Keeping Families Together: Validating the Skills Assessment for Parents with Intellectual Disability

Background and Purpose: Parents with intellectual disabilities (ID) are overrepresented in the child welfare system. Because of this, it is important for child welfare practitioners and researchers to have access to valid instruments to assess parenting skills in this population. These instruments can be used to assess parents’ levels of functioning, identify strengths and areas for improvement, and evaluate change over time.

While a myriad of instruments has been used to assess parenting skills in parents with ID, it is not clear whether those designed for parents without ID are relevant for those with this disability. In this study we evaluated the psychometric properties of the newly developed Skills Assessment for Parents with Intellectual Disability (SAPID), an instrument completed by social workers in their assessment and work with parents with ID with child welfare involvement.

Methods: Data from 133 clients who participated in a prevention program for parents with ID who had open child welfare cases between 2006 and 2017 were included in the sample. Data were extracted from client files, and out-of-home care status were obtained from the County’s Department of Social Services one year after program completion. Parenting skills were measured at four time points: at intake, after a skill was taught, and three months post-termination. Confirmatory factor analysis (CFA) was conducted to determine the validity of the SAPID. Goodness-of-fit was assessed by examining a range of fit statistics. Predictive validity was evaluated by comparing baseline and final scores for each construct in a series of dependent samples t-tests.

Results: The best fitting CFA model consisted of three latent constructs, Overall Safety, Parent-child Interactions, and Daily Life Skills, consisting of between two and five observed exogenous variables each (X²(41)=66.53; p=0.01). As X²:df<2, additional examination of the model was warranted. The RMSEA for this model was 0.07 (90% CI: 0.04 – 0.10) with a CFI=0.94 and TLI= 0.92, indicating good overall model fit.

Subscales were developed by taking a sum of each observed variable loading onto each construct. A total parenting skills score was obtained by taking a sum of the subscales. There was a significant improvement at the p<0.01-level in each of the subscales and the total scale for those who completed the program while there was no significant improvement for those not completing the program. Additionally, parents whose families remained intact one year after program termination saw a significant improvement at the p<0.01-level on all parenting subscales from baseline to final assessment while those whose children were placed in out-of-home care saw no significant change.

Conclusions and Implications: This study adds to the knowledge base by assessing the validity of SAPID, an instrument to expediently assess parenting skills in parents with ID who have child welfare involvement. High factor loadings in the CFA combined with small standard errors are indicative of convergent validity. Additionally, the covariances between each of the latent construct indicate the independence of one from the other, indicative of discriminant validity between the subscales. Other implications for practice will be discussed.