Instability for Children in Guardianship Versus Adoption Placements with Kin: Approximating the Right Counterfactual

Background and Purpose:  Over the past two decades there has been a rapid increase in the number of former foster youth living in legally permanent guardianship and adoptive homes. While the shift from foster care to permanency is heralded as a major child welfare achievement, little is known about the long-term stability of guardianship and adoptive placements. In addition, previous post-permanency research has been hampered by the use of short study windows and small convenience samples, and given limited attention to potential selection bias. Finally, few studies have rigorously examined what would happen to children who exit to guardianship if guardianship were not a permanency option.

To address these gaps in the literature, this study examined placement instability, or discontinuity, for foster youth living in kinship subsidized guardianship or adoption placements. A population of current and former foster youth was tracked for at least 10 years, or until children reached age 18. Placement changes were examined for children in subsidized guardianship homes versus two different counterfactual conditions: adoption only or adoption and long-term-foster-care (LTFC).

Methods:  Administrative data was obtained from the Illinois Integrated Database, maintained by the Illinois Department of Child and Family Services. The study population included all children between ages 3 and 16 placed in permanent kinship adoptive or guardianship homes in Illinois between 1998 and 2002 who spent at least 3 years in foster care (N=17,047). Also, 4,840 children in LTFC with similar placement histories were combined with the adopted children for the second comparison group.

Descriptive statistics were estimated for the population, and then nearest neighbor within caliper propensity score matching (PSM) was used to match children on eight selected covariates. The two counterfactual comparisons were examined with matched groups, and post-matching analyses included chi-square tests of independence and bivariate regression models to examine differences in discontinuity, or placement changes, by placement type.

Results: Descriptive statistics indicated that less than 20% of children in guardianship or adoption placements experienced discontinuity, as compared to over half of children placed in LTFC. No statistically significant differences were found between the matched groups on covariates in bivariate tests. Regression models estimated using the matched groups indicated that, when compared to adopted children only, children in guardianship placements were 66% more likely to experience discontinuity, OR=1.66; 95% CI [1.48, 1.86]. However, when compared to children in both adoptive placements and LTFC, children in guardianship placements were 30% less likely to experience discontinuity, OR=0.70; 95% CI [0.64, 0.78].

Conclusions: This study has implications for child welfare policy and practice. Findings indicate that placement changes are less likely for children in both guardianship and adoptive homes as compared to children in LTFC. Further, this study provides evidence that careful counterfactual reasoning is needed to examine complex relationships between child and family characteristics, placement type, and child outcomes over time. Results generally support the use of guardianship as a potential permanency solution for families who are not planning to adopt, in order to provide long-term stability for children in LTFC.