Introduction

Bullying is a perpetual issue within schools that affects children from all grade levels. During the 2011 school year, approximately 28 % of students aged 12–18 reported being victims of bullying (Institute of Education Sciences 2011). Bullying is defined as a repeated negative behavior that occurs over time in a relationship characterized by an imbalance of strength and power (Olweus 1994). Although the definition of bullying centers on the imbalance of power between perpetrators and victims, very few researchers have investigated this phenomenon with regard to racial power differentials within schools. Further, researchers fail to look at how the imbalance of power can differentially affect different types of bullying.

The consequences of bullying and peer victimization, particularly during adolescence, are dire. During adolescence, awareness of self and exploration of where one fits in society is critical in the formation of identity (Erikson 1968). A substantial amount of research finds that peer victimization leads to deleterious psychological, social, and academic outcomes. In particular, researchers find that victimization negatively affects students’ self-worth, self-esteem, psychological health, social self-concept, academic achievement, and feelings of belonging (Felix et al. 2009; Van Lier et al. 2012). Determining the interaction between individual and school level variables on types of bullying can help school administrators identify children at risk for specific types of bullying and link identification efforts to interventions that specifically target the mechanism behind the bullying behavior. The purpose of the present study is to investigate the impact of school diversity on overall bullying and race-based victimization among diverse schools and adolescents.

Student Race and Victimization

Researchers have identified several risk factors that contribute to peer victimization. Much of the research has identified person-centered factors including internalizing and externalizing behaviors such as lack of self-esteem and behavioral problems (Hanish and Guerra 2000a; Hong and Espelage 2012; Olweus 1978). Another individual factor given more attention in recent years is student race/ethnicity (Hong and Espelage 2012). Research on racial differences in bullying experiences is varied and conflicting. Some researchers find that African American youth report being bullied less frequently than their Caucasian or Hispanic counterparts (Nansel et al. 2001; Spriggs et al. 2007). Other findings contradict this fact with Hispanic youth bullied less than Caucasian and African Americans (Hanish and Guerra 2000b). Overall, although differences have been found in the literature, they are inconsistent. The inconsistency in experiences of bullying across racial groups has led researchers to conclude that race alone is not directly related to victimization (Vervoort et al. 2010).

School Diversity and Victimization

Although individual factors are important when investigating bullying, they only explain a small amount of the variance when looking at targets of peer victimization (Hanish and Guerra 2000a). In addition, individual factors such as student ethnicity are less relevant when not accounting for their embedded contexts, such as schools (Graham 2006). Thus, researchers have suggested a need to examine bullying from an ecological perspective (Thijs et al. 2014; Vervoort et al. 2010). The relationship between race and bullying is multifaceted and influenced by the racial make-up of schools and classrooms (Juvonen et al. 2006; Graham 2006). It is important to examine perceptions of incidences of bullying with a consideration of the race of the victims and the racial composition of the school. Several theories such as the imbalance of power thesis (Graham 2006), group threat theory (Blalock 1967), and constrict theory (Putnam 2007), explain the dynamics of intergroup racial contact and potential contribution to peer victimization. The present study focuses on the imbalance of power hypothesis, which states that the power of a certain group in a school context is partially determined by the relative number of group members (Graham 2006). The imbalance of power hypothesis holds particularly true with markers of privilege and/or disadvantage such as race and ethnicity (Thijs et al. 2014; Vervoort et al. 2010). Thus, the numerical balance or imbalance between ethnic groups in a school can lead to higher rates of peer victimization (Felix and You 2011). This theory overlaps with the definition of bullying that highlights that bullying involves an imbalance of power between the bully and the victim (Graham 2006; Juvonen et al. 2006). Although the present study focuses on the imbalance of power hypothesis, other hypotheses such as group threat theory (Blalock 1967) help explain the social psychological aspects of bullying across races. For example, Schlueter and Scheepers (2010) found that larger proportions of minority populations incite perceptions of threatened majority group interests and in turn foster anti-minority group attitudes.

