A Multi-Level Bayesian Analysis of Racial Bias in Police Shootings at the County-Level in the United States, 2011–2014

A Multi-Level Bayesian Analysis of Racial Bias in Police Shootings at the County-Level in the United States, 2011–2014

by Cody T. Ross

Published: November 5, 2015


A geographically-resolved, multi-level Bayesian model is used to analyze the data presented in the U.S. Police-Shooting Database (USPSD) in order to investigate the extent of racial bias in the shooting of American civilians by police officers in recent years. In contrast to previous work that relied on the FBI’s Supplemental Homicide Reports that were constructed from self-reported cases of police-involved homicide, this data set is less likely to be biased by police reporting practices. County-specific relative risk outcomes of being shot by police are estimated as a function of the interaction of: 1) whether suspects/civilians were armed or unarmed, and 2) the race/ethnicity of the suspects/civilians. The results provide evidence of a significant bias in the killing of unarmed black Americans relative to unarmed white Americans, in that the probability of being {black, unarmed, and shot by police} is about 3.49 times the probability of being {white, unarmed, and shot by police} on average. Furthermore, the results of multi-level modeling show that there exists significant heterogeneity across counties in the extent of racial bias in police shootings, with some counties showing relative risk ratios of 20 to 1 or more. Finally, analysis of police shooting data as a function of county-level predictors suggests that racial bias in police shootings is most likely to emerge in police departments in larger metropolitan counties with low median incomes and a sizable portion of black residents, especially when there is high financial inequality in that county. There is no relationship between county-level racial bias in police shootings and crime rates (even race-specific crime rates), meaning that the racial bias observed in police shootings in this data set is not explainable as a response to local-level crime rates.


In 2014, Kyle Wagner began an open contribution campaign [1] to compile all records of police-involved shootings in the United States between 2011 and 2014 in an attempt to better record the use of lethal force by police [2]. The U.S. Police-Shooting Database (USPSD) collects information on the race/ethnicity of civilians shot by police, their status as armed or unarmed, the identity of the officer(s) involved, relevant geographic information, and citations to detailed descriptions of the events.

While other databases on police shootings have been published by the government, for example through the FBI’s Supplementary Homicide Report [3], or the CDC’s National Vital Statistics System [4], these records are often censored of critical information (such as the names of the officers involved), lack independent evaluation of the justification for the shooting, and are selectively published. The FBI data, for instance, are not only incomplete, but may be structurally biased by the reporting behaviors of police, as the majority of the 17,000+ police departments in the United States do not file fatal police shooting reports, or do so only selectively [5]. According to Gabrielson et al. [5], Florida departments have failed to file reports since 1997. The data collected thus far by the USPSD help to shed light on racial bias in police shootings in Florida, which has some of the most racially-biased police shooting rates in the nation. In Miami-Dade, for example, unarmed black individuals are estimated to be more than 22 times as likely to be shot by police than unarmed white individuals. Such patterns in police violence have been immune to public scrutiny until now.

The failure of the nation’s police to critically evaluate their own use of force, has led the United Nations Committee Against Torture [6] to sharply criticize the ever growing militarization of police departments in the United States, especially as evidence of significant race-based and sexuality-based brutality and excessive use of force has been uncovered, including bonafide acts of torture (e.g., those committed by Chicago Police Commander Jon Burge and others under his command, between 1972 and 1991). The UN Committee Against Torture specifically noted that it: “regrets the lack of statistical data available on allegations of police brutality and the lack of information on the result of the investigations undertaken in respect of those allegations” (pp. 13, [6]). This paper provides a response to the first of these two concerns.

Moving Forward

The work of documenting police violence in the United States, has recently begun through several open-contribution, public-access projects in addition to the USPSD. The Stolen Lives Project started by the Anthony Baez Foundation and the National Lawyers Guild [7], the Fatal Encounters Database started by Brian Burghart [8], and the Killed By Police database [9] are examples, as is the Mapping Police Violence project [10], which emphasizes visualization of the raw data from the above-mentioned databases. Additionally, Wikipedia.org [11], the Washington Post [12], and the Guardian [13] have begun keeping rigorous statistics on police shootings in specific years. Unlike the censored data released by official sources, the data in the USPSD and other grassroots databases allow for fine-scale evaluation of the use of lethal force, including investigation of department-specific and even officer-specific patterns. It is, for instance, possible to identify police departments and officers who kill unarmed black individuals at disproportionate rates. With the previously-used SHR data, lack of reporting and/or selective-biases in reporting of police shootings, could have masked underlying racial biases in police shootings, or masked the rates at which unarmed individuals are shot by police.

USPSD data will provide the public and federal agencies within the United States with much needed information describing where external review of police procedures, training, and practices may be needed to protect the civil rights of American citizens. Additionally, the data may be of use to: 1) communities during local elections of mayors, city council members, and police chiefs, 2) organizations, like the United Nations Committee Against Torture, reviewing allegations of racially motivated homicide and torture, and 3) academics seeking to understand the structural drivers of race-based violence and homicide by police.

full report can be read at http://journals.plos.org/plosone/article?id=10.1371%2Fjournal.pone.0141854

Leave a Reply

Your email address will not be published. Required fields are marked *