Testing and Validation of the Adapted Color-Blind Racial Attitudes (CoBRAS-CW) Scale Among Child Welfare Practitioners

Background: Over the past decade, substantial research has documented ongoing racial disparities for children of color in child welfare. Experts have posited that disparities in child welfare may be perpetuated by implicit, interpersonal, institutional, and structural biases. Yet, to date, few scales have been adapted to measure practitioner attitudes and recognition of implicit, institutional, and structural biases that perpetuate discrimination and inequity against families of color. As part of an ongoing project to develop, implement, and evaluate an innovative model to improve family engagement and increase awareness of racial bias, the authors of this study administered and adapted the Color-Blind Racial Attitudes Scale for a sample of child welfare practitioners in a large Southern state.

Methods: This study tested and piloted the CoBRAS-CW scale among child welfare practitioners as part of a multi-year, multi-site study to improve outcomes for children in foster care in a large Southern state. Through a series of meetings with multiple researchers with expertise in racial disparities in child welfare, the team identified two factors (unawareness of racial privilege and unawareness to blatant racial issues) to include in the pilot. Based on initial feedback from a team of stakeholders that piloted the survey, we modified two of the items from the original scales. This resulted in a final adapted scale of 11 items. We administered the modified CoBRAS-CW scale to practitioners and supervisors who worked at one of the child placing agencies where the project was implemented. A total n = 56 practitioners completed all CoBRAS-CW items. The team conducted a principal component factor analysis utilizing Varimax rotation to identify factors and structure of the 11 items. Second, Cronbach’s Alpha was calculated to identify internal reliability of the scale.

Results: Assumptions of adequate sample size, multivariate normality, and linearity for principal components and principal axis components factor analyses were met and satisfied, and Bartlett’s test of Sphericity for the EFA (0.82) was statistically significant. Any items with factor loadings under 0.40 were deleted from the model. Two main factors were identified that explained 68.47% of the total variance. Thus, the final model included two subscales with a total of 11 items. Internal consistency was analyzed for each factor and good reliability was observed (subscale 1 α =0.87; subscale 2 α=0.83).

Conclusion: This study adapted and piloted the CoBRAS-CW scale to measure unawareness of racial bias and discrimination among child welfare practitioners. Preliminary findings suggest that the adapted two-factor scale presents good internal reliability. Future research could conduct a confirmatory factor analysis with a larger sample and examine convergent and discriminant validity. As the larger project is rolled out over the next two years, the research team plans to build the evidence and validity for CoBRAS-CW among child welfare practitioners and examine changes over time among practitioners that receive the practice model.