Introduction

Despite significant progress in the understanding of prejudice in intergroup relations research within the last decade, minority group members, especially the ones living in countries where social norms are still set by majority group members, continue to be the targets of discrimination and power imbalances based on group membership still define lower and higher status group members’ positions in the society (e.g. Peterson & Runyan, 1993; Sidanius & Pratto, 1999). The effects of being the targets of discrimination in multiethnic settings have detrimental consequences for ethnic minorities not only in relation to their psychological and physical well-being (see Schmitt, Branscombe, Postmes, & Garcia, 2014 for a recent review), but also in relation to their attitudes towards the majority group. Perceptions of discriminatory behaviours towards the ingroup often challenge minorities’ efforts to develop positive intergroup relationships with the majority group by keeping them socially distant and segregated in the society they live in (e.g. Major, Quinton, & McCoy, 2002; Schmitt & Branscombe, 2002). Although previous research has focused on the causes and consequences of minority group members’ perceived level of discrimination in intergroup relations research, few studies have examined how majority group members perceive discrimination targeted towards minorities and to what extent such perceptions relate to majority group members’ attitudes towards minority group members (e.g. Shelton, Richeson, & Salvatore, 2005). In the current research, we examined first how majority status group members’ (Turkish) beliefs about the prevalence of ethnic discrimination towards ethnic minority status group members (Kurdish) were related to the majority group’s attitudes and support for multiculturalism towards the target minority group and whether the association between these constructs would be mediated by perceived level of threat from the target minority group (Study 1). In Study 2, we tested experimentally whether presenting Turkish participants with bogus news reports about the prevalence of ethnic discrimination towards Kurdish group members would affect the level of affective outgroup attitudes towards this group, support for multiculturalism, and attitudes towards minority culture maintenance and intergroup contact.

Perception of Discrimination and Attitudes

Previous research has shown that dominant group members’ negative judgments and evaluations of the minority group have impacts on minority group members’ attitudes towards the majority group. These meta-perceptions—perceptions of what the other group thinks or wants—have been shown to influence outgroup attitudes such as outgroup prejudice and trust (Shelton et al., 2005; Vorauer, Main, & O’Connell, 1998). For example, minority group members are usually aware of prejudicial and discriminatory attitudes from the majority group and often expect negative attitudes from these group members (Adams, Garcia, Purdie-Vaughns, & Steele, 2006; Pinel, 1999; Shelton, 2003; Shelton & Richeson, 2005; Shelton et al., 2005). Such negative expectations from the majority group in turn lead minority group members to anticipate anxiety and threat in further interactions with outgroup members, impairing future possibilities to build positive intergroup relationships for both sides (e.g. Mendoza-Denton, Downey, Purdie, Davis, & Pietrzak, 2002; Tropp, 2003).

Although dominant group members are often less aware of minority group members’ experiences of discrimination in the society (e.g. Tatum, 1994), previous research has shown that majority group members’ perceptions of the experiences of the minority group may also influence majority group members’ attitudes towards the target minority group. Zagefka, Brown, Broquard, and Martin (2007) investigated whether the perception of acculturation preferences of minority group members affected majority group members’ attitudes towards immigrants and found that minority group members’ desire for culture maintenance was related to more negative outgroup attitudes among majority group members. Biernat, Vescio, and Theno (1996) demonstrated that perceived value violation from the part of the outgroup was related to negative outgroup judgments among majorities. Relatively less studied is how majority group members’ perceptions of the prevalence of discrimination towards minority group members are reflected in their attitudes towards minority group members. It is now known that perceived ethnic discrimination has detrimental consequences for ethnic minorities (e.g. Schmitt et al., 2014) and undermines the development of positive intergroup attitudes and behaviours (e.g. Shelton et al., 2005), but how the prevalence of discrimination targeted towards minorities affects outgroup attitudes among the majority group members is relatively understudied. Through two studies, we therefore aimed to understand the perspective of ethnic majority group members in relation to ethnic discrimination towards minorities and how such perceptions are associated with outgroup attitudes and support for multiculturalism towards ethnic minorities.

The Turkish–Kurdish Context

We tested our research questions in the unique setting of Turkish–Kurdish interethnic context. Compared to many Western countries, Turkey is generally ranked among countries where group-based inequalities are high. In Hofstede’s cultural dimensions, Turkey has been found to be relatively high in power distance and uncertainty avoidance (Hofstede, 1980) which suggests that hierarchy and power relations in Turkey are fundamental aspects of group-based memberships in the society. For example, recent statistical research in Turkey has suggested that group memberships such as ethnic and gender group memberships are critical factors in being successful in the society and upward social mobilization is often a difficult process for lower status group members (Çarkoğlu & Kalaycıoğlu, 2009).

