In my podcast episode with Dr. Phil Goff I outlined some rough estimates