Methodological Challenges for Research on Racial Bias in Police Shootings

Abstract

hootings of racial minority civilians by police officers have long been a major social issue in the United States, the subject of extensive media coverage, and a reality that has galvanized social movements, such as Black Lives Matter. Police in the United States shoot people at far higher rates than police in other economically developed countries (Zimring, 2017). The victims of these shootings are disproportionately Black and, to a lesser extent, Hispanic, compared with their shares of the population. These shootings occur in a society that continues to grapple with its oftentimes racist history, one that, to this day, remains beset with pervasive racial inequalities. This is especially true of the relationship between police in the United States and the country's minority communities, which, although complex, has long been characterized by high degrees of both mistreatment and mistrust. It is perhaps no surprise, then, that this issue has become a national flashpoint and that high-profile police shootings weigh so heavily on the psyche of Americans, especially Black Americans.In this context, researchers have increasingly sought to answer a fundamental question: To what extent are police shootings of civilians driven by racial bias? This is research that is high profile and high stakes, being published in prestigious journals, getting coverage in leading media KEY FINDINGS■ The impact of police shootings is severe-particularly for communities of color in the United States-and more-nuanced research is needed to understand the dynamics of racial bias.■ Thinking of racial bias in police shootings as a process with a series of stages can improve research in this area. These stages include officers' encounters with civilians, use of force, and the type of force used.■ Fully unpacking how racial bias operates in police shootings is challenging due to data limitations. These limitations include unreliable federal and crowdsourced datasets, and the problem is particularly severe for non-fatal shootings.■ The most common forms of tests-benchmark and outcome tests-face significant challenges in inferring bias that research has only started to confront.■ We recommend several ways to improve understanding of the dynamics of racial bias in police shootings: better data, clearer definitions and estimands, better methods, and the use of bounding and sensitivity analyses.