We specifically focused on the perception of ethnic discrimination against the Kurdish ethnic group in Turkey who, despite constituting the largest ethnic minority group in the society (18% of the population, Konda, 2011), has been an oppressed minority group as a result of a strong Turkification process and the assimilationist policies of the Turkish government over the years (e.g. Yeğen, 1996). Reinforced by the cycle of violent attacks between the PKK (Kurdistan Workers Party, Partiya Karkerên Kurdistan) and the Turkish State which resulted in the death of thousands of people from both sides since the 1990s (Göçek, 2011), today, interethnic conflict between Turkish and Kurdish ethnic groups is an intractable socio-political issue which is reflected on many aspects of Kurdish group members’ lives. Especially, the emergence of strong anti-Kurdish discourses on various sources of media has led Kurdish ethnic group members to become a stigmatized group (Bora, 2006). As a result, this ethnic group, among other ethnic minority groups in Turkey, has been found to report highest levels of perceived ethnic discrimination (Duman & Alacahan, 2011; Duman, 2013) and both Turkish and Kurdish group members seem to maintain strong negative stereotypes towards each other, leading to further segregation in the Turkish society (Bilali, Çelik, & Ok, 2014). In particular after 2014, the stability of socio-political relationships between the Turkish State and the Kurdish minority group has deteriorated. Overall, these recent events call for further research that would provide ways to improve Turkish–Kurdish relationships at the intergroup and interpersonal level.

Study 1

Study 1 aimed to explore associations between the extent to which the majority group believed the target minority group to be discriminated against in the society and the level of outgroup attitudes and support for multiculturalism towards the target minority group. We specifically argued that the more Turkish majority group members believed Kurdish minority group members to be discriminated in the Turkish society, the more positive they would be towards the outgroup, since the existence of higher levels of discrimination targeting the Kurdish minority group would imply the maintenance of power imbalances in the Turkish society and would reinforce the dominant role of the higher status Turkish group. Social Identity and Self-Categorization theories (Tajfel & Turner, 1986; Turner, Hogg, Oakes, Reicher, & Wetherell, 1987) suggest that people strive to maintain their self-esteem by affirming a positive group identity and often tend to favour their own group in relation to the outgroup. Prejudice and discrimination have often roots in social group identification processes such that people engage in discriminatory behaviours in order to maintain their group’s well-being which in turn boosts their own well-being. Therefore, it is possible that the more Turkish participants believed there was ethnic discrimination towards the Kurdish minority group, the more they would be able to maintain group distinctiveness and power and consequently the more positive and supportive they would be towards this group. Based on these empirical findings, we proposed that perceived discrimination towards minority group members would be positively associated with outgroup attitudes and support for multiculturalism among majority group members (Hypothesis 1).

We further aimed to explore the role of perceived threat from the outgroup in these associations. We suggested that one route whereby perceived prevalence of discrimination towards the minority group is associated with outgroup attitudes and support for multiculturalism would be through decreased levels of perceived threat from the outgroup. Previous research has been limited to examine majority group members’ perception of discrimination against minority group members. Perceiving oneself discriminated has been found to be related to negative psychological outcomes such as lower self-esteem and higher depression (e.g. Pascoe & Richman, 2009). Perceiving an outgroup member discriminated against is likely to represent this outgroup as powerless and vulnerable in the society and may therefore lead to reduced levels of threat from this group. On the other hand, the key role of perceived threat and competition in outgroup attitudes has been well established in social psychology literature (e.g. Brown, Maras, Masser, Vivian, & Hewstone, 2001; Esses, Dovidio, Jackson, & Armstrong, 2001; LeVine & Campbell, 1972). For example, Group Threat Theory (Blalock, 1967) demonstrated that large numbers of outgroups lead ingroup members to feel threatened and thereby increase conflict and hostility towards outgroup members. Both realistic and symbolic threats explain negative outgroup attitudes and behaviours; while realistic threats refer to threats targeting economic resources and employment opportunities, symbolic threats refer to threats based on group differences on norms, cultural beliefs, and values (Ward & Masgoret, 2006).

Empirically, high levels of perceived threat have been found to be related to negative outgroup attitudes and behaviours (Riek, Mania, & Gaertner, 2006; Stephan & Stephan, 2000) and lower levels of support for multiculturalism (Verkuyten, 2007). Social identity threat often occurs when the ingroup position in the society is challenged by the increasing power gained by the outgroup, since such challenges to the ingroup would question the power and dominance of the majority status group in the society (e.g. Branscombe, Ellemers, Spears, & Doosje, 1999; Ellemers, Spears, & Doosje, 2002). For example, increasing numbers of ethnic minorities which imply the increase in the outgroup power and consequently the decrease in the ingroup power are likely to increase the level of perceived threat from the outgroup (e.g. Putnam, 2007). In the case of ethnic/racial minorities and immigrants, especially in Western countries, it has been found that higher concentration of minorities leads to greater threat perception which is often the result of prejudice and discrimination (Bobo, 1988; Oliver & Mendelberg, 2000; Quillian, 1995). Previous research has also found direct associations between power threat and recognition of discrimination. Investigating Dutch participants’ recognition of discrimination towards immigrants, Verkuyten and Martinovic (2015) found that there was a negative relationship between recognition of immigrant discrimination and power threat. In the light of these empirical and theoretical findings, we predicted that the more participants believed that ethnic discrimination towards the minority group was prevalent, the lower levels of threat they would perceive from the outgroup and in turn the more favourable outgroup attitudes and support for multiculturalism they would report (Hypothesis 2).

