Lery, B. (2009). Neighborhood structure and foster care entry risk: The role of spatial scale in defining neighborhoods. Children and Youth Service Review, 31(3), 331-337.
Out-of-home placement prevention programs have traditionally focused on a range of support programs targeting individual families, although there appears to be a renewal in interest in the potential of community level initiatives. Spatial analyses have found that neighbourhood characteristics (e.g., poverty, residential instability) are associated with child maltreatment rates. However, researchers have not yet examined the link between neighbourhood risk factors with foster care placements. This study examined the link between neighbourhood risk factors and out-of-home placements with all first-entries to care (lasting longer than four days) across three years in Alameda County in California (N= 3311).
Results indicate that ìresidential instability, impoverishment and childcare burden are significantly positively related to increased risk of neighbourhood rates of foster care entryî (p. 334). Autocorrelations and spatial lag analyses also indicate that contiguous (or adjacent) neighbourhood vulnerability significantly increases the likelihood of child placements in nearby neighbourhoods. Out-of-home placements are significantly more likely in neighbourhoods with comparatively more disadvantage, and close proximity to disadvantaged neighbourhoods also significantly raises the risk of placements. The magnitude of this finding is large: 69% of the variance in placements is accounted for by the neighbourhood characteristics measured in the study. Findings suggest that preventive efforts to improve neighbourhood social capital may reduce out-of-home placements.
The data was extracted from Californiaís Child Welfare Services Case Management System. Limiting sample to first-time placements eliminates the problem of children with multiple placements during the studyís frame of study. In California, counties are the child welfare administration jurisdictions; Alameda contains variation in relevant variables such as poverty and ethnic differences. At the time of the study Alameda had a population of about 1.4 million residents, approximately 360,000 of which were children; about 15% of maltreatment allegations were substantiated, and about two-thirds of these resulted in a placement (child placements per year ranged from 841-1033 during the study period). Analyses were conducted using three different spatial scales ó census tracts, census blocks and zip codes ó to ìminimize the tendency to commit the ecological fallacy, whereby conclusions are drawn about smaller units or individuals based on results from aggregated areasî (p.332). Data are cross-sectional and do not take into account individual case characteristics and cannot infer the causal link of the statistical relationship.