7_2 Equal rights for homosexuals is a threat for our norms and values. .83  q7_3 Homosexual men should not make such a fuss about their sexual orientation. .61  q7_4 Gay and straight people should be treated equally. −.70  q7_5 Lesbian women are abnormal. .79  q8_1 Migrants are generally not trustworthy. .83  q8_2 Migrants come here to profit from our social security model. .88  q8_3 Migrants are a threat to our culture and customs. .89  q8_4 The presence of different cultures is an enrichment of our society. −.59  q15_2 Jews have too much influence in our country. .70  q15_3 Jews think they are better than all others. .68  q15_4 One should avoid Jews; you cannot trust them. .88  q15_5 Jews are only after money. .88  q24_2 Muslim men dominate Muslim women. .55  q24_3 When it comes down, Islamic countries will turn against Europe. .75  q24_4 Islamic history and culture is more violent than others. .81  q24_5 Islamic values are a threat to Europe. .91 Correlations between factors    Homonegativity 1.00    Anti-immigrant attitudes .47 1.00    Anti-Semitism .29 .45 1.00    Islamophobia .31 .73 .33 1.00 Model fit: Chi-square: 622.03; Df: 113; RMSEA: .067; CFI: .967; TLI: .960 Homonegativity Anti-immigrant Attitudes Anti-Semitism Islamophobia Standardized factor loadings  q7_1 Children should learn that being homosexual is completely normal. −.67  q7_2 Equal rights for homosexuals is a threat for our norms and values. .83  q7_3 Homosexual men should not make such a fuss about their sexual orientation. .61  q7_4 Gay and straight people should be treated equally. −.70  q7_5 Lesbian women are abnormal. .79  q8_1 Migrants are generally not trustworthy. .83  q8_2 Migrants come here to profit from our social security model. .88  q8_3 Migrants are a threat to our culture and customs. .89  q8_4 The presence of different cultures is an enrichment of our society. −.59  q15_2 Jews have too much influence in our country. .70  q15_3 Jews think they are better than all others. .68  q15_4 One should avoid Jews; you cannot trust them. .88  q15_5 Jews are only after money. .88  q24_2 Muslim men dominate Muslim women. .55  q24_3 When it comes down, Islamic countries will turn against Europe. .75  q24_4 Islamic history and culture is more violent than others. .81  q24_5 Islamic values are a threat to Europe. .91 Correlations between factors    Homonegativity 1.00    Anti-immigrant attitudes .47 1.00    Anti-Semitism .29 .45 1.00    Islamophobia .31 .73 .33 1.00 Model fit: Chi-square: 622.03; Df: 113; RMSEA: .067; CFI: .967; TLI: .960 Table A1 CFA Results for the Specific Prejudice Factors Homonegativity Anti-immigrant Attitudes Anti-Semitism Islamophobia Standardized factor loadings  q7_1 Children should learn that being homosexual is completely normal. −.67  q7_2 Equal rights for homosexuals is a threat for our norms and values. .83  q7_3 Homosexual men should not make such a fuss about their sexual orientation. .61  q7_4 Gay and straight people should be treated equally. −.70  q7_5 Lesbian women are abnormal. .79  q8_1 Migrants are generally not trustworthy. .83  q8_2 Migrants come here to profit from our social security model. .88  q8_3 Migrants are a threat to our culture and customs. .89  q8_4 The presence of different cultures is an enrichment of our society. −.59  q15_2 Jews have too much influence in our country. .70  q15_3 Jews think they are better than all others. .68  q15_4 One should avoid Jews; you cannot trust them. .88  q15_5 Jews are only after money. .88  q24_2 Muslim men dominate Muslim women. .55  q24_3 When it comes down, Islamic countries will turn against Europe. .75  q24_4 Islamic history and culture is more violent than others. .81  q24_5 Islamic values are a threat to Europe. .91 Correlations between factors    Homonegativity 1.00    Anti-immigrant attitudes .47 1.00    Anti-Semitism .29 .45 1.00    Islamophobia .31 .73 .33 1.00 Model fit: Chi-square: 622.03; Df: 113; RMSEA: .067; CFI: .967; TLI: .960 Homonegativity Anti-immigrant Attitudes Anti-Semitism Islamophobia Standardized factor loadings  q7_1 Children should learn that being homosexual is completely normal. −.67  q7_2 Equal rights for homosexuals is a threat for our norms and values. .83  q7_3 Homosexual men should not make such a fuss about their sexual orientation. .61  q7_4 Gay and straight people should be treated equally. −.70  q7_5 Lesbian women are abnormal. .79  q8_1 Migrants are generally not trustworthy. .83  q8_2 Migrants come here to profit from our social security model. .88  q8_3 Migrants are a threat to our culture and customs. .89  q8_4 The presence of different cultures is an enrichment of our society. −.59  q15_2 Jews have too much influence in our country. .70  q15_3 Jews think they are better than all others. .68  q15_4 One should avoid Jews; you cannot trust them. .88  q15_5 Jews are only after money. .88  q24_2 Muslim men dominate Muslim women. .55  q24_3 When it comes down, Islamic countries will turn against Europe. .75  q24_4 Islamic history and culture is more violent than others. .81  q24_5 Islamic values are a threat to Europe. .91 Correlations between factors    Homonegativity 1.00    Anti-immigrant attitudes .47 1.00    Anti-Semitism .29 .45 1.00    Islamophobia .31 .73 .33 1.00 Model fit: Chi-square: 622.03; Df: 113; RMSEA: .067; CFI: .967; TLI: .960 Table A2 CFA Results for the Attitudinal Predictors of Prejudice Authoritarianism Gender Traditionalism Relative Deprivation Standardized factor loadings  q12_3 Most of our social problems could be solved, if we could somehow get rid of the immoral, crooked people. .63  q12_4 Obedience and respect for authority are the two most important virtues children have to learn. .72  q12_5 Laws should become stricter, because too much freedom is not good for people. .69  q14_1 It isn’t really as important for a girl to get a good education as it is for a boy. .81  q14_2 In general, boys can be brought up freer than girls. .72  q14_3 It’s unnatural if women give guidance to men in a company. .87  q14_4 It is the most natural thing for a man to be a breadwinner and for a women to take care of the household and the children. .70  q30_1 If we need something from the government, people like me have to wait longer than others. .87  q30_2 People like me are being systematically neglected, whereas other groups received more than they deserve. .94  q30_3 In times of economic crises people like me are always the first victims. .82 Correlations between factors    Authoritarianism 1.00    Gender traditionalism .37 1.00    Relative deprivation .42 .26 1.00 Model fit: Chi-square: 106.07; Df: 32; RMSEA: .048; CFI: .993; TLI: .990 Authoritarianism Gender Traditionalism Relative Deprivation Standardized factor loadings  q12_3 Most of our social problems could be solved, if we could somehow get rid of the immoral, crooked people. .63  q12_4 Obedience and respect for authority are the two most important virtues children have to learn. .72  q12_5 Laws should become stricter, because too much freedom is not good for people. .69  q14_1 It isn’t really as important for a girl to get a good education as it is for a boy. .81  q14_2 In general, boys can be brought up freer than girls. .72  q14_3 It’s unnatural if women give guidance to men in a company. .87  q14_4 It is the most natural thing for a man to be a breadwinner and for a women to take care of the household and the children. .70  q30_1 If we need something from the government, people like me have to wait longer than others. .87  q30_2 People like me are being systematically neglected, whereas other groups received more than they deserve. .94  q30_3 In times of economic crises people like me are always the first victims. .82 Correlations between factors    Authoritarianism 1.00    Gender traditionalism .37 1.00    Relative deprivation .42 .26 1.00 Model fit: Chi-square: 106.07; Df: 32; RMSEA: .048; CFI: .993; TLI: .990 Table A2 CFA Results for the Attitudinal Predictors of Prejudice Authoritarianism Gender Traditionalism Relative Deprivation Standardized factor loadings  q12_3 Most of our social problems could be solved, if we could somehow get rid of the immoral, crooked people. .63  q12_4 Obedience and respect for authority are the two most important virtues children have to learn. .72  q12_5 Laws should become stricter, because too much freedom is not good for people. .69  q14_1 It isn’t really as important for a girl to get a good education as it is for a boy. .81  q14_2 In general, boys can be brought up freer than girls. .72  q14_3 It’s unnatural if women give guidance to men in a company. .87  q14_4 It is the most natural thing for a man to be a breadwinner and for a women to take care of the household and the children. .70  q30_1 If we need something from the government, people like me have to wait longer than others. .87  q30_2 People like me are being systematically neglected, whereas other groups received more than they deserve. .94  q30_3 In times of economic crises people like me are always the first victims. .82 Correlations between factors    Authoritarianism 1.00    Gender traditionalism .37 1.00    Relative deprivation .42 .26 1.00 Model fit: Chi-square: 106.07; Df: 32; RMSEA: .048; CFI: .993; TLI: .990 Authoritarianism Gender Traditionalism Relative Deprivation Standardized factor loadings  q12_3 Most of our social problems could be solved, if we could somehow get rid of the immoral, crooked people. .63  q12_4 Obedience and respect for authority are the two most important virtues children have to learn. .72  q12_5 Laws should become stricter, because too much freedom is not good for people. .69  q14_1 It isn’t really as important for a girl to get a good education as it is for a boy. .81  q14_2 In general, boys can be brought up freer than girls. .72  q14_3 It’s unnatural if women give guidance to men in a company. .87  q14_4 It is the most natural thing for a man to be a breadwinner and for a women to take care of the household and the children. .70  q30_1 If we need something from the government, people like me have to wait longer than others. .87  q30_2 People like me are being systematically neglected, whereas other groups received more than they deserve. .94  q30_3 In times of economic crises people like me are always the first victims. .82 Correlations between factors    Authoritarianism 1.00    Gender traditionalism .37 1.00    Relative deprivation .42 .26 1.00 Model fit: Chi-square: 106.07; Df: 32; RMSEA: .048; CFI: .993; TLI: .990 REFERENCES Abts K. , Kochuyt T. . 2013 . “De Vreemde Bedreiging Van de Verzorgingsstaat.” Tijdschrift voor Sociologie 34 ( 3-4 ): 227 - 49 . Adorno Theodor W. , Frenkel-Brunswick Else , Levinson Daniel K. , Nevitt Sanford R. . 1950 . The Authoritarian Personality . New York : Harper . Aerts Saskia , Dewaele Alexis , Cox Nele , Van Houtte Mieke . 2014 . “Homonegativity in the Technical and Vocational Track: A Survey of Secondary School Students in Flanders.” Journal of LGBT Youth 11 4 : 364 - 87 . Google Scholar CrossRef Search ADS Agnew Christopher R. , Thompson Valda D. , Gaines Stanely O. . 2000 . “Incorporating Proximal and Distal Influences on Prejudice: Testing a General Model Across Outgroups.” Personality and Social Psychology Bulletin 26 4 : 403 - 18 . Google Scholar CrossRef Search ADS Akrami Nazar , Ekehammar Bo , Bergh Robin . 2011 . “Generalized Prejudice: Common and Specific Components.” Psychological Science 22 1 : 57 - 59 . Google Scholar CrossRef Search ADS PubMed Allport Gordon W. 1958 . The Nature of Prejudice . New York : Doubleday Anchor . Altemeyer Bob. 1998 . “The Other ‘Authoritarian Personality.’” Advances in Experimental Social Psychology 30 : 47 - 92 . Google Scholar CrossRef Search ADS Altemeyer Bob , Hunsberger Bruce . 1992 . “Authoritarianism, Religious Fundamentalism, Quest, and Prejudice.” International Journal for the Psychology of religion 2 : 113 - 33 . Google Scholar CrossRef Search ADS Asbrock Frank , Sibley Chris G. , Duckitt John . 