The Influence of Interethnic Ideologies on Intention to Discriminate through In-Group Favoritism

This study examines the effects of two interethnic ideologies (assimilation and multiculturalism) on in-group favoritism and discrimination intention toward immigrants. Specifically, this study aims to test the concomitant impact of these two ideologies on intergroup biases in order to affirm whether these two paths are related to intergroup bias. Moreover, this study is designed to extend previous work that found relationships between interethnic ideologies and in-group favoritism to discrimination intention. Graduate students in management programs (N = 182) answered a questionnaire. The findings show that both interethnic ideologies are concomitantly related to in-group favoritism. In particular, while assimilation is positively related to in-group favoritism, multiculturalism is negatively related to in-group favoritism. Additionally, it shows evidence of indirect relationships between interethnic ideologies and the discrimination intention through in-group favoritism. The results are discussed in light of interethnic ideologies literature and presents directions for future research.


Introduction
Societies today are increasingly culturally diverse, meaning that people from different origins are living together (e.g., [1]).Such diversity is often accompanied by discrimination toward specific groups despite the fact that anti-discrimination legislation has been widely adopted since the nineties OPEN ACCESS (e.g., [2,3]).To improve our knowledge regarding this impediment, scholars working in the field of intergroup relationships have proposed that interethnic ideologies of diversity (i.e., beliefs about how to manage diversity in diverse societies) can reduce discrimination and even more broadly reduce intergroup bias 1 .These ideologies have already been examined within educational and national contexts but only a few studies have directly made use of them within workplaces (e.g., [5,6]).Moreover, while attitudes have been used as several times dependent variables, behavioral intention outcomes are still scarce, if not non-existent.A threefold contribution will be made to interethnic ideologies literature regarding (1) the concomitant impact of interethnic ideologies; (2) the investigated dependent variables; and (3) the role of bias as an underlying process.