Method

Participants and Procedure

A total of 356 university students who self-identified as Turkish ethnic group members completed questionnaires (222 females, 133 males, 1 unknown, M age = 20.81, SD = 2.91). Data were collected with the help of research collaborators in five different universities located in Istanbul, Turkey. The ethnic composition of university campuses varied, but usually represented the national average (Kurdish people forming the 18% of the population; Konda, 2011). Most of the participants were from middle socio-economic backgrounds. The majority of mothers and fathers completed a high school educational level and above (64 and 79%, respectively). Pen and paper questionnaires were completed in the campus setting during lectures. All scales were back-translated into Turkish. Participants were given informed consents that strictly stated ethical procedures of the study, and they were told they could withdraw from the study if they felt uncomfortable.

Measures

Control Variables

We included two control variables that may be potentially associated with outgroup attitudes and support for multiculturalism. We included previous contact with the minority group which has been previously shown to be positively associated with attitudes and multiculturalism in the Turkish–Kurdish context (Çelebi, Verkuyten, & Smyrnioti, 2016). This was measured by one item asking participants to report the number of Kurdish people they knew (“How many Kurdish people do you know?”) with a response scale ranging from 1 (none) to 7 (more than 30). We also included perceived status of the minority group as a control variable, since status perception has been found to be critical in the evaluation of the outgroup (e.g. Tausch, Hewstone, Kenworthy, Cairns, & Christ, 2007). This was measured by a single item asking participants to report the level of perceived socio-economic status of the Kurdish ethnic group (“How would you rate the socio-economic status of Kurdish ethnic group in general in Turkey?”). The response scale ranged from 1 (very low) to 7 (very high).

Perceived Prevalence of Discrimination

Two items adapted from the studies of Swim and Miller (1999) and Verkuyten and Martinovic (2015) were used for measuring perceived prevalence of discrimination targeted towards the Kurdish ethnic group. Participants were asked to report how frequently they believed that the Kurdish group in Turkey was ethnically discriminated in the work place and how frequently they believed that the Kurdish group in Turkey was ethnically discriminated against in social life. The response scale ranged from 1 (never) to 5 (all the time). The reliability coefficient of the scale was .91.

Perceived Threat from the Minority Group

We used a six-item scale for measuring perceived threat from the outgroup (Florack, Piontowski, Rohmann, Balzer, & Perzig, 2003). Participants were asked to indicate how threatening and enriching they found minority groups on different topics (economy, unemployment, cultural values, political attitudes, religion, and language). The response scale ranged from 1 (threatening) to 7 (enriching). A sample item was “How threatening/enriching do you find Kurdish people living in Turkey on Turkish economy?”. We recoded all items such that higher scores indicated higher perception of threat. Since the scale contained items for both realistic and symbolic threat, we conducted an initial factor analysis to determine whether the scale had a unidimensional structure. An exploratory factor analysis with varimax rotation demonstrated that the perceived threat scale was unidimensional; therefore, all items were averaged to represent a general perceived threat scale (Cronbach’s alpha = .91).

Attitudes Towards the Minority Group

Attitudes towards the minority group were assessed by three items that were adapted from previous research (e.g. Feddes, Noack, & Rutland, 2009). Items were “I find Kurdish people nice”, “I find Kurdish people kind”, and “I think Kurdish people are friendly”. The response scale ranged from 1 (strongly disagree) to 5 (strongly agree). The reliability of the scale was good with a Cronbach’s alpha value of .86.

Support for Multiculturalism

Support for multiculturalism in this research was conceptualized as the positive evaluation of differences across ethnic groups. Five items adapted from Verkuyten (2007) assessed the extent to which participants endorsed a multiculturalist ideology towards the Kurdish ethnic minority group. Example items were: “I can learn a lot from the Kurdish ethnic group” and “It is never easy to understand people from the Kurdish ethnic group (R)”. The response scale ranged from 1 (strongly disagree) to 5 (strongly agree). The reliability of the scale was good, demonstrated by a Cronbach’s alpha value of .82.

Data Strategy

We estimated model parameters with Structural Equation Modelling using Mplus software (Muthén & Muthén, 1998–2016) to present a systematic observation of the relationships between variables. Latent variables were used to represent all variables, except contact and perceived status of the minority group which were both measured by a single item, and therefore represented as observed variables. Since initial analyses did not reveal any age, gender, and socio-economic status effects on outcomes measures, these demographic variables were not included in the model.

The following goodness-of-fit indices were used: the Chi-square test, the root-mean-square error of approximation (RMSEA), standardized root-mean-square residual (SRMR), and the comparative fit index (CFI). A good fit is achieved by a nonsignificant Chi-square test, a CFI value greater than 0.95, an RMSEA of less than .06, and an SRMR of less than .08 (Hu & Bentler, 1999). The maximum likelihood estimation was used in competitive structural models and the Chi-square difference test was used to compare the fit of models (e.g. Wagner, Christ, Pettigrew, Stellmacher, & Wolf, 2006). Indirect effects were computed to test the significance of the suggested mediational pathways.

Results

Descriptive statistics and bivariate correlations are presented in Table 1. The mean score for the perception of ethnic discrimination against Kurdish minority group members (M = 3.10, SD = 1.07) was average as shown by a nonsignificant one-sample t test, t(354) = 1.57, p = .12. A confirmatory factor analysis (CFA) was performed to test the fit of our measurement model. Including all latent variables in the model, the measurement model yielded good fit, χ2(48) = 130.91, p < .05, CFI = .96, RMSEA = .07, SRMR = .04.

