Intergroup Threat as a Mediator of Ethnic Identification and Intergroup Orientations

Ethnolinguistic Vitality of Language Groups in the Study

The last census in Croatia showed that slightly less than 10% of the population was made up of minority groups (Croatian Bureau of Statistics, 2011), while the Croatian Constitution recognizes 22 ethnic minorities across different regions. We conducted our study in four multiethnic communities in Croatia, where ethnic minorities are large enough to exercise their constitutional right to be educated in their own minority languages (the Serb–Croat context in the town of Vukovar and surroundings, the Italian–Croat context in several towns in Istria County, the Hungarian–Croat context in the town of Osijek and surroundings, and the Czech–Croat context in the town of Daruvar and surroundings). In each of the regions, a single minority has been more concentrated than others: in Vukovar these are Serbs (15.5% vs. 80.2% of Croats), in Istria Italians (6.03% vs. 83.82% of Croats), in Osijek region Hungarians (2.7% vs. 87% of Croats), and in Daruvar region Czechs (5.25% vs. 86.82% of Croats).

Minority groups in these regions either go to separate schools from the Croatian majority or are schooled in separate classrooms (Čorkalo Biruški et al., 2019). Though all four minority groups have preserved their ethnolinguistic vitality, a variety of historical and cultural factors have shaped different majority–minority dynamics in these regions. One dimension to compare our research contexts is the linguistic similarities between Croatian and the respective minority language. Similar languages might ensure easier integration into Croatian society, but also may lead to a lack of positive distinctiveness between the two groups in question (Turner, 1975). From that perspective, Serbian and Czech as Slavic languages are more similar to the Croatian language (especially Serbian), whereas Hungarian and Italian are very different from Croatian. Furthermore, status of the minority language is also important: the higher the status of the language, the more the ethnolinguistic vitality of the groups is likely to be preserved. For example, the Italian language has a high status (Hržica et al., 2011) and many children in Croatia learn Italian as their second language (Balen, 2015). Conversely, Hungarian is considered more difficult to learn. Moreover, the linguistic community of Hungarians is rather small and fairly isolated and therefore majority members may perceive little social benefits from learning Hungarian. However, at the same time physical isolation has made it easier for Hungarians to preserve their ethnolinguistic vitality while also challenging their integration into wider society. As for the Czech language, its native speakers are also concentrated in one particular region; however, the majority and minority exchange has been much more intensive over the course of history. Thus, contemporary Czech language used by minority in Daruvar region is strongly influenced by the Croatian language.

The third dimension of comparison is the geographical distance of minority group members’ kin-state. Whereas Italy, Serbia, and Hungary are neighboring countries to Croatia, this is not a case with the Czech Republic. Thus, being close to the kin-state can make minority more attached to its ethnolinguistic identity and less motivated to integrate into Croatian society. This strategy may be perceived by majority Croats as threatening. Finally, the fourth dimension of comparison is the history of conflict between groups. From this point of view, the Vukovar region is different from the rest because of the recent conflict between Croats and Serbs, and the tension between the two groups is still present. On the opposite end is the Czech–Croat context with no history of majority–minority conflict. As for Italians and Hungarian, the past conflicts with Croats are likely considered to be irrelevant.

These different multiethnic contexts offer an opportunity to gain a deeper understanding of the role of minority language education in shaping intergroup relations of pupils. We argue that the language of education further enhances identification with one’s ethnic group and makes it more salient in everyday life. Moreover, it is also possible that by making ethnic identity more salient, it emphasizes the role of ethnonationalism in intergroup relations. Moreover, minority language education might enhance perceived intergroup threat in both majority and minority groups living in the same community.

Aim of the Study

Drawing on social identity theory and ITT perspectives in the context of minority education in Croatia, we looked at the mediating effect of symbolic threat and intergroup anxiety in the relationship between in-group identification (ethnic identification and ethnonationalism), and negative and positive out-group orientation (i.e., behavioral intentions to discriminate and to act prosocially toward out-group members). Furthermore, we explored if the group status moderates this relationship (see Figure 1). We tested these models in four social contexts in Croatia.

OPEN IN VIEWER

Figure 1.

Theoretical model of moderated mediation.

Based on previous studies suggesting that ethnonationalism, rather than ethnic identity per se, is detrimental for intergroup relations (Brewer, 1999; Čorkalo Biruški & Kamenov, 2003; Jelić et al., 2014; Penic et al., 2017), we expect ethnonationalism to be positively related to out-group derogation and negatively to prosocial tendencies toward the out-group via greater perception of symbolic threat and intergroup anxiety. The inclusion of two different types of intergroup threat will enable us to see whether cognitive or affective path plays a more important role in intergroup relations.

