Explaining the disparity in placement instability among African-American and white children in child welfare: A Blinder–Oaxaca decomposition

Abstract

African-American children in the child welfare system are at disproportionate risk of adverse experiences including placement instability. This article compares placement instability among African-American and white children in the National Survey of Child and Adolescent Well-Being and identifies mechanisms underlying racial disparities using a Blinder–Oaxaca decomposition. The type of initial out-of-home placements contributes significantly to the racial gap in placement instability. However a large amount of racial disparity remains unexplained. Additional factors, not captured by these analyses, apparently explain African-American's increased risk of placement instability. Predictors of placement instability differ between racial groups. Among African-Americans, older age, initial placement in a setting other than kinship care, and having a higher externalizing CBCL score at baseline are associated with greater instability. Among white children, however, only initial placement in a foster care setting predicted placement instability.

Introduction

Over-representation of African-Americans in the child welfare system has been a persistent concern since the 1950s (Billingsley & Giovannoni, 1972, Roberts, 2002, Select Committee on Children, Y., & Families, 1989, Smith & Devore, 2004). African-Americans comprise 34% of the children in foster care settings even though they represent only 15% of the U.S. child population (US Government Accountability Office, 2007). The reasons for this over-representation are unclear. One possibility is that abuse and neglect are more common among African-American children, and recent data from the National Incidence Study of Child Abuse and Neglect (NIS-4) support this possibility (Sedlak, McPherson, & Das, 2010).¹ Another explanation involves their experience in that system once they enter it; African-American children remain in the system longer, and this difference contributes to their over-representation (Courtney & Wong, 1996, Wulczyn et al., 2007). According to 2005 data from the Adoption and Foster Care Analysis and Reporting System (AFCARS), 45% of children leaving foster care were non-Hispanic white, while only 28% of exiting children were African-American (US Department of Health Human Services, 2007).

In addition, African Americans in foster care experience more frequent placement changes (Huebner, 2007, Zinn et al., 2006). This difference is especially troubling because frequent changes harm children's development (Harden, 2004, Leslie et al., 2000, Wulczyn et al., 2002). Multiple placements increase children's sense of rejection and decrease their ability to form emotional ties with their caregivers (Webster, Barth, & Needell, 2000), which in turn reduces the chance that subsequent placements will be stable (Harden, 2004, Webster et al., 2000).

Racial differences in time in out-of-home care, the likelihood of exiting from care, and placement instability are poorly understood. On one hand, African-American children have higher levels of factors known to predict adverse experiences in child welfare: lower family income (Jones, 1998), the presence of health conditions and disabilities (Connell et al., 2006, James et al., 2004), and reduced access to supportive services (DeNavas-Walt et al., 2007, Newacheck et al., 2003, US Government Accountability Office, 2007). Alternatively, longer length of stay in out-of-home care, placement instability, and decreased likelihood of exit for African-American children could reflect differences in the way the child welfare system treats children of different races (Derezotes & Poertner, 2005, Dettlaff & Rycraft, 2008, Hill, 2004, US Government Accountability Office, 2007). Personnel may respond to different children differently with racial biases operating in ways that they themselves do not recognize. Research has not determined how these mechanisms contribute to racial disparities in child welfare. Complicating matters is the fact that the processes of interest may operate differently in different communities (e.g. rural v. urban) (Johnson et al., 2007, Needell et al., 2003, Wells & Guo, 1999, Yampolskaya et al., 2007). To understand community-level variation, data including multiple communities is needed.

This study uses nationally representative data from the National Survey of Child and Adolescent Well-Being (NSCAW) to compare placement instability among black and non-Hispanic white children in the child welfare system and to identify mechanisms underlying these differences. A statistical decomposition is employed that quantifies the extent to which racial differences in placement instability reflect (1) differences in the prevalence of risk factors associated with instability and (2) differences in the relationships between these risks and instability. Distinguishing these two explanations is critically important since they imply different solutions. For example, if children of different races enter the system with different needs (levels of risk), then one would not expect them to have the same experiences within that system. On the other hand, if two otherwise similar children of different races enter the system and have different experiences, one needs to explore differences in the way in which the system cares for each (i.e., the relationship between these factors and instability).