Karen McCallum1, An-Lin Cheng1. 1. School of Nursing, University of Missouri-Kansas City, Kansas City, Missouri.
Abstract
OBJECTIVE: In response to criticisms of the traditional investigative model of child protective services (CPS) as adversarial, a Differential Response Model has emerged, with investigative and noninvestigative alternative response (AR) paths. The purpose of this study was to identify relationships of county-level community variables to response paths. DESIGN AND SAMPLE: Secondary analysis used data from the National Child Abuse and Neglect Data System linked to county-level variables from the American Community Survey. The final dataset included 62,499 cases and 98 counties from five states. MEASURES: Multilevel modeling was used to analyze the binary outcome variable of CPS response path (AR, non-AR). Predictor variables included indicators at child, county, and state levels. RESULTS: County-level variables (housing vacancy, child poverty, unemployment, and households with public assistance) were significant predictors (p < .05) of CPS response path and accounted for 12.30% of variability in the final three-level model. Individual variables (report source, maltreatment type, child age, race, and number of children in the report) were also significant predictors. CONCLUSION: County-level community variables have significant relationships with CPS response paths and impact how CPS units respond to new referrals. Research is needed to apply advanced multilevel analytic procedures to more accurately model nested relationships.
OBJECTIVE: In response to criticisms of the traditional investigative model of child protective services (CPS) as adversarial, a Differential Response Model has emerged, with investigative and noninvestigative alternative response (AR) paths. The purpose of this study was to identify relationships of county-level community variables to response paths. DESIGN AND SAMPLE: Secondary analysis used data from the National Child Abuse and Neglect Data System linked to county-level variables from the American Community Survey. The final dataset included 62,499 cases and 98 counties from five states. MEASURES: Multilevel modeling was used to analyze the binary outcome variable of CPS response path (AR, non-AR). Predictor variables included indicators at child, county, and state levels. RESULTS: County-level variables (housing vacancy, child poverty, unemployment, and households with public assistance) were significant predictors (p < .05) of CPS response path and accounted for 12.30% of variability in the final three-level model. Individual variables (report source, maltreatment type, child age, race, and number of children in the report) were also significant predictors. CONCLUSION: County-level community variables have significant relationships with CPS response paths and impact how CPS units respond to new referrals. Research is needed to apply advanced multilevel analytic procedures to more accurately model nested relationships.