Perceived peer victimization and ethnic school context: the impact of interethnic climate

In western countries, there is concern about ethnic school segregation, as recent studies find indications that it is unfavorable for educational achievement (eg. Sweden: Szulkin & Jonsson, 2006; the Netherlands: Karsten et al., 2006). However, other studies have pointed to the flip side of this picture, arguing that school desegregation might have negative consequences for non-cognitive outcomes, such as peer victimization (Hanish & Guerra, 2000). When educational policies do not consider these potential adverse consequences, they run the risk to fail. Therefore, educational research should identify and explain these negative effects of desegregation, in order to counteract them and make desegregation policies work. In this contribution, we focus on perceived peer victimization as a possible effect of ethnic school composition. Studies investigating this, have found widely divergent results, and have still not come to an understanding of this relationship (Graham, 2006; Hanish & Guerra, 2000; Juvonen, et al., 2006; Verkuyten & Thijs, 2002; Vervoort, Scholte, & Overbeek, 2008). One of the factors responsible for this, can be the differing operationalization of ethnic composition – some use the proportion of students from a certain ethnic group (Vervoort et al., 2008), others a calculated index of heterogeneity (Graham, 2006; Juvonen et al., 2006). To counter this, we use both constructs simultaneously. For our analyses, we draw upon three theoretical frameworks. Firstly, we use the ‘Imbalance of Power’-thesis, that states that victimization is most likely to occur when ethnic groups are not equally large (Graham, 2006; Juvonen et al., 2006). This explanation starts from the most defining characteristic of victimization: the imbalance of power that exists between bully and victim. These actors are said to draw their power from the numerical strength of their ethnic group: if the own group is larger, they have more power, and vice versa. However, this new line of reasoning counters the well-established approach of group threat theory (Blalock, 1967), that states that interethnic conflict is the highest in schools where the ethnic groups are roughly as large: in those contexts, it is not clear which group is in charge, and, as bullying is a strategy to acquire social dominance (Demanet, 2008), it is used to resolve the matter. As a third theory, we use constrict theory (Putnam, 2007). According to this theory, the amount of ethnic diversity in a given context triggers social anomie or social isolation. As such, this theory can be used to hypothesize that victimization rates are higher in schools with a larger ethnic heterogeneity. The above stated theories contend that the relation between ethnic composition and peer victimization is not a mechanical one, but is moderated by intermediate processes. Especially certain factors of the school interethnic climate are said to mediate the relationship between ethnic composition of the school and victimization. In our contribution, we test these hypotheses as well. More specifically, we test whether the amount of interethnic conflict at school, the number of interethnic friendships in a school, and the amount of multicultural education at school indeed act as mediating mechanisms. We used data gathered in 2008–2009 from 2485 pupils in a sample of 68 primary schools in Flanders as part of Segregation in Primary Education in Flanders (SIPEF) project. Given the data at different levels - variables of ethnic school composition as determinants at school level and perceived peer victimization as outcome at pupil level, the use of multilevel modeling is appropriate. The models are estimated separately for migrants and natives. The determinants are entered stepwise in the model, to determine if selection and mediating effects occur. In the first step, we estimate effects of the composition variables – ethnic minority concentration and heterogeneity -, followed by control variables in the second. In the third model we add the amount of interethnic conflict and friendships at school; in the fourth the amount of multicultural education. In the fifth, these interethnic climate variables are entered simultaneously. The results indicate that ethnic composition indeed has its influences on the rates of peer victimization, if only for migrants: for natives, it exerts no influence whatsoever. Migrants have a significantly lower chance of being victimized in schools where they comprise 80% to 100% of the student population, compared to schools where they are largely outnumbered by native students. This supports the expectations from the ‘Imbalance of Power’-thesis. The ethnic heterogeneity appears to have no relation to victimization, contradicting our expectations from constrict theory. There are indications that interethnic climate is in part responsible for the observed effects. Especially the amount of interethnic conflict at school appears as an important moderator: there seems to be less interethnic conflict in schools with a majority of migrant students, which is responsible for the decreased rates of victimization in those contexts. Counter intuitively, the more multicultural education there is in a school, the more students report to be bullied, probably because the presence of multicultural education in those schools raise the students’ awareness to racist victimization. This finding is in line with the study by Verkuyten and Thys (2002), who give the same plausible explanation