Despite the relationship between bullying and the imbalance of power hypothesis, few studies have investigated the relationship between student race, school diversity, and subsequent peer victimization outcomes. In addition, the majority of the research has been conducted with populations outside of the United States. These studies have found partial support for the imbalance of power hypothesis. For example, Durkin et al. (2011) found that minority and majority children reported more peer victimization in schools with more ethnic minorities. In Belgian schools, the number of minority students in a school predicted less victimization for minority children and had no effect for ethnic majority children (Agirdag et al. 2011). The authors attribute this finding to the buffering effect of being part of the majority culture within the broader society. Verkuyten and Thijs (2002) also found support for the imbalance of power hypothesis, finding Dutch children with low numbers of other Dutch children in their classes were more often victimized than when in classes with high proportions of Dutch children. On the contrary, research conducted in the Netherlands by Vervoort et al. (2010) did not support the interaction between race and school composition on peer victimization. Their findings suggested that ethnic majority groups were not significantly more victimized when in the numerical minority than when in the numerical majority. Further, higher levels of school diversity led to more reports of bullying, but not necessarily aimed at the minority ethnic group in the class.

Although these findings are important steps in understanding these phenomena, research is needed on populations within the United States (US). Conducting research within the US is important as racial groups within the US have a history of privilege and oppression that may differ from other countries. Indeed, the history of slavery, segregation, and ongoing acts of racism and discrimination in the US may affect interracial relationships in schools. The three studies that we know of, conducted in the United States provide partial or full support for the imbalance of power hypothesis. Hanish and Guerra (2000b) find support for the interaction between school diversity, student race, and bullying among elementary school students. More specifically, Caucasian children in the numerical minority were at a greater risk of being victimized than Caucasian children in the numerical majority in their schools. In contrast, African American children were slightly more likely to be victimized in predominately African American schools than in predominately non-African American schools. Juvonen, Nishina, and Graham (2006) and Felix and You (2011) both found that ethnic diversity in schools was associated with lower levels of peer victimization with very few effects found by ethnic group. Juvonen and Colleagues conclude that power relations between groups are more balanced in more ethnically diverse schools and contribute to fewer bullying behaviors. Overall, research supporting the imbalance of power hypothesis is varied in its sample populations and findings for various ethnic groups.

Race-Based Victimization

Despite the recent attention given to school diversity and student race in the research on peer victimization, few studies have investigated how the interaction between diversity and student race influence the type of bullying a child experiences. It is plausible that if students are more likely to experience victimization because of their lack of ethnic fit within the norms of the school setting, it could also be the subject of the bullying. Thus, ethnic minorities within a school setting could be more likely to experience racist victimization. Research conducted by Agirdag et al. (2011), supports this assertion, finding in Belgium schools that diversity is more connected to race-based victimization than victimization in general. They call for future research to investigate the impact of school diversity on both race-based victimization and bullying in general. In the Netherlands, Verkuyten and Thijs (2002) found that ethnic minority children (Turkish, Moroccan, and Surinamese) are more often the victim of racist name-calling than Dutch children. School context played a role in the likelihood of experiencing race-based name-calling. For Dutch children, a higher proportion of Dutch children was associated with less racist victimization. For minority groups, however, a higher proportion of Dutch children was related to higher levels of racist victimization. Lastly, Verkuyten and Thijs examined the frequency of racist name-calling, finding that ethnic minorities (Turkish children) were more likely to experience race-based victimization than the ethnic majority group (Dutch children). Although research has investigated school diversity, student race and bullying behaviors, further exploration is necessary to determine the interactions between student race, school diversity, bullying and race-based victimization in particular (Vervoort et al. 2010).

