The contributions of community, family, and school variables to student victimization.

An ecological perspective was used to predict school violence from a number of variables describing the students and their families based on nationally representative sample of 10,400 students in grades 7-11 in 162 schools across Israel. Self-administered anonymous questionnaires that included a scale for reporting victimization by serious and moderate physical violence, threats, and verbal-social victimization were filled out during class. Hierarchical linear modeling (HLM) examined the relationships between students' reports of victimization and student level variables (gender, age) and the school-level variables--cultural affiliation (Jewish vs. Arab), the socio-economic status (SES) of the school's neighborhood and students' families, school and class size, school level (junior high vs. high), and school climate. Variance between schools accounted for 9-15% of the variance in student victimization (major factors being school climate characteristics and percent of boys at the school). Boys reported higher victimization than girls for all forms of school violence. Students in junior high schools reported more victimization than high school students. Higher levels of victimization were reported in overcrowded classes, while school size was not associated with students' reports of victimization. Although there were almost no differences between Jewish and Arab schools in students' reports of victimization, the SES of the school's neighborhood and students did have a moderate effect. The discussion highlights the importance of improving school climate and the need to allocate more resources to schools in low SES contexts in order to protect students from school violence.