Heterogeneity of Ingroup Identity and Anti-Immigrant Prejudice: The Moderating Role of RWA and Outgroup Homogeneity

have a homogeneous identity—i.e., a singular and distinct identity built on a common origin, culture and goal. Socially constructed ingroup homogeneity, paired with a perception of national essentialism, plays a fundamental role in delimitating clear-cut borders and in legitimizing the very existence of national entities (Gellner, 1997; Pehrson, Brown and Zagefka, 2009). Interestingly, expectations of homogeneity are also found on a transnational level, concerning more recently constructed social identities, such as the European identity. European identity is perceived as a shared and singular European identity (based on common ethnic roots, history and culture), which has made heterogeneity within Europe more difficult to accept for its members. Indeed, challenges to such a homogeneous perception of European identity, through the admission of nations that are perceived as too culturally or politically different (such as Turkey) is a longstanding debate within European countries (Caglar, 1997). The expectation of homogeneous social identities is congruent with one of the most influential theoretical models in social psychology. Distinctiveness is a fundamental motive underlying not only inter-personal but also inter-group processes (Brewer, 1993; Jetten, Spears and Manstead, 1997; Jetten and Spears, 2003). A key postulate of social identity theory (SIT) is that individuals are motivated to attain and maintain a positive and distinct social identity (Tajfel and Turner, 1979, 1986). A group is perceived as distinct from other groups when intra-group variability is lower than inter-group variability (Turner, Oakes, Haslam and McGarthy, 1994; Turner, Hogg, Oakes, Reicher and Wetherell, 1987; Tajfel, 1969; Tajfel and Wilkes, 1963; see also Campbell, 1958; Simon and Brown, 1987). Accordingly, increasing intergroup similarity challenges ingroup distinctiveness (Gabarrot and FalomirPichastor, 2017). Similarly, accentuating the perception of ingroup homogeneity contributes to the attainment of a positively distinct social identity (Simon and Pettigrew, 1990). Of particular relevance for the present research, as a corollary, accentuating intra-group variability (i.e., heterogeneity) may reduce the perceived inter-group variability and therefore the ingroup distinctiveness. Thus, ingroup heterogeneity may challenge the fundamental psychological motive for positive ingroup distinctiveness. How can group members deal with such a threat to ingroup distinctiveness? Overall, the definition of an ingroup identity can affect how we perceive and treat outgroup members (Reicher and Hopkins, 2001). Because a heterogeneous ingroup identity challenges the very essence of the ingroup and members’ motive for a positive and distinct identity, ingroup heterogeneity may lead to ingroup members’ increasing intergroup differentiation to reaffirm the threatened ingroup distinctiveness Frederic, N. S. and Falomir-Pichastor, J. M. (2018). Heterogeneity of Ingroup Identity and Anti-Immigrant Prejudice: The Moderating Role of RWA and Outgroup Homogeneity. International Review of Social Psychology, 31(1): 13, 1–12, DOI: https://doi.org/10.5334/irsp.152

There is a strongly held and pervasive belief that nations have a homogeneous identity-i.e., a singular and distinct identity built on a common origin, culture and goal. Socially constructed ingroup homogeneity, paired with a perception of national essentialism, plays a fundamental role in delimitating clear-cut borders and in legitimizing the very existence of national entities (Gellner, 1997;Pehrson, Brown and Zagefka, 2009). Interestingly, expectations of homogeneity are also found on a transnational level, concerning more recently constructed social identities, such as the European identity. European identity is perceived as a shared and singular European identity (based on common ethnic roots, history and culture), which has made heterogeneity within Europe more difficult to accept for its members. Indeed, challenges to such a homogeneous perception of European identity, through the admission of nations that are perceived as too culturally or politically different (such as Turkey) is a longstanding debate within European countries (Caglar, 1997).
The expectation of homogeneous social identities is congruent with one of the most influential theoretical models in social psychology. Distinctiveness is a fundamental motive underlying not only inter-personal but also inter-group processes (Brewer, 1993;Jetten, Spears and Manstead, 1997;Jetten and Spears, 2003). A key postulate of social identity theory (SIT) is that individuals are motivated to attain and maintain a positive and distinct social identity (Tajfel andTurner, 1979, 1986). A group is perceived as distinct from other groups when intra-group variability is lower than inter-group variability (Turner, Oakes, Haslam and McGarthy, 1994;Turner, Hogg, Oakes, Reicher and Wetherell, 1987;Tajfel, 1969;Tajfel and Wilkes, 1963;see also Campbell, 1958;Simon and Brown, 1987). Accordingly, increasing intergroup similarity challenges ingroup distinctiveness (Gabarrot and Falomir-Pichastor, 2017). Similarly, accentuating the perception of ingroup homogeneity contributes to the attainment of a positively distinct social identity (Simon and Pettigrew, 1990). Of particular relevance for the present research, as a corollary, accentuating intra-group variability (i.e., heterogeneity) may reduce the perceived inter-group variability and therefore the ingroup distinctiveness. Thus, ingroup heterogeneity may challenge the fundamental psychological motive for positive ingroup distinctiveness.
How can group members deal with such a threat to ingroup distinctiveness? Overall, the definition of an ingroup identity can affect how we perceive and treat outgroup members (Reicher and Hopkins, 2001). Because a heterogeneous ingroup identity challenges the very essence of the ingroup and members' motive for a positive and distinct identity, ingroup heterogeneity may lead to ingroup members' increasing intergroup differentiation to reaffirm the threatened ingroup distinctiveness (Tajfel and Turner, 1979;Jetten, Spears and Postmes, 2004). Consequently, ingroup heterogeneity may increase outgroup derogation, at least among those group members who are particularly motivated to maintain clear and distinctive ingroup boundaries. Though enlightening, the existing literature on the impact of identity heterogeneity on prejudice is still sparse. To our knowledge, only two studies confirm such a prediction. A heterogeneous (vs. homogeneous) ingroup identity leads to more outgroup derogation amongst people high on conservative values (Roccas and Amit, 2011) and among high identifiers (Falomir-Pichastor and Frederic, 2013).
In the present research we aimed at increasing our understanding of the negative consequences of ingroup heterogeneity on outgroup derogation in two ways. Firstly, we wanted to extend the moderating role of individual differences in Right Wing Authoritarianism (RWA), given that individuals high in RWA may be particularly challenged by ingroup heterogeneity. Secondly, we also wanted to examine whether this pattern is moderated by the perception of the outgroup as more or less heterogeneous.