Current Study and Research Questions

Although the victimization literature provides some insight into the relationship between student race, school diversity, and bullying there are several limitations. First, the majority of the research conducted on school diversity and bullying has been conducted with populations outside of the United States. The United States has a distinct history with race relations that is uniquely different from the other countries studied and could potentially influence the relationship between school diversity and bullying. Second, the studies that have been conducted in the United States have focused on elementary or high school aged children (Felix and You 2011; Hanish and Guerra 2000b), or did not account for the experiences of Caucasian youth (Juvonen et al. 2006). Although important, more information is needed on bullying behavior with middle school aged children, a developmental period that often spurs more bullying behavior (Nansel et al. 2001). Last, to our knowledge, no study has investigated the relationship between school diversity, student race, and race-based victimization in a sample of middle school students. The purpose of the present study is to investigate the impact of school diversity on overall bullying and race-based victimization. Due to contradictory findings in the literature, research questions rather than specific hypotheses are presented. The first research question examines whether the prevalence of bullying changes with the racial diversity of schools. The second research question examines whether race-based victimization differs from patterns of general bullying.

Method

Participants

Participants were 4,581 middle school students in grades 6–8, with an average age of 12.75 years (SD = 1.09). The sample consisted of more female (53.4 %) than male (46.5 %) participants (three students who represented .1 % of the sample did not indicate their gender). Among the participants, 89.4 % were Caucasian, and 10.6 % were African American. Complete descriptive data are presented in Table 1.

Table 1 Descriptive statistics for the variables of study
Full size table

Procedures

The data were retrieved from data collected as a part of a larger research study. A research team from a large Midwestern university and county convened to create a survey instrument called the Coordinated Community Student Survey (C2S2). The initial deployment was in 2005. The C2S2 pilot study invited all students in grades fourth through twelfth to participate. Approximately 50 % of students (about 12,000) provided written consent to participate. Students have continued to fill out the C2S2 in the spring of each subsequent year, for the following 9 years (2014). Research staff administered the surveys each year during school hours, via paper and Internet-based completion methods depending on the school. Cross-sectional data from year 2 (2006) of the project were selected due to data availability. Data for individuals were de-identified and included the original items and scale scores relevant to this project.

Measures

Coordinated Community Student Survey (C2S2)

The C2S2 is an instrument designed to assess self-reported student’s attitudes, beliefs, and behaviors. The C2S2 is a 208-item questionnaire that assesses student physical, social, and psychological functioning using reliable and valid multi-item scales selected from public and peer-reviewed sources. Data from the bullying construct was used for this study.

Bullying

Children completed 13 items related to the frequency of their personal experiences with social exclusion, physical aggression, and name-calling. Students were asked “in the past year, how often did the following things happen” and were given 13 bullying related stems (M = 20.44, SD = 6.22). Bullying items were rated on a 4 point likert scale (1 = never and 4 = a lot). A sample item included “A kid at my school teased me about my body.” This subscale has a reliability of α = .84.

Race-Based Victimization

Race-based victimization was measured with a single item about children’s perception of race-based victimization “A kid at my school teased me about my race/ethnicity or the color of my skin.” Children rated race-based victimization on a 4 point likert scale (1 = never and 4 = a lot).

School Diversity

The ethnic composition variable was calculated by dividing the number of students from a particular racial group by the total number of students in the school (Vervoort et al. 2010). Additionally, after examining the percentages, Asian, American Indian, and Hispanic students were dropped from the analysis because the largest percentage of students in any of those three groups was 7 %. As such, the demographic context of the school based on these ethnicities was not examined due to limited variability.

Data Analyses

A hierarchical linear model (HLM; Bryk and Raudenbush 1992) was utilized to address the first research question. HLM was used because students were nested within schools. Thus, the school level effect needed to be examined to ensure the results could be interpreted at the individual student level. The school level variable was ethnic composition, and the student level variables were race, ethnic identity, and bullying score. The test for the random intercept was insignificant so all results are reported at the student level. A multilevel ordered logistic regression was utilized to address the second research question regarding race-based victimization because it was divided into four ordered categories: never, sometimes, often, and a lot. A multilevel analysis was also conducted due to student nesting within schools.