2010 . “Right-Wing Authoritarianism and Social Dominance Orientation and the Dimensions of Generalized Prejudice: A Longitudinal Test.” European Journal of Personality 24 : 324 - 40 . Bäckström Martin , Björklund Fredrik . 2007 . “Structural Modeling of Generalized Prejudice: The Role of Social Dominance, Authoritarianism, and Empathy.” Journal of Individual Differences 28 : 10 - 17 . Google Scholar CrossRef Search ADS Bergamaschi Alessandro. 2013 . “Adolescents and Prejudice: A Comparative Study of the Attitudes of Two European Adolescent Populations Regarding the Issues that are Raised by Increasing Cultural and Religious Pluralism.” International Journal of Intercultural Relations 37 3 : 302 - 12 . Google Scholar CrossRef Search ADS Bierly Margaret M. 1985 . “Prejudice Toward Contemporary Outgroups as a Generalized Attitude.” Journal of Applied Social Psychology 15 2 : 189 - 99 . Google Scholar CrossRef Search ADS Bilsky Leora. 2009 . “Muslim Headscarves in France and Army Uniforms in Israel: A Comparative Study of Citizenship as a Mask.” Patterns of Prejudice 43 4 : 287 - 311 . Google Scholar CrossRef Search ADS Blalock Hubert M. 1967 . Toward a Theory of Minority-Group Relations . New York : John Wiley & Sons . Bleich Erik. 2011 . “What is Islamophobia and How Much is There? Theorizing and Measuring an Emerging Comparative Concept.” American Behavioral Scientist 55 12 : 1581 - 1600 . Google Scholar CrossRef Search ADS Blumer Herbert. 1958 . “Race prejudice as a Sense of Group Position.” The Pacific Sociological Review 1 : 3 - 7 . Google Scholar CrossRef Search ADS Bobo Lawrence. 1999 . “Prejudice as Group Position: Microfoundations of a Sociological Approach to Racism and Race Relations.” Journal of Social Issues 3 55 : 445 - 72 . Google Scholar CrossRef Search ADS Bobo Lawrence , Hutchings Vincent . 1996 . “Perceptions of Racial Group Competition: Extending Blumer’s Theory of Group Position to a Multiracial Social Context.” American Sociological Review 6 ( 61 ): 951 - 72 . Google Scholar CrossRef Search ADS Bourdieu Pierre. 1990 . The Logic of Practice . Cambridge, UK : Polity Press . Bourdieu Pierre. 1984 . Distinction: A Social Critique of the Judgement of Taste . London, UK : Routlegde . Bourdieu Pierre. 1985 . “The Social Space and the Genesis of Groups.” Theory and Society 6 14 : 423 - 44 . Bracke Sarah. 2012 . “From ‘Saving Women’ to ‘Saving Gays’: Rescue Narratives and their Dis/continuities.” European Journal of Women's Studies 19 2 : 237 - 52 . Google Scholar CrossRef Search ADS Bratt Christopher. 2005 . “The Structure of Attitudes Toward Non-Western Immigrant Groups: Second-Order Factor Analysis of Attitudes Among Norwegian Adolescents.” Group Processes & Intergroup Relations 8 4 : 447 - 69 . Google Scholar CrossRef Search ADS Choma Becky L. , Hodson Gordon . 2008 . “And So the Pendulum Swings: A Framework for Conceptualizing the Causes of Prejudice.” Pp. 1 - 25 in The Psychology of Modern Prejudice , edited by Melanie A. Morrison , Morrison Todd G. . New York : Nova Science Publishers . Connell R. W. 2005 . Masculinities . Cambridge, UK : Polity . Duckitt John. 2006 . “Differential Effects of Right-Wing Authoritarianism and Social Dominance Orientation on Outgroup Attitudes and their Mediation by Threat from and Competitiveness to Outgroups.” Personality and Social Psychology Bulletin 32 5 : 684 - 96 . Duckitt John , Sibley Chris G. . 2007 . “Right Wing Authoritarianism, Social Dominance Orientation, and the Dimensions of Generalized Prejudice.” European Journal of Personality 21 : 113 - 30 . Google Scholar CrossRef Search ADS Ekehammar Bo , Akrami Nazar , Gylje Magnus , Zakrisson Ingrid . 2004 . “What Matters Most to Prejudice: Big Five Personality, Social Dominance Orientation, or Right-Wing Authoritarianism?” European Journal of Personality 18 : 463 - 82 . Google Scholar CrossRef Search ADS Fekete Liz. 2009 . A Suitable Enemy: Racism, Migration, and Islamophobia in Europe . New York : Pluto Press . Fiske Susan T. , Cuddy Amy C. , Glick Peter , Xu Jun . 2002 . “A Model of (Often Mixed) Stereotype Content: Competence and Warmth Respectively Follow from Perceived Status and Competition.” Journal of Personality and Social Psychology 82 6 : 878 - 902 . Google Scholar CrossRef Search ADS PubMed Ganzeboom Harry , Treiman Donald J. . 1996 . “Internationally Comparable Measures of Occupational Status for the 1988 International Standard Classification of Occupations.” Social Science Research 25 : 201 - 39 . Google Scholar CrossRef Search ADS Hagendoorn Louk. 1995 . “Intergroup Biases in Multiple Group Systems: The Perception of Ethnic Hierarchies.” European Review of Social Psychology 6 : 199 - 228 . Google Scholar CrossRef Search ADS Herek Gregory M. 1988 . “Heterosexuals' Attitudes Toward Lesbians and Gay Men: Correlates and Gender Differences.” The Journal of Sex Research 25 4 : 451 - 77 . Google Scholar CrossRef Search ADS Hjerm Mikael , Nagayoshi Kikuko . 2011 . “The Composition of the Minority Population as a Threat: Can Real Economic and Cultural Threats explain Xenophobia.” International Sociology 26 6 : 815 - 43 . Google Scholar CrossRef Search ADS Hodson Gordon , Dhont Kristof . 2015 . “The Person-Based Nature of Prejudice: Individual Difference Predictors of Intergroup Negativity.” European Review of Social Psychology 26 1 : 1 - 42 . Google Scholar CrossRef Search ADS Hövermann Andrea , Gross Eva M. , Zick Andreas , Messner Steven F. . 2015 . “Understanding the Devaluation of Vulnerable Groups: A novel Application of Institutional Anomie Theory.” Social Science Research 52 : 408 - 21 . Google Scholar CrossRef Search ADS PubMed Hudson Walter W. , Ricketts Wendell A. . 1980 . “A Strategy for the Measurement of Homophobia.” Journal of Homosexuality 5 4 : 357 - 72 . Google Scholar CrossRef Search ADS PubMed Hunsberger Bruce , Jackson Lynne M. . 2005 . “Religion, Meaning, and Prejudice.” Journal of Social Issues 61 : 807 - 26 . Google Scholar CrossRef Search ADS Key Valdimer O. 1949 . Southern Politics in State and Nation . New York : Knopf . Kinder Donald. R. , Kam Cindy M. . 2010 . Us Against Them: Ethnocentric Foundations of American Opinion . Chicago : The University of Chicago Press . Koopmans Ruud. 2015 . “Religious Fundamentalism and Hostility Against Outgroups: A Comparison of Muslims and Christians in Western Europe.” Journal of Ethnic and Migration Studies 44 1 : 33 - 57 . Google Scholar CrossRef Search ADS Kumar Deepa. 2012 . Islamophobia and the Politics of Empire . Chicago : Haymarket Books . Kunovich Robert M. 2004 . “Social Structural Position and Prejudice: An Exploration of Cross-National Differences in Regression Slopes.” Social Science Research 33 1 : 20 - 44 . Google Scholar CrossRef Search ADS Meer Nasar. 2013 . “Semantics, Scales and Solidarities in the Study of Antisemitism and Islamophobia.” Ethnic and Racial Studies 36 3 : 500 - 515 . Google Scholar CrossRef Search ADS Merton Robert K. [1957] 1968 . Social Theory and Social Structure . New York : The Free Press . Meuleman B. , Davidov E. , Billiet J. . 2009 . “Changing Attitudes Toward Immigration in Europe, 2002-2007: A Dynamic Group Conflict Theory Approach.” Social Science Research 38 2 : 352 - 65 . Google Scholar CrossRef Search ADS PubMed Olzak Susan. 1994 . The Dynamics of Ethnic Competition and Conflict . Stanford, CA : Stanford University Press . Pargament Kenneth I. , Trevino Kelley , Mahoney Anette , Silberman Israela . 2007 . “They Killed our Lord: The Perception of Jews as Desecrator of Christianity as a Predictor of Anti-Semitism.” Journal of the Scientific Study of Religion 46 2 : 143 - 58 . Google Scholar CrossRef Search ADS Plummer David. 1999 . One of the Boys: Masculinity, Homophobia, and Modern Manhood . New York : Harrington Park Press . Podsakoff Philip M. , MacKenzie Scott B. , Lee Jeong-Yeon , Podsakoff Nathan P. . 2003 . “Common Method Biases in Behavioral Research: A Critical Review of the Literature and Recommended Remedies.” Journal of Applied Psychology 88 5 : 879 - 903 . Google Scholar CrossRef Search ADS PubMed Ray John J. , Lovejoy F. H. . 1986 . “The Generality of Racial Prejudice.” The Journal of Social Psychology 126 4 : 563 - 64 . Google Scholar CrossRef Search ADS Riek Blake M. , Mania Eric. W. , Gaertner Samuel L. . 2006 . “Intergroup Threat and Outgroup Attitudes: A Meta-Analytic Review.” Personality and Social Psychology Review 10 4 : 336 - 53 . Google Scholar CrossRef Search ADS PubMed Runciman Walter G. 1966. Relative Deprivation and Social Justice: A Study of Attitudes to Social Inequality in Twentieth Century England . Berkeley : University of California Press . Schiefer David. 2013 . “Cultural Values and Group-Related Attitudes: A Comparison of Individuals with and without Migration Background Across 24 Countries.” Journal of Cross-Cultural Psychology 44 2 : 245 - 62 . Google Scholar CrossRef Search ADS Schiffer Sabine , Wagner Constantin . 2011 . “Anti-Semitism and Islamophobia—New Enemies, Old Patterns.” Race & Class 52 : 377 - 84 . Google Scholar CrossRef Search ADS Schneider Anne , Ingram Helen . 1993 . “Social Construction of Target Populations: Implications for Politics and Policy.” American Political Science Review 87 02 : 334 - 47 . Google Scholar CrossRef Search ADS Spruyt Bram , Elchardus Mark . 2012 . “Are Anti-Muslim Feelings More Widespread than Anti-Foreigner Feelings? Evidence from Two Split-Sample Experiments.” Ethnicities 12 6 : 800 - 20 . Google Scholar CrossRef Search ADS Stephan Walter G. , Ybarra Oscar , Martínez Martínez Carmen , Schwarzwald Joseph , Tur-Kaspa Michal . 1998 . “Prejudice Toward Immigrants to Spain and Israel: An Integrated Threat Theory Analysis.” Journal of Cross-Cultural Psychology 29 4 : 559 - 76 . Google Scholar CrossRef Search ADS Strabac Zan , Listhaug Ola . 2008 . “Anti-Muslim Prejudice in Europe: A Multi-Level Analysis of Survey Data from 30 Countries.” Social Science Research 37 : 268 - 86 . Google Scholar CrossRef Search ADS Swyngedouw Marc. 1995 . “The ‘Threatening Immigrant’ in Flanders 1930-1980: Redrawing the Social Space.” New Community 21 3 : 325 - 40 . Taras Ray. 2013 . “ ‘Islamophobia Never Stands Still’: Race, Religion, and Culture.” Ethnic and Racial Studies 36 3 : 417 - 33 . Google Scholar CrossRef Search ADS Thornton Arland , Alwin Duane F. , Camburn Donald . 1983 . “Causes and Consequences of Sex-Role Attitudes and Attitude Change.” American Sociological Review 48 2 : 211 - 27 . Google Scholar CrossRef Search ADS PubMed Whitley Bernard E. 2009 . “Religiosity and Attitudes Towards Lesbians and Gay Men: A Meta-analysis.” The International Journal for the Psychology of Religion 19 1 : 21 - 38 . Google Scholar CrossRef Search ADS Zick Andreas , Küpper Beate , Hövermann Andreas . 2011 . Intolerance, Prejudice and Discrimination: A European Report . Berlin, Germany : Friedrich-Ebert-Stiftung . Zick Andreas , Wolf Carina , Küpper Beate , Davidov Eldad , Schmidt Peter , Heitmeyer Wilhelm . 2008 . “The Syndrome of Group-Focused Enmity: The Interrelation of Prejudices Tested with Multiple Cross-Sectional and Panel Data.” Journal of Social Issues 64 2 : 363 - 83 . Google Scholar CrossRef Search ADS © The Author(s) 2018. Published by Oxford University Press on behalf of the Society for the Study of Social Problems.All rights reserved. For permissions, please email: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Social Problems Oxford University Press

Differentiated Threat and the Genesis of Prejudice: Group-Specific Antecedents of Homonegativity, Islamophobia, Anti-Semitism, and Anti-Immigrant Attitudes

Loading next page...
 
/lp/ou_press/differentiated-threat-and-the-genesis-of-prejudice-group-specific-E9tij8C0Sv
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the Society for the Study of Social Problems.All rights reserved. For permissions, please email: journals.permissions@oup.com
ISSN
0037-7791
eISSN
1533-8533
D.O.I.
