| Abstract: | Relations among witnessing violence, victimization, and physical aggression were investigated within a high‐risk sample of 1,156 sixth graders. Longitudinal, multilevel analyses were conducted on two waves of data from two cohorts of students in 37 schools from four communities. The sample was 65% male and 67% African American. Neighborhood concentrated disadvantage, witnessing violence, victimization, and physical aggression were strongly and positively correlated at the school level. Contrary to hypothesis, exposure to violence did not mediate the effects of neighborhood concentrated disadvantage on changes in physical aggression. As expected, witnessing violence and physical aggression had bidirectional longitudinal effects on each other at the student level. In contrast, there were no cross‐variable relations between changes in violent victimization and aggression over time. |