The Moderating Role of RWA
Not all ingroup members are similarly sensitive and concerned by the definition of their social identity. On the one hand, research has shown that high identifiers not only perceive their group to be more homogeneous (as a default), they also homogenize their group in response to threats (Castano and Yzerbyt, 1998;Doosje, Spears, Ellemers, and Koomen, 1999;Ellemers, Spears and Doosje, 1997;Simon and Klandermans, 2001). On the other hand, the more people value conservation the more they emphasize self-restriction, order and are resistant to change (Schwartz, 1992). They are also prone to value ingroup homogeneity as a resource and as an important attribute to their identity (Roccas and Amit, 2011). Accordingly, individuals scoring high on conservatism (Roccas and Amit, 2011) or on ingroup identification (Falomir-Pichastor and Frederic, 2013) express greater prejudice towards an outgroup when confronted with information describing their ingroup identity as heterogeneous (vs. homogeneous).
In the present research we contend that individuals high in Right Wing Authoritarianism (RWA) may be equally prone to perceive ingroup heterogeneity as threatening. Indeed, individuals with high scores on RWA value conservatism, authoritarianism and traditionalism (the three sub-dimensions of RWA) to a higher degree in order to preserve ingroup homogeneity and entitativity which is believed to protect the group from potential harm (Altemeyer, 1998;Duckitt, Wagner, du Plessis and Birum, 2002). Additionally, people scoring high on RWA tend to have a more negative perception of diversity in general (Kauff, Asbrock and Wagner, 2013), and hold a strong belief that their surrounding world is generally dangerous and therefore must be controlled (Onraet, Dhont and Van Hiel, 2014). Thus, homogeneous ingroup identities may be perceived as an important resource for individuals scoring high on RWA. These individuals seem to prefer their ingroup identity to be homogeneous and may thus be particularly unsatisfied with a heterogeneous ingroup identity. Accordingly, we expected RWA to moderate the effect of ingroup heterogeneity on outgroup derogation, and more specifically that ingroup heterogeneity will increase outgroup derogation in particular among individuals high in RWA (Hypothesis 1).