10.1093/socpro/spy002
Publisher site
See Article on Publisher Site

Abstract

Abstract In this article, we argue that an exclusive focus on the generalized aspect of prejudice limits understanding of the structure and genesis of prejudice towards particular outgroups. In order to conceptualize the specific nature of particular prejudices, we propose the differentiated threat approach. This framework postulates that different outgroups challenge diverse realistic and symbolic interests, and that these outgroup specific threats affect various socioeconomic strata and cultural groups differentially. The differentiated threat approach is applied to analyse majority-group Belgians’ attitudes towards immigrants, Muslims, Jews, and homosexuals. The results show that a common denominator of prejudice can be distinguished, but that the prejudices towards the various outgroups contain substantively relevant unique components that are influenced by socio-demographic and attitudinal predictors in diverging ways. Gender traditionalism is found to reinforce Homonegativity and temper Islamophobia at the same time. Feelings of relative deprivation are more strongly related to Islamophobia than to other forms of prejudice, and are unrelated to homonegativity. Religious involvement plays a more decisive role in the formation of anti-Semitism and Homonegativity than it does in the other forms of prejudice. Anti-immigration attitudes show a class gradient that is absent in attitudes towards other outgroups. Our results evidence that the concrete realization of attitudes towards a specific outgroup cannot be understood without paying attention to structural and contextual factors, such as social positions, the nature of intergroup relations, power balances, and elite discourses. generalized prejudice, anti-Semitism, homonegativity, Islamophobia, differentiated threat approach In early research on intergroup relations, scholars put forward the idea that negative feelings tend to generalize across various outgroups (Allport 1958; Adorno et al. 1950). Numerous studies have evidenced how this so-called “generalized prejudice” (Akrami, Ekehammar, and Bergh 2011; Bäckström, and Björklund 2007; Duckitt and Sibley 2007) or “group-focused enmity” (Zick et al. 2008) is related to personality traits and ideological orientations such as right-wing authoritarianism and social dominance orientation (Asbrock, Sibley, and Duckitt 2010; Ekehammar et al. 2004). The discovery of a common denominator in negative attitudes towards various outgroups has contributed to our knowledge of intergroup relations. In this article, however, we argue that the exclusive focus on the common denominator of prejudices tends to obscure the particular causes of prejudice towards specific outgroups. Our argument is that—while the formation of general negative outgroup attitudes takes place via more or less universal social-psychological mechanisms—the concrete realization of attitudes towards particular outgroups is crucially dependent on both structural and contextual factors (see Akrami et al. 2011). Social positions of ingroup individuals and their sense of group position, intergroup relations characterized by differential social contact, socioeconomic competition and power balances, as well as available media and elite discourses might explain why prejudice is triggered toward one outgroup but not towards others (Hagendoorn 1995). Although this is acknowledged in the literature (Agnew, Thompson, and Gaines 2000; Ray and Lovejoy 1986; Zick et al. 2008), a systematic theoretical elaboration and empirical test of the specificity of prejudices towards various outgroups is still lacking. In this article, we develop a theoretical framework—the differentiated threat approach—that conceptualizes how different outgroups may be prejudiced against for different reasons by different strata of the ingroup. Subsequently, we provide an empirical application of this approach by engaging with two generalizability questions: (1) To what extent do negative attitudes generalize across different outgroups? Can we ascertain generalized prejudice or do individuals hold differentiated views regarding various social outgroups? (2) Can a single general causal model explain negative attitudes towards various outgroups or are specific prejudices triggered by particular sets of predictors, as the differentiated threat approach predicts? These questions are answered by articulating them in the concrete intergroup context of Belgium, where attitudes of Belgian majority-group members towards four outgroups—namely immigrants, Muslims, Jews and sexual minorities—are scrutinized. THEORY AND HYPOTHESES Generalized Prejudice: Personality and the Common Denominator of Prejudices A dominant paradigm in social-psychological studies of intergroup relations stresses that the inclination to hold prejudices generalizes across outgroups. This idea of “generalized prejudice” was already present in Gordon Allport’s (1958) seminal study, where he noted that “people who reject one outgroup will tend to reject other outgroups. If a person is anti-Jewish, he is likely to be anti-Catholic, anti-Negro, anti any outgroup” (p. 66). More recently, Andreas Zick and colleagues (2008) introduced the concept of “group-focused enmity” (GFE) to refer to a generalized devaluation of outgroups with an ideology of inequality as its common core. Hostile attitudes toward different outgroups, who are considered to be unequal, socially threatening, and/or culturally deviant, are understood as the components of a single syndrome, namely GFE (see also Zick et al. 2008). Over the years, numerous studies have accumulated empirical evidence for this hypothesis. Prejudices towards very diverse social groups—including various ethnic and religious groups, but also women, sexual minorities, elderly people, disabled persons, and homeless people—are indeed found to correlate very strongly (Agnew et al. 2000; Altemeyer 1998; Bierly 1985; Bratt 2005; Ekehammar et al. 2004; Zick et al. 2008). The GFE approach not only implies a stable structure of interrelated prejudices, but also argues that the various components are triggered by a single set of antecedents. Explanations why individuals are so consistent in their acceptance or rejection of various outgroups are often sought in person-based variables such as personality traits, ideological dispositions, and cognitive styles (Hodson and Dhont 2015). Theodor Adorno and colleagues (1950), for example, maintained that general outgroup negativity is rooted in an authoritarian personality structure. Building on this theoretical claim, contemporary scholars argue that individuals with high levels of right-wing authoritarianism tend to reject outgroups they consider as a threat to the social-cultural order—namely ingroup values, norms, stability, and security. Individuals with strong social dominance orientations, on the other hand, are more inclined to devalue groups of perceived lower social status (Duckitt and Sibley 2007). Similarly, generalized prejudice has been linked to Big Five personality traits like agreeableness and openness to experience (Ekehammar et al. 2004). Personality is not the only factor put forward to explain generalized prejudice, however. Low educated, unemployed, and non-manual workers, men, and elderly persons are generally more prone to hostility towards outgroups (Altemeyer 1998; Schiefer 2013; Zick et al. 2011). Also, religious involvement is claimed to play a—albeit somewhat paradoxical—role in the genesis of generalized prejudice. Christian fundamentalism is positively related to feelings of GFE (Koopmans 2015), but other measures of Christian religiosity, such as church attendance, are only weakly associated with general outgroup hostility (Altemeyer and Hunsberger 1992). In summary, the generalized prejudice approach (1) focuses primarily on the common component of generalized prejudice rather than on the particularity of the specific components; and (2) seeks the origins of generalized prejudice in characteristics of the individual, such as cognitive capacities, personality structures, and socio-demographics, rather than in the social context of specific intergroup relations and features of the target group. The Differentiated Threat Approach: Social Context and the Outgroup Specificity of Prejudice An exclusive focus on the generalized character of prejudice limits our understanding of prejudice. Which outgroups become subject to derogation and negative imagery depends crucially on “the options a specific society offers” (Zick et al. 2008:367). Although early (Allport 1958) and more recent (Fiske et al. 2002) social-psychological studies have recognised the importance of situational factors for understanding prejudice (but see also Choma and Hodson 2008), several sociological theories stressed the importance of more insight in the structural and contextual factors of prejudice—including patterns of social contact and competition, economic and power relations, as well as media and elite discourses positions—in order to understand why negative attitudes towards specific (and not all) outgroups are triggered (see, for example, Agnew et al. 2000; Akrami et al. 2011; Ray and Lovejoy 1986). Herbert Blumer’s (1958) concept of “sense of group position” explicitly stresses the paramount importance of the concrete contexts and social structures in which intergroup relations are shaped: “[race] prejudice exists basically in a sense of group position rather than in a set of feelings which members of one racial group have toward the members of another racial group. This different way of viewing race prejudice shifts study and analysis from a preoccupation with feelings as lodged in individuals to a concern with the relationship of [racial] groups” (p. 3). This is to say, prejudice is a way to describe, justify, and understand one’s own structural and group position in relation to that of outgroup members. The development of prejudices, according to this logic, is fundamentally a social and collective process originating from dominant group members’ perception that their proprietary claims over certain resources and privileges as well as their shared beliefs and social hierarchies are being challenged by a subordinate group. In a similar vein, scholars like Lawrence Bobo (1999; Bobo and Hutchings 1996) and Pierre Bourdieu (1985, 1990) characterize social space as a universe of ongoing struggle between individuals and groups, making clear that it is meaningless to study particular expressions of prejudice without reference to the specific context of relative group positions and intergroup struggles within different social fields. Such a relational perspective posits prejudices as structured by the ingroup’s relative positioning vis-à-vis outgroups in the power configuration and their ability to mobilize their capital (social, cultural, economic, and/or symbolic) to sustain or improve status (Bobo 1999; Bourdieu 1984). As such, the particular characteristics of the relations between two groups facilitate certain prejudices and inhibit others. Similarly, pioneers of group conflict theory (GCT), such as Valdimer Key (1949) and Hubert Blalock (1967), observe that anti-Black prejudice and discrimination were more prevalent in U.S. regions where blacks made up considerable proportions of the resident population. This led to the idea that prejudices towards outgroups are essentially rooted in the perception that they threaten prerogatives of the own social group (Meuleman, Davidov, and Billiet 2009). In this logic, prejudice is seen as a defensive reaction to a sense of threat caused by conflicts between two or more social groups. This approach emphasizes the role of social comparisons—essentially not egoistic, but fraternal or group relative deprivation—to explain the genesis of prejudice. Individuals who perceive that their ingroup is worse off or disadvantaged in comparison to a relevant outgroup are more likely to express prejudice toward that group (Merton [1957] 1968; Runciman 1966). In order to investigate the outgroup specific antecedents of prejudice, the general idea that outgroup threats cause prejudice among the dominant population needs to be qualified in two ways—we call this the differentiated threat approach: (1) the type of perceived threat is idiosyncratic to specific intergroup relations; and (2) outgroup threats do not affect a society as a whole, but disproportionally affect specific socioeconomic strata and cultural segments of the ingroup. Consequently, the differentiated threat approach—as a theoretical lens—can only be understood and articulated with explicit reference to the specific social contexts. The first component of the differentiated threat approach acknowledges that the nature of the perceived conflict/threat can be very diverse. Outgroups can challenge the collective interest, identity, and power status of the ingroup, resulting in different types of threat (Riek, Mania, and Gaertner 2006). On the one hand, realistic or socioeconomic threat is induced by the (subjective) experience of relative deprivation, caused by competition for material goods such as well-paid jobs, affordable housing, or resources of the welfare state (Olzak 1994). Symbolic or cultural threat, on the other hand, originates in intergroup conflict over the established social order, cultural traditions, and shared beliefs, norms, and values (Stephan et al. 1998). Importantly, distinct outgroups can be perceived either as a realistic threat, a symbolic threat, or a combination of both (Hjerm and Nagayoshi 2011). Based on the nature of perceived threat (realistic versus symbolic) posed by outgroups, various types of “others” can be distinguished (see Figure 1): (1) deviant groups (low realistic threat, high symbolic threat): non-subordinated outgroups who are violating the social norms and threatening the established social-cultural order and societal stability. These groups are perceived as socially/culturally deviant, but not struggling for collective resources and not necessarily of lower social status (e.g., LGBT, Ku Klux Klan members, or atheists in the United States); (2) competing groups (high realistic threat, low symbolic threat): socially subordinated outgroups with low status and power who compete actively for scarce resources and challenge social inequalities, but not culturally deviant. These groups can be classified as disadvantaged norm-conformists since they are not challenging the values of the ingroup, but the distribution of collective resources (e.g., unemployed or homeless people); and (3) dissident groups (high realistic threat, high symbolic threat): outgroups seen as competing with but also threatening the ingroup since they tend to struggle actively for scarce resources and are challenging the ingroup’s values and norms at the same time. These groups could be described as usurpers trying to seize status and power by claiming scarce resources and shifting established cultural order (e.g., immigrants). Figure 1. View largeDownload slide Typology of Outgroups According to the Nature of Threat Posed Figure 1. View largeDownload slide Typology of Outgroups According to the Nature of Threat Posed This categorization of groups bears resemblance to Anne Schneider and Helen Ingram’s (1993) theory on the social construction of target groups that distinguishes groups on the basis of their power resources and the valence of the social constructions (positive versus negative). In line with the dual-process motivational model (Duckitt 2006) and the stereotype content model (Fiske et al. 2002), we argue that different types of outgroups may be prejudiced against for distinct reasons, according to different group-specific mechanisms such as the nature of the perceived threat and the content of ascribed stereotypes. Prejudices against outgroups that are symbolically threatening (e.g., deviant groups) are predominantly driven by authoritarian dispositions. Feelings of relative deprivation, on the other hand, are expected to induce prejudices towards socially subordinate groups competing for scarce resources (e.g., competing groups) (Asbrock et al. 2010). Similarly, Susan Fiske and colleagues (2002) show that prejudices cluster along two dimensions, namely “perceived competence” defined by the relative socioeconomic status of groups (low versus high status) and “perceived warmth” defined by the type of interdependence between groups (competitive versus