Racial Matching and the Likelihood of Aging out of Foster Care: A Logistic Regression Analysis

Background and Purpose:

Aging out of foster care is associated with a range of negative outcomes, including housing instability, lower educational attainment, and poor mental health. Understanding the factors that influence whether a child exits foster care to permanency versus aging out is critical to improving outcomes. One such factor may be racial matching between foster youth and caregivers. Racially matched placements may promote cultural continuity, stronger attachment, and increased placement stability. These factors could reduce the likelihood of youth aging out of care without permanent connections. This study investigates whether same-race placements are associated with reduced odds of aging out, with attention to racial/ethnic disparities across placement experiences.

Method:

This study utilized data from the Adoption and Foster Care Analysis and Reporting System (AFCARS), child file FY 2019, to compare the likelihood of aging out for youth in racially matched foster family placements compared to those in transracial foster family placements. The sample included 398,354 children from all 50 U.S. states plus Washington DC, who entered care in FY 2016. Logistic regression was used to assess the relationship between racial matching and aging out, controlling for demographics and maltreatment history. Two supplementary analyses were conducted: one with race/ethnicity as a moderator and a second to adjust for potential state effects.

Results:

Racial matching was not significantly associated with the likelihood of aging out (OR=1.04, p =.358). The strongest correlate was age at exit (OR = 5.83, p < .001), with older youth having dramatically higher odds of aging out. Youth with more placements (OR = 1.02, p < .001) and more removal episodes (OR = 1.05, p < .001) had a higher likelihood of aging out. As part of a supplementary analysis, a model was run with an interaction between child race/ ethnicity and racially matched placements; the interaction terms were not significant, indicating a similar relationship between racially matched placements and aging out of care among all racial/ethnic groups. In the second supplementary analysis, the intraclass coefficient was high (0.38), highlighting over one-third of the variance in aging out to be at the state level.

Conclusions and Implications:

This study found no evidence that racially matched placements significantly influence aging-out outcomes. Furthermore, race/ethnicity did not moderate the relationship between racially matched placements and aging out. However, when we ran another binary regression model, we observed a strong significant effect at the state level. This suggests that further investigation is needed to explore the significance of the state-level variable. Future research should also explore other potential “matching” factors in addition to race/ethnicity (e.g., language, religion, national origin) that may improve outcomes for youth in care.