The Intersection of Race, Poverty, and Risk in Child Welfare Decision-Making

Background

Racial disproportionality in the child welfare system refers to the overrepresentation of a racial group compared to their percentage in the child population. African American children, for whom disproportionality has been documented at every stage in the child welfare system, have been the most significantly affected. Although multiple studies have examined the causes of disproportionality at each decision-making point in the child welfare process, many have not sufficiently accounted for the potential confounds of race/ethnicity with income and assessed risk level. To address disproportionality, it is important to understand both where in the child welfare decision-making process it is most likely to occur and all factors that may impact it in addition to race. This paper presents a study examining disproportionality at three key decision-making points: substantiation, opening a case for services, and removing children into substitute care versus providing family-based services.

Methods

The analytic sample was cases processed by the Texas Department of Family and Protective Services from September 2003 through February 2005, resulting in 177,936 cases at the investigation (substantiation) stage and 123,621 cases at the service provision stage. Three analyses were conducted to examine separately the decisions to substantiate a case, open a case for services, and to remove a child from the home versus provide family-based services. Each analysis used a binary logistic regression model in which predictors included race, household income, risk assessment score, socio-demographic variables, and report source variables.

Results

Race had minimal impact on overall rates of substantiation, although there was evidence that race had a slight effect on substantiation for certain forms of maltreatment (physical abuse OR=1.405, p<.01; sexual abuse OR=1.213, p<.01). However, poverty played a strong role in the disproportionate rate of substantiation among African American families (OR=1.892, p<.01). Race did have a significant effect on the decisions to open a case for services and to remove a child from the home; African Americans were more likely to have cases opened for services (OR=1.208, p<.01) and to be removed (OR=1.77, p<.01). Additionally, poverty was associated with increased odds of case opening (OR=2.40, p<.01).

Implications

Findings suggest that income is a stronger predictor of maltreatment substantiation than is race, implying that previous findings of disproportionality at this decision point may be due to confounds between race/ethnicity and income. At the service decision stage, even when other factors were covaried, racial disproportionality existed for both case opening and removal decisions, but poverty also appeared to serve as an indicator of risk for decision-makers. These findings regarding the complex interplay between race, assessed risk, and income should be included in future studies of disproportionality and further, highlight a need for additional research into the mechanisms by which race/ethnicity and income influence child welfare decision-making processes. The potential for race/ethnicity and income to impact child welfare decision-making must be fully understood by caseworkers, and should be addressed in caseworker training. These findings should likewise be used to guide practice and policy interventions addressing disproportionality.