cooperative). As a consequence, groups that are perceived to be similar on stereotype content are targeted for similar reasons. The second component of the differentiated threat approach is its focus on the social stratification and cultural diversity of the dominant society. The dominant group cannot be reduced to a homogenous bloc, but is instead segmented into different socioeconomic classes (with distinct risks and resources) and cultural groups (with distinct value patterns and ideological dispositions). Therefore, intergroup competition for realistic and symbolic goods does not affect all groups in society to the same degree. Realistic threats emanating from specific outgroups will only affect certain socioeconomic segments of the dominant group, while symbolic threats posed by distinct outgroups will only appeal to some cultural sections of dominant society. Conversely, prejudiced attitudes are not necessarily directed towards all possible outgroups, but are rather focused on those outgroups that are socially constructed as a challenge to realistic and/or symbolic interests (Schneider and Ingram 1993). The differentiated threat approach can thus be translated into a two general propositions: (1) socioeconomic segment A of the dominant society will develop prejudice towards outgroup B if outgroup B threatens the realistic interests of segment A; (2) cultural segment C of the dominant society will develop prejudice towards outgroup D if outgroup D threatens the symbolic interests of segment C. Below, these two propositions are translated into testable hypotheses regarding prejudices towards four outgroups in Belgium. Contextualizing Specific Prejudices: The Case of Belgium The quintessence of the differentiated threat approach is that prejudices cannot be understood in abstracto, but instead they need to be situated in social space (cf. Bourdieu 1985), i.e., the specific structural positions, competitive contexts, and cultural-discursive contexts in which intergroup relations are embedded. The general propositions of the differentiated threat approach cannot be tested directly but provide an operational lens by which prejudice can be studied in a particular social space. This contribution illustrates how this can be done by applying the theoretical framework of the differentiated threat approach to the particular context of Belgium. We focus on prejudices towards four outgroups that figure prominently in current debates in Belgium: immigrants, Muslims, homosexuals, and Jews. Below, we contextualize the societal positions of the four groups and derive hypotheses on the group-specific antecedents of prejudice that indirectly allow us to assess the basic premises of the differentiated threat approach. In the Belgian context the notion of “immigrants” is predominantly associated with Turkish and Moroccan communities that have their roots in the post-war labor migration. The popular use of the term immigrants often also includes the second and third generation descendants of these labor migrants. Persons of Turkish and Moroccan descent generally occupy disadvantaged socioeconomic positions and are often perceived as a threat for low-skilled jobs and social welfare provisions of the native Belgians (Abts and Kochuyt 2013). As a culturally visible and distinct outgroup, immigrants of Turkish and Moroccan origin are perceived as a threat to the established social-cultural order as well (Swyngedouw 1995). In this sense, this target group can be described as “dissident”: both competing with low-skilled natives for resources and status, and threatening the ingroup’s norms and values. Although many persons of Turkish and Moroccan descent are Muslim, it is of crucial importance to distinguish anti-immigration attitudes from a second form of prejudice, namely Islamophobia. Islamophobia refers to indiscriminate negative attitudes or hostility toward Islam and Muslims as a religious community (Bleich 2011). In Western Europe, since 9/11 the supposed historical incompatibility of European and Islamic values is central to the rise of Islamophobia (Strabac and Listhaug 2008). Anti-Islam discourse depicts Islam as a monolithic, inherently violent, and uniquely sexist religion, whose followers are seen as the ultimate cultural “other” that will never be able to cope with democratic and liberal values of Western society (Kumar 2012; Taras 2013). Islam as a denomination and Muslims as a religious-cultural group are in the eyes of prejudiced individuals a cultural rather than an economic threat, and are therefore considered as a “deviant” group. Third, homonegativity refers to the degree of prejudicial biases towards gay, lesbian, bisexual, and transsexual (LGBT) people (Hudson and Ricketts 1980). Although Belgium is known as a frontrunner in terms of LGBT rights and the social climate has become more positive in the last decade (Aerts et al. 2014), homonegativity is still present in Belgian society. The core of homonegative discourse is that homosexuality poses a threat to the social-cultural order: Homosexuals are a “deviant group,” threatening the Christian value system, violating traditional morality, and challenging conventional gender role patterns. Fourth, anti-Semitism refers not to hostility toward Israel, but to the cultivation of resentments against an imagined and generalized “collective Jew” who is a cultural other that does not make any effort to assimilate, and simultaneously is a powerful other with a strong grip on the economy and competing with domestic actors—thus “dissident” (Meer 2013; Schiffer and Wagner 2011). Yet, the nature of dissidence ascribed to Jews is quite specific. Most dissident groups are precisely blamed for not contributing enough to the economy and being in competition with low-skilled natives for scarce resources (e.g., the Turkish and Moroccan communities). The prototypical ultra-orthodox Haredi Jews (who are strongly concentrated in Antwerp and dominate its diamond industry) are contrarily blamed for trying to control domestic economy and world finances. As such, whereas immigrant communities are dissidents of low social status, Jews are imagined as dissidents of high status (Fiske et al. 2002). Note that our selection of outgroups does not cover all theoretically possible types of others (see Figure 1) since our study comprises two deviant (homosexuals and Muslims) and two dissident (immigrants and Jews) groups. Although it does not include an example of competing others, our study contains sufficient variety in intergroup contexts to illustrate the differentiated threat approach empirically. This contextualization, combined with the propositions of the differentiated threat approach, leads us to expect that the different forms of prejudice are triggered—at least partially—by group-specific antecedents. The first proposition of the differentiated threat approach relates to the axis of realistic threat and stipulates that socioeconomic segments of dominant society will develop prejudice towards outgroups that threaten their specific realistic interests. Since immigrants as well as Jews are seen as an economic threat—albeit each in their distinct way—people with strong feelings of relative deprivation are especially expected to develop prejudice towards these groups. We anticipate relative deprivation to be a more relevant antecedent for anti-immigrant attitudes and anti-Semitism than for the other forms of prejudice like homonegativity or Islamophobia, which are predominantly rooted in perceptions of cultural threat. However, the socioeconomic status and hence nature of the economic competition posed by immigrants and Jews is radically different. As a result of the specific form of ethno-stratification in Belgium, low-skilled natives experience persons of Turkish and Moroccan descent as competitors for scarce economic goods (especially jobs, housing, and welfare provisions). As a result, anti-immigration attitudes are expected to crystallize along class lines, with low-skilled manual workers and the unemployed showing higher levels of anti-immigrant attitudes (Kunovich 2004). In the case of anti-Semitism, such class gradient is not expected. The second proposition of the differentiated threat approach relates to the symbolic axis, and states that the dominant cultural group will develop prejudice towards outgroups that threaten their specific symbolic interests. Islamophobia and homonegativity are both embedded in cultural rather than in economic threat perceptions. Muslims as well as sexual minorities are perceived as a threat to social order and dominant value schemes, but the specific sets of values these outgroups challenge are far from identical. As a result, Islamophobia and homonegativity are triggered according to group-specific mechanisms, albeit in quite different social groups characterized by different value patterns. The relationship between traditional gender role attitudes (Thornton, Alwin, and Camburn 1983) and the formation of prejudices exemplifies this crucial point. Regarding sexual minorities, one of the most persistent stereotypes about homosexuality relates to the violation of (patriarchal) gender norms (Aerts et al. 2014). Consequently, homonegativity is expected to be more prevalent among individuals identifying strongly with traditional gender roles. Similarly, males generally hold stronger homonegative beliefs than women (Herek 1988). This gender dimension of homonegativity is not just an effect of sex per se, but is closely related to the gender role expectations within society, especially the hegemonic (traditional) conceptions of masculinity (Connell 2005; Plummer 1999). Islamophobia is linked to threats to gender role expectations as well, but with different (if not opposite) underlying values and mechanisms. The idea that Islam is an inherently sexist religion is one of the constituting myths of Islamophobic discourse (Kumar 2012). Because the critique that Islam oppresses women resonates among persons who endorse gender equality, negative attitudes against Muslims are especially strong among persons rejecting traditional gender norms, which imply male domination and patriarchy. The paradoxical situation that gender traditionalism reinforces homonegativity and at the same time tempers Islamophobia is also apparent from discourse of political elites. Extreme right-wing political actors regularly mobilize the frame of the “homonegative Muslim community” in order to exhort anti-Islam sentiments (Bracke 2012). Besides gender equality, Islam is also perceived as a threat for certain value sets—such as separation of church and state, individual rights/freedom, democracy, and tolerance—that are strongly endorsed by the higher educated. In the eyes of this presumably tolerant group, Islam skepticism becomes legitimate as a defense of tolerance, liberal democracy, and Enlightenment (Bilsky 2009; Spruyt and Elchardus 2012). In other words, the “Muslim other” is transformed into a “suitable enemy,” i.e., a more legitimate object of prejudice (Fekete 2009). Hence, highly educated citizens might be more likely to hold negative attitudes towards Muslims than towards other outgroups, like immigrants, Jews, or homosexuals. The impact of religiosity on group-specific prejudices can be understood from the framework of cultural threats, but is highly complex. Allport (1958:444) already asserted that religion can both make and unmake prejudice. Although most religions preach tolerance toward outgroups, they permit some forms of prejudice against people who are perceived to violate the religion’s value system. A literature review supports the distinction between proscribed prejudices (race/ethnicity) and permitted prejudices (homosexuality) (Hunsberger and Jackson 2005). Following the Catholic teaching that identifies homosexuality as opposing religious morality and their value system, the Church enables prejudice against LGBT, as it is consistent with religious beliefs (Whitley 2009). Similarly, the portrayal of the role of Jews in the crucifixion of Jesus Christ or, more generally, their threat to Christianity might promote anti-Semitism among church-involved Christians (Pargament et al. 2007). Hypotheses Based on the generalized prejudice and differentiated threat approaches, we can derive two sets of hypotheses regarding the commonality (H1-H9) and specificities (H10-H20) of prejudice towards immigrants, homosexuals, Muslims, and Jews (see Table 1 for a summary). According to social-psychological theory, a strong common denominator—namely GFE or generalized prejudice—can be distinguished in prejudices towards the four outgroups (H1). The respective prejudices are anticipated to be shaped by the same antecedents. Higher levels of generalized prejudice are expected among men (H2), elderly persons (H3), lower educated individuals (H4), lower social classes (H5), and religious persons (H6). Authoritarianism (H7), feelings of relative deprivation (H8), and gender traditionalism (H9) are hypothesized to reinforce generalized prejudice. Table 1. Overview of Hypotheses Regarding the Common and Specific Determinants of Prejudices General Pattern Deviations from the General Pattern GFE Homonegativity Anti-immigrant Attitudes Islamophobia Anti-Semitism Gender (being male) + (H2) + (H11) Old age + (H3) Low education + (H4) − (H17) Low social class + (H5) + (H15) Strong religious involvement + (H6) + (H12) + (H19) Authoritarianism + (H7) Relative deprivation + (H8) − (H13) + (H16) + (H20) Gender traditionalism + (H9) + (H14) − (H18) General Pattern Deviations from the General Pattern GFE Homonegativity Anti-immigrant Attitudes Islamophobia Anti-Semitism Gender (being male) + (H2) + (H11) Old age + (H3) Low education + (H4) − (H17) Low social class + (H5) + (H15) Strong religious involvement + (H6) + (H12) + (H19) Authoritarianism + (H7) Relative deprivation + (H8) − (H13) + (H16) + (H20) Gender traditionalism + (H9) + (H14) − (H18) Table 1. Overview of Hypotheses Regarding the Common and Specific Determinants of Prejudices General Pattern Deviations from the General Pattern GFE Homonegativity Anti-immigrant Attitudes Islamophobia Anti-Semitism Gender (being male) + (H2) + (H11) Old age + (H3) Low education + (H4) − (H17) Low social class + (H5) + (H15) Strong religious involvement + (H6) + (H12) + (H19) Authoritarianism + (H7) Relative deprivation + (H8) − (H13) + (H16) + (H20) Gender traditionalism + (H9) + (H14) − (H18) General Pattern Deviations from the General Pattern GFE Homonegativity Anti-immigrant Attitudes Islamophobia Anti-Semitism Gender (being male) + (H2) + (H11) Old age + (H3) Low education + (H4) − (H17) Low social class + (H5) + (H15) Strong religious involvement + (H6) + (H12) + (H19) Authoritarianism + (H7) Relative deprivation + (H8) − (H13) + (H16) + (H20) Gender traditionalism + (H9) + (H14) − (H18) The differentiated threat approach, on the other hand, predicts that prejudices towards immigrants, homosexuals, Muslims, and Jews cannot be reduced completely to a single common denominator and expects considerable particularities for the different prejudices (H10). Furthermore, the four prejudices are assumed to originate along specific mechanisms. Note that the hypotheses regarding group-specific antecedents are formulated in terms of deviations from the general pattern (i.e., the explanatory model for GFE). Homonegativity: Gender is expected to affect homonegativity even more strongly than GFE. While men are more prejudiced than women towards outgroups in general (see H2), the gender gap is expected to be more outspoken when homonegativity is concerned (H11). Similarly, religious involvement is anticipated to have a stronger impact on homonegativity (H12), while feelings of relative deprivation are assumed to be less relevant when explaining negative attitudes towards LGBT people (H13). Homonegativity is expected to be more outspoken among gender traditionalists (H14). Anti-immigrant attitudes: Belonging to the lower social classes (H15) and relative deprivation (H16) are expected to stimulate anti-immigrant sentiments more strongly than generalized prejudice. Islamophobia: Whereas a low educational level is supposed to be a strong predictor of general prejudice (see H4), we expect a less outspoken educational gradient for Islamophobia (H17). Furthermore, gender traditionalists are hypothesized to show lower levels of Islamophobia (H18). Anti-Semitism: Religious involvement (H19) as well as relative deprivation (H20) are assumed to