Results

Overall Bullying

The results of the HLM analysis for the effects of race and ethnic composition on bullying indicated several statistically significant outcomes. First, a significant relationship emerged between bullying and race (β = .283, p = .003), finding that, in general, Caucasian students were bullied significantly more than were African American students. The second significant relationship involved the interaction between race and ethnic composition (β = −.003, p = .026), finding that Caucasian students were bullied more often when in the ethnic minority. The relationship between school ethnic composition and bullying for African Americans was not significant. Additionally, significant differences in the amount of bullying experienced between genders emerged (β = −0.07, p < .001). Female students experienced significantly less bullying than male students (see Table 2; Fig. 1).

Table 2 Fixed effects coefficients (top) and variance estimates (bottom) for the predictors of bullying
Full size table
Fig. 1
figure 1

Amount of bullying experienced and school ethnic composition

Full size image

Race-Based Victimization

A multilevel ordered logistic regression was conducted to address whether race and ethnic composition predicted race-based victimization (never, not much, sometimes, a lot). The results indicated that all effects were statistically significant. A statistically significant relationship emerged between race-based victimization frequency and race (β = 1.25, p < .001) as well as race-based victimization frequency and gender (χ2 (3) = 19.50, p < .001). With gender and ethnic composition of the school held constant, Caucasian students faced race-based victimization more frequently than did African American students, and males faced more race-based victimization than females with ethnic composition of the school and race held constant. The second significant finding was between race-based victimization frequency and ethnic composition (β = −0.01, p < .001), indicating that, in general, students faced more race-based victimization if their ethnic group was in the minority at the school. The third significant finding was the interaction between race and ethnic composition (β = −0.02, p < .001). Caucasian students in the minority faced race-based victimization more than did African American students in the minority. On the other hand, African American students in the majority faced race-based victimization more than did their Caucasian counterparts (see Table 3).

Table 3 Fixed effects coefficients (top) and variance estimates (bottom) for the predictors of race-based victimization
Full size table

To further describe race-based victimization, percentages were grouped into three categories (0–24, 25–49, and 50 % and above). Only 16 students were in the 50–74 % range so the top two categories were combined. Among the African American sample, 46.9 % attended schools comprised of 0–24 % African-Americans, 18.3 % attended schools comprised of 25–49 % African Americans, and 34.7 % attended schools comprised of 50 % or more African Americans. Among the Caucasian students, only .9 % attended schools where they represented 0–24, 4.4 % attended schools where they represented 25–49%, and 94.7 % attended schools where they represented 50 % or more of the student body (see Table 4).

Table 4 Percentage of race-based victimization among African American and Caucasian Students
Full size table

Discussion

Inconsistencies exist within the bullying literature on the relationship between individual factors (i.e. race), the racial make-up of schools and incidences of bullying. In addition, little research exists that examines the prevalence, existence, and effects of race-based victimization on American youth. Researchers outside of the United States find school diversity has a stronger impact on race-based victimization compared with other types bullying (Agirdag et al. 2011). This investigation is important as we continue to understand the victims of bullying, under what context the bullying behavior exists, and the type of bullying experienced within school contexts. The goal of the present study was to discern salient individual and contextual contributors to the prevalence and incidence of bullying amongst youth. To understand the effects of victimization, we must consider the contexts in which individuals are more likely to experience peer victimization. The present study aimed to address the influence of the social context on bullying by examining the intersections of race, school racial diversity, and race-based victimization.

Hierarchical linear modeling was used to determine if school context moderates the relationship between race and bullying. Similar with past research, Caucasian students reported being bullied more than did African American students (Nansel et al. 2001; Spriggs et al. 2007). When considering the effects of school context, the present study furthers previous research on elementary children suggesting that Caucasian students are bullied more when in the ethnic minority in a school compared with other ethnic groups (Hanish and Guerra 2000b). One explanation for this finding is consistent with the imbalance of power hypothesis. In the present study within more diverse schools, traditional minority students are the group in “power” because of their size. In addition, definitions of bullying behaviors may differ across groups. For example, within African American communities “playing the dozens” or “signifying” is often common. This is often a game of insults and put downs deeply rooted in the African American culture (Jemie 2003). Although African Americans may understand this teasing as funny banter, those from other cultures, such as Caucasian youth may perceive it as bullying. Thus, cultural inoculation can communicate messages about what is considered “bullying behavior.” Further, White privilege, or the set of unearned privileges and advantages Whites receive in the broader society (McIntosh 1988), further complicates the imbalance of power hypotheses. In this situation, Whites find themselves with privilege in the broader society that may or may not transfer into settings in which they are the numerical minority.

