The study reported in this article is based on a nationally representative sample of 10,400 students in grades 7 through 11 in 162 schools across Israel. The authors used hierarchical linear modeling to examine the differences between Jewish and Arab schools in the relationships between school-level variables--socioeconomic status (SES) of the school's neighborhood and students' families, school size and class size, school level (junior high and high), and school climate--and students' victimization reports (serious physical victimization, threats, moderate physical victimization, and verbal-social victimization). The results show that whereas school climate and school size seem to operate similarly across different cultural contexts, the SES of a school's neighborhood and students' families were associated with victimization for students in Arab schools only. Theoretical implications of these findings for school violence research in other cultures are discussed.
KEY WORDS: community; culture; family; hierarchical linear modeling; Israel; student victimization
The concept of "risk factors" is widely accepted in the violence and social work research literature; however, few studies have examined whether different within-country ethnic groups have similar risk factors for victimization. Furthermore, most risk factor studies on victimization have not included school variables. Israeli Arab and Jewish schools are embedded in different ethnic, cultural, and economic contexts. In this study, we examined whether the relationships among school, family, and community risk factors for student victimization are similar for Jewish and Arab students. We examined students' reports of school victimization and compared the effects of community, family, and school risk factors between Jewish and Arab schools.
Risk factors are defined as variables that promote or increase the probability of onset, severity, and duration of a problem (Coie et al., 1993). Risk factors may be individual characteristics, family factors, specific life experiences, or contextual factors (Pollard, Hawkins, & Arthur, 1999). The more a child is exposed to multiple risk factors, the more he or she is hkely to engage in violent behavior (Herrenkohl et al., 2000; Saner & Ellickson, 1996). Clearly, there are many types of risk factors embedded in different contexts. Nevertheless, compared with other normative contexts such as the family or the community, most school violence studies and theories have not incorporated risk factors that emerge from the school setting (Astor, Meyer, & Behre, 1999; Baker, 1998). Moreover, studies rarely examine school-based risk factors concurrently with community and family variables in diverse cultural settings. Knowing how the community, family, and school contribute to student victimization in different cultures could enhance theory and practice.
Considerable progress has been made toward identifying risk factors for violent behavior among children and youths. However, very few models explicate the role of the school in relation to the family and the community. For example, Fraser (1996) suggested social environmental conditions that might promote or mitigate aggressive behavior. Social context that provides opportunity for adequate training in cognitive and social skills can reduce children's tendency to use violence to achieve social goals. Variables such as consistent parental supervision, rewards for pro-social behavior, engagement with pro-social peers, and attendance at schools with positive climates are seen as protective and enhancing. Fraser presented several social conditions that might enhance children's use of violence, such as family conditions, processes, experiences that may reinforce children's aggression (such as failure to set limits and coercive style of parent-child interaction), peer rejection, living in poverty, and individual factors such as biological conditions. According to Fraser, all these variables might limit a child's opportunity to engage and learn social and developmental skills. As a result, children who live in such conditions are at higher risk to turn to aggression to achieve social goals.
Other studies show that peer rejection and membership in a deviant peer group are related to child antisocial behavior and disorder at school and at home (Dishion, Patterson, Stoolmiller, & Skinner, 1991; Natvig, Albrektsen, & Qvarnstrom, 2001; Patterson, DeBaryshe, & Ramsey, 1989).
Still other inquiries show that low academic performance and school failure are identified frequently as risk factors for involvement in antisocial and violent behavior in multiple environments (Dishion et al., 1991; Herrenkohl et al., 2000; Patterson et al., 1989; Saner & Ellickson, 1996).
These examples mention school factors but focus mainly on one ecological setting and do not explicate the relationships among school, family, and community variables, or their impact on children's use of violence. There is a growing consensus that school violence should be treated as a multidetermined phenomenon that can be explained by many factors on multiple levels (Astor, Pitner, & Duncan, 1996; Catalano & Hawkins, 1996; Fraser, 1996; Goldstein, 1994; Hellman & Beaton, 1986; Lindstrom, 2001; Welsh, Greene, & Jenkins, 1999). Only a handful of studies have examined risk factors on multiple levels and dimensions (see, for instance, Welsh et al., 1999). The present study expands on one of the most extensive studies of risk factors for school violence, conducted in Israel by Benbenishty and associates (Benbenishty, Zeira, & Astor, 2000; Khoury-Kassabri, Benbenishty, Astor, & Zeira, 2004). The study examined the relationships between school victimization (serious physical victimization, moderate physical victimization, threats, and verbal-social victimization) and many aspects of the school's ecology in a nationally representative sample of students in Israel (162 schools and almost 10,400 students).
Risk Factors for School Violence in Israel
Community and Family Context. Higher levels of serious physical victimization and threats are associated with low socioeconomic status (SES) of the schools' neighborhood and students' families. These results are consistent with findings that indicate that the neighborhood in which children and adolescents grow and develop influences young people's relationships with violence. Poverty, discrimination, and lack of opportunities for education and employment have all been identified as important community risk factors for interpersonal violence (Attar, Guerra, & Tolan, 1994; Hamburg, 1998; Tolan, Guerra, & Kendall, 1995).
Junior High and High Schools. Junior high students report being victimized in school more than students in high schools do. This tendency for younger students to report greater victimization than older students is reported in numerous studies (Borg, 1999; Boulton & Underwood, 1992; Whitney & Smith, 1993).
School Size and Class Size. Larger classes are associated consistently with higher levels of each victimization type examined. School size, on the other hand, is not associated with victimization. According to Walker and Gresham (1997), large schools and classrooms pose a major challenge. In large schools and classes teachers have difficulties developing and maintaining meaningful relationships with students, especially at-risk students who have more intense needs for attention and involvement.
School Climate. Three aspects of school climate have been examined: (1) school policy against violence that includes clear, consistent, and fair rules; (2) teacher support of students; and (3) students' participation in decision making and in the design of interventions to prevent school violence. All three aspects of the school climate are associated with all types of victimization studied by Khoury-Kassabri et al., 2004. Khoury-Kassabri and associates' study (2004) showed that school violence could be predicted by several contextual school factors.
Interplay between Cultural Variables and Risk Factors for Violence
Most studies exploring the role of risk factors on youth violence have tried to identify universal factors that might promote or impede violent behavior. Only a few studies tried to examine whether the relationships between risk factors and violent behaviors are similar across different social-cultural groups (Brownfield, 1987; Paschall, Ennett, & Flewelling, 1996). Comparisons between groups are important because the effects of risk factors on violent behavior may interact with a variety of demographic and socioeconomic characteristics such as race, ethnicity, and culture (Brownfield; Fraser, 1996; Guerra, Huesmann, Tolan, Van Acker, & Eron, 1995; Pollard et al., 1999). Thus, a risk factor that should be the target of preventive interventions in one group may be less important for another group. Efforts to design culturally sensitive interventions should be informed by these differences among groups in risk factors for violence.
Overall, norms of certain groups may cross cut multiple settings. For instance, in Japanese society, often characterized as an interdependent culture, factors such as the power of the group to make individuals conform to the group standards and norms and the need for homogeneity were found to promote bullying among youths (Sugimori, 1998). Also, the lack of communication within the family and the inflexible and hierarchical social structure of the Japanese society in many cases impede the normal development of young people and might enhance youth involvement in social problems (Feldman, 1998). The way in which Japanese children interpret or experience these family or societal norms at school is unclear because researchers often see schools as mere reflections of family and cultural norms (Kim et al., 2000; Shumba, 2001; Wilson, 1982). It is possible that norms in the school are slightly different and have greater...