have a stronger positive effect/relation with on anti-Semitism than on GFE. Importantly, these hypotheses illustrate that generalized prejudice and differentiated threat approaches are not mutually exclusive, but rather complement each other. Whereas the idea of generalized prejudice provides the general underlying mechanism through which prejudices originate, the differentiated threat approach provides more specific insights into which outgroups are targeted by which socioeconomic and cultural categories. DATA AND METHODS Data Set To test the hypotheses, the Belgian data collection for the fourth round of the European Values Study is used. In 2009, a two-stage probability sample of the Belgian population older than 18 years were approached by means of a computer assisted personal interview (CAPI). The realized sample size is 1,509 (response rate: 50.0 percent). Upon completion of the interview, participants were asked to fill out a drop-off questionnaire and send it back by mail. Seventy-five percent of the participants in the CAPI (i.e., 1,136 respondents) returned a completed questionnaire. Non-response analysis shows that certain groups (such as persons aged 40 to 54 and those with a tertiary education) are overrepresented in the realized sample, while others (persons without secondary education, individuals younger than 25 and those over 70) are underrepresented. To remediate this imbalance, we apply post-stratification weights correcting for the joint distribution of gender, age, and education. Because this study focuses on the structure of prejudices among majority-group members, respondents of immigrant background (operationalized as not having Belgian nationality or being born abroad) are excluded from the analysis, leading to a sample size of 1,007. Indicators Measures of Prejudice Each form of prejudice is measured by four to six Likert-items. These items were included in the drop-off questionnaire with a maximal spread in order to minimize memory effects and contamination. Homonegativity is measured by means of statements referring to the (ab)normality of homosexuality (q7_1; q7_5), homosexuality as a normative threat (q7_2), openness of homosexual identities (q7_3), and non-discrimination of homosexuals (q7_4). Anti-immigrant attitudes are operationalized by items regarding the trustworthiness of immigrants (q8_1), social security threat posed by immigrants (q8_2), and the impact of immigration on cultural life (q8_3; q8_4). The introductory text to this battery states that the term immigrants refers in the first place to persons of Turkish of Moroccan origin living in Belgium. Four statements referring to the presumed influence (q15_2), self-complacency (q15_3), trustworthiness (q15_4), and greediness (q15_5) of Jewish persons indicate anti-Semitism. Finally, Islamophobia is measured by means of items on the role of women in Islam (q24_2), the presumed violent nature of Islam (q24_4), and Islam as a geopolitical (q24_3) or cultural (q24_5) threat to Western society. The scales are partially balanced. Besides statements expressing a negative attitude, positively worded items are also included. To answer these statements, respondents were offered five-point scales, ranging from “agree completely” (1) to “disagree completely” (5). Validity, reliability, and dimensionality of the items were tested and confirmed by means of confirmatory factor analysis (CFA) (see Table A1 in the Appendix). Explanatory Variables Various indicators of social-structural position are included, namely gender, age (in years), and educational level (up to lower secondary degree, higher secondary degree, tertiary degree). In the operationalization of religious involvement, we make a distinction between Christians (mostly Catholics) who attend religious services on a regular basis (i.e., at least once per month); persons who consider themselves as Christians but who do not or only occasionally attend services, and non-believers and atheists. Employment status is based on the Erikson-Goldthorpe-Portocarero (EGP) (Ganzeboom and Treiman 1996) social class scheme, and distinguishes between (1) service class I and the self-employed; (2) non-manual workers (i.e., service class II and routine non-manuals); (3) manual workers (skilled and unskilled); and (4) those who never worked (and for which the EGP category can consequently not be determined). A second set of predictors concerns three attitudinal dimensions, each operationalized by means of multiple indicators (five-point agree-disagree statements). Authoritarianism is measured by means of three items on the importance of obedience and respect (q12_4), strict laws (q12_5), and punishment of immoral persons (q12_3). Three items (q30_1; q30_2; q30_3) referring to the feeling of being disadvantaged as a group by government and institutions indicate group relative deprivation. Finally, gender traditionalism is measured by means of items regarding gender-stereotypical education for boys and girls (q14_1; q14_2) and gender roles on the labor market (q14_3; q14_4). CFA showed that these items are sufficiently valid and reliable measurements for the intended concepts (see Appendix B). Statistical Modelling The statistical analysis is carried out in two major steps. First, in order to study how strong the common and group-specific components of prejudice are, we estimate a second-order confirmatory factor model (see also Bratt 2005; Zick et al. 2008). The point of departure of this approach is that a respondent’s response to a statement regarding a specific outgroup reflects three components: (1) second-order factor GFE, i.e., the common component of prejudices; (2) group-specific prejudice, i.e., prejudice towards the outgroup explicitly mentioned in the items; and (3) random measurement error at the item level.1 1 Unfortunately, our model only assumes random measurement error, and does not take systematic measurement error into account. This has important consequences for the distinction between common and specific prejudices. Insofar as the different items are affected by the same bias—i.e., so-called common method variance (CMV) (Podsakoff et al. 2003)—the common factor will absorb the systematic measurement errors. Thus, the second-order CFA approach overestimates rather than underestimates the commonality of prejudices—a fact that is often neglected in empirical research. The model thus specifies a so-called second-order factor (i.e., GFE) that captures the common variance of the first-order factors (i.e., the group-specific prejudices) (see Figure 2). Figure 2. View largeDownload slide Second-Order Factor Model with Standardized Factor Loadings Note: N = 1,007 Figure 2. View largeDownload slide Second-Order Factor Model with Standardized Factor Loadings Note: N = 1,007 Second, in order to study whether the roots of prejudice are common or group-specific (the second generalizability question), structural and attitudinal variables are added to the second-order CFA model (see Figure 3). This model captures the similarity of antecedents by estimating effects of the predictors on the second-order factor GFE (H2-H9). Group-specific explanations (H11-H20) are incorporated by allowing direct effects from the predictors on the various forms of prejudice (the first-order factors). These direct effects indicate deviations from the common explanatory model, i.e., that the impact of the respective predictor on the specific prejudice is different from that on the GFE factor. For reasons of parsimony, only effects from the predictors to the specific prejudices that lead to a significant improvement are introduced in the final model. Figure 3. View largeDownload slide GFE and Forms of Prejudice Explained by Means of Structural and Attitudinal Predictors – Significant Direct Effects Note: Direct arrows from predictors on first-order prejudices are striped. Figure 3. View largeDownload slide GFE and Forms of Prejudice Explained by Means of Structural and Attitudinal Predictors – Significant Direct Effects Note: Direct arrows from predictors on first-order prejudices are striped. All reported models are estimated using Mplus version 7.1. The ordered categorical nature of the indicators is taken into account by using the robust weighted least squares estimator (WLSMV). All reported parameters below are standardized, apart from the effects of the dummy variables, which are semi-standardized. As a result, the effects of dummy variables refer to the difference with the reference category in terms of standard deviations on the dependent variable. RESULTS The Common Denominator of Prejudices: A Second-Order Factor Model In order to analyse to what extent group-specific prejudices share a common denominator, we estimate a model with four first-order factors (anti-immigrant attitudes, homonegativity, Islamophobia, and anti-Semitism) loading on a second-order factor, capturing the communality of prejudices or GFE (see Figure 2). The anti-immigrant factor is allowed to correlate with Islamophobia, since the immigrant group referred to—people of Turkish and Moroccan origin—is predominantly Muslim. The first-order factor loadings are quite high (mostly >.60; many >.80), indicating that items are sufficiently valid and reliable measurements. This second-order factor model has an acceptable model fit (see Model 1 in Table 2): The root mean squared error of approximation (RMSEA) equals .065, and both the comparative fit index (CFI; .968) and Tucker-Lewis index (TLI; .962) are sufficiently close to 1. The second-order factor model has a substantially better fit than a model in which the four specific prejudices are not related to each other (implying that there is no communality; see Model 2 in Table 2) or a model in which all items load on a single first-order factor (meaning that prejudice has no group-specific components; see Model 3 in Table 2). This confirms our hypotheses: overlap indeed exists between the specific prejudices (H1), but at the same time they contain specific components and can therefore not be reduced to a single GFE dimension (H10). Table 2. Fit Indices for Various Decompositions of the Common and Specific Components of Prejudice Chi-square Df RMSEA CFI TLI Model 1 Specific and common components: second-order factor model 602.48 114 .065 .968 .962 Model 2 Only specific components: four unrelated first-order factors 4669.59 119 .195 .705 .663 Model 3 Only common component: all items load on a single factor 4761.76 119 .197 .699 .656 Chi-square Df RMSEA CFI TLI Model 1 Specific and common components: second-order factor model 602.48 114 .065 .968 .962 Model 2 Only specific components: four unrelated first-order factors 4669.59 119 .195 .705 .663 Model 3 Only common component: all items load on a single factor 4761.76 119 .197 .699 .656 Table 2. Fit Indices for Various Decompositions of the Common and Specific Components of Prejudice Chi-square Df RMSEA CFI TLI Model 1 Specific and common components: second-order factor model 602.48 114 .065 .968 .962 Model 2 Only specific components: four unrelated first-order factors 4669.59 119 .195 .705 .663 Model 3 Only common component: all items load on a single factor 4761.76 119 .197 .699 .656 Chi-square Df RMSEA CFI TLI Model 1 Specific and common components: second-order factor model 602.48 114 .065 .968 .962 Model 2 Only specific components: four unrelated first-order factors 4669.59 119 .195 .705 .663 Model 3 Only common component: all items load on a single factor 4761.76 119 .197 .699 .656 The second-order factor loadings contain information on the relative importance of the specific and common components. All factor loadings are high (.54, .53, .86, and .60; see Table 3), expressing that the specific prejudices are strongly influenced by the common GFE factor. Interestingly, the loading for anti-immigrant attitudes is substantially stronger than for the other three forms of prejudice. A general inclination to prejudice informs attitudes towards all outgroups, but it does even more so for anti-immigrant attitudes. The factor loading approaches 1, indicating that it becomes very hard to distinguish between GFE and anti-immigrant attitudes. In the Belgian public opinion, the category “immigrant” is the master template onto which processes of the formation of generalized prejudice are projected. Immigrants, especially of Turkish and Moroccan descent, have become the Belgian archetypical object of othering. Table 3. Common and Specific Components of the Four Forms of Prejudice Loading on 2nd Order Factor GFE Communality Uniqueness Homonegativity .54 .29 .71 Anti-immigrant attitudes .86 .74 .26 Anti-Semitism .53 .28 .72 Islamophobia .60 .36 .64 Loading on 2nd Order Factor GFE Communality Uniqueness Homonegativity .54 .29 .71 Anti-immigrant attitudes .86 .74 .26 Anti-Semitism .53 .28 .72 Islamophobia .60 .36 .64 Table 3. Common and Specific Components of the Four Forms of Prejudice Loading on 2nd Order Factor GFE Communality Uniqueness Homonegativity .54 .29 .71 Anti-immigrant attitudes .86 .74 .26 Anti-Semitism .53 .28 .72 Islamophobia .60 .36 .64 Loading on 2nd Order Factor GFE Communality Uniqueness Homonegativity .54 .29 .71 Anti-immigrant attitudes .86 .74 .26 Anti-Semitism .53 .28 .72 Islamophobia .60 .36 .64 Based on the second-order factor loadings, the common (communalities) and specific (uniqueness) components of the four forms of prejudice can be calculated. The communalities represent the proportion of variance that the forms of prejudice share with the common GFE factor. Table 3 shows that the different prejudices have substantial unique components. Homonegativity, anti-Semitism, and Islamophobia share roughly one third of their variance with GFE, while two thirds of the variance is outgroup specific. The overlap between anti-immigration attitudes and GFE is considerably larger (namely 74 percent). But even in this case, a non-negligible portion of variation (26 percent) is specific for the outgroup referred to. By consequence, an exclusive focus on this communality—that constitutes only the smaller share of the variance—leads by definition to a limited understanding of the genesis of prejudices. Common or Differential Antecedents: A Full Structural Equation Model To study whether the different antecedents precede the specific forms of prejudice, structural and attitudinal predictors are added to the second-order factor model (see Figure 3). Table 4 displays the direct effects of the predictors on GFE (showing the common impact of the antecedents on prejudice in general) as well as the direct effects on the four specific prejudice-components (i.e., the deviations from the general pattern). Sometimes, these direct effects can be misleading as they represent impacts controlling for all other predictors (also those further in the causal chain). Therefore, Table 5 additionally displays the total effects of the predictors on the prejudices. These total effects are the sum of the direct effects shown in Table 4 and the indirect effects that run through the mediating variables. Detailed insight in the explanatory model requires information on both effects. While the direct effects are especially useful to uncover differential impacts of predictors, the total effects provide insight in the general patterns in the data. Table 4. Direct Effects of Structural and Attitudinal Characteristics on Forms of Prejudice GFE Homonegativity Anti-immigrant attitudes Islamophobia Anti-Semitism Par. Est. p-value Par. Est. p-value Par. Est. p-value Par. Est. p-value Par. Est. p-value Gender (male, ref. cat.)  Female −.143 .082 −.169 .013 Age −.155 .009 .178 .000 .183 .000 Education (lower secondary, ref. cat.)  Higher secondary .036 .707  Tertiary −.039 .718 Social class (manual workers, ref. cat.)  Service class/self-employed .165 .280 −.068 .538  Non-manual workers .169 .117 −.182 .015  Never worked −.018 .904 −.156 .209 Religious involvement (regular church attendance, ref. cat)  Occasional church attendance .070 .527 −.329 .002 −.236 .034  Non-believers and atheists .039 .661 −.181 .015 −.025 .776 Authoritarianism .594 .000 Relative deprivation .297 .000 −.158 .001 .109 .009 Gender traditionalism .130 .007 .364 .000 −.207 .000 Explained variance (%) 58.9 53.9 69.8 50.1 28.3 Chi-square: 1226.29; Df: 506; RMSEA: .038; CFI: .967; TLI: .961 N = 1,007 GFE Homonegativity Anti-immigrant attitudes Islamophobia Anti-Semitism Par. Est. p-value Par. Est. p-value Par. Est. p-value Par. Est. p-value Par. Est. p-value Gender (male, ref. cat.)  