The Role of Outgroup Homogeneity
The perceived homogeneity versus heterogeneity of the ingroup, and its related consequences (e.g., outgroup derogation), does not occur in a vacuum. Ingroup identity is created by comparing the definition and perception of the ingroup to relevant outgroups (e.g., Barth, 1969;Tajfel and Turner, 1986). Therefore, in the present research we also aimed at investigating whether outgroup heterogeneity (vs. homogeneity) moderates the predicted pattern of results.
Despite that little is known about such a potential moderating role of the outgroup perception, a considerable body of research has shown a tendency to perceive the outgroup as homogeneous (as compared with the ingroup), and the consequences of this perception in terms of increased stereotyping and prejudice towards outgroup members (e.g., Linville, Salovey and Fischer, 1986;Messick and Mackie, 1989;Park, Judd and Ryan, 1991;Quattrone, 1986). Some authors argue that the outgroup homogeneity effect is due to the application of outgroup stereotype to each outgroup member, which facilitates the development of generalized negative attitudes (Park et al., 1991) and aggressive behaviour towards outgroups (Wilder, 1978).
While past research examined the effects of ingroup and outgroup heterogeneity on intergroup relations separately, it might seem relevant to examine the effect of these two factors conjointly. Indeed, the perception of the ingroup is constructed through the perception of the outgroup (e.g., Barth, 1969;Tajfel and Turner, 1986). As Simon stated (1992, p. 4), 'Consequently, if the perception of group homogeneity is to be investigated from an intergroup perspective, group members' perceptions of ingroup homogeneity should be examined in relation to their perceptions of outgroup homogeneity, and vice versa'.
Accordingly, in the present research we contend that the perceived homogeneity (vs. heterogeneity) of the outgroup should moderate ingroup members' reactive responses to ingroup distinctiveness. More specifically, whereas individuals high in RWA are expected to increase outgroup derogation when ingroup identity is heterogeneous (H1), this pattern should appear in particular when the outgroup is perceived as homogeneous. Indeed, in such circumstances individuals high in RWA may feel even more threatened by a highly homogeneous outgroup that constitutes a relevant target of comparison. This rationale is consistent with SCT, according to which, a collection of individuals constitutes a salient entity to the extent that they have a low degree of intra-group variability. Thus, outgroup homogeneity (vs. heterogeneity) makes the outgroup appear as a highly salient comparison target, which accentuates the inter-group contrast. By mere contrast, outgroup homogeneity emphasizes the heterogeneity of ingroup identity and thus the motivation among individuals high on RWA to restore ingroup distinctiveness by increasing outgroup derogation. Furthermore, as previously stated, homogeneous outgroups tend to be more prone to be stereotyped (Park et al., 1991;Wilder, 1978). A heterogeneous identity implies that the criteria used to define group membership is more flexible, and therefore challenges inter-group boundaries. At its most extreme, the flexibility in defining the ingroup could be interpreted by some as meaning that there is ultimately no identity to speak of, as there is a clear lack of consensus. By contrast, a homogeneous outgroup is highly entitative and could be perceived as having a distinct and clear identity to fill the void left by a heterogeneous ingroup identity. Thus, people high in RWA belonging to a group with a heterogeneous identity might be particularly motivated to create a clear distinction from a homogeneous outgroup as they might wish to avoid being associated with outgroup stereotypes as well as deterring any potential influence this outgroup could have in re-defining the ingroup identity. In sum, our second hypothesis established that outgroup heterogeneity moderates the predicted interaction between RWA and ingroup heterogeneity on outgroup derogation (H2). More specifically, group members particularly threatened by ingroup heterogeneity (i.e., those with high scores on RWA) may be particularly prone to increase derogation of homogeneous (vs. heterogeneous) outgroups.

The Present Studies
The goal of the present research is twofold: we seek to examine if RWA moderates the effect of ingroup heterogeneity on outgroup derogation (H1), and whether outgroup homogeneity (vs. heterogeneity) moderates this pattern (H2). In order to do so, we conducted three studies, two of which were set in a national context (in which we manipulated the heterogeneity of Swiss national identity), and one was set in the supra-national European context (in which we manipulated the heterogeneity of the European national identity). We initially assessed RWA in all studies, either manipulated ingroup heterogeneity (Studies 1-2) or kept it constant (Study 3), and also manipulated outgroup heterogeneity (Studies 2-3). The main dependent variable was outgroup derogation, as assessed through three different scales: subtle prejudice (Study 1), modern prejudice (Study 2) and intergroup discrimination (Study 3). We postulated that a heterogeneous (vs. homogeneous) ingroup identity should lead to more outgroup derogation in particular among individuals scoring high on RWA (H1), and that this effect should appear particularly when the outgroup is homogeneous (vs. heterogeneous) (H2).

Study 1
Study 1 examined the first hypothesis in the Swiss national context. We expected a heterogeneous (vs. homogeneous) Swiss national identity to lead to higher levels of prejudice toward immigrants particularly among people high in RWA.

Method
Participants. The study was conducted in 2013. Participants were recruited in different public spaces and university campuses in Geneva and filled in a paper and pencil questionnaire without compensation. We only retained those with Swiss nationality, which resulted in a final sample of 83 participants (47 women and 35 men, one participant did not declare their gender; M age = 28.24 years, SD = 13.91).
Procedure. Participants were asked to participate in a survey ostensibly regarding several socio-political themes concerning Swiss nationals. They initially indicated their level of RWA and were then randomly assigned to one of the two experimental conditions (ingroup identity: homogeneous vs. heterogeneous). Finally we introduced a manipulation check and assessed prejudice toward immigrants as our main dependent variable. Unless otherwise mentioned, answers to all questions in this study and all the following studies were collected on 7-points scales ranging from 1 ('Not at all') to 7 ('Absolutely'). At the end of the survey participants were fully debriefed. 1