Table 1 Means, standard deviations, and correlations among main study variables
Full size table

First, a partially mediated model was constructed. Perceived threat was regressed on perceived prevalence of discrimination, and outgroup attitudes and support for multiculturalism were regressed on threat and perceived prevalence of discrimination. For this initial model, direct associations between discrimination and outcome measures were maintained. Findings indicated a good fit, χ2(68) = 168.85, p < .05, CFI = .95, RMSEA = .07, SRMR = .06. Among control variables, previous contact with Kurdish group members was significantly and positively associated with both outgroup attitudes (B = .14, SE = .04, p < .001) and support for multiculturalism (B = .13, SE = .03, p < .001). Perceived prevalence of discrimination was significantly and negatively associated with perceived threat (B = −.13, SE = .06, p < .05), and perceived threat was significantly and negatively related to outgroup attitudes and support for multiculturalism (B = −.44, SE = .05 and B = −.68, SE = .06, respectively, both p < .001). While the direct association between perceived prevalence of discrimination and support for multiculturalism was significant (B = .22, SE = .05, p < .001), this association was not significant in relation to outgroup attitudes (B = .01, SE = .05, p > .05). A second fully mediated model where direct associations between perceived prevalence of discrimination and outcome measures were omitted demonstrated good fit, χ2(70) = 197.68, p < .05, CFI = .94, RMSEA = .07, SRMR = .07. The Chi-square difference test demonstrated that the fully mediated model was significantly worse than the partially mediated model, ∆χ2(2) = 28.83, p < .001. Therefore, the partially mediated model was retained.

We further tested additional alternative models by changing the associations between variables. First, a mediation model where support for multiculturalism was treated as a mediator rather than an outcome measure was conducted. The model with two mediators indicated a good fit; χ2(70) = 183.10, p < .05, CFI = .95, RMSEA = .07, SRMR = .07. Comparing this model with the baseline model (partial mediation), it was found that the alternative model yielded a significantly worse fit, ∆χ2(2) = 14.25, p < .001. Next, we tested a model where attitudes and support for multiculturalism were treated as independent variables and perceived prevalence of discrimination was treated as the dependent variable. This model also revealed good fit, χ2(73) = 194.76, p < .05, CFI = .94, RMSEA = .07, SRMR = .08, but model fit was significantly worse compared to the baseline model, demonstrated by a significant Chi-square difference test, ∆χ2(5) = 26.91, p < .001. Figure 1 displays the final mediation model. Indirect effects of the final model indicated that the mediational routes from perceived prevalence of discrimination to both outcomes through perceived threat were significant (B = .06, SE = .03 for outgroup attitudes and B = .09, SE = .04 for multiculturalism, both p < .05).

Fig. 1
figure 1

Final mediation model showing associations between perceived prevalence of discrimination, perceived threat and outgroup attitudes, and support for multiculturalism. Note Final model fit: χ2(68) = 168.85, p < .05, CFI = .95, RMSEA = .07, SRMR = .06. Note † p < .10, *p < .05, **p < .01, ***p < .001

Full size image

Study 2

Study 1 showed that perceived prevalence of discrimination towards Kurdish ethnic minorities was related to more positive outgroup attitudes and support for multiculturalism among Turkish majority group members (Hypothesis 1). Furthermore, as expected, these associations were significantly mediated by perceived threat from the outgroup (Hypothesis 2). Although this study provided initial evidence for the associations between beliefs about the prevalence of discrimination towards minorities and outgroup attitudes, the prevalence of ethnic discrimination towards ethnic minorities was assessed subjectively. It is possible that participants who were already highly prejudiced towards the Kurdish minority group or who already held negative outgroup attitudes towards the Kurdish ethnic group or participants who held higher ethnic and/or national identities were more biased against the Kurdish group and consequently over or underestimated the prevalence of discrimination towards ethnic minority groups (e.g. Operario & Fiske, 2001). For example, prior studies have demonstrated that Whites’ guilt may in fact lead to greater estimation of the prevalence of discrimination towards Blacks (Swim & Miller, 1999). Verkuyten and Martinovic (2015) also found that a high sense of shared national identity was related to greater recognition of discrimination against immigrants. Moreover, in Study 1 we used a cross-sectional design which limits our conclusion about the causality between study variables.

Therefore, in Study 2, rather than the subjective perception of ethnic discrimination towards a target group, we aimed to understand how a more objective understanding of ethnic discrimination would influence majorities’ attitudes. Using an experimental design, we presented participants with bogus newspaper reports stating information about the prevalence of ethnic discrimination targeting Kurdish group members in Turkey. More specifically, we examined whether Turkish participants who were given information about the prevalence of ethnic discrimination towards the Kurdish group would be more positive towards this outgroup compared to participants in the neutral condition (no information received).