In addition, the results revealed no significant differences in overall bullying for African Americans across contexts. Although no significant differences were found, the type of bullying in each setting could look vastly different. Based on the imbalance of power hypothesis, students in the ethnic minority are vulnerable to peer victimization (Felix and You 2011). This may have been the case for African Americans attending more diverse schools. A potential explanation for similar levels of bullying in school with large percentages of African Americans might be the result of similarities that exist within a social milieu (i.e. classroom or school). Students who are different from this norm will more than likely experience peer victimization regardless of whether they share the same racial and ethnic background. Furthermore, it could be that individuals with a negative view of their race have a greater propensity to victimize individuals in that particular race. Researchers have identified different profiles of racial identity that may illuminate the incidence and prevalence of within-group bullying (Chavous et al. 2003). Based on this research, students who believe that race is central in their identity, and have low views of their race, publically or privately, tend to have maladaptive achievement and behavior outcomes (Chavous et al. 2003). Thus, the racial identity of an individual may influence how they treat others within the same racial or ethnic groups. Future research must consider the racial identity of students as an individual factor that may contribute to incidences of bullying rather than race alone.

When looking at race-based victimization specifically, on average, Caucasian students experienced almost three times the amount of bullying than African American students (30.4 and 11.9 % respectively). Though this is true, students’ experiences with race-based victimization varied significantly by individual student race as well as the diversity of their school context. Relative to themselves, both groups faced more race-based victimization when in the small minority at their school versus when in the majority. This finding is consistent with the imbalance of power hypothesis and research that asserts that a limited number of like peers can lead to peer victimization (Graham 2006). Additionally, relative to African Americans, Caucasian students reported being bullied more frequently because of their race when a numerical minority. One reason for this may be a devaluation of the culture of the majority by traditional minority students such as African Americans. Fordham and Ogbu (1986) found that African Americans who embodied characteristics of the culture of the majority were often criticized and ridiculed by peers for acting White. Although their research has been controversial (Sohn 2011; Stinson 2011), it certainly seems plausible that Caucasian students within a predominately African American setting would also be ridiculed for displaying stereotypical Caucasian behaviors. Other research has substantiated this claim finding that Caucasian students in predominately African American settings are expected to acculturate to the majority culture within the school (Morris 2006).

On the other hand, in settings with more students of their own race, African American students reported experiencing race-based victimization “a lot” double their Caucasian peers (3.4 and 1.7 % respectively). The presence of race-based victimization among African American youth at a rate double their counterparts could be explained by the plethora of research conducted on colorism within the African American community (Heckstall 2013). The present study asked whether students had ever been teased because of their race/ethnicity or the color of their skin. Within a Eurocentric society, comments about the color of other individual’s skin and its relationship to status, and privilege are ever present in the African American community and other communities of color. For example, historically within the United States, African American’s worth in society was often judged based on the passing of the “brown paper bag test.” African Americans who were lighter than a brown paper bag were often afforded privileges that were not afforded to those who were darker. This sentiment holds true today with colorism impacting everything from identity development (Thompson and Keith 2004) to socioeconomic status (Herring 2004).

Additionally, the results indicated significant gender differences in the amount of overall bullying and race-based victimization. Males were significantly more likely to experience bullying in general and experience race-based victimization specifically. Research shows that, when measuring bullying with an overall bullying measure (as done in this study), males score higher on levels of bulling behavior (Hong and Espelage 2012). This finding, in conjunction with studies that find that the majority of bullying behavior occurs within same sex dyads (O’Brien 2011), also supports the claim that males are the targets of this bullying behavior as found in the present study.