Female −.143 .082 −.169 .013 Age −.155 .009 .178 .000 .183 .000 Education (lower secondary, ref. cat.)  Higher secondary .036 .707  Tertiary −.039 .718 Social class (manual workers, ref. cat.)  Service class/self-employed .165 .280 −.068 .538  Non-manual workers .169 .117 −.182 .015  Never worked −.018 .904 −.156 .209 Religious involvement (regular church attendance, ref. cat)  Occasional church attendance .070 .527 −.329 .002 −.236 .034  Non-believers and atheists .039 .661 −.181 .015 −.025 .776 Authoritarianism .594 .000 Relative deprivation .297 .000 −.158 .001 .109 .009 Gender traditionalism .130 .007 .364 .000 −.207 .000 Explained variance (%) 58.9 53.9 69.8 50.1 28.3 Chi-square: 1226.29; Df: 506; RMSEA: .038; CFI: .967; TLI: .961 N = 1,007 Notes: The parameters displayed are semi-standardized when the independent variable is a dummy (thus representing the difference with the reference category in terms of standard deviations of the dependent variable) and fully standardized in all other cases. The analysis is weighted for gender, age, and education. Table 4. Direct Effects of Structural and Attitudinal Characteristics on Forms of Prejudice GFE Homonegativity Anti-immigrant attitudes Islamophobia Anti-Semitism Par. Est. p-value Par. Est. p-value Par. Est. p-value Par. Est. p-value Par. Est. p-value Gender (male, ref. cat.)  Female −.143 .082 −.169 .013 Age −.155 .009 .178 .000 .183 .000 Education (lower secondary, ref. cat.)  Higher secondary .036 .707  Tertiary −.039 .718 Social class (manual workers, ref. cat.)  Service class/self-employed .165 .280 −.068 .538  Non-manual workers .169 .117 −.182 .015  Never worked −.018 .904 −.156 .209 Religious involvement (regular church attendance, ref. cat)  Occasional church attendance .070 .527 −.329 .002 −.236 .034  Non-believers and atheists .039 .661 −.181 .015 −.025 .776 Authoritarianism .594 .000 Relative deprivation .297 .000 −.158 .001 .109 .009 Gender traditionalism .130 .007 .364 .000 −.207 .000 Explained variance (%) 58.9 53.9 69.8 50.1 28.3 Chi-square: 1226.29; Df: 506; RMSEA: .038; CFI: .967; TLI: .961 N = 1,007 GFE Homonegativity Anti-immigrant attitudes Islamophobia Anti-Semitism Par. Est. p-value Par. Est. p-value Par. Est. p-value Par. Est. p-value Par. Est. p-value Gender (male, ref. cat.)  Female −.143 .082 −.169 .013 Age −.155 .009 .178 .000 .183 .000 Education (lower secondary, ref. cat.)  Higher secondary .036 .707  Tertiary −.039 .718 Social class (manual workers, ref. cat.)  Service class/self-employed .165 .280 −.068 .538  Non-manual workers .169 .117 −.182 .015  Never worked −.018 .904 −.156 .209 Religious involvement (regular church attendance, ref. cat)  Occasional church attendance .070 .527 −.329 .002 −.236 .034  Non-believers and atheists .039 .661 −.181 .015 −.025 .776 Authoritarianism .594 .000 Relative deprivation .297 .000 −.158 .001 .109 .009 Gender traditionalism .130 .007 .364 .000 −.207 .000 Explained variance (%) 58.9 53.9 69.8 50.1 28.3 Chi-square: 1226.29; Df: 506; RMSEA: .038; CFI: .967; TLI: .961 N = 1,007 Notes: The parameters displayed are semi-standardized when the independent variable is a dummy (thus representing the difference with the reference category in terms of standard deviations of the dependent variable) and fully standardized in all other cases. The analysis is weighted for gender, age, and education. Table 5. Total Effects of Structural and Attitudinal Characteristics on Forms of Prejudice GFE Homonegativity Anti-immigrant Attitudes Islamophobia Anti-Semitism Par. Est. p-value Par. Est. p-value Par .Est. p-value Par .Est. p-value Par. Est. p-value Gender (male, ref. cat.)  Female −.087 .303 −.344 .000 −.072 .303 .019 .742 −.046 .306 Age .067 .227 .294 .000 .055 .221 .178 .000 .036 .221 Education (lower secondary, ref. cat.)  Higher secondary −.221 .040 −.129 .045 −.182 .037 −.137 .068 −.117 .040  Tertiary −.682 .000 −.350 .000 −.562 .000 −.449 .000 −.363 .000 Social class (manual workers, ref. cat.)  Service class/self-employed −.186 .239 −.112 .223 −.221 .010 −.118 .255 −.099 .235  Non-manual workers .008 .945 −.016 .813 −.176 .029 .011 .888 .004 .945  Never worked −.002 .991 .177 .053 −.157 .224 −.110 .282 −.001 .991 Religious involvement (regular church attendance, ref. cat.)  Occasional church attendance .130 .319 −.330 .002 .107 .316 .124 .166 −.167 .130  Non-believers and atheists −.077 .423 −.328 .000 −.063 .419 .014 .836 −.066 .455 Authoritarianism .594 .000 .288 .000 .489 .000 .386 .000 .316 .000 Relative deprivation .297 .000 −.014 .730 .245 .000 .303 .000 .158 .000 Gender traditionalism .130 .007 .427 .000 .107 .004 −.122 .005 .069 .005 Explained variance (%) 58.9 53.9 69.8 50.1 28.3 Chi-square: 1226.29; Df: 506; RMSEA: .038; CFI: .967; TLI: .961 N = 1,007 GFE Homonegativity Anti-immigrant Attitudes Islamophobia Anti-Semitism Par. Est. p-value Par. Est. p-value Par .Est. p-value Par .Est. p-value Par. Est. p-value Gender (male, ref. cat.)  Female −.087 .303 −.344 .000 −.072 .303 .019 .742 −.046 .306 Age .067 .227 .294 .000 .055 .221 .178 .000 .036 .221 Education (lower secondary, ref. cat.)  Higher secondary −.221 .040 −.129 .045 −.182 .037 −.137 .068 −.117 .040  Tertiary −.682 .000 −.350 .000 −.562 .000 −.449 .000 −.363 .000 Social class (manual workers, ref. cat.)  Service class/self-employed −.186 .239 −.112 .223 −.221 .010 −.118 .255 −.099 .235  Non-manual workers .008 .945 −.016 .813 −.176 .029 .011 .888 .004 .945  Never worked −.002 .991 .177 .053 −.157 .224 −.110 .282 −.001 .991 Religious involvement (regular church attendance, ref. cat.)  Occasional church attendance .130 .319 −.330 .002 .107 .316 .124 .166 −.167 .130  Non-believers and atheists −.077 .423 −.328 .000 −.063 .419 .014 .836 −.066 .455 Authoritarianism .594 .000 .288 .000 .489 .000 .386 .000 .316 .000 Relative deprivation .297 .000 −.014 .730 .245 .000 .303 .000 .158 .000 Gender traditionalism .130 .007 .427 .000 .107 .004 −.122 .005 .069 .005 Explained variance (%) 58.9 53.9 69.8 50.1 28.3 Chi-square: 1226.29; Df: 506; RMSEA: .038; CFI: .967; TLI: .961 N = 1,007 Notes: The parameters displayed are semi-standardized when the independent variable is a dummy (thus representing the difference with the reference category in terms of standard deviations of the dependent variable) and fully standardized in all other cases. The analysis is weighted for gender, age, and education. Table 5. Total Effects of Structural and Attitudinal Characteristics on Forms of Prejudice GFE Homonegativity Anti-immigrant Attitudes Islamophobia Anti-Semitism Par. Est. p-value Par. Est. p-value Par .Est. p-value Par .Est. p-value Par. Est. p-value Gender (male, ref. cat.)  Female −.087 .303 −.344 .000 −.072 .303 .019 .742 −.046 .306 Age .067 .227 .294 .000 .055 .221 .178 .000 .036 .221 Education (lower secondary, ref. cat.)  Higher secondary −.221 .040 −.129 .045 −.182 .037 −.137 .068 −.117 .040  Tertiary −.682 .000 −.350 .000 −.562 .000 −.449 .000 −.363 .000 Social class (manual workers, ref. cat.)  Service class/self-employed −.186 .239 −.112 .223 −.221 .010 −.118 .255 −.099 .235  Non-manual workers .008 .945 −.016 .813 −.176 .029 .011 .888 .004 .945  Never worked −.002 .991 .177 .053 −.157 .224 −.110 .282 −.001 .991 Religious involvement (regular church attendance, ref. cat.)  Occasional church attendance .130 .319 −.330 .002 .107 .316 .124 .166 −.167 .130  Non-believers and atheists −.077 .423 −.328 .000 −.063 .419 .014 .836 −.066 .455 Authoritarianism .594 .000 .288 .000 .489 .000 .386 .000 .316 .000 Relative deprivation .297 .000 −.014 .730 .245 .000 .303 .000 .158 .000 Gender traditionalism .130 .007 .427 .000 .107 .004 −.122 .005 .069 .005 Explained variance (%) 58.9 53.9 69.8 50.1 28.3 Chi-square: 1226.29; Df: 506; RMSEA: .038; CFI: .967; TLI: .961 N = 1,007 GFE Homonegativity Anti-immigrant Attitudes Islamophobia Anti-Semitism Par. Est. p-value Par. Est. p-value Par .Est. p-value Par .Est. p-value Par. Est. p-value Gender (male, ref. cat.)  Female −.087 .303 −.344 .000 −.072 .303 .019 .742 −.046 .306 Age .067 .227 .294 .000 .055 .221 .178 .000 .036 .221 Education (lower secondary, ref. cat.)  Higher secondary −.221 .040 −.129 .045 −.182 .037 −.137 .068 −.117 .040  Tertiary −.682 .000 −.350 .000 −.562 .000 −.449 .000 −.363 .000 Social class (manual workers, ref. cat.)  Service class/self-employed −.186 .239 −.112 .223 −.221 .010 −.118 .255 −.099 .235  Non-manual workers .008 .945 −.016 .813 −.176 .029 .011 .888 .004 .945  Never worked −.002 .991 .177 .053 −.157 .224 −.110 .282 −.001 .991 Religious involvement (regular church attendance, ref. cat.)  Occasional church attendance .130 .319 −.330 .002 .107 .316 .124 .166 −.167 .130  Non-believers and atheists −.077 .423 −.328 .000 −.063 .419 .014 .836 −.066 .455 Authoritarianism .594 .000 .288 .000 .489 .000 .386 .000 .316 .000 Relative deprivation .297 .000 −.014 .730 .245 .000 .303 .000 .158 .000 Gender traditionalism .130 .007 .427 .000 .107 .004 −.122 .005 .069 .005 Explained variance (%) 58.9 53.9 69.8 50.1 28.3 Chi-square: 1226.29; Df: 506; RMSEA: .038; CFI: .967; TLI: .961 N = 1,007 Notes: The parameters displayed are semi-standardized when the independent variable is a dummy (thus representing the difference with the reference category in terms of standard deviations of the dependent variable) and fully standardized in all other cases. The analysis is weighted for gender, age, and education. We start the interpretation with the common impact of antecedents, i.e., the effects on GFE. Each of the three attitudinal variables is significantly related to GFE (see Table 4). As hypothesized, persons with feelings of relative deprivation (H8) and an authoritarian disposition (H7) tend to hold more prejudiced views in general. Especially authoritarianism is strongly linked to GFE: An increase of one standard deviation on the authoritarianism scale goes hand in hand with an increase of .594 standard deviations of the GFE-score. Also the endorsement of traditional gender roles turns out to be positively linked to GFE, although its impact is considerably weaker (.130). Of all structural variables, only education and age are significantly related to GFE. As expected, education has a strong tempering effect on generalized prejudice (H4). Looking at the total effects (Table 5), those with a degree from tertiary education score .682 standard deviations lower on GFE than persons with a lower secondary degree (i.e., the reference category). Also between individuals with a higher secondary and a lower secondary degree a significant difference (of -.221) is found. Education has no significant direct effect on GFE, however (see Table 4), indicating that the educational effect is fully mediated by the three attitudinal variables. The fact that the higher educated report less prejudice is completely accounted for by the fact that this group has a less authoritarian worldview, feels less deprived, and is more critical of traditional gender roles. In contrast to H3, age does not have a significant overall impact on GFE. When the attitudinal characteristics of the older cohorts are controlled for, we detect a negative net age effect (-.155; see Table 4). Gender (H2), social class (H5), and religious involvement (H6) are not significantly related to GFE, which contradicts our hypotheses as well. Taken together, the structural and attitudinal predictors explain almost 60 percent of the total variance of GFE. The primary purpose of this analysis, however, is to find out whether certain predictors have group-specific effects that deviate from this general pattern. The fact that several predictors have a direct effect on the first-order prejudices indicates that such differential effects are indeed present. It was already mentioned that relative deprivation had a moderate, positive effect on GFE. Direct effects of relative deprivation on homonegativity and Islamophobia, however, evidence that attitudes towards these two specific outgroups deviate from this general effect of relative deprivation. In case of homonegativity, the direct effect is negative (−.158 in Table 4). Compared to other forms of prejudice, the impact of relative deprivation on homonegativity is significantly weaker, which is exactly what was anticipated (H13). The total effects in Table 5 show that, overall, relative deprivation is not significantly related to homonegativity. Clearly, the perception of being deprived can translate into prejudiced attitudes, but not towards homosexuals. On the contrary, a positive direct effect of relative deprivation on Islamophobia (.109) illustrates that relative deprivation instigates anti-Muslim sentiments even more that it does so for prejudice towards outgroups. Gender traditionalism has a differentiated impact on homonegativity and Islamophobia. Traditional gender attitudes have a strong positive direct effect on homonegativity (.364). While gender traditionalism has a quite modest positive impact on prejudice in general, attitudes towards homosexuals are much more clearly structured along gender role expectations, thereby confirming H14. Of all predictors of homonegativity, gender traditionalism has the most explanatory power. For those adhering to traditional gender roles, homosexuality clearly presents a strong value threat. The situation is quite different in the case of Islamophobia: As expected (H18) traditional gender attitudes have a negative direct effect (−.207) that counteracts the overall prejudice-inducing effect of gender traditionalism completely. The total effect of gender traditionalism on Islamophobia is significantly negative. Thus, persons stressing the importance of gender traditionalism tend to be more prejudiced in general, but are found to score lower on Islamophobia. This is not surprising, given that the subordinate position of Muslim women is one of the constitutive myths of Islamophobia. As expected, the impact of authoritarianism, finally, is far less differentiated. For each of the specific forms of prejudice, authoritarianism is the single most important predictor, and no prejudice-specific effects are detected. Differentiated effects are also found for several structural variables. While no significant gender differences exist regarding GFE, women do report less homophobic attitudes than men (direct effect: −.169; see Table 4), which confirms H11. Also as expected, social class has a differential impact on anti-immigrant attitudes (H15). In terms of the direct effects, there is a significant gap in anti-immigrant attitudes between non-manual and manual workers. In the total effects it can be seen that anti-immigrant attitude is the only form of prejudice that is structured among class lines: The higher service class/self-employed and the non-manual workers report significantly less anti-immigrant sentiments than the manual workers (who serve as reference category in this analysis). Also religious involvement has differential impacts on homonegativity (H12) and anti-Semitism (H19). As expected, the gap between church-goers and marginal Christians is more outspoken regarding homonegativity and anti-Semitism than for other prejudices. This pattern illustrates the role of homosexuals and Jews as permitted targets of prejudice among church-involved Christians. In addition to the hypothesized effects, we also find a small but significant group-specific impact of age. In the case of homonegativity and Islamophobia, age has a direct positive effect (.178 and .183 respectively; see Table 4). The same pattern can also be seen in the total effects (Table 5): Homonegativity and Islamophobia are the only two forms of prejudice for which a significant age effect is found. Homosexuality and Islam are the only issues for which an opinion gap between younger and older cohorts is observed. Finally, the hypothesis regarding the specific effect of education on Islamophobia (H17) was not confirmed. Although the hypotheses tested here do not provide a direct test of the differentiated threat approach, this analysis indirectly renders clear support for its basic propositions. Our results provide a differentiated and nuanced view on the predictors of various prejudices. Besides a set of common antecedents, such as authoritarianism and educational level, the four forms of prejudice have quite different relationships to several predictors in the model. Furthermore, the pattern of group-specific antecedents largely confirms the key idea that prejudices are embedded in realistic and symbolic intergroup conflicts; prejudices toward a particular outgroup develop among specific