Previous research has shown that exposure to media and news can represent influential tools in the formation of social stereotypes and stereotypic attitudes (e.g. Mastro & Kopacz, 2006; Schemer, 2012; Tan, Fujioka, & Tan, 2000). In particular, negative depictions of ethnic minorities and immigrants in the news often play a key role in the activation of negative attitudes in audience (Dixon & Azocar, 2007; Domke, 2001; Entman, 1990; Power, Murphy, & Coover, 1996), and minority group members are generally associated with news involving crime and terrorism in these media tools (e.g. Dixon & Linz, 2000; Van Dijk, 1992). Recently, the Kurdish ethnic group in Turkey has been similarly negatively portrayed in various aspects of media (e.g. Bora, 2006). According to Sezgin and Wall (2005), the Turkish media often discredits the existence of Kurdish cultural values and displays this ethnic group as a socio-political issue that constitutes a major source of concern for the Turkish society.

Previous experimental research has tested the effects of exposure to news reports on the evaluation of outgroups. For example, Vergeer, Lubbers, and Scheepers (2000) found that Dutch participants who were presented with newspapers characterized by negative reporting about an ethnic crime conceived ethnic minorities more as a threat. In another experiment, Power et al. (1996) showed that attitudes towards African Americans are changeable, depending on whether participants were exposed to stereotypic or counter-stereotypic portrayal of African Americans in media. Peffley, Shields, and Williams (1996) presented White participants with a televised crime story in which the perpetuator was either White or Black. The experiment showed that participants who were presented with the Black individual as the perpetuator of a violent crime had more negative attitudes towards the Black suspect compared to the case with White individuals as the suspect. Overall, these studies have shown that media portrayals of ethnic minorities are likely to influence dominant group members’ attitudes towards ethnic minorities by activating negative stereotypes (e.g. Mastro & Kopacz, 2006).

Based on the findings of Study 1 which indicated that perceived prevalence of discrimination towards the Kurdish promoted Turkish participants’ outgroup attitudes and support for multiculturalism towards this group and previous research showing the positive effects of discrimination awareness on positive outgroup attitudes (e.g. Case, 2007; Powell, Branscombe, & Schmitt, 2005), we suggested that presenting Turkish participants with bogus news reports that provided information about the prevalence of ethnic discrimination towards Kurdish ethnic group members would increase positive attitudes and support towards this outgroup.

In Study 2, in addition to outgroup attitudes and support for multiculturalism, we extended our dependent variables to attitudes towards minority group culture maintenance and intergroup contact. Previous research has shown that it is not only ethnic minorities who hold acculturative strategies such as assimilation, integration, separation, or marginalization (Berry, 1997), but dominant group members also prefer various acculturation strategies such as integration or separation for ethnic minority groups (Zagefka & Brown, 2002). More importantly, the acculturative strategies adopted by dominant group members are likely to influence attitudes towards the minority group. For example, Zagefka and Brown (2002) demonstrated that Germans’ attitudes towards minorities’ culture maintenance and intergroup contact were significantly related to lower ingroup bias. Dominant group members who endorsed minority culture maintenance and intergroup contact also reported they perceived intergroup relationships more favourably than the ones who did not endorse culture maintenance and contact. Similar to outgroup attitudes and support for multiculturalism, it is possible that attitudes towards minority culture maintenance and intergroup contact would be more positive when participants are provided with discrimination information, since individuals who learn about minority discrimination are likely to consider this outgroup as powerless and thereby perceive less threat from this outgroup (e.g. Powell et al., 2005; Spanierman et al., 2008). In turn, this could lead to higher levels of support for integration. Therefore, we proposed that presenting participants with discrimination information would also increase positive attitudes towards minority culture maintenance and intergroup contact.

Participants were randomly assigned to three conditions: neutral (no news reports), lower ethnic discrimination (few companies have been found to discriminate against the Kurdish), and higher ethnic discrimination (most companies have been found to discriminate against the Kurdish). We specifically hypothesized that participants presented with bogus news reports that provided information about a high rate of discrimination towards the Kurdish group would report more positive affective outgroup attitudes and more favourable attitudes towards minorities’ culture maintenance and intergroup contact, and support multiculturalism more compared to participants presented with news reports that provided information about a low rate of discrimination and the neutral condition (Hypothesis 3).

Method

Participants

A total of 82 undergraduate psychology university students (16 males and 66 females M age = 21.91, SD = 1.90) studying at a private university in Istanbul (Turkey) were recruited for the study. All participants’ nationality and ethnic group were self-reported as “Turkish”. The subjective socio-economic status (SES) of participants was measured by one item (“How would you rate your socio-economic status?”) measured on a scale from 1 (very poor) to 7 (very wealthy). The mean SES of participants was upper-middle class (M = 4.69, SD = .79).

Procedure and Materials

Data were collected in a psychology laboratory by research assistants, and participants were randomly assigned to three different conditions (neutral, low discrimination, and high discrimination). At the beginning of the research, participants were told that the aim of the study was to explore intergroup processes in the Turkish society and they received course credits for participation.

Demographics

The questionnaire started with the demographic information section which assessed age (in years), gender (0 = Male, 1 = Female), subjective SES, and self-reported ethnic background (0 = Other, 1 = Turkish). We further assessed contact quantity by asking participants to report the number of people they knew from various ethnic backgrounds including Turkish, Kurdish, Armenian, Syrian, Laz, and Cherkes reported on a response scale ranging from 1 (none) to 7 (more than 31) and contact quality (“How would you rate your contact with these ethnic group members?”) reported on a response scale ranging from 1 (very negative) to 7 (very positive).