Interethnic Ideologies and Intergroup Bias
Two interethnic ideologies have been conceptualized and contrasted in intergroup literature (e.g., [7][8][9]).They are defined as systems of ideas that suggest "how to best organize a diverse society" ( [10], p. 338).The first ideology, called assimilation, is a kind of colorblindness ideology.It suggests that group differences should be ignored and that minority groups should adopt the majority group's perspective (e.g., [9]).The second ideology, called multiculturalism, suggests valuing group differences (e.g., [11]).Regarding their respective effects on intergroup relationships, assimilation endorsement is positively linked to evaluative bias whereas multiculturalism is negatively related to evaluative bias [11][12][13].For example, attitudes toward multiculturalism are positively related to positive judgments about Muslims Turkish for Dutch adolescents [11] and negatively linked to perceived cultural distance of different ethnic groups for Dutch participants [14].Ideologies differ from acculturation framework, which refers to "the dual process of cultural and psychological change that takes place as a result of contact between two or more cultural groups" ( [15], p. 698).Acculturation refers to a process of change, while ideologies are the attitudes and normative expectations about what should be done to manage diversity.Despite this conceptual distinction, acculturation orientation can be linked with interethnic ideologies.In particular, similarities exist between the following: (1) assimilation acculturation orientation and assimilation interethnic ideology; and (2) multiculturalism acculturation orientation and multiculturalism interethnic ideology.
Based on past literature, our study aims to extend research on interethnic ideologies and intergroup biases in three directions regarding the following: (1) the conjoint impact of both ideologies; (2) investigated dependent variable; and (3) underlying processes.
First, interethnic ideology studies rarely simultaneously introduce multiculturalism and assimilation within the same analyses.To affirm that both interethnic ideologies play a conjoint role in intergroup bias, we need to show that they have a concomitant effect on the criteria.Indeed, it is possible that one ideology contributes more than the other when explaining in-group favoritism where only one of the two would be significant when both are simultaneously introduced.This issue is central.If only one ideology is related to intergroup bias, it means that there is only one option for acting on intergroup 1 The term "intergroup bias" is broadly used in this article referring to "as an unfair evaluative, emotional, cognitive, or behavioral response toward another group in ways that devalue or disadvantage the other group and its members either directly or indirectly by valuing or privileging members of one's own group" ( [4], p. 1084).It refers to stereotypes, prejudice, intention and discrimination.bias through interethnic ideologies.In the opposite case, a conjoint effect implies two available ways to do so.Some recent and uncommon studies investigated both interethnic ideologies [4,11,16,17], but more research is still needed in this direction.
Second, studies on interethnic ideologies have frequently examined stereotypes or attitudes (i.e., intergroup bias) toward immigrants (e.g., [18]).What has been less examined are the behavioral outcomes or intentions.According to the theory of reasoned action (e.g., [19]), attitude is not the best predictor of behavior.Compatibility argument proposed in this theory states that a behavior is better explained by proximal attitudes toward the behavior, like intentions, rather than a more distal attitude toward the object.The intention to perform a behavior is more specific than a general positive or negative attitude toward an object or a group.It is a motivational factor that indicates the effort and willingness to perform a behavior, and it is "the closest cognitive antecedent of actual behavioral performance" ( [19], p. 188).This theory is largely supported in the literature regarding different fields such as physical activity [20] or sun-safe behaviors [21].In the specific field of intergroup relationships, this is also supported by assumptions of the "bias map model" [22].The empirical test of this model supports the relationships between attitudes and behavioral tendencies.Finally, a meta-analysis shows that the relationship between the intention to discriminate and discriminatory behaviors is stronger than the relationship between attitudes and discriminatory behaviors [23].
Beyond empirical support for the relationship between attitudes and intention, two initial attempts to investigate the intention to discriminate have been made concerning acculturation strategies [24,25].Despite their contributions to the literature, these studies have certain limitations and call for further investigation.We present the two studies successively and discuss their limits.Wagner, Tisserant, and Bourhis [24] aimed to investigate relationships between acculturation strategies and discrimination intention (i.e., called "propensity to discriminate" in their paper) by students and workers toward North African immigrants of a Muslim background.They showed that the assimilation acculturation orientation is positively related to the propensity to discriminate within a sample of students but not the others acculturation orientations (i.e., individualism, integration, integration-transformation, segregation and exclusion).For a sample of workers, the integration-transformation acculturation orientation is negatively related to the propensity to discriminate but this is not the case for others acculturation orientations.Their results also indicate that identification with the endogroup (i.e., French people) was not related to acculturation orientations.Additionally, a general measure (including threat, negative evaluations, and social dominance orientation) was related to assimilation for students and to integration-transformation for workers.Particularly, the acculturation orientation mediated the relationship between the general measure and the propensity to discriminate.Despite the interesting conceptualization of discrimination intentions suggested by Wagner et al. [24], the fact that participants received explicit instructions to hire neither a Muslim nor a woman somewhat undermines the conclusions.Given this incitement to discriminate, participants may have conformed to the instructions rather than reported their actual individual levels of discrimination intention.This fact may have exacerbated degree of discrimination intention that was reported.The second study used another measure that does not suffer from this issue.Tiboulet et al. [25] examined the relationships between ethnocentric acculturation orientations (including assimilation, segregation, and exclusion), non-ethnocentric acculturation orientations (including integration and individualism), prejudice and behavioral intentions during recruitment.They showed a positive relation between ethnocentric acculturation orientations on the one hand, and then prejudice and the intention to discriminate on the other hand.However, they created two categories of acculturation orientations (i.e., non-ethnocentric and ethnocentric) rather than address each strategy separately.Moreover, they worked on the acculturation framework, whereas our focus is on diversity ideologies.
The third contribution to interethnic literature is related to underlying processes between interethnic ideologies and the intention to discriminate.To our knowledge, no study (neither Wagner et al. [24] nor Tiboulet et al. [25]) has shown that in-group favoritism is a mediator of the relationships between interethnic ideologies and the intention to discriminate.Based on Tiboulet et al.'s [25] work, such an indirect effect can nevertheless be expected.By regressing the intention to discriminate on ethnocentric acculturation orientation and prejudice they found that only prejudice, not ethnocentric acculturation orientation, remained a significant predictor of the intention to discriminate.This points the possibility of mediation.
In view of the forementioned results, our study aims to investigate the relationships between both interethnic ideologies on the intention to discriminate through in-group favoritism.Previous work on interethnic ideologies leads us to hypothesize that multiculturalism (assimilation) endorsement is negatively (positively) related to in-group favoritism toward immigrants which, in turn, is positively related to the intention to discriminate during hiring decisions, H1 and H2, respectively.

