Adoptive Disengagement Study

OCFS Research Proposal Review Submittal for Approval

Names:

Barbara Rittner, PhD., Associate Professor, Director PhD Program

Marsha Schwam-Harris, MSW, Doctoral Student

School of Social Work

New York State University at Buffalo

Federal Funding: None

Description of Research Proposal Review by IRB: Approval from OCFS will be required to submit to IRB because the data are from OCFS and we will need your permission to conduct the study.

Purpose and Nature of the Research: This study will examine variables associated with the disruption and dissolution of subsidized adoptions in children residing in the State of New York. The research will be conducted using data supplied by OCFS that does not include identifying case information.

Questions to be Addressed:

This study will examine factors associated with children who had been adopted returning to foster care from adoption. All children in this study had been adjudicated dependent and at some point parental rights were terminated. They were placed for adoption in both kinship and non-kinship families as subsidized adoptions with ACS. At some point these children left their adoptive families and were placed back in foster care. This study examined those characteristics most predictive of adoption disengagement.

Specific questions will include:

What are the frequencies of dissolution of an adoption initiated by child, adoptive parent or state agency?

Is age at adoption a factor in disruption?

Is race a factor in disruption?

Is gender a factor in disruption?

Is the relationship of the adoptive family to the child a factor (kinship v. non-kinship)?

Is location of adoptive family residence a factor in dissolution (zipcode rates of runaway behaviors)?

Is adoptive family composition a factor in dissolution?

Is age at adoption correlated with age at disruption? That is, are older children at adoption likely to have shorter duration of adoptions?

Is type of service category (basic, special, exceptional) correlated with the age at adoption?

Is type of service category (basic, special, exceptional) a factor in the rate of dissolution in these cases?

Are types of services provided to adoptive family a factor in disruption?

Are the placements in foster care prior to adoption a factor in dissolution?

Are specific school problems associated with disruption?

Are psychiatric problems of the child associated with disruption?

Are physical health problems associated with disruption?

Are disruptive and threatening behavioral problems associated with disruption?

Is a pattern of runaway behavior by adopted child a factor in dissolution of adoption?

Research Design:

This study will utilize a survey design. In the first wave, existing aggregate demographic WMS/CCRS data of the children who returned to foster care after finalization of adoption will be used to determine which factors or combination of factors are most predictive of adoption disengagement. In the second wave, surveys will be developed to obtain additional information from case files to determine if there are child specific factors, family related or service related factors that contribute to adoption disengagement.

Sampling Approach:

Wave 1 will utilize the WMS/CCRS data provided by OCFS on the children who have reentered foster care from adoptions. Stage 2 review of case files will utilize a survey form. A stratified representative randomly selected sample of cases will be used. The strata will reflect service categories, age at adoption, gender, and race.

Kinds of Data:

See above

Source of Data:

See above

Analytic Approach:

Wave 1:

Step 1 will be to analyze the distribution of the variables. Descriptive analyses of the distribution of age at adoption, age at disengagement, gender, race, service categories will be provided.

Step 2 will include chi-square analyses of distribution of variables.

Step 3 will include ANOVA to determine whether there are significant variances by category in disengagement. In particular where the age at adoption is related to the age at disengagement. In additional correlation will be done by category of ages at adoption to service categories to determine the relationship to disengagement.

Step 4 will be survival analysis to determine the rate of disengagement at ages of adoption.

Step 5 logistic regression will be dome to determine the likelihood of disengagement based on variables at adoption.

Wave 2:

Step 1 will be to code and enter data

Step 2 will be to analyze the distribution of the variables. Descriptive analyses of the distribution of age at adoption, age at disengagement, gender, race, service categories will be provided.

Step 3 will include chi-square analyses of distribution of variables, in particular to determine which factors (kinship v nonkinship, family composition, services, prior placements, child behaviors, etc.) are most likely to be significantly distributed or related to disengagement.

Step 4 will include ANOVA to determine whether there are significant variances by category in disengagement including the additional child, family and community variables. In additional correlation will be done by category and will include the additional child, family and community variables.

Step 5 will be survival analysis to determine the rate of disengagement at ages of adoption and will include the additional child, family and community variables.

Step 6 logistic regression will be dome to determine the likelihood of disengagement and will include the additional child, family and community variables.

Access and confidentiality:

All data will be maintained in secure files on a server with limited password protected access. All survey forms will be maintained in a locked cabinet with sign out keys.

Confidentiality agreement

Maintain anonymity of respondents

All data will be coded without identifying case information and will be maintained as above

Procedures for informed consent:

Risks

none

Data Security

As above

Data Presentation in aggregate

As above