On Causal Inference and the Limits of Disproportionality as a Construct: The Case of Foster Care Placement

Abstract

The unwarranted disparities in our foster care system are the result of historical and structural causes. If we hope to correct these disparities by addressing their root causes, we must measure them carefully. In this article, I ask whether our usual measures of Black child/White child placement differences tell us what we need to know to act on those disparities with greater certainty. I specifically examine the limits of disproportionality as a measure. As a way to observe over- and underrepresentation, disproportionality is useful. When the discussion turns to the causes of differential treatment of Black children (or any other children) within the child protection system, there are more useful ways to measure those differences. I demonstrate this point in technical terms, and then I suggest alternative approaches to the measurement problems. Other approaches, I argue, offer a more vivid picture of the problem social workers are trying to solve.