RWA.
Right Wing authoritarianism was assessed using a reduced nine-item version of the Duckitt, Bizumic, Krauss and Helen's (2010) RWA scale, which was translated in French and adapted to the Swiss context. This multidimensional scale is composed of three classical subcomponents of RWA: authoritarianism (e.g., 'Our society does NOT need tougher government and stricter laws'), conservatism (e.g., 'The "old-fashioned ways" and "old-fashioned values" still show the best way to live') and traditionalism (e.g., 'This country will flourish if young people stop experimenting with drugs, sex, and alcohol.'). 2 However, for the purpose of the present research we computed an overall RWA score so that a higher score would indicate higher levels of RWA (α = 0.81, M = 3.79, SD = 1.08).
National identity heterogeneity. In order to ensure a manipulation of ingroup heterogeneity, participants were focused on the inherent heterogeneity (vs. homogeneity) of Swiss national identity, while the outgroup (i.e., immigrants) was excluded from the definition and mention of said heterogeneous (or homogeneous) national identity. Participants were presented with the same induction of Swiss national identity heterogeneity used by Falomir-Pichastor and Frederic (2013). More specifically, participants were presented with an excerpt of a fictitious report written by social scientists from various fields. In the homogeneous condition [heterogeneous condition in bracket] the report was titled 'The Swiss identity: an example of cultural singularity [diversity]' and stated that 'A recent survey of a representative sample of Swiss citizens showed that the Swiss identity is characterized by cultural singularity [ We talk about the Swiss identity in the singular [plural] for a number of reasons.' The report continued saying that several arguments supported this statement, namely that one of Switzerland's defining characteristics is its federalism, in that 'the country consists of a set of regions with the same [different] roots, customs, laws, objectives and views. This gives the confederation a homogeneous [heterogeneous] identity.' The final conclusion of the report was that 'Swiss identity is characterized by its homogeneity [heterogeneity] and its cultural singularity [diversity]' (see appendix 1 for the original induction).
As a reinforcement for this induction we asked participants to give personal arguments of what makes Swiss national identity homogeneous [heterogeneous].

Dependent variables
Manipulation check. The experimental manipulation was followed by a manipulation check item. We asked participants to indicate, according to their personal opinion, how they would describe Swiss identity on a scale ranging from 1 = homogeneous to 7 = heterogeneous (M = 3.79, SD = 1.08).
Prejudice. We finally assessed participants' level of prejudice through a ten-item scale. This scale was adapted from the Pettigrew and Meertens' (1995) subtle prejudice scale to the Swiss context and translated into French (e.g., Immigrants living here should not push themselves where they are not wanted; Many other groups have come to Switzerland and overcome prejudice and worked their way up; Immigrants should do the same without special favours; Immigrants have family values that are different to Swiss values; I have sympathy for Immigrants living here in Switzerland). A Prejudice score was computed such that a higher score indicates higher levels of prejudice (α = 0.85, M = 3.73, SD = 1.01).

Results and Discussion
We regressed the dependent variables on RWA (standardized scores), national identity heterogeneity (-1 = homogeneity; 1 = heterogeneity) and their interaction.  prejudice specifically among participants with high RWA. Unexpectedly, this study also showed that a homogeneous national identity increased prejudice among participants low in RWA. However, given that this effect was not replicated in Study 2 we will not discuss it here (see General Discussion). In the next two studies we tested hypothesis 2 according to which this pattern should appear, specifically when the outgroup is homogeneous.

Study 2
The design of this study was similar to that of Study 1 except that we also manipulated the level of heterogeneity of the outgroup identity. According to hypothesis 2, we expected that participants with high scores on RWA will show higher levels of prejudice in the heterogeneous (vs. homogeneous) ingroup identity specifically when the outgroup is described as homogeneous (vs. heterogeneous).

Method
Participants and procedure. The procedure was similar to the one in Study 1, with the only exception being that outgroup heterogeneity was manipulated after the ingroup heterogeneity manipulation. The final sample was 82 Swiss nationals recruited in public spaces and on university campuses in Geneva and Lausanne (49 women and 33 men; M age = 34.16 years, SD = 13.99). They were randomly assigned to one of the four experimental conditions in a 2 (ingroup: homogeneous vs. heterogeneous) × 2 (outgroup: homogeneous vs. heterogeneous) experimental design including RWA as a continuous variable. We assessed RWA (α = 0.79, M = 4.33, SD = 0.95) and used the same manipulation of the heterogeneity of national identity as in Study 1. Outgroup homogeneity. In order to manipulate outgroup homogeneity, participants read a fictitious report established by a panel of experts describing immigrants from ex-Yugoslavia as either homogeneous (an entity where the similarities and cohesion between the different countries composing it were highlighted) or heterogeneous (a group where the differences between, and independence of, the different countries were highlighted). The focus on a specific outgroup made it easier to develop relatable and credible arguments in favour of or against outgroup homogeneity. The heterogeneous outgroup condition included a drawn map drawn of the former republic of Yugoslavia highlighting the boarders between the different countries and mentioning the different countries different names. In the homogeneous outgroup condition the drawn map de-emphasized the boarders between the different countries and the whole territory was labelled as 'Ex-Yugoslavia'.