Participants and Procedure
The questionnaire was distributed to graduate students in management (Belgium universities) as part of a larger questionnaire on diversity and relationships between workers through an online survey (N = 41) or paper-and-pencil questionnaires during class (N = 141).Because the focus was on intergroup relationships between Belgians (i.e., the endogroup) and immigrants (i.e., the exogroup), participants whose reported origins were outside the European Union of 15 were removed (e.g., Moroccan, Congolese, or Polish students).This criterion was chosen because these people encounter the most difficulties when entering the labor market (e.g., [26,27]).Individuals whose countries of origin were part of the EU-15 and who did not suffer from discrimination or suffered less were kept in the sample.Among the final sample, 54% of the participants were male and the average age was 24.Participants were required to complete the scales in the order that they appeared, namely in the order of the hypothesized model.A cover letter ensured confidentiality and provided researchers' contact information.Individuals were invited to continue after they giving their free and informed consent.

Measures
Except for in-group favoritism and control variables, participants indicated their agreement with items using a 7-point Likert-type scale.
Multiculturalism.Five items adapted from Wolsko et al. [12] and Iweins, Desmette, Yzerbyt, and Stinglhamber [28] were designed to measure multiculturalism (e.g., "I find it important to take into account features of workers from each origin".Please see the Appendix, which provides all the items).
Assimilation.The six-item scale of assimilation from Wolsko et al. [12] was adapted to measure assimilation (e.g., "I believe that all workers, irrespective of their ethnic origin, have to embrace the work values of Belgian society".Please see the Appendix, which provides all the items).
In-Group favoritism.Participants were asked to evaluate how they feel regarding six groups, including Belgians and immigrants, using a thermometer ranging from 0 to 100.In-group favoritism was the difference between the Belgian and immigrant evaluations (e.g., [8]).
Discrimination intention in hiring decisions.The Tiboulet et al. [25] scale was used to measure discrimination intention in hiring decisions.Participants have to imagine themselves in a situation where they are in charge of recruiting of an immigrant candidate.Participants had to indicate whether a set of five criteria (e.g., origin, neighborhood of birth, wearing religious symbols or from the community of belonging, foreign appearance, and personal choices due to a national belonging) would influence their decision.We left out two criteria from the initial seven because they were specific to a consultant recruiter function.
Control variables.Participants were required to indicate their gender, age, and education level.Because we have recourse to two kinds of procurement, online surveys and paper-and-pencil questionnaires, we controlled for procurement.Finally, because participants were either Belgian or immigrants from the EU-15, we used participants' origin as a control variable.

Exploratory Factor Analysis and Preliminary Analyses
To determine the factor structure of the items from different scales, an exploratory factor analysis with principal components was used (SPSS22).Given the likelihood of correlations among constructs, an oblique rotation was conducted.Kaiser-Guttman's "eigenvalues greater than one" [29,30] and scree plot criteria indicated a four-factor structure.All factor loadings were more than 0.40 [29,31], and the amount of variance extracted from the four-factor solution was 61.31%.None of the item communalities were greater than 0.80 [29].Based on the exploratory factor analysis, we computed the items' means for each factor: multiculturalism, assimilation and intention to discriminate.For in-group favoritism, we computed a difference score (see the "measures" section).The factor loadings, eigenvalues, and percentages of explained variance are presented in Table 1.The difference between the Belgian and immigrant evaluations was statistically significant (t(181) = 9.73, p < 0.001), therefore we computed the in-group favoritism index (i.e., the difference score).
Descriptive statistics, Cronbach's alpha, and correlations are shown in Table 2.As indicated in Table 2, age is negatively related to in-group favoritism (r = −0.19,p = 0.01): older people have less in-group favoritism than younger people.Procurement is negatively related to discrimination intention.As recommended by Becker [32], control variables that are correlated to our dependent variables were entered in the relevant analyses to ensure that our expected relationships remain beyond the effects of sociodemographic variables.Interestingly, the correlation between multiculturalism and assimilation is not significant (r = −0.07,p = 0.330) (see the discussion section).Note: N = 182; α coefficients are reported on the diagonal (written in brackets).Gender was coded 1 = male and 2 = female.Procurement was coded 1 = paper-and-pencil and 2 = online.Origin was coded 1 = Belgian and 2 = immigrants from EU-15 (cf.participants and procedure section); * p < 0.05, ** p < 0.01, *** p < 0.001.
As seen in Table 2, the correlations between both interethnic ideologies and the intention to discriminate are not significant.The relationships between independent and dependent variables are no longer needed to show a mediation effect when variables are distal or when small size effects are expected; significant indirect effects allow mediation effects to be observed (e.g., [33]).In this respect, it is preferable to use the terminology "indirect effect" rather than mediation [34]; however, terminology does not impact the empirical results.

