Not All Neutrals Are Equal: Unpacking the Neutral Point in Stereotype Measures

• In Event: Identity and Measurement in Racial Politics

Abstract

The endorsement of negative racial stereotypes (lazy, unintelligent, violent) has been used as a measure of explicit racial prejudice. Typically, respondents are asked to rate African Americans (or other racial minorities) on these dimensions relative to whites on a seven-point scale. Respondents are also asked to rate whites on these dimensions relative to Blacks or other racial minorities. Subsequently, researchers construct a difference measure that captures how respondents view African Americans relative to whites. Researchers generally interpret a difference score of zero as evidence of a lack of expressed explicit racial prejudice. However, there are many different ways by which a respondent may arrive at a difference score of zero. For example, a respondent may rate both groups at the midpoint of “four,” which means that they perceive both groups as equally hardworking and lazy. However, a respondent could also rate both groups as a “seven” (equally lazy) and arrive at a difference score of zero. Still, respondents could also give both groups a rating of “one” (equally hardworking) and arrive at a difference score of zero. While all the aforementioned scenarios would result in a difference score of zero, many scholars using racial stereotypes measures typically treat all difference scores of “zero” equivalently. In this paper, we argue and demonstrate that this operationalization of the measure may be problematic because it overlooks different ways a respondent can be classified as having a neutral difference score. Using survey data from the 1992-2020 American National Election Studies, we demonstrate that respondents who provide a difference score of zero by rating both groups at the midpoint of the scale (“four”) are more supportive of racially liberal policies than respondents who provide a difference score of zero by rating both groups equally lazy (above 4). Respondents who rate both groups at the midpoint of the scale (four) are also less supportive of racially liberal policies than respondents who provide a difference score of zero by rating both groups equally hardworking. In other words, even if all difference scores of zero are treated as the absence of an expression of racial prejudice, how respondents arrived at a difference score of zero matters. We also find that difference scores of zero, or neutral ratings, have increased over time and that this increase has been largely driven by white Democrats. Finally, we couple our observational data with an experiment, in which we manipulate the options provided to respondents for the stereotype measure. Our study has implications for how we interpret difference scores of zero, and for studies of racial attitudes that rely on this measure.

.

Authors

• Lafleur Stephens

• Burcu Kolcak, Rutgers University

• Darius Cozart, Princeton University