News Reports Priming

After the demographic section, participants received the bogus news report. In the neutral condition, participants were given a non-relevant scene to imagine; they were asked to imagine the latest holiday they went to and were further asked questions about this holiday. In the low discrimination condition, participants were told that they would read a newspaper report about small to moderate companies in Turkey. The following instruction was given:

The A&G research team has conducted an impressive research with 200 small to moderate companies in Turkey at the end of 2015 in order to explore the demographic characteristics of people who were recruited by these companies. Data came from human resource centres of the companies involved. Companies were specifically chosen from diverse cities in Turkey to represent the national average. A total of 10.559 recruitment processes have been examined. One of the topics the research team explored was the extent at which there was ethnic discrimination during the recruitment process. According to the results, only 13% of the companies have been discriminatory towards the Kurdish minority group in applications with similar experiences and educational backgrounds, whereas 87% of the companies have not been found to ethnically discriminate towards the Kurdish minority group. Please indicate your thoughts about this finding with five–six sentences.

In the high discrimination condition, the rate of ethnic discrimination was reversed such that many of the companies have been found to discriminate towards the Kurdish ethnic minority group:

…According to the results, only 13% of the companies have not been discriminatory towards the Kurdish minority group in applications with similar experiences and educational backgrounds, whereas 87% of the companies have been found to ethnically discriminate towards the Kurdish minority group….

Dependent Measures

Unless otherwise stated, all response scales ranged from 1 (strongly disagree) to 7 (strongly agree). Support for multiculturalism was measured by five items adapted from Verkuyten (2007) and assessed the extent at which participants endorsed a multiculturalist ideology towards ethnic minority groups (see Study 1, α = .70). Affective outgroup attitudes were measured by three affective bipolar items adapted from Wright, Aron, McLaughlin-Volpe, and Ropp (1997). Participants were asked to report how cold/warm they feel towards ethnic minority group members, how negative/positive they feel towards ethnic minority group members, and how hostile/friendly they feel towards ethnic minority group members. Higher scores on this scale showed more positive feelings towards ethnic minorities (α = .94). Attitudes towards minority culture maintenance was measured by three items used by Zagefka and Brown (2002). We replaced the word “immigrants” in the original scale with “ethnic minorities” (“I do not mind if ethnic minorities in Turkey maintain their own culture”, “I do not mind if ethnic minorities maintain their own religion, language and clothing”, and “I do not mind if ethnic minorities in Turkey maintain their own way of living”). Attitudes towards minority contact (Zagefka & Brown, 2002) was measured by two items which assessed the extent to which participants supported ethnic minority group members in forming contact with ethnic majority group members (“I think it is important that members of ethnic minority groups also spend time with ethnic majority groups” and “I think that members of ethnic minority groups should have friends from ethnic majority groups”). The reliability of both scales was high with Cronbach’s alpha values of .97 and .92, respectively.

Results

As preliminary analyses, we checked for associations between demographic variables (SES, gender, age, and previous contact quantity and quality) and main dependent variables. Findings showed that subjective SES, age, and gender were not significantly related to outcome measures. However, previous contact quantity and quality with Kurdish ethnic group members were significantly and positively associated with all outcome measures. We conducted a MANCOVA test (multivariate tests of covariance) with condition as the fixed factor, previous contact quantity and quality as covariates, and affective outgroup attitudes, support for multiculturalism, and attitudes towards minority culture maintenance and intergroup contact as dependent variables.

Results demonstrated that the multivariate effect of previous contact quantity was nonsignificant, F(4, 68) = 1.25, Wilks’ Lambda = .93, p = .30, η 2 p  = .07, whereas the multivariate effect of previous contact quality was significant, F(4, 68) = 14.79, Wilks’ Lambda = .54, p < .001, η 2 p  = .47. We further observed a significant multivariate effect of condition, F(8, 136) = 3.09, Wilks’ Lambda = .72, p < .01, η 2 p  = .15.

An examination of the effect of condition on each dependent measure revealed that there was a significant main effect of condition on multiculturalism, F(2, 71) = 6.51, p < .01, η 2 p  = .16. A further observation of LSD (least significant difference) post hoc tests indicated that the difference between the higher discrimination (M = 4.68, SD = 1.59) and lower discrimination (M = 4.93, SD = 1.59) conditions was nonsignificant (p = .27). The neutral condition yielded lower rates of support for multiculturalism (M = 3.94, SD = .44) compared to both lower discrimination and higher discrimination conditions (p < .01 and p = .02, respectively). The main effect of condition on affective outgroup attitudes was not significant, F(2, 71) = 1.58, p = .21, η 2 p  = .04. Condition had a significant effect on attitudes towards culture maintenance, F(2, 71) = 3.60, p = .03, η 2 p  = .09. The difference between the lower discrimination (M = 6.01, SD = 1.46) and higher discrimination (M = 5.19, SD = 1.61) conditions was marginally significant (p = .08); lower discrimination condition led to more positive attitudes towards culture maintenance. The difference between the lower discrimination and neutral condition (M = 4.77, SD = 1.73) was also significant, p = .01, whereas the difference between higher discrimination and neutral conditions was not significant, p = .36. Finally, the main effect of condition on attitudes towards contact was significant, F(2, 71) = 7.59, p < .01, η 2 p  = .18. Accordingly, the neutral condition (M = 4.73, SD = 1.87) yielded significantly less favourable attitudes for minority–majority contact compared to the higher discrimination (M = 5.52, SD = 1.50) and the lower discrimination (M = 6.29, SD = .86) conditions, p = .04 and p < .001, respectively. The difference between the lower and higher discrimination conditions was marginally significant, p = .06. Figure 2 displays means and standard deviations across conditions.Footnote 1