In sum, minority status within schools is significantly related to victimization in general and race-based victimization specifically. In addition to the reasons provided above, a number of school level and societal factors could also explain the phenomena discovered. First, other factors such as the presence of active support groups, minority school staff, and school climate may indeed lessen the amount of victimization minority groups experience. For example, research has found a decrease in race-based victimization in supportive schools with higher levels of teacher diversity (Larochette et al. 2010). Thus, it is important to cultivate school environments that accept, foster, and encourage diversity and acceptance.

Although the present study provides valuable insight into bullying behaviors in varying school contexts, several limitations should be noted. First, students provided self-report information regarding bullying behaviors. Although self-report provides an accurate picture of student perceptions regarding bullying and is preferred for data collected over multiple waves (Espelage and Swearer 2003), some assert that it may vary from what actually happens in the school setting (Juvonen et al. 2006). Researchers have recently moved to using peer nominations as a measure of victimization status rather than self-report (Vervoort et al. 2010). In addition, the sample used for the present study was largely Caucasian and attended predominately-Caucasian schools. The lack of diversity in the schools that Caucasian students attended could have contributed to our small effect sizes. Future research could focus on separately investigating within group differences. Additionally, due to data being drawn from a larger dataset, our analyses are limited in scope and do not take into consideration other school level variables such as school climate, socioeconomic levels or urbanicity. Although other research suggests that the majority of the risk for peer victimization occurs at the student level (Felix and You 2011), further investigation of school level factors could help researchers understand the potential complexities/interactions that exist in peer victimization. Lastly, the sample used is only a subset of the population, with data collected on youth in the Midwest. Thus, results are not broadly generalizable.

Future research should continue to investigate the impact of school context on the bullying behavior of youth. Other factors that contribute to race-based victimization should also be investigated such as school climate, student status (popular, neglected, rejected, etc.) and student level of acculturation. Variables related to the racial attitudes of students and how this impacts bullying behavior have also been investigated and should be investigated further (Thijs et al. 2014). Lastly, researchers should take this line of inquiry a step further by not only investigating risk factors for the engagement in this behavior but also identifying protective factors for youth in contexts likely to promote race-based victimization. The identification of protective factors is imperative in the planning of effective intervention programs and strategies (Hanish and Guerra 2000b).

Conclusion

Increases in cognitive maturity (Bigler and Liben 2007) and advances in brain development (Moore et al. 2012) make adolescence a developmental period where race becomes more salient. Exposure to more diverse peers through the transition to middle school facilitates the exploration of social categories such as race (Roberts et al. 1999). Finally, a greater awareness of racial stereotypes and the attempt to maintain power in social settings where one is a part of the numerical majority can contribute to the risk of victimization of racial minorities (Apfelbaum et al. 2008; Graham 2006). The aforementioned factors make adolescence a prime time to investigate the interaction between levels of school diversity, race, and race-based victimization.

The present study makes several important contributions to the literature on adolescence and provides valuable insight into the experiences of bullying across ethnic groups within varying diverse contexts. The findings partially support the imbalance of power hypothesis finding that Caucasian American middle school youth who are numerical minorities within their school context are more likely to be bullied. It is particularly important to note that race-based victimization has a different explanatory mechanism for Caucasian and African American students. Although both groups are more likely to experience race-based victimization when in the minority, race-based victimization among African Americans in settings with more African American students also poses a problem. Understanding these differences, can assist schools in identifying and implementing school climate interventions. For example, schools with high numbers of African American students may want to consider adopting interventions that improve upon students’ racial and ethnic identity to address within group racial victimization. Further, to improve school climate, schools with a strong imbalance of power between racial groups would benefit from creating an open and inclusive environment that allows the voices of all children to be heard (Madsen and Mabokela 2013). Overall, the findings of the present study underscore the importance of addressing how both race and school diversity influence peer victimization and more specifically race-based victimization during adolescence.