Reducing Racial Polarization in Perceptions of Police-Involved Shootings

• In Event: Systems of Exclusion and Belonging among Minorities

Abstract

Americans appear to form particularly strong opinions about racialized events, and individuals in different racial groups often reach seemingly irreconcilable views on these issues. Earlier research suggests that this divergence stems from how individuals process information, because the same information appears capable of leading different individuals to distinct conclusions. One prominent explanation of this difference is that people reach diverging conclusions in part because the beliefs and attitudes they already hold render some of the new information more credible than other pieces of that information (a form of biased Bayesian updating). This raises the possibility that manipulations influencing preexisting beliefs and attitudes could help mitigate racial differences in information processing.

To test this possibility, we present the results of a preregistered two-wave survey experiment conducted on a broad national sample of Black and White Americans (N=1,000) wherein individuals were asked to evaluate witness statements about a fictitious officer-involved shooting. Prior to the scenario, respondents were asked their preexisting beliefs about the prevalence of racial bias in police shootings and their perceptions of criminality rates among Black vs. White Americans. Some respondents received an intervention where they were exposed to the best data on the actual rates of bias and criminality to see if manipulating preexisting beliefs would result in smaller racial differences in scenario evaluation. Although we found evidence that individuals’ preexisting beliefs were susceptible to manipulation and that these beliefs were closely related to how individuals made sense of the scenario, manipulating preexisting beliefs did not have a significant effect on scenario evaluations either directly or indirectly. The results suggest that changing preexisting beliefs is not sufficient to reduce racial biases in information evaluation. We consider a few reasons why this might be the case.

Authors

• Josh Pasek, University of Michigan

• Fabian Guy Neuner, Arizona State University

• Hakeem Jerome Jefferson, Stanford University

• Brianna Zichettella, University of Michigan