Dependent variables
Manipulation checks. After each experimental induction we introduced a manipulation check. Regarding ingroup heterogeneity, we used the same manipulation check (M = 4.74, SD = 1.72) as in Study 1. Regarding outgroup heterogeneity, we asked participants to indicate, accord-ing to their personal opinion, how they would describe immigrants from Ex-Yougoslavia on a scale ranging from 1 = homogeneous to 7 = heterogeneous (M = 3.76, SD = 1.90).
Prejudice. In this study, participants' level of prejudice towards immigrants from Ex-Yugoslavia was assessed using the nine-item modern prejudice scale (Akrami, Ekehammer, and Araya, 2000), which was adapted to the Swiss context and translated into French (e.g. Xenophobic and racist groups are no longer a threat to immigrants from Ex-Yugoslavia living in Switzerland; Special programs are needed to create jobs for immigrants from Ex-Yugoslavia; Immigrants from Ex-Yugoslavia are getting too demanding in their push for equal rights). A prejudice score was computed such as a higher score indicates higher levels of prejudice (α = 0.76, M = 3.76, SD = 0.86).

Results and Discussion
We regressed the dependent variables on RWA (standardized scores), ingroup heterogeneity and outgroup heterogeneity (both coded: -1 = homogeneity; 1 = heterogeneity) and their interactions. Accordingly, the results of this study provided consistent evidence in support of our second hypothesis. Ingroup heterogeneity increased prejudice towards immigrants among participants high in RWA specifically when the outgroup was homogeneous. In order to provide further support to Hypothesis 2, we conducted a third study with a more simplified design using a different paradigm.

Study 3
In this study we introduced several methodological changes. To simplify the experimental design, we assessed RWA and experimentally manipulated outgroup homogeneity, but kept constant the critical ingroup heterogeneity condition. Furthermore, we also wanted to examine if our predictions still hold in a different context, such as the E.U. Accordingly, the European identity was presented as heterogeneous for all participants, while outgroup heterogeneity was induced by manipulating the heterogeneity of immigrants from non-European countries. Finally, we assessed outgroup derogation through a measure of intergroup discrimination. In this study, according to Hypothesis 2, and given that the ingroup heterogeneity condition was kept constant, we expected a RWA × outgroup heterogeneity interaction effect. More specifically, individuals with higher scores on RWA will show higher discrimina-tion against immigrants, and this relationship would be stronger when immigrants are described as homogeneous (vs. heterogeneous).

Method
Participants and procedure. The study was conducted in 2017. We recruited participants through Foule factory online platform, and only those with French (i.e., E.U.) nationality were retained for the analyses. They were compensated 1.5 euros and the 76 recruited participants (39 women and 37 men; M age = 36.63 years, SD = 12.16) were randomly assigned to one of two experimental conditions in a 2 (outgroup: homogeneous vs. heterogeneous) with RWA (as a standardized covariate) design.
Ingroup heterogeneity. The heterogeneity of the European identity was induced and kept constant across conditions. Thus, all participants read a excerpt of a report allegedly written by social scientists from various fields about the heterogeneity of the European identity. The report was titled 'The European identity: An example of cultural diversity' and stated that 'A recent survey of a representative sample of European citizens showed that the European identity is characterized by cultural diversity. Because of the existence of different countries and languages, the European Union constitutes a heterogeneous conglomerate in which there are different ways of feeling European. These findings are in line with the opinions of numerous researchers (sociologists, economists, psychologists, and political scientists). We talk about the European identity in plural for a number of reasons.' The report continued to say that several arguments supported this statement namely that the 'E.U. consists of a set of countries with different roots, customs, laws, objectives and views. This gives the E.U. a heterogeneous identity.' The  In-group heterogeneity In-group homogeneity Prejudice Out-group heterogeneity Out-group homogeneity final conclusion of the report was that 'European identity is characterized by its heterogeneity and its cultural diversity'. We then finally manipulated outgroup homogeneity and assessed discrimination.

