Background: Although youth are exposed to many forms of violence, most studies have been concentrated on only one type of violence exposure and focused on older adolescents or very young children. Little is known about direct and indirect effects of violent stressors on the health of African American adolescents in urban middle schools or the cumulative effect of multiple forms of exposures.
Objective: To test theoretically derived relationships between the types and levels of violence exposure and experiences; coping; and physical, behavioral, and mental health outcomes.
Methods: A structural equation modeling approach was used in this cross-sectional predictive correlational model testing design. Youth's experiences with exposure to and witnessing of violence were examined on three levels-community, family, and peer-in relation to physical, behavioral, and mental health outcomes. The sample (n = 309) consisted almost entirely of African American seventh graders from four urban middle schools. Forty-two percent of students were boys. More than 80% said that they had been in a boyfriend or girlfriend relationship, and 55% were currently in such a relationship.
Results: Eight of the 15 paths tested in the hypothesized model were found to be statistically significant, indicating an average fit (χ2 = 133.06, df = 40, ratio of 3.3, p <.001, root mean square error of association =.087, normed-fit index =.89, comparative-fit index =.92, goodness-of-fit index =.93). Removing nonsignificant paths statistically improved model fit, resulting in an adequate fit (χ2 = 146.78, df = 47, χ2 / df = 3.1, p <.001, root mean square error of association =.083, normed-fit index = .88, comparative-fit index =.91, goodness-of-fit index =.93). Although coping did not demonstrate a mediating effect on health outcomes, it had a direct effect on physical and mental health.
Discussion: The findings partially support the hypothesized model. Violence at home, personal violence, and coping had significant direct effects on health outcomes. Community violence did not have a significant effect, and coping was not an intermediary variable in this sample.