Relationships among Variables
We tested our hypotheses by modeling inferred causal processes using multiple regressions.As recommended by Hayes [34], we used a bootstrapping procedure to compute a bias-corrected and accelerated confidence interval around the indirect effect.
The results 2 (see Table 3) indicate that multiculturalism is negatively related to in-group favoritism (β = −0.22,p = 0.002) which, in turn, is positively related to the intention to discriminate in hiring (β = 0.20, p = 0.009), thereby controlling for assimilation.Because the confidence interval of the bootstrap analyses did not include zero, we concluded an indirect effect of multiculturalism on the intention to discriminate during hiring (BCa95%CI = [−0.160;−0.007]).These results confirm hypothesis 1.As formulated in hypothesis 2, assimilation is positively related to in-group favoritism (β = 0.16, p = 0.022), which in turn is linked to the intention to discriminate in hiring.The indirect effect of assimilation on the intention to discriminate in hiring decisions through in-group favoritism is significant (BCa95%CI = [0.010;0.120]) 3 .Table 4 summarizes total, direct, and indirect effects for each regression, i.e., with multiculturalism as independent variable (IV) and with assimilation as IV.In the macros used to compute the bootstrapping procedure, we can introduce control variables.Because some control variables are correlated with our dependent variables (in-group favoritism and intention to discriminate), we introduce them in our analyses with in-group favoritism and the discrimination intention as dependent variables.These results are presented in the text.When only the significant control variables are introduced, namely age for in-group favoritism and procurement for the hiring discrimination intention, p-values and betas are slightly different, but the results and conclusions are similar.

Conclusions
Based on interethnic ideologies literature, our study investigated the relationships between interethnic ideologies on the one hand and in-group favoritism and the intention to discriminate on the other.
Beyond reaffirming the positive (negative) relationship between assimilation (multiculturalism) and in-group favoritism, our findings shed new light on the literature on interethnic ideologies by making three contributions.First, the results support the concomitant impact of both interethnic ideologies on intergroup bias.The negative relationship between multiculturalism and in-group favoritism remains significant when assimilation is introduced as a control; the same result is observed for assimilation controlling for multiculturalism.Both interethnic ideologies are concomitantly related to in-group favoritism indicating two ways to decrease intergroup bias (cf.practical implication).Second, they extend the well-established relationships between interethnic ideologies and attitudinal measures of bias to the intention to discriminate.These results indicate that interethnic ideologies are related to behavioral intention outcomes, and that they can contribute to decrease organizational discrimination.However, the direct relationships between interethnic ideologies and the discrimination intention are not significant.While our design does not allow for the conclusion that intentions are more predictive of discrimination than attitudes, our results are in line with the "theory of reasoned action" (e.g., [19]) and the "bias map model", which both show that attitudes are an antecedent of behavioral tendencies [22].Finally, the effects of ideologies on the discrimination intention are indirect through in-group favoritism, indicating that it is an underlying process.
Because multiculturalism values group differences while assimilation ignores them, a negative correlation can conceptually be expected.However, these ideologies are not mutually exclusive [35] and the preferences for interethnic ideologies depend on the context in which they are assessed [36].A debate exists in the literature on the valence-positive or negative-of the relationship between these two ideologies.Previous works highlighted that this correlation can depend on the group (majority versus minority) and the chosen measure of interethnic ideologies [11].Moreover, despite interethnic ideologies being different from acculturation strategies, some scholars also did not find correlations between integration and assimilation (e.g., [24]).This is similar to our findings regarding ideologies.
Beyond its contributions, this study has some limitations and leaves unsolved issues for future research.The cross-sectional design of our study does not allow for causal conclusions.However, the direction of the proposed relationships has already been established, particularly through experimental designs (e.g., [22,37]).Some authors suggest that while the ideologies-intergroup bias relationships are supported, the inverse relationship is also relevant (e.g., [38]).Future research needs to test these relationships through panel data that allow for testing the direction of causation and, if the two are significant, determine which direction is the strongest.
The sample of management students in our study threatens the ecological validity which is another limitation.The debate in the literature surrounding the use of students as surrogate managers [39,40] has not yet been resolved.The similarity of the results between the two populations depends on the following: (1) the sample being considered, i.e., graduate students in management versus others majors [41]; (2) the considered outcome (e.g., [41,42]); and (3) the context, i.e., cooperative versus competitive [43].Regarding our study, the sample comprised of management students in a non-competitive context.Our outcomes were attitudinal measures, which can differ from those of working managers [39].Despite previous studies with similar findings among workers (e.g., [24,25]), caution should be taken when generalizing our findings to working managers, and the study should be replicated using a sample of workers.The generalizability of our study is also threatened by our small sample size and the fact that we investigated the hypothesized relationships in only one context.To increase the generalizability of our findings, future research should replicate these results among various work contexts.
Our research allows us to provide insights into directions of future research.First, more research is needed to further investigate the effects of interethnic ideologies on behavioral measures of discrimination.Despite intentions already being better predictors of discrimination than attitudes, intentions and behaviors remain different.Some scholars have proposed procedures aimed at directly measuring discrimination (e.g., [44]).Second, the focus of our study was on two frequently contrasted interethnic ideologies, namely assimilation and multiculturalism.Two other ideologies are suggested in the literature: colorblindness and polyculturalism.Future research should add these two conceptions to the two examined in this paper to provide a more detailed portrayal of the ideologies-bias relationships.Regarding colorblindness, this ideology suggests ignoring group differences.Such an ideology can be implemented in different forms.Assimilation is one of them; it specifies that minority groups should adopt the majority group's perspective (e.g., [8,10]).The link between colorblindness (in any other form than assimilation) and intergroup relationships is ambiguous (e.g., [45,46]).For example, participants in a multiculturalism or colorblindness condition have less bias than participants in a control condition with no difference between multiculturalism and colorblindness conditions [46].Conversely, Richeson and Nussbaum [37] show that compared to the multiculturalism condition, the colorblindness condition generated more bias (explicit and implicit).Regarding the proposition of a polyculturalist ideology, it is based on the observation that none of the usual ideologies fully meet the needs for affiliations of majority and minority groups' members simultaneously [7,47].Polyculturalism "recognizes people's racial and ethnic backgrounds (similar to the multicultural approach in all of its forms), but instead of focusing on the differences among different cultural groups […] [it] focuses on the many connections among groups due to past and present interactions and mutual influence" ( [7], pp. .Initial studies on polyculturalism are promising, indicating a negative relationship with intergroup outcomes beyond the influence of assimilation, colorblindness, and multiculturalism [11,48].They show the effect of the four diversity ideologies on evaluative bias, though not on behavioral measures yet [11]. Our results provide several practical implications.The non-significant direct effect between interethnic ideologies and the intention to discriminate implies that if managers and organizations want to reduce hiring discrimination intention, they need to first act on in-group favoritism.In this regard, our findings show that both interethnic ideologies play a complementary role in in-group favoritism, which means that two ways exist to reduce organizational discrimination, i.e., strengthening multiculturalism and reducing the endorsement of assimilation.To this end, organizations should favor (avoid) a multiculturalism (assimilation) context.The perception of multiculturalist (assimilationist) context is indeed positively related to the personal endorsement of multiculturalism (assimilation), which in turn is negatively (positively) related to prejudice [4].

Appendix
Multiculturalism items [28,46] As a manager, I find it important to take into account the specifics of workers from each origin.I find it important to be knowledgeable about how workers from different origins deal with professional issues.I find it important to bring workers to have a better knowledge of the competences of workers from different origins.I find it important to take into account features of workers from each origin.I believe that workers from different origins each have strengths that can be identified.Assimilation items [28,46] As manager, I consider that all workers, irrespective of their ethnic origin, have to embrace the work values of Belgian society.I believe that all workers, irrespective of their ethnic origin, have to embrace the work values of Belgian society.I consider that all workers, irrespective of their origin, should have the same Belgian work values.I consider that immigrants should adopt the work habits of the Belgian people.I expect immigrants to meet work norms to the same degree as Belgian workers.I think that all workers, irrespective of their origin, have to comply with the organization' regulation.

Table 1 .
Exploratory factor analysis: eigenvalues, percentage of explained variance by each factor and factor loadings.
Note: N = 182; Bold numbers indicate on which factor the item loads.

Table 2 .
Descriptive statistics and simple intercorrelations among variables.

Table 3 .
Regression coefficients for model testing the effects of interethnic ideologies on the intention to discriminate through in-group favoritism.

Table 4 .
Summary of total, direct and indirect effects.