How Did County-Level Context Influence Foster Care Exit Rate Disparity?

Background and Purpose: The first paper found heterogeneous disparities, prompting a closer examination of county-level contextual variables to identify causal influences on exit disparities. The county random effects models and state fixed effects models along with child characteristics in the first paper revealed large county and state variations. In this paper, we investigate which county factors influence exit disparities and build causal inferences from that analysis. Among the variables included in the model, we focus on race-specific poverty rates, urbanicity, and the counties’ adjusted admission rates.

Methods: To analyze the impacts of contextual variables on exit disparities, we considered race-specific county-level measures of child poverty, including both Black child poverty and White child poverty, using 5-year estimates from the American Community Survey (ACS) for calendar years 2011-2015. We also included urbanicity data, using the 2013 National Center for Health Statistics (NCHS) Urban-Rural Classification Scheme for Counties. Finally, we used counties’ adjusted placement rate to see the relationship between placement rate and exit disparity. This method tests whether counties’ placement rate residuals have an impact on exit disparities. The residuals were measured by Empirical Bayes estimates from a separate county random effects model using admission as the dependent variable.

Results: We identified the relationship between race-specific poverty and exit disparity. As White child poverty increases, exit rate disparity decreased significantly. The impact of Black poverty was relatively small. The study found that exit disparity was narrower in places with greater levels of White poverty, which is counterintuitive. The exit disparity also varied by urbanicity. In rural counties, the Black child/White child differences are smaller. We also identified the relationship between adjusted placement rates and exit rates; however, in contrast to previously published work, a county’s adjusted placement rate does not have an impact on racial disparity in the fully specified model.

Conclusions and implications: This is a substantial empirical challenge if, in the end, the goal is to understand how system forces such as structural bias generate disparate outcomes within the child welfare system. By using robust measures to link poverty and socio-ecological diversity with exit disparity, we were able to find evidence showing an actionable set of relationships. Specifically, the study highlights the need to consider the differential impact of poverty on different racial and ethnic groups and the importance of examining how the different urbanicity settings impact how we understand disparities. Again, the findings from this study reveal why it is important to move the disparity discussion away from generalizations that see disparity in monographic terms. How structural bias manifests is likely different depending on where one is looking. Our study reinforces this idea.