However, the link between ethnic identification and intergroup orientation is less easy to predict due to the mixed results in previous studies (Costarelli & Calli, 2004; Ruttenberg et al., 1996). We argue that introducing both measures of ethnic identification and ethnonationalism simultaneously would separate their unique contribution to intergroup behavior, and that ethnic identification might even be associated with positive out-group outcome via lesser perception of symbolic threat and intergroup anxiety.

In addition, we propose that the group status (majority or minority) moderates the link between intergroup threats and intergroup behavior. Based on ITT’s assumptions, we could expect our models to account for more variance among members of the majority group. However, results corroborating this assumption have not been consistent suggesting that it all depends on the type of threat, type of behavior, and specific social context (Löw Stanić, 2014).

Finally, following the recent developments in intergroup contact theory (Hodson et al., 2013), we decided to include measures of both positive and negative intergroup behavior instead of focusing on just negative behaviors.

Method

Results and Discussion

Mediation Models

Since we found no significant differences in regression coefficients between majority and minority group members, we will interpret our models for the combined sample.

The Model A column in Table 2 ³ shows the total, direct, and indirect effects for the tendency to discriminate against the out-group. Overall, the mediation path analysis was conducted on data from N = 986 and the model explained 24.5% of the variance in the outcome variable. As for direct effects, ethnonationalism was significantly positively related to the tendency to discriminate, while the effect of ethnic identity was not significant. Furthermore, the indirect effects of both ethnic identification and ethnonationalism via both forms of intergroup threat were significant. However, these mediation effects were in the opposite direction. That is, the more participants were attached to their ethnic group, the less threatened and anxious they felt when interacting with members of the other ethnic group. These lower levels of threat and anxiety reduced their tendency to discriminate against the out-group, suggesting that a sense of ethnic belonging might even serve as a protective factor in multiethnic context. However, the more participants believed their ethnic group was superior to other groups, the more threatened and anxious they felt toward the out-group and consequently were more prone to discriminate against it (Figure 2).

OPEN IN VIEWER

Figure 2.

Path diagram of mediation model for the whole sample, Model A.

The Model B column in Table 2 shows the total, direct, and indirect effects for the tendency for prosocial behavior toward the out-group. Overall, the mediation path analysis was conducted on data from N = 985 and the model explained 22.5% of the variance in the outcome variable. Looking at the direct effects, ethnonationalism was negatively and ethnic identity was positively related to prosocial behavior. Furthermore, the indirect effects of both ethnic identity and ethnonationalism via both forms of intergroup threat were significant. Again, these mediation effects were in the opposite direction. Results showed that the more participants were attached to their ethnic group, the less anxiety they would feel when interacting with members of the other ethnic group, which in turn motivated them to act prosocially toward them. This finding further emphasizes protective role of the ethnic identity. On the other hand, the more participants believed their ethnic group was superior to other groups, the more threatened and anxious they felt toward the out-group and consequently were less prone to act prosocially (Figure 3).

OPEN IN VIEWER

Figure 3.

Path diagram of mediation model for the whole sample, Model B.

To summarize, the overall models provided convincing evidence for the partial mediation of intergroup threats on the relationship between in-group identification and out-group orientations. However, the results indicate the importance of distinguishing between different forms of ethnic attachment: sense of belonging to one’s ethnic group (i.e., ethnic identification) and viewing the in-group as superior to other groups (i.e., ethnonationalism). It seems that it is ethnonationalism that is detrimental for intergroup relations, partially due to its connection with the perception of the out-group threat. Perceived symbolic intergroup threat and intergroup anxiety are serious obstacles for intergroup relations—by facilitating out-group discrimination and by inhibiting prosocial behavior toward the out-group. These effects did not differ between majority and minority.

General Discussion

In this article, we explore multiethnic communities where minority groups exercise their right to minority education. We hypothesized that this context would influence ethnic identification and intergroup attitudes and behavior. The present study contributes to the field by exploring both positive and negative out-group orientations. As expected, perceived threat reduces positive and facilitates negative intergroup behavior. Our results also confirmed that perceived intergroup threat mediates the relationship between in-group identification and positive and negative out-group orientation. Moreover, intergroup anxiety and symbolic threat were both mediators of this relationship.

The study also confirmed the importance of distinguishing between two types of in-group attachment. Our findings show that ethnic identity and ethnonationalism have different impacts on the perception of intergroup threat. When controlling for ethnonationalism, ethnic identity as a measure that is not confounded with in-group bias and superiority issues is related to a lower out-group threat. Moreover, while ethnonationalism has a positive direct effect on out-group discrimination, ethnic identity does not. The opposite holds for prosocial behavior toward the out-group members. Thus, our findings are in line with others emphasizing that it is ethnonationalism, rather than sense of belonging per se, that is detrimental for intergroup relations (Brewer, 2007; Roccas et al., 2006).

