Development and Initial Validation of the Disavowal of Racial Bias Scale (DRB)

While research suggests that blatant expressions of racism are on the decline, more subtle forms of bias persist (Dovidio & Gartner, 2004; Nosek, Banaji, & Greenwald, 2002). These biases can be automatic and unintentional, often occurring outside conscious awareness. Studies suggest that developing awareness is the first step to moderating discriminatory thoughts and behaviors (Divine & Monteith, 1993; Dovidio & Gaertner, 2004; Durrheim, Hook, & Riggs, 2009; Monteith & Voils, 1998). When White Americans are aware of their biases, they are more likely to adjust their attitudes and alter their behaviors. Crucially, when biases go unacknowledged, there are fewer opportunities to combat unintentional racism. As a result, the tendency to disavow racial biases demands scholarly attention. In order to further research in this area, a way of measuring awareness of racial bias is needed. The purpose of this dissertation was to develop and initially validate the Disavowal of Racial Bias Scale (DRB). A review of the research on racial bias helped generate 38 initial items. An empirical approach was then used to determine an optimal version of the scale. In Phase 1, an exploratory factor analysis (EFA) of data from 579 participants suggested a 2-factor model with a total of 24 items. The first factor was named Bias Examples because it included statements referring to specific examples of racial bias. The second factor was named Bias Existence because it included statements referring to the general phenomenon of racial bias. In Phase 2, a confirmatory factor analysis (CFA) of data from a second subsample of 579 participants was used to confirm the factor structure identified in Phase 1. Both subscales demonstrated high internal consistency, providing evidence of the DRB's reliability. Further psychometric evaluations provided evidence of convergent and discriminant validity. However, the 2-factor model did not appear to be reasonably consistent with the data as evidenced by a poor model fit. Although there are many promising aspects of the final 24-item DRB, more work is needed to make it a valid measure for future use. Limitations of this study and recommendations for future scale development in this area will be discussed.