socioeconomic (or cultural) segments of society who see their realistic (or symbolic) interest threatened. For homonegativity, this point is most obvious: The impact of 5 out of 8 predictors deviates significantly from the general pattern, and homonegativity is consistently stronger among the elderly, males, churchgoers and gender traditionalists—i.e., groups experiencing homosexuality as a stronger threat to the moral order. Also for Islamophobia, several differential effects are found that interestingly enough often go in to opposite direction as those for homonegativity. The fact that Islamophobia and homonegativity correlate positively (see section above titled “The Common Denominator of Prejudices”) does not exclude that both forms of prejudice have partially different social roots and are shaped by different processes. Anti-immigrant attitudes deviate least from the common explanatory model. This is hardly surprising, given the large overlap between anti-immigrant prejudice and GFE. The particularity of the antecedents’ prejudices is also evidenced by the sizeable differences in explanatory power of the predictors: The model explains 69.8 percent of the variation in anti-immigrant attitudes versus 28.3 percent in the case of anti-Semitism. CONCLUSION In this article, we argue that an exclusive focus on the generalized aspect of prejudices limits our understanding of the structure and genesis of particular prejudices. The realization of attitudes towards specific outgroups is contingent on structural and contextual factors, such as social positions, the nature of intergroup relations, power balances, and elite discourses. In order to conceptualize the outgroup specific nature of prejudices, we develop a theoretical framework—the differentiated threat approach—that hinges on two crucial principles. First, the model acknowledges that the interests that are challenged by outgroups are diverse (realistic versus symbolic threats), and that various types of outgroups—deviant, competing, and dissident—can be distinguished. Second, the differentiated threat approach stresses that societies are structurally and culturally stratified, and that perceived outgroup threats are not equally distributed across society, but disproportionally affect specific socioeconomic strata and cultural groups depending on the nature of conflict. As a result, individuals’ prejudices are not necessarily directed towards all possible outgroups, but are instead focused on outgroups that are perceived to challenge the realistic and/or symbolic interests of the relevant reference group. Paramount to the differentiated threat approach is that outgroup threats are embedded in intergroup relations and have to be interpreted within specific contexts. Hence, in-depth knowledge about the intergroup relations in particular settings are required. As such, the relevance of the differentiated threat approach was tested empirically by analysing the attitudes of Belgian majority-group members towards four outgroups, namely immigrants, Muslims, Jews, and sexual minorities. Our study shows that, although a common denominator of generalized prejudice can be distinguished, negative attitudes towards the various outgroups contain substantively relevant unique components. The degree to which a specific prejudice is linked with GFE varies across outgroups. A very strong overlap exists between GFE and anti-immigration attitudes, which suggests that in Belgian public opinion the category “immigrant” is the master template onto which processes of the formation of generalized prejudice are projected. Our analysis furthermore reveals that certain factors—such as low education, feelings of relative deprivation, and an authoritarian disposition—consistently contribute to generalized prejudice. But we also find that this general explanation model ignores that specific prejudices are partly triggered by particular sets of predictors. Clearly, there are particular reasons and motivations to be prejudiced idiosyncratic to the outgroup. The differential impact of gender traditionalism on Islamophobia and homonegativity illustrates this point perfectly. Gender traditionalism is the strongest factor inducing homonegativity, but tends to reduce anti-Islam feelings at the same time. The reason for this paradoxical effect should be sought in the specific content of prejudices towards homosexuals and Muslims. One of the most persistent stereotypes about homosexuality relates to the violation of traditional gender norms (including heteronormativity and patriarchy), while the critique that Islam is an inherently sexist religion oppressing women is precisely one of the constitutive elements of Islamophobia. Homosexuals as well as Muslims are disliked because they are perceived as a threat to certain values. But the specific value sets concerned and segments of society that feel threated are fundamentally different. Various other predictors where found to have an outgroup specific impact that deviates from the explanation model for generalized prejudice. Feelings of relative deprivation, for example, are more strongly related to Islamophobia than to other forms of prejudice, and are unrelated to homonegative views. Religious involvement plays a more decisive role in the formation of anti-Semitism and homonegativity than it does in the other forms of prejudice. And anti-immigration attitudes show a class gradient that is absent in attitudes towards other outgroups. Notably, the analyses showed that especially the origins of homonegativity deviated from those of GFE. As such, homonegativity can be considered as a rather “exceptional” case of prejudice. Looking at the position of LGBT in the Belgian context, it is not hugely surprising that homosexuality—as opposed to immigration or the increasing visibility of Islam—is more accepted by public opinion and almost a non-issue in political discourse. This finding once more stresses the necessity of a contextualized approach. The structure of prejudices and the classification of outgroups (into dissident, deviant, and competing groups) are not universal givens, but vary across societies and could undergo modifications over time. Zick and colleagues (2011), for instance, found that sexism and homophobia were much less associated with prejudice in Hungary and Poland, compared to more Western countries. Besides, prejudice towards certain outgroups might move away from the GFE component, while attitudes towards others might move into the center or even become the pinnacle of prejudice. This is illustrated by the finding that, due to the increased marketization of Western societies, prejudices toward unprofitable groups (e.g., unemployed and homeless) are claiming a more prominent position in the structure of prejudice (Hövermann et al. 2015). Our results confirm various hypotheses that were derived from the differentiated threat approach, and as such illustrate the necessity of a contextualized approach. These findings do not invalidate the theory of generalized prejudice, however. The generalized prejudice and differentiated threat approaches complement each other in the understanding of prejudice and excluding one from the other would provide a distorted and one-sided picture. As such, future research examining the origins of prejudice should try to differentiate between the group-specific and generalized component of prejudice and emphasize the contextualisation of the group-specific component. Furthermore, because the differentiated threat approach critically highlights the importance of context, future research should evaluate its theoretical underpinnings in different social settings. Doing so, future research could include examples of “competing” groups—a type of others that was not included in our empirical analysis. However, it cannot be overemphasized that studies using the differenitated threat approach need to be mindful that the outgroups that were analyzed in this article might hold different social positions and evoke different types of threat in other societies (for example, immigrants as a predominantly economic threat in Italy versus cultural threat in France; Bergamaschi 2013). Finally, although our approach has a strong focus on threat, it has shown that prejudice is not solely based on uncontrollable personality traits but largely depends on situational group membership and the search for a positive self-identity and personal/community security (Kinder and Kam 2010). Hence, it can easily be translated for the study of group identities, which is related to the study of prejudices. For example, the differentiated threat approach, and its focus on the position of groups vis-à-vis each other, can be used by social identity researchers interested in how the positionality of the different others to the ingroup may invoke differenitated othering mechanisms. Alternatively, social identity researchers can expand our differenitated threat model by adding the layer of social identitification to the intergroup positionality. The authors would like to thank the participants of the Leuven-Tilburg Sociology seminar as well as three anonymous reviewers for their helpful comments. This study was made possible by grants from KU Leuven research council (OT/13/30), the Belgian National Lottery, and the National Science Foundation FWO-Vlaanderen (Grant Number ZKC6622-Asp/13). APPENDIX Table A1 CFA Results for the Specific Prejudice Factors Homonegativity Anti-immigrant Attitudes Anti-Semitism Islamophobia Standardized factor loadings  q7_1 Children should learn that being homosexual is completely normal. −.67  q7_2 Equal rights for homosexuals is a threat for our norms and values. .83  q7_3 Homosexual men should not make such a fuss about their sexual orientation. .61  q7_4 Gay and straight people should be treated equally. −.70  q7_5 Lesbian women are abnormal. .79  q8_1 Migrants are generally not trustworthy. .83  q8_2 Migrants come here to profit from our social security model. .88  q8_3 Migrants are a threat to our culture and customs. .89  q8_4 The presence of different cultures is an enrichment of our society. −.59  q15_2 Jews have too much influence in our country. .70  q15_3 Jews think they are better than all others. .68  q15_4 One should avoid Jews; you cannot trust them. .88  q15_5 Jews are only after money. .88  q24_2 Muslim men dominate Muslim women. .55  q24_3 When it comes down, Islamic countries will turn against Europe. .75  q24_4 Islamic history and culture is more violent than others. .81  q24_5 Islamic values are a threat to Europe. .91 Correlations between factors    Homonegativity 1.00    Anti-immigrant attitudes .47 1.00    Anti-Semitism .29 .45 1.00    Islamophobia .31 .73 .33 1.00 Model fit: Chi-square: 622.03; Df: 113; RMSEA: .067; CFI: .967; TLI: .960 Homonegativity Anti-immigrant Attitudes Anti-Semitism Islamophobia Standardized factor loadings  q7_1 Children should learn that being homosexual is completely normal. −.67  q7_2 Equal rights for homosexuals is a threat for our norms and values. .83  q7_3 Homosexual men should not make such a fuss about their sexual orientation. .61  q7_4 Gay and straight people should be treated equally. −.70  q7_5 Lesbian women are abnormal. .79  q8_1 Migrants are generally not trustworthy. .83  q8_2 Migrants come here to profit from our social security model. .88  q8_3 Migrants are a threat to our culture and customs. .89  q8_4 The presence of different cultures is an enrichment of our society. −.59  q15_2 Jews have too much influence in our country. .70  q15_3 Jews think they are better than all others. .68  q15_4 One should avoid Jews; you cannot trust them. .88  q15_5 Jews are only after money. .88  q24_2 Muslim men dominate Muslim women. .55  q24_3 When it comes down, Islamic countries will turn against Europe. .75  q24_4 Islamic history and culture is more violent than others. .81  q24_5 Islamic values are a threat to Europe. .91 Correlations between factors    Homonegativity 1.00    Anti-immigrant attitudes .47 1.00    Anti-Semitism .29 .45 1.00    Islamophobia .31 .73 .33 1.00 Model fit: Chi-square: 622.03; Df: 113; RMSEA: .067; CFI: .967; TLI: .960 Table A1 CFA Results for the Specific Prejudice Factors Homonegativity Anti-immigrant Attitudes Anti-Semitism Islamophobia Standardized factor loadings  q7_1 Children should learn that being homosexual is completely normal. −.67  q7_2 Equal rights for homosexuals is a threat for our norms and values. .83  q7_3 Homosexual men should not make such a fuss about their sexual orientation. .61  q7_4 Gay and straight people should be treated equally. −.70  q7_5 Lesbian women are abnormal. .79  q8_1 Migrants are generally not trustworthy. .83  q8_2 Migrants come here to profit from our social security model. .88  q8_3 Migrants are a threat to our culture and customs. .89  q8_4 The presence of different cultures is an enrichment of our society. −.59  q15_2 Jews have too much influence in our country. .70  q15_3 Jews think they are better than all others. .68  q15_4 One should avoid Jews; you cannot trust them. .88  q15_5 Jews are only after money. .88  q24_2 Muslim men dominate Muslim women. .55  q24_3 When it comes down, Islamic countries will turn against Europe. .75  q24_4 Islamic history and culture is more violent than others. .81  q24_5 Islamic values are a threat to Europe. .91 Correlations between factors    Homonegativity 1.00    Anti-immigrant attitudes .47 1.00    Anti-Semitism .29 .45 1.00    Islamophobia .31 .73 .33 1.00 Model fit: Chi-square: 622.03; Df: 113; RMSEA: .067; CFI: .967; TLI: .960 Homonegativity Anti-immigrant Attitudes Anti-Semitism Islamophobia Standardized factor loadings  q7_1 Children should learn that being homosexual is completely normal. −.67  q7_2 Equal rights for homosexuals is a threat for our norms and values. .83  q7_3 Homosexual men should not make such a fuss about their sexual orientation. .61  q7_4 Gay and straight people should be treated equally. −.70  q7_5 Lesbian women are abnormal. .79  q8_1 Migrants are generally not trustworthy. .83  q8_2 Migrants come here to profit from our social security model. .88  q8_3 Migrants are a threat to our culture and customs. .89  q8_4 The presence of different cultures is an enrichment of our society. −.59  q15_2 Jews have too much influence in our country. .70  q15_3 Jews think they are better than all others. .68  q15_4 One should avoid Jews; you cannot trust them. .88  q15_5 Jews are only after money. .88  q24_2 Muslim men dominate Muslim women. .55  q24_3 When it comes down, Islamic countries will turn against Europe. .75  q24_4 Islamic history and culture is more violent than others. .81  q24_5 Islamic values are a threat to Europe. .91 Correlations between factors    Homonegativity 1.00    Anti-immigrant attitudes .47 1.00    Anti-Semitism .29 .45 1.00    Islamophobia .31 .73 .33 1.00 Model fit: Chi-square: 622.03; Df: 113; RMSEA: .067; CFI: .967; TLI: .960 Table A2 CFA Results for the Attitudinal Predictors of Prejudice Authoritarianism Gender Traditionalism Relative Deprivation Standardized factor loadings  q12_3 Most of our social problems could be solved, if we could somehow get rid of the immoral, crooked people. .63  q12_4 Obedience and respect for authority are the two most important virtues children have to learn. .72  q12_5 Laws should become stricter, because too much freedom is not good for people. .69  q14_1 It isn’t really as important for a girl to get a good education as it is for a boy. .81  q14_2 In general, boys can be brought up freer than girls. .72  q14_3 It’s unnatural if women give guidance to men in a company. .87  q14_4 It is the most natural thing for a man to be a breadwinner and for a women to take care of the household and the children. .70  q30_1 If we need something from the government, people like me have to wait longer than others. .87  q30_2 People like me are being systematically neglected, whereas other groups received more than they deserve. .94  q30_3 In times of economic crises people like me are always the first victims. .82 Correlations between factors    Authoritarianism 1.00    Gender traditionalism .37 1.00    Relative deprivation .42 .26 1.00 Model fit: Chi-square: 106.07; Df: 32; RMSEA: .048; CFI: .993; TLI: .990 Authoritarianism Gender Traditionalism Relative Deprivation Standardized factor loadings  q12_3 Most of our social problems could be solved, if we could somehow get rid of the immoral, crooked people. .63  q12_4 Obedience and respect for authority are the two most