Contrary to our expectations, group status did not moderate the relationship between threat and intergroup orientations. It is possible that group status is more relevant for intergroup contact than for in-group attachment. Another possibility is that the children from the majority and the minority group in the same community are more similar, as they have not yet adhered to social norms as adults did (Low Stanić, 2014). Future studies should compare results from children and adults from both groups to investigate whether children function more as the adults from the majority or as the adults from minority group.

We also tested the intergroup threat model on the four subsamples, in each of the four contexts of our study. Findings show that context does play a role, but the main assumptions of the model still hold, except in the Czech–Croat context. This context is highly integrated, and it is even possible that assimilation of the minority is in progress. Namely, this was the only context in which minority group respondents chose to complete the survey in Croatian, which in itself might have influenced their results by making them aware of their identity as Croatian citizens. In addition, ethnic identification as a Czech seems to be more of a cultural fact than an identity issue. We argue that in the Czech–Croat context there are no status differences, as present in other contexts of our study. Hence, it may be that in this context, there are more harmonious minority–majority relations. Therefore, the intergroup threat model may not be relevant here because perceived intergroup threat is low and thus does not affect intergroup relations.

In the remaining three contexts, ITT model was confirmed (i.e., we found no substantial differences between specific minorities, which is itself a finding that highlights the salience of ethnonationalism effect compared with ethnic identity effect). This finding suggests a strong link between the glorification of one’s group and negative out-group emotions (i.e., intergroup anxiety), which in turn results in a stronger tendency to discriminate against out-group members, and a lower tendency to act prosocially toward the out-group. The effect of ethnic identity was less consistent and opposite to the effects of ethnonationalism. In the same vein, it seems that symbolic threat has weaker impact than intergroup anxiety; indirect effects via intergroup anxiety were found in all contexts for both outcome variables, while indirect effects via perceived symbolic threat were significant only in the Serb–Croat context for the tendency to act prosocially, and in the Italian–Croat context for the tendency to discriminate. This suggests that emotions may be more important in intergroup contexts than cognitive factors, but this assumption should be tested in future studies.

Furthermore, the greatest variance in intergroup orientation was explained in the Serb–Croat context (the most conflicted of the four communities), followed by the Hungarian–Croat context, and the least in the Italian–Croat context. Moreover, for the most harmonious Czech–Croat context the model did not hold. So, it seems that contextual differences are based on the dimension of conflict versus harmony, and the model is best suited for more conflicted and not for more harmonious setting. These findings reflect different intergroup dynamics of language groups in the four different contexts. It is possible that in the more integrated settings both negative and positive intergroup behaviors are determined by factors we did not include in this study, such as school and community norms, social identity based on attachment to the community, town, or region, instead of one’s ethnic group. However, this should be further explored in future studies.

Conclusions

Based on these results, we have several recommendations for changes in school systems in multiethnic communities. First, schools should aim to encourage a healthy attachment to one’s own ethnic group, without undermining others, and at fostering a dual identity, where each group may show how it contributes to a unified identity, (e.g., belonging to a common school, place, or country). For instance, minority students should be able to decorate schools for their main holidays and make some presentations or plays to talk about their culture to Croats, and vice versa. Teachers should talk with their students about how other cultures enrich community and organize multicultural events. Students should feel free to talk about their ethnicity and find similarities and differences between them. Such an approach, as shown by Riek et al. (2010), leads to a lower level of perceived intergroup threat. Furthermore, schools in a multiethnic community should be a place of appreciation of diversity (thus reducing the symbolic threat) and learning about diversity (thus reducing intergroup anxiety because the out-group is no longer unknown) as well as a safe environment for children’s development (Childs, 2017). For this to be accomplished, it is necessary to develop joint extracurricular activities where children from both groups can learn from and about each other and work together as suggested by intergroup contact theory. Especially useful would be to encourage learning of both languages in multiethnic communities and to treat both these languages as the languages of the region, as a cultural heritage and a resource, and a source of community pride (Čorkalo Biruški et al., 2019). A first step in that direction would be to investigate attitudes toward majority and minority languages in each community, so the space for attitudinal change can be established.

This study has shown that ethnonationalism and not ethnic identity is associated with negative intergroup outcomes. Thus, as ethnonationalism is based on the incorrect and generalized belief that one’s own group is better than the others, we believe that critical thinking should be highly encouraged and taught in the schools. Reducing ethnonationalism and at the same nurturing appreciation of one’s own ethnic group might then lead to a positive feeling of ethnic pride without the negative consequences, such as intergroup threat and anxiety connected to the out-group.

Supplemental Material