Fig. 2
figure 2

Means of support for multiculturalism, outgroup attitudes, and attitudes towards culture maintenance and intergroup contact across three conditions. Note LD lower discrimination, HD higher discrimination

Full size image

General Discussion

The current research aimed to understand how the prevalence of ethnic discrimination towards a target minority group would affect outgroup attitudes and support for multiculturalism among majority group members in the unique context of Turkish–Kurdish interethnic relationships in Turkey. In Study 1, we aimed to assess associations between the majority group’s subjective beliefs about the prevalence of ethnic discrimination towards minority group members and outgroup attitudes and support for multiculturalism. We further tested whether perceived threat from the minority group would mediate these associations. As expected, we found that the more Turkish participants believed there was ethnic discrimination towards the Kurdish group in the society, the lower threat they felt, and consequently the more favourable and supportive attitudes they held towards minorities. These findings are in line with previous research that has shown that awareness of prejudice is likely to influence outgroup attitudes positively (e.g. Powell et al., 2005; Spanierman et al., 2008) and prevalence of discrimination is related to higher levels of support for affirmative action (Aberson & Haag, 2003; Harrison, Kravitz, Mayer, Leslie, & Lev-Arey, 2006). We extended this research literature by showing that subjective beliefs about the prevalence of ethnic discrimination towards minorities may indeed improve majorities’ attitudes by decreasing perceived levels of threat from the outgroup.

Study 2 aimed to examine whether it was possible to improve Turkish participants’ affective outgroup attitudes, support for multiculturalism, and attitudes towards minority culture maintenance and intergroup contact by presenting participants with bogus news reports providing information about the prevalence of ethnic discrimination towards Kurdish minority group members. Supporting Hypothesis 3, we demonstrated that the most negative attitudes towards the Kurdish minority group members were reported in the neutral condition where participants were asked to imagine a non-relevant neutral scene. In both lower and higher discrimination conditions, we found increases in positive attitudes towards the Kurdish group in relation to all outcome measures, except affective outgroup attitudes. One reason for this may be that compared to other attitudes such as attitudes towards culture maintenance, affective outgroup attitudes may be more challenging to change, especially in the current intractable Turkish–Kurdish conflict context. Recent research in Turkey has shown that intergroup trust is generally low between Turkish and Kurdish group members (Çelebi, Verkuyten, Köse, & Maaliepard, 2014), and negative stereotypes against Kurdish group members continue to be part of Turkish–Kurdish interethnic relationships (Bilali et al., 2014). In such a setting, a brief news report may not have been an effective strategy to improve affective outgroup attitudes.

Interestingly, we found that the lower discrimination condition led to more positive attitudes compared to the higher discrimination condition, especially in relation to attitudes towards minority culture maintenance and intergroup contact. This finding seems counterintuitive and opposed to our initial suggestion, because higher discrimination perception is likely to produce lower levels of threat compared to lower discrimination perception and in turn more likely to promote the formation of positive intergroup relationships. One possibility is that the relationship between the prevalence of ethnic discrimination towards minority group members and outgroup attitudes among majority group members is not linear; the most positive attitudes towards ethnic minorities may have been produced in the lower discrimination condition, since the perception of discrimination at a lower level may suggest that the outgroup is discriminated to some extent so that the majority group status is maintained, but at the same time hostility and exclusionary attitudes towards ethnic minorities is at an acceptable level. Consequently, participants in the lower discrimination condition may have been more tolerant towards multiculturalism, support for culture maintenance and contact. The higher discrimination condition, on the other hand, may imply that the Kurdish group is extensively discriminated and socially excluded, which may make the conflictual aspect of the Turkish–Kurdish interethnic relationships more salient and thereby lead to less favourable outgroup attitudes. One can also posit that the higher discrimination condition may have simply influenced participants in the way that they believed that ethnic discrimination was legitimate and common in the society. Previous research has shown that people are in general vulnerable to ingroup norms regarding discriminatory attitudes (Crandall, Eshleman, & O’Brien, 2002; Jetten, Spears, & Manstead, 1996). Another possibility is the potential effect of perceived realism of the scenarios on attitudes. Although we aimed to present participants with objective statistical research results, the higher discrimination condition may have been perceived to be more unrealistic compared to the lower discrimination condition and therefore may have been less effective in improving attitudes.

Overall, findings have critical implications for the prejudice and discrimination literature. First, previous research has shown that perceived discrimination among minority group members is detrimental for minority groups’ efforts to build positive relationships with the majority group (e.g. Major et al., 2002; Mendoza-Denton et al., 2002; Schmitt & Branscombe, 2002). From the perspective of the majority group, we found that beliefs about the prevalence of discrimination were in fact beneficial for the formation of positive outgroup attitudes among ethnic majority group members. This also fits previous theoretical research stating the importance of the awareness about discrimination among the dominant group members (e.g. Case, 2007) and collective guilt studies. For example, Kernahan and Davis (2007) demonstrated that an educational course that examined prejudice and racism increased White college students’ awareness of racism and their feelings of guilt and discomfort about equality of opportunities, which in turn led to more positive intergroup relationships (e.g. Branscombe, Slugoski, & Kappen, 2004).