important virtues children have to learn. .72  q12_5 Laws should become stricter, because too much freedom is not good for people. .69  q14_1 It isn’t really as important for a girl to get a good education as it is for a boy. .81  q14_2 In general, boys can be brought up freer than girls. .72  q14_3 It’s unnatural if women give guidance to men in a company. .87  q14_4 It is the most natural thing for a man to be a breadwinner and for a women to take care of the household and the children. .70  q30_1 If we need something from the government, people like me have to wait longer than others. .87  q30_2 People like me are being systematically neglected, whereas other groups received more than they deserve. .94  q30_3 In times of economic crises people like me are always the first victims. .82 Correlations between factors    Authoritarianism 1.00    Gender traditionalism .37 1.00    Relative deprivation .42 .26 1.00 Model fit: Chi-square: 106.07; Df: 32; RMSEA: .048; CFI: .993; TLI: .990 Table A2 CFA Results for the Attitudinal Predictors of Prejudice Authoritarianism Gender Traditionalism Relative Deprivation Standardized factor loadings  q12_3 Most of our social problems could be solved, if we could somehow get rid of the immoral, crooked people. .63  q12_4 Obedience and respect for authority are the two most important virtues children have to learn. .72  q12_5 Laws should become stricter, because too much freedom is not good for people. .69  q14_1 It isn’t really as important for a girl to get a good education as it is for a boy. .81  q14_2 In general, boys can be brought up freer than girls. .72  q14_3 It’s unnatural if women give guidance to men in a company. .87  q14_4 It is the most natural thing for a man to be a breadwinner and for a women to take care of the household and the children. .70  q30_1 If we need something from the government, people like me have to wait longer than others. .87  q30_2 People like me are being systematically neglected, whereas other groups received more than they deserve. .94  q30_3 In times of economic crises people like me are always the first victims. .82 Correlations between factors    Authoritarianism 1.00    Gender traditionalism .37 1.00    Relative deprivation .42 .26 1.00 Model fit: Chi-square: 106.07; Df: 32; RMSEA: .048; CFI: .993; TLI: .990 Authoritarianism Gender Traditionalism Relative Deprivation Standardized factor loadings  q12_3 Most of our social problems could be solved, if we could somehow get rid of the immoral, crooked people. .63  q12_4 Obedience and respect for authority are the two most important virtues children have to learn. .72  q12_5 Laws should become stricter, because too much freedom is not good for people. .69  q14_1 It isn’t really as important for a girl to get a good education as it is for a boy. .81  q14_2 In general, boys can be brought up freer than girls. .72  q14_3 It’s unnatural if women give guidance to men in a company. .87  q14_4 It is the most natural thing for a man to be a breadwinner and for a women to take care of the household and the children. .70  q30_1 If we need something from the government, people like me have to wait longer than others. .87  q30_2 People like me are being systematically neglected, whereas other groups received more than they deserve. .94  q30_3 In times of economic crises people like me are always the first victims. .82 Correlations between factors    Authoritarianism 1.00    Gender traditionalism .37 1.00    Relative deprivation .42 .26 1.00 Model fit: Chi-square: 106.07; Df: 32; RMSEA: .048; CFI: .993; TLI: .990 REFERENCES Abts K. , Kochuyt T. . 2013 . “De Vreemde Bedreiging Van de Verzorgingsstaat.” Tijdschrift voor Sociologie 34 ( 3-4 ): 227 - 49 . Adorno Theodor W. , Frenkel-Brunswick Else , Levinson Daniel K. , Nevitt Sanford R. . 1950 . The Authoritarian Personality . New York : Harper . Aerts Saskia , Dewaele Alexis , Cox Nele , Van Houtte Mieke . 2014 . “Homonegativity in the Technical and Vocational Track: A Survey of Secondary School Students in Flanders.” Journal of LGBT Youth 11 4 : 364 - 87 . Google Scholar CrossRef Search ADS Agnew Christopher R. , Thompson Valda D. , Gaines Stanely O. . 2000 . “Incorporating Proximal and Distal Influences on Prejudice: Testing a General Model Across Outgroups.” Personality and Social Psychology Bulletin 26 4 : 403 - 18 . Google Scholar CrossRef Search ADS Akrami Nazar , Ekehammar Bo , Bergh Robin . 2011 . “Generalized Prejudice: Common and Specific Components.” Psychological Science 22 1 : 57 - 59 . Google Scholar CrossRef Search ADS PubMed Allport Gordon W. 1958 . The Nature of Prejudice . New York : Doubleday Anchor . Altemeyer Bob. 1998 . “The Other ‘Authoritarian Personality.’” Advances in Experimental Social Psychology 30 : 47 - 92 . Google Scholar CrossRef Search ADS Altemeyer Bob , Hunsberger Bruce . 1992 . “Authoritarianism, Religious Fundamentalism, Quest, and Prejudice.” International Journal for the Psychology of religion 2 : 113 - 33 . Google Scholar CrossRef Search ADS Asbrock Frank , Sibley Chris G. , Duckitt John . 2010 . “Right-Wing Authoritarianism and Social Dominance Orientation and the Dimensions of Generalized Prejudice: A Longitudinal Test.” European Journal of Personality 24 : 324 - 40 . Bäckström Martin , Björklund Fredrik . 2007 . “Structural Modeling of Generalized Prejudice: The Role of Social Dominance, Authoritarianism, and Empathy.” Journal of Individual Differences 28 : 10 - 17 . Google Scholar CrossRef Search ADS Bergamaschi Alessandro. 2013 . “Adolescents and Prejudice: A Comparative Study of the Attitudes of Two European Adolescent Populations Regarding the Issues that are Raised by Increasing Cultural and Religious Pluralism.” International Journal of Intercultural Relations 37 3 : 302 - 12 . Google Scholar CrossRef Search ADS Bierly Margaret M. 1985 . “Prejudice Toward Contemporary Outgroups as a Generalized Attitude.” Journal of Applied Social Psychology 15 2 : 189 - 99 . Google Scholar CrossRef Search ADS Bilsky Leora. 2009 . “Muslim Headscarves in France and Army Uniforms in Israel: A Comparative Study of Citizenship as a Mask.” Patterns of Prejudice 43 4 : 287 - 311 . Google Scholar CrossRef Search ADS Blalock Hubert M. 1967 . Toward a Theory of Minority-Group Relations . New York : John Wiley & Sons . Bleich Erik. 2011 . “What is Islamophobia and How Much is There? Theorizing and Measuring an Emerging Comparative Concept.” American Behavioral Scientist 55 12 : 1581 - 1600 . Google Scholar CrossRef Search ADS Blumer Herbert. 1958 . “Race prejudice as a Sense of Group Position.” The Pacific Sociological Review 1 : 3 - 7 . Google Scholar CrossRef Search ADS Bobo Lawrence. 1999 . “Prejudice as Group Position: Microfoundations of a Sociological Approach to Racism and Race Relations.” Journal of Social Issues 3 55 : 445 - 72 . Google Scholar CrossRef Search ADS Bobo Lawrence , Hutchings Vincent . 1996 . “Perceptions of Racial Group Competition: Extending Blumer’s Theory of Group Position to a Multiracial Social Context.” American Sociological Review 6 ( 61 ): 951 - 72 . Google Scholar CrossRef Search ADS Bourdieu Pierre. 1990 . The Logic of Practice . Cambridge, UK : Polity Press . Bourdieu Pierre. 1984 . Distinction: A Social Critique of the Judgement of Taste . London, UK : Routlegde . Bourdieu Pierre. 1985 . “The Social Space and the Genesis of Groups.” Theory and Society 6 14 : 423 - 44 . Bracke Sarah. 2012 . “From ‘Saving Women’ to ‘Saving Gays’: Rescue Narratives and their Dis/continuities.” European Journal of Women's Studies 19 2 : 237 - 52 . Google Scholar CrossRef Search ADS Bratt Christopher. 2005 . “The Structure of Attitudes Toward Non-Western Immigrant Groups: Second-Order Factor Analysis of Attitudes Among Norwegian Adolescents.” Group Processes & Intergroup Relations 8 4 : 447 - 69 . Google Scholar CrossRef Search ADS Choma Becky L. , Hodson Gordon . 2008 . “And So the Pendulum Swings: A Framework for Conceptualizing the Causes of Prejudice.” Pp. 1 - 25 in The Psychology of Modern Prejudice , edited by Melanie A. Morrison , Morrison Todd G. . New York : Nova Science Publishers . Connell R. W. 2005 . Masculinities . Cambridge, UK : Polity . Duckitt John. 2006 . “Differential Effects of Right-Wing Authoritarianism and Social Dominance Orientation on Outgroup Attitudes and their Mediation by Threat from and Competitiveness to Outgroups.” Personality and Social Psychology Bulletin 32 5 : 684 - 96 . Duckitt John , Sibley Chris G. . 2007 . “Right Wing Authoritarianism, Social Dominance Orientation, and the Dimensions of Generalized Prejudice.” European Journal of Personality 21 : 113 - 30 . Google Scholar CrossRef Search ADS Ekehammar Bo , Akrami Nazar , Gylje Magnus , Zakrisson Ingrid . 2004 . “What Matters Most to Prejudice: Big Five Personality, Social Dominance Orientation, or Right-Wing Authoritarianism?” European Journal of Personality 18 : 463 - 82 . Google Scholar CrossRef Search ADS Fekete Liz. 2009 . A Suitable Enemy: Racism, Migration, and Islamophobia in Europe . New York : Pluto Press . Fiske Susan T. , Cuddy Amy C. , Glick Peter , Xu Jun . 2002 . “A Model of (Often Mixed) Stereotype Content: Competence and Warmth Respectively Follow from Perceived Status and Competition.” Journal of Personality and Social Psychology 82 6 : 878 - 902 . Google Scholar CrossRef Search ADS PubMed Ganzeboom Harry , Treiman Donald J. . 1996 . “Internationally Comparable Measures of Occupational Status for the 1988 International Standard Classification of Occupations.” Social Science Research 25 : 201 - 39 . Google Scholar CrossRef Search ADS Hagendoorn Louk. 1995 . “Intergroup Biases in Multiple Group Systems: The Perception of Ethnic Hierarchies.” European Review of Social Psychology 6 : 199 - 228 . Google Scholar CrossRef Search ADS Herek Gregory M. 1988 . “Heterosexuals' Attitudes Toward Lesbians and Gay Men: Correlates and Gender Differences.” The Journal of Sex Research 25 4 : 451 - 77 . Google Scholar CrossRef Search ADS Hjerm Mikael , Nagayoshi Kikuko . 2011 . “The Composition of the Minority Population as a Threat: Can Real Economic and Cultural Threats explain Xenophobia.” International Sociology 26 6 : 815 - 43 . Google Scholar CrossRef Search ADS Hodson Gordon , Dhont Kristof . 2015 . “The Person-Based Nature of Prejudice: Individual Difference Predictors of Intergroup Negativity.” European Review of Social Psychology 26 1 : 1 - 42 . Google Scholar CrossRef Search ADS Hövermann Andrea , Gross Eva M. , Zick Andreas , Messner Steven F. . 2015 . “Understanding the Devaluation of Vulnerable Groups: A novel Application of Institutional Anomie Theory.” Social Science Research 52 : 408 - 21 . Google Scholar CrossRef Search ADS PubMed Hudson Walter W. , Ricketts Wendell A. . 1980 . “A Strategy for the Measurement of Homophobia.” Journal of Homosexuality 5 4 : 357 - 72 . Google Scholar CrossRef Search ADS PubMed Hunsberger Bruce , Jackson Lynne M. . 2005 . “Religion, Meaning, and Prejudice.” Journal of Social Issues 61 : 807 - 26 . Google Scholar CrossRef Search ADS Key Valdimer O. 1949 . Southern Politics in State and Nation . New York : Knopf . Kinder Donald. R. , Kam Cindy M. . 2010 . Us Against Them: Ethnocentric Foundations of American Opinion . Chicago : The University of Chicago Press . Koopmans Ruud. 2015 . “Religious Fundamentalism and Hostility Against Outgroups: A Comparison of Muslims and Christians in Western Europe.” Journal of Ethnic and Migration Studies 44 1 : 33 - 57 . Google Scholar CrossRef Search ADS Kumar Deepa. 2012 . Islamophobia and the Politics of Empire . Chicago : Haymarket Books . Kunovich Robert M. 2004 . “Social Structural Position and Prejudice: An Exploration of Cross-National Differences in Regression Slopes.” Social Science Research 33 1 : 20 - 44 . Google Scholar CrossRef Search ADS Meer Nasar. 2013 . “Semantics, Scales and Solidarities in the Study of Antisemitism and Islamophobia.” Ethnic and Racial Studies 36 3 : 500 - 515 . Google Scholar CrossRef Search ADS Merton Robert K. [1957] 1968 . Social Theory and Social Structure . New York : The Free Press . Meuleman B. , Davidov E. , Billiet J. . 2009 . “Changing Attitudes Toward Immigration in Europe, 2002-2007: A Dynamic Group Conflict Theory Approach.” Social Science Research 38 2 : 352 - 65 . Google Scholar CrossRef Search ADS PubMed Olzak Susan. 1994 . The Dynamics of Ethnic Competition and Conflict . Stanford, CA : Stanford University Press . Pargament Kenneth I. , Trevino Kelley , Mahoney Anette , Silberman Israela . 2007 . “They Killed our Lord: The Perception of Jews as Desecrator of Christianity as a Predictor of Anti-Semitism.” Journal of the Scientific Study of Religion 46 2 : 143 - 58 . Google Scholar CrossRef Search ADS Plummer David. 1999 . One of the Boys: Masculinity, Homophobia, and Modern Manhood . New York : Harrington Park Press . Podsakoff Philip M. , MacKenzie Scott B. , Lee Jeong-Yeon , Podsakoff Nathan P. . 2003 . “Common Method Biases in Behavioral Research: A Critical Review of the Literature and Recommended Remedies.” Journal of Applied Psychology 88 5 : 879 - 903 . Google Scholar CrossRef Search ADS PubMed Ray John J. , Lovejoy F. H. . 1986 . “The Generality of Racial Prejudice.” The Journal of Social Psychology 126 4 : 563 - 64 . Google Scholar CrossRef Search ADS Riek Blake M. , Mania Eric. W. , Gaertner Samuel L. . 2006 . “Intergroup Threat and Outgroup Attitudes: A Meta-Analytic Review.” Personality and Social Psychology Review 10 4 : 336 - 53 . Google Scholar CrossRef Search ADS PubMed Runciman Walter G. 1966. Relative Deprivation and Social Justice: A Study of Attitudes to Social Inequality in Twentieth Century England . Berkeley : University of California Press . Schiefer David. 2013 . “Cultural Values and Group-Related Attitudes: A Comparison of Individuals with and without Migration Background Across 24 Countries.” Journal of Cross-Cultural Psychology 44 2 : 245 - 62 . Google Scholar CrossRef Search ADS Schiffer Sabine , Wagner Constantin . 2011 . “Anti-Semitism and Islamophobia—New Enemies, Old Patterns.” Race & Class 52 : 377 - 84 . Google Scholar CrossRef Search ADS Schneider Anne , Ingram Helen . 1993 . “Social Construction of Target Populations: Implications for Politics and Policy.” American Political Science Review 87 02 : 334 - 47 . Google Scholar CrossRef Search ADS Spruyt Bram , Elchardus Mark . 2012 . “Are Anti-Muslim Feelings More Widespread than Anti-Foreigner Feelings? Evidence from Two Split-Sample Experiments.” Ethnicities 12 6 : 800 - 20 . Google Scholar CrossRef Search ADS Stephan Walter G. , Ybarra Oscar , Martínez Martínez Carmen , Schwarzwald Joseph , Tur-Kaspa Michal . 1998 . “Prejudice Toward Immigrants to Spain and Israel: An Integrated Threat Theory Analysis.” Journal of Cross-Cultural Psychology 29 4 : 559 - 76 . Google Scholar CrossRef Search ADS Strabac Zan , Listhaug Ola . 2008 . “Anti-Muslim Prejudice in Europe: A Multi-Level Analysis of Survey Data from 30 Countries.” Social Science Research 37 : 268 - 86 . Google Scholar CrossRef Search ADS Swyngedouw Marc. 1995 . “The ‘Threatening Immigrant’ in Flanders 1930-1980: Redrawing the Social Space.” New Community 21 3 : 325 - 40 . Taras Ray. 2013 . “ ‘Islamophobia Never Stands Still’: Race, Religion, and Culture.” Ethnic and Racial Studies 36 3 : 417 - 33 . Google Scholar CrossRef Search ADS Thornton Arland , Alwin Duane F. , Camburn Donald . 1983 . “Causes and Consequences of Sex-Role Attitudes and Attitude Change.” American Sociological Review 48 2 : 211 - 27 . Google Scholar CrossRef Search ADS PubMed Whitley Bernard E. 2009 . “Religiosity and Attitudes Towards Lesbians and Gay Men: A Meta-analysis.” The International Journal for the Psychology of Religion 19 1 : 21 - 38 . Google Scholar CrossRef Search ADS Zick Andreas , Küpper Beate , Hövermann Andreas . 2011 . Intolerance, Prejudice and Discrimination: A European Report . Berlin, Germany : Friedrich-Ebert-Stiftung . Zick Andreas , Wolf Carina , Küpper Beate , Davidov Eldad , Schmidt Peter , Heitmeyer Wilhelm . 2008 . “The Syndrome of Group-Focused Enmity: The Interrelation of Prejudices Tested with Multiple Cross-Sectional and Panel Data.” Journal of Social Issues 64 2 : 363 - 83 . Google Scholar CrossRef Search ADS © The Author(s) 2018. Published by Oxford University Press on behalf of the Society for the Study of Social Problems.All rights reserved. For permissions, please email: journals.permissions@oup.com

Journal

Social ProblemsOxford University Press

Published: Feb 13, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off