Family Instability and Exposure to Violence in Childhood and Adolescence

Two contemporary social trends in the U.S. the growing instability in children's family structures, the continuing exposure of many children to violence in their homes and neighborhoodsare each the subject of significant attention from researchers, policymakers, and the public at large. Yet, they are rarely connected to each other in empirical research, even though such research can expand understanding of why family instability is a source of inequality, elucidate the etiology of violence and its deleterious effects on young people, and shed light on the ramifications of other potential social crises (e.g., the effects of mass incarceration on families, growing inequality by race and social class). The purpose of this proposed project is to leverage extant data from the Project on Human Development in Chicago Neighborhoods (PHDCN) and federal data bases (e.g., U.S. Census) to examine the degree to which past and contemporaneous family structure changes predict greater and increasing exposure to violence among children and adolescents. The conceptual model to be tested elaborates on this basic longitudinal association in several ways, including by: 1)measuring family instability by changes in parents, partnership statuses but also more broadly through the movement of other adults and children through the home, 2)exploring how parental incarceration predicts exposure to violence indirectly through family instability as well as more directly, 3) assessing the variability in links among family instability, parental incarceration, and exposure to violence by race, poverty, age, and gender, and 4) examining such variability across neighborhoods differing in social disadvantages and cultural resources. The PHDCN includes three waves of child, parent, and community data for multiple cohorts (ages 3, 6, 9, 12, and 15 to be used here; n = 16% White, 35% African American, 45% Latino/a, 4% other race/ethnicity). These data will be analyzed through lagged and cross-lagged structural equation modeling linked to other techniques designed to improve causal inference in the face of observable confounds (e.g., propensity score weighting) and unobservable confounds (e.g., fixed effects, robustness indices). Three manuscripts will be produced for submission to peer-reviewed journals, with policy briefs, press releases, and teaching resources linked to each one. ca/ncf