Evaluating the Indian Child Welfare Act: A within‑State Quasi‑Experimental Study Using Propensity Score Matching

Background/purpose: Indigenous children remain disproportionately represented in foster care at more than twice their share of the general population. Although the Child Welfare System has long harmed Indigenous children’s mental health and identity development, recent shifts suggest that sovereignty‑affirming policies such as the Indian Child Welfare Act (ICWA) may help protect their cultural and Tribal connections. However, despite ICWA’s importance, no study has directly evaluated its effectiveness. Because some states have enacted their own ICWA statutes to clarify federal mandates, this exploratory study examines states that passed and ICWA, comparing child welfare outcomes before and after enactment.

Methods: This study is among the first to use the RAPIDS dataset (Kim et al., 2025). We analyzed the first foster care spell of Indigenous only children from 2005 to 2020. Our primary independent variables are states that passed an ICWA (LA, MI, NE, WA, WI, AK), comparing outcomes before (non-experimental) and after (experimental) enactment. States analyzed both aggregated and separately. We used propensity matching to derive our non-experimental group. Each dependent variable was analyzed separately, including placed, discharged with relative, foster care length of stay, reunification, adoption, lost parental rights, and case transferred to agencies nearer a child’s Tribe or relatives. Using regression analyses, we estimated the average treatment effect.

Findings: Aggregated, the experimental group (n=3039) was more likely to be placed with relatives (ATE=6.02) and have case transferred (ATE=6.53) than the non-experimental group (n=11669). The group also spent 189 fewer days in care and were less likely to be reunified with parents, adopted, or experienced termination of parental rights. At the state level, AK, NE, and WI showed increases in relative placements, and all states saw reductions in time spent in care. AK, MI, and NE had higher probabilities of case transfers, while WA showed a decrease. MI and NE were the only states with notable declines in reunifications, whereas WA experienced an increase. AK, NE, and WI had lower rates of parental rights termination, and WI was the only state with a meaningful decline in adoptions following its ICWA enactment.

Conclusion/implications: Results suggest, when aggregated and separately, ICWAs may reduce time spent in foster care, a favorable outcome given the harms associated with prolonged stays. ICWAs may also increase the likelihood of case transfers, indicating that children are being placed with relatives while in the system. States had fewer losses of parental rights, and fewer adoptions across variants, which may indicate efforts to preserve caregiver-children relationships and, in turn, Indigenous communities. Experimental states consistently showed lower likelihood of reunification, diverging from our hypothesis. This may reflect ICWA’s preventative effects, as lower‑risk cases that might otherwise escalate to removal are mitigated, leaving mostly higher‑risk cases compared to before the passage of ICWA. Variance in outcomes across states may reflect differences in how states implement or define their state-specific ICWA mandates. Overall, this study suggests ICWA can shape outcomes for children likely protected under the law, indicating that similar policies can reinforce Tribal Sovereignty.