Examining Spatial Regimes of Child Maltreatment Allegations in a Social Vulnerability Framework

Abstract

The present research examines child maltreatment allegations (CMAs) in San Diego County, California, exploring spatial patterns of Child Protective Services involvement and multiple, multidimensional measures of neighborhood social vulnerability. Results showed significant patterns of spatial clustering (i.e., hot and cold spots) of CMAs across the county (Moran's I = .316, p < .001). A geographically weighted regression (GWR) was implemented to examine the relationship between CMAs and social vulnerability at the census-tract level, thereby overcoming the deficiencies of global models. Nonstationarity was detected across four indices of vulnerability (socioeconomic status, race/ethnicity, household composition, and health vulnerability) as well as proximity to on-premise alcohol outlets, percentage of residents in each census tract affected by food deserts, and population density, in some cases showing countervailing effects depending on spatial location. A hierarchical clustering was performed on the GWR coefficients to identify spatial regimes, or clusters, across the county. The results yielded six spatial regimes of social vulnerability differentially related to CMA rates. The present study demonstrates the novelty of GWR in combination with a hierarchical cluster analysis for exploring how local contextual processes influence child maltreatment reporting rates across the county.

Keywords: Child Protective Services; child maltreatment; ecological models; methodology; risk factors.