Evaluating the color-blind racial attitudes scale using item response theory

The Color-Blind Racial Attitudes Scale (CoBRAS; Neville et al., 2000) is a widely used measure of color-blind racial ideology, a belief system that minimizes the significance of race and race-based systemic inequities. Despite its widespread use and promise, questions remain regarding the CoBRAS's dimensionality and the validity of its scores across race and gender. This preregistered study evaluated the CoBRAS's psychometric properties using full-information factor analysis within an item response theory framework. Measurement invariance across race and gender was evaluated at both the scale and item level using multiple-group confirmatory factor analysis and differential item functioning tests. Data were drawn from four cross-sectional samples (N_total = 4,010). Participants who identified as Asian, Black, or White and cisgender were included in the main analyses to optimize sample size for comparisons. Factor analytic results supported a bifactor structure comprising a general dimension of color-blind racial attitudes with three specific dimensions. A reduced 18-item set excluding two misfitting items performed comparably to the full-scale overall, with modest fit gains in race-group models. Configural invariance was supported across race and gender. However, metric and scalar invariance were not supported across racial groups, suggesting that individuals from different racial backgrounds may interpret and respond to CoBRAS items differently. Across gender, the 20-item scale demonstrated configural and metric invariance but not scalar invariance, whereas the reduced item set demonstrated full configural, metric, and scalar invariance. The findings highlight the value of advanced psychometric methods in enhancing the precision of measures of sociocultural phenomena across diverse populations. (PsycInfo Database Record (c) 2026 APA, all rights reserved).