Outgroup heterogeneity.
In order to manipulate outgroup heterogeneity, participants read an excerpt of a fictitious interview allegedly conducted with a professor and specialist in the field of migration to the E.U. This report described immigrants from non-European countries as either homogeneous (an entity where the similarities and cohesion between the different countries composing it were highlighted) or heterogeneous (a group where the differences between, and independence of, the different countries were highlighted). To strengthen the induction, we took the example of immigrants from North Africa as having similarities (vs. differences) in political, religious and cultural practices. The report concluded with the fact that generally immigrants from non-European countries coming to the E.U. are either homogeneous or heterogeneous (depending on the experimental condition).

Dependent variables
Manipulation check. Right after the induction of the outgroup heterogeneity we included two items as manipulation check. We asked participants to indicate, according to their personal opinion, how they would describe non-European immigrants on a scale ranging from 1 = homogeneous to 7 = heterogeneous and from 1 = singular to 7 = diverse. We created a composite score by averaging these two items (α = 0.75, M = 4.51, SD = 1.96), the higher the score the more the outgroup was perceived to be heterogeneous.
Intentions to discriminate. In this study we developed a new 10-item intergroup discrimination scale assessing the extent to which Europeans favour their ingroup against non-European immigrants. The scale items were adapted from a measure of intergroup discrimination in which nationals had to distribute different resources between the ingroup and immigrants (Falomir-Pichastor, Munoz-Rojas, Invernizzi and Mugny, 2004). The new scale includes items such as: In the E.U., European citizens should have a priority on non-European immigrants in the job market; In the E.U, European citizens should benefit from more social services and higher benefits than non-European immigrants (see appendix 2 for overview of the ten items). A discrimination score was computed such that a higher score would indicate higher levels of discrimination (α = 0.92, M = 3.66, SD = 1.48).

Results and Discussion
We regressed the dependent variables on RWA (standardized values), outgroup (-1 = heterogeneity; 1 = homogeneity), as well as their interaction.

