Using Rigor In Evaluating a Child-Focused Adoption Recruitment Model
Methods. An experimental impact evaluation was carried out in 25 WWK sites. Using child welfare administrative data, intent-to-treat analyses compared the outcomes of 517 children assigned to the treatment group with those of 497 children assigned to the control group. Non-experimental data were obtained via a web-based case management system that tracked individual WWK recruiter activities with children served in over 100 WWK sites. Visits to experimental program sites were conducted twice and provided information about program context. Surveys of prospective adoptive parents and foster youth provided data on client perspectives of the program.
Results. The evaluation found that the likelihood of adoption for children in the treatment group is more than one-and-a-half times greater than the likelihood for children in the control group. Subpopulation analyses found the program yields the largest improvement in the likelihood of achieving adoption for older children and those with mental health disorders. No differential impact was found for child gender or for race and Hispanic origin. Non-experimental implementation analyses found the program was implemented unevenly across children. Overall, greater consistency in implementing model components was correlated with a higher likelihood of adoption, and two model components--building a relationship between the recruiter and child and diligent recruitment—seemed to drive this association. This exploratory finding was supported by recruiters’ view that building a relationship with the children is the most important component of the model.
Conclusions and implications. Overall, the study yields evidence that the WWK model is more effective than other services in achieving adoptions for foster youth, particularly for older youth and those with mental health disorders. The findings also demonstrate that adoptive families can be found for children for whom it has traditionally been difficult to find adoptive homes. Non-experimental findings warrant further research to explore what specific activities within the model are responsible for its success. Preliminary findings point to the importance of adoption recruiters establishing one-on-one relationships with children. Finally, the successful implementation of the randomized control design in 25 sites indicates that many agencies are amenable to such research.