Data were obtained from a larger study of school violence involving 6,189 students in grades 5-12 who attended 26 schools in two suburban school districts of southern California. The sample contained 45.7 percent males and 54.3 percent females. In addition to victimization information, the study sought information on perceived danger, hostility, interpersonal trust, feeling of belonging to school, like or dislike of school, peer and teacher connections, preoccupation with school violence, and average course grades. Results indicated significant differences between the two study groups on indexes of school violence victimization, perceived danger at school, and average course grades. The most critical finding was that students with invalid and/or unreliable responses reported significantly more violence than the comparison group. The need for research on the accuracy of school violence self-reports and concerns about the accuracy of existing school violence prevalence information are examined. 30 references and 1 table