Second, findings provide important insights in relation to the specific context of Turkish–Kurdish interethnic conflict. Within the last decade, Kurdish people living in Turkey have been substantially stigmatized and the public discourse has often concentrated on the violent acts that are perpetuated by ethnic minority group members, whereas discrimination towards ethnic minority groups, especially towards the Kurdish group, has been relatively less discussed in the mainstream public sphere (e.g. Bora, 2006). For example, Sezgin and Wall (2005) reported that the Turkish press stating the existence of the Kurdish issue in Turkey often focus on terrorism related events instead of problems specifically experienced by the Kurdish ethnic minority group. The current study demonstrated that beliefs about the prevalence of prejudice in the society may be related to how ethnic minority group members are perceived and how attitudes towards these groups are formed among dominant group members.

Although these studies have provided important insights about the various links between discrimination towards Kurdish minority group members and attitudes in the context of Turkish–Kurdish intergroup relationships, a number of limitations should be stated. First, in Study 1, we were unable to assess any temporal and/or causal relationship between beliefs about discrimination and outgroup attitudes, as this was a single time-point study. Moreover, the prevalence of discrimination was based on self-reports which substantially limits any statement about the directionality of the effects. Study 2 aimed to overcome this limitation by using an experimental design in which a relatively more objective ethnic discrimination assessment was used. Nevertheless, the use of media news reports may still represent subjectivity in the evaluation of ethnic discrimination, since different sources of media may provide unreliable information. For example, during Gezi Protests in 2014, social media has been used to communicate fabricated information which has substantially decreased overall trust in communication systems (Haciyakuboglu & Zhang, 2015).

Other questions have also remained unanswered. Although we assessed perceived threat as a mediating mechanism in Study 1, we failed to demonstrate how exactly lower and/or higher discrimination conditions in Study 2 led to increases in positive outgroup attitudes among majority group participants. Further research should include various affective and cognitive measures that would potentially explain the processes whereby perceived discrimination towards minorities would reflect upon majorities’ outgroup attitudes and support for multiculturalism. For example, it has been shown that when dominant group members think about privileged group members’ advantages and minority group members’ disadvantages, they become more positive towards minority groups and support affirmative action more through increased collective guilt (e.g. Branscombe et al., 2004; Powell et al., 2005). In addition to inducing collective guilt, the perception of discrimination towards a minority group is likely to result in other affective, cognitive, and social reactions in dominant group members (Spanierman et al., 2008). For example, discrimination towards minority groups in the society has been found to trigger feelings of empathy towards the minority group and lead participants to hold more positive outgroup attitudes and provide greater support for minority groups’ rights (Spanierman & Heppner, 2004; Stephan & Finlay, 1999).

A further research avenue is to explore whether outgroup attitudes may lead to the perception of discrimination against a minority group. It is possible that negative attitudes towards a minority group would influence how discrimination against this group would be perceived. Experimental research that would manipulate outgroup attitudes could investigate this assumption. Similarly, social dominance orientation (SDO) may be an important variable to incorporate in this context. Previous research has found that higher SDO was related to higher support for multiculturalism through decreased perceptions of group inequality (Kteily, Sheehy-Skeffington, & Ho, 2016). In a similar vein, SDO is likely to explain how outgroup attitudes and perception of discrimination towards a minority group may be interrelated.

We should also note that in Study 2 we have only concentrated on group-based institutional discrimination towards Kurdish participants. Further research should be conducted to understand how personal discrimination accounts in various settings may affect attitudes towards minority group members. It is now known that acknowledging disadvantages experienced by minority groups helps in the reduction of prejudice among dominant group members (e.g. Powell et al., 2005), yet how minorities’ ethnic discrimination experiences at the personal level influence majority group’s perception is relatively less studied. Previous research has shown that both majority and minority group members tend to overestimate group-based discrimination compared to individual-based discrimination (Moghaddam, Stalkin, & Hutcheson, 1997; Operario & Fiske, 2001), which could be a further research question to investigate within the current context. Study 2 also included generalized attitudes towards ethnic minorities rather than the specific Kurdish ethnic minority. Finally, we failed to include a manipulation check in Study 2, which should be included in further studies to ensure that participants in the high discrimination condition actually perceived more discrimination compared to the neutral and low discrimination conditions.

In conclusion, the current studies contribute to the intergroup relationships literature on discrimination and prejudice by showing the associations between beliefs about the prevalence of ethnic discrimination towards minority group members and outgroup attitudes and support for multiculturalism among majority group members and demonstrating the role of prevalence of ethnic discrimination in promoting favourable attitudes towards multiculturalism, minority culture maintenance, and intergroup contact. Findings suggest the importance of prejudice and discrimination in the formation of intergroup attitudes not only among ethnic minorities, but also among ethnic majorities who evaluate the prevalence of ethnic discrimination towards ethnic minorities in the society.