Reverse Subtyping: The Effects of Prejudice Level on the Subtyping of Counterstereotypic Outgroup Members

Abstract

Although many anti-bias interventions try to overcome stereotypes by presenting positive and/or counterstereotypic members of the outgroup, people often subtype these members and refuse to see them as typical of the outgroup. Although subtyping has been shown to be a common phenomenon, it is unclear if preexisting attitudes moderate this process. The current study examined whether preexisting prejudice levels would moderate the subtyping process. Specifically, it was found that although high-prejudiced individuals subtyped a positive racial outgroup member, low-prejudiced individuals demonstrated the opposite pattern and subtyped negative outgroup members as being atypical. This suggests that although the subtyping phenomenon may be universal, its expression is moderated by preexisting intergroup attitudes.

Notes

¹We chose to use the upper and low quartiles in order to get the biggest difference between high and low prejudiced groups. Although this does limit our ability to infer how the subtyping processes may work for those in the mild to moderate prejudice range, our rational was to first look for differential subtyping between the two extremes of prejudice because this is where there are clear predictions. It is less clear what the motivations of midrange prejudiced individuals may be and how this would impact the subtyping process.

²Two different pictures were used for each racial group in order to control for the effects of any specific picture. Within each racial group, there were no significant differences between the pictures on the dependent measures.

³One possibility is that these subtyping effects could merely be due to differences in the perceived valance of the negative Black target by the high- and low-prejudiced individuals. In other words, the negative Black target may be seen more positively by the low-prejudiced group. To examine this, a brief posttest was run on a separate sample of participants. Using a similar procedure in the main study, a group of high-prejudiced and low-prejudiced participants were presented with the negative Black target and asked to rate the general valance of the target (e.g., How negative or positive do you feel James's characteristics are?). Results indicated that there were no significant differences in perceived valance between the high-prejudiced (M = 3.92) and low-prejudiced (M = 4.21) groups, t(32) = .937, p = .36. This suggests that the findings in the main study are not simply due to differences in perceived valance among high- and low-prejudiced individuals.

Note. All reported coefficients are unstandardized; 95% confidence intervals were obtained using 5,000 bootstrap resamples.

*p < .05.