An examination of neighborhood context and risk for youth violence
Analyses combined measures from the 1990 census for Seattle with data from the Seattle Social Development Project (SSDP), a developmental longitudinal study of health-risk behaviors among urban youths. In the first set of analyses, multi-level models were constructed using the HLM program of Bryk, Raudenbush, & Congdon (1996). These models addressed the nested structure of individuals within neighborhoods (defined by block groups) and examined relationships between context measures derived from the 1990 census and individual-level outcomes. In the second set of analyses, standard logistic regression models were used to examine relationships between risk factor constructs and violence outcome measures.Results from the multi-level regression models revealed that youths' perceptions of neighborhood disorganization and attachment to neighborhoods varied between block group areas and that variation was associated with levels of neighborhood disadvantage, measured by the 1990 census. Residential stability, a second census measure, was also related to youths' levels of attachment to their neighborhoods. Further, analyses showed that neighborhood disadvantage may be related to gang involvement and violence during adolescence.Logistic regressions were used to estimate the prediction of violence (at ages 15, 16, and 18) using risk factor variables, all from youth-reports, representing the neighborhood, family, school, and peer domains. On the basis of theory, variables were entered hierarchically by blocks according to their domain of influence. An examination of the overall contribution of each domain to the prediction of violence and the unique effect of each risk factor was carried out. Analyses revealed that in all three hierarchical regressions (i.e. for violence at each age) each block contributed significantly to the overall prediction of violence. It was also determined that variables with unique effects were similar for violence at ages 15 and 16. Similarity between models was less apparent for violence at age 18. Generally, relationships between risk factors and violence at ages 16 and 18 remained consistent after controlling for violence at age 15. Implications of these findings for preventive interventions are discussed.
- Social welfare