General Discussion
The present research aimed at examining the effects of a heterogeneous ingroup identity on outgroup derogation by focusing on two moderators: RWA and outgroup heterogeneity. Overall, the three studies provide support for the hypotheses that the perception of a heterogeneous ingroup identity increases outgroup derogation particularly among individuals scoring high on RWA (H1; Study 1). Furthermore, this pattern is observed specifically when the outgroup is homogeneous (as compared to heterogeneous, H2 and H3; Studies 2-3). All three studies also showed that participants high in RWA overall perceive both their ingroup and the immigrant outgroups as relatively homogeneous. Thus, challenging the perception of ingroup homogeneity increases outgroup derogation amongst a certain subset of ingroup members who tend to perceive homogeneity as a fundamental attribute of a satisfying and distinct social identity. Furthermore, this effect appears specifically when the outgroup is perceived as homogeneous.
Before discussing the present findings, several limitations of the present research should be acknowledged. First, the three conducted studies included a relatively small sample size, which suggests they are underpowered. Second, another limitation relates to the fact that we kept the heterogeneity of the ingroup identity constant in Study 3, which challenges a comprehensive comparison of findings between Study 2 and Study 3. Despite the relevance of these limitations, we should highlight that the present results were consistent across studies and were observed across two different contexts. Whereas Studies 1 and 2 were conducted in the context of national identity, Study 3 was conducted in the context of European identity, which suggests that these findings can be generalized to other constructed identities. However, it is worth noting that these contexts can also bring cultural differences that were not investigated in the present research. For instance, Swiss national identity, and to an even greater degree the E.U. identity, may overall be perceived as heterogeneous identities because of the strong political, cultural and linguistic differences between sub-groups. One may wonder whether a similar pattern of findings could be observed in more homogeneous contexts (e.g., French national identity). At the same time, given that the present findings were observed in a context in which ingroup identity is strongly perceived as heterogeneous, one may assume this pattern of findings to be replicated, or even strengthened, in contexts where expectations about ingroup heterogeneity are lower. Future studies should examine this issue, as well as investigate whether the present findings will be replicated regarding alternative intergroup contexts such as those related to gender, sexual orientation, religion or political identities.
The present findings are of relevance for research on intergroup relations. First, we also assumed that information about the homogeneity of ingroup identity may be reassuring for people scoring high on RWA, namely because ingroup homogeneity preserves the clear and expected ingroup distinctiveness (Jetten et al., 2004). However, whereas the results of Study 1 fully support our first hypothesis regarding participants high in RWA, the opposite pattern was found for participants scoring low on RWA. Indeed, people low on RWA express more prejudice towards immigrants when ingroup identity is homogeneous (vs. heterogeneous). This finding was not replicated in Studies 2 and 3, but these studies introduced a different and more complex experimental design. Therefore, we should consider the possibility that, in contrast to people high on RWA, people low on RWA may perceive or even prefer ingroup identity as more heterogeneous.
This unexpected finding is in line with different fields of research. Firstly, research suggests that the lower the score on RWA the higher openness to diversity (Kauff et al., 2013). Thus, individuals scoring low on RWA may perceive a homogeneous ingroup identity as more unsatisfying than a heterogeneous ingroup identity, and therefore increase outgroup prejudice. Secondly, this understanding is also consistent with the observed link between RWA and the perceived ingroup heterogeneity (manipulation check): the lower the score on RWA the less participants tend to perceive ingroup identity as homogeneous. Thus, RWA seems to reflect an initial preference (or default perception) regarding the heterogeneity of ingroup identity, which subsequently disposes group members to react against inconsistent information by increasing outgroup derogation. Finally, this finding is also consistent with literature showing that high identifiers tend to engage in a process of reactive distinctiveness, while low identifiers tend to engage in a process of reflective distinctiveness, meaning that greater inter-group distinctiveness leads to more outgroup derogation (Jetten et al., 2004). Thus, people low on RWA may have reacted to ingroup homogeneity through a reflective (rather than reactive) distinctiveness processes. Further research is needed to extend our knowledge about the possibility that participants low in RWA increase outgroup derogation specifically when ingroup identity is described as homogeneous.
Second, the present findings provide consistent support for hypothesis H2. More specifically, individuals high in RWA increased outgroup derogation in the heterogeneous ingroup identity condition specifically when the immigrant outgroup was described as homogeneous (vs. heterogeneous). However, we should acknowledge that the mechanism underlying this effect was basically assumed but was not directly explored in the present research. However, this prediction is consistent with social identity and intergroup relations perspectives. Indeed, homogeneous outgroups can be more prone to deindividuation and stereotyping, and ingroup members can perceive them as more likely to be a target for derogation (Park et al., 1991). Moreover, SIT and SCT suggest that ingroup perception is influenced by outgroup perception. More specifically, a homogeneous outgroup may exacerbate the salience of ingroup heterogeneity and its potential challenges, which would drive ingroup members to express greater outgroup derogation in order to restore a positive and distinct identity. Globally our effects are believed to be due to a threat to ingroup distinctiveness. However, two mechanisms underlying our effects can be derived from the aforementioned literature. The first one suggests that a homogeneous outgroup would be more prone to being derogated because it is more easily stereotyped. The motivation to distinguish oneself from this group could be greater as people high in RWA would avoid having their ingroup being associated with said stereotypes. The second interpretation suggests that outgroup homogeneity merely emphasizes the heterogeneity of the ingroup via a contrast effect which prompts people high in RWA to restore the lack of distinctiveness that comes with a heterogeneous ingroup identity by derogating said outgroup. In the first instance the outgroup plays a more active role in contributing to the threat as people high in RWA might fear that the outgroup can contribute negatively to how the ingroup define themselves, particularly when they are left with a self-definition that is open to interpretation and influence (when ingroup identity is heterogeneous). In the second instance, the outgroup plays a more passive role as it is simply a salient target of comparison and derogation, a means to an end, namely restoring ingroup distinctiveness at all costs. Future research should investigate this question further, but as for the present research, it is possible that these two mechanisms operate conjointly as they are not mutually exclusive. Indeed, a homogeneous outgroup could be threatening because it simultaneously accentuates the heterogeneity of ingroup identity and because it is more likely (compared to a heterogeneous outgroup) to be perceived as an unwanted entitative and stereotypical force that could fill the void left by a heterogeneous ingroup identity. Although the present research did not directly seek to prove that one mechanism was at play over the other, our research at least does suggest that it is wise to examine ingroup identity heterogeneity in contrast to the outgroup homogeneity effect.
Third, the present research focused specifically on outgroup heterogeneity, but there are reasons to believe that other outgroup characteristics may be of relevance. For instance, in Study 1 the immigrant group was generic (without any reference to a specific immigrant group), and past research has showed that when the origins of an immigrant group are not explicit they are perceived as more incompetent and untrustworthy-i.e., as to be of lower status (Cuddy, Fiske, Demoulin and Leyens, 2000;Eckes, 2002). Furthermore, in Study 2 immigrants were specified to be from former Yugoslavia, and in Study 3 they were specified to be from non-European countries but the emphasis in the outgroup homogeneity induction was on immigrants from North Africa. Therefore, across the studies immigrants could have been perceived to be of lower status than, and culturally dissimilar to, ingroup members. Since past research suggests that low status (vs. high status) homogeneous groups are more likely to be targets of prejudice (Badea, Brauer, and Rubin, 2012;Er-Rafiy andBrauer, 2012, 2013), one may suggest that the present findings are relevant in explaining prejudice specifically towards low-status immigrants. However, research also shows that high-status outgroups are not necessarily exempt from prejudice. Indeed, when ingroup members face an outgroup of higher status they can find themselves in an unfavourable comparison to this outgroup, and, according to the postulate of SIT, the motivation for positive distinctiveness could be threatened. This is in line with the literature on relative deprivation demonstrating that higher-status outgroups could be the target of prejudice because ingroup members feel deprived compared to higher-status outgroup members due to their lower status (Vanneman and Pettigrew, 1972;Dambrun and Guimond, 2001;Pettigrew and Meertens, 1995;Dambrun, Maisonneuve, Duarte, and Guimond, 2002). Thus, certain ingroup members facing an unsatisfying heterogeneous ingroup identity could express more negative attitudes towards a higher-status outgroup.
Third, research has shown that culturally dissimilar outgroups are more likely to be a target of prejudice (Mahfud, Badea, and N'gbala, 2015). However, past research has also shown that perceived (cultural) similarity between the ingroup and the outgroup does not moderate the effect of ingroup heterogeneity on outgroup derogation (Falomir-Pichastor and Frederic, 2013). This finding is consistent with research within the SIT framework showing that high identifiers may be threatened by both low and high cultural similarity but for different reasons (Jetten, Spears and Manstead, 2001). This could suggest that outgroup homogeneity and not similarity has a potential moderating role on the effects of ingroup identity heterogeneity on anti-immigrant prejudice.
Finally, one remaining question regards the extent to which the moderating role of RWA activates the same processes as alternative individual's differences such as group identification and SDO. Despite some similarities, different mechanisms can also be advanced. On the one hand, high identifiers and people high in RWA may react differently to the threats introduced by a heterogeneous ingroup identity, as well as threats introduced by specific outgroups. While high identifiers are particularly concerned by maintaining ingroup distinctiveness at all costs (even when facing a valued group; see Jetten et al., 2003), people high on RWA might be less sensitive to distinctiveness threats and more concerned by the outgroup threatening characteristics. Indeed, past research suggests that people high on RWA are particularly sensitive to external threat (Onraet, Dhont and Van Hiel, 2014). Similarly to high identifiers, research has shown that people high on SDO are motivated to maintain positive distinctiveness and clear-cut inter-group boundaries even from immigrants who seek to assimilate (vs. those who do not) (Thomsen, Green and Sidanius, 2008). One could expect people high on SDO to exhibit similar attitudes to high identifiers when facing a heterogeneous ingroup identity as they might also be concerned by maintaining ingroup distinctiveness at all costs rather than being concerned with outgroup characteristics. Future research should look into the diverging manners in which individual differences moderate the investigated processes.

Conclusion
The perception of national or supra-national (i.e., European) identities, as well as the perception of outgroups such as immigrants are continuously constructed through various socio-political agents as well as in everyday discourse. Given that how we define and perceive ingroup identities and outgroups has the potential to impact prejudice and discrimination against immigrants, it is important to understand how these perceptions could conjointly affect attitudes and policy making. While the present research showed that promoting a homogeneous ingroup identity reduced immigrant derogation amongst people high in RWA, the policies resulting from these findings would not necessarily benefit immigrants. Indeed, research has shown that positive attitudes towards an outgroup does not necessarily guarantee better conditions for minority outgroups (Wright and Baray, 2012). Furthermore, a homogeneous ingroup identity may guarantee clear intergroup boundaries and ingroup distinctiveness. Consequently, if immigrants were to seek to integrate into a national or European group, they could be faced with a homogeneous ingroup identity that marginalizes them as it leaves little wiggle room for diverging definitions of the ingroup identity. Therefore, ingroup homogeneity would not only encourage assimilationist ideologies and policies but also simultaneously encourage policies that make it more difficult for immigrants to integrate the ingroup. If immigrants are perceived to form a distinct group from nationals they could be perceived and treated as second-class citizens and be expected to abide by different rules compared to ingroup members in order to make it more difficult for them to integrate the in-group and re-appropriate the in-group identity (e.g., expulsion of criminal immigrants, stricter criteria to become part of the ingroup, higher expectations of immigrants to contribute positively to society, obligation for immigrants to choose between the nationality of the in-group and their original nationality, etc.). Conversely, the present findings suggest that simultaneously promoting a heterogeneous ingroup identity and an image of immigrants as heterogeneous may help reduce prejudice and discrimination against immigrants while simultaneously encouraging an environment that could be more inclusive of immigrants who wish to integrate the ingroup (Roblain, Malki, Azzi and Licata, 2017).

Additional Files
The additional files for this article can be found as follows: • Appendix 1. Original Induction of National identity heterogeneity. https://doi.org/10.5334/irsp.152.s1 • Appendix 2. Original items of the ten-item discrimination scale. https://doi.org/10.5334/irsp.152.s1 Notes 1 For this study and the following studies, participants were fully debriefed at the end of the survey. 2 Given that we were interested in RWA as a whole concept rather than focusing on the subdimensions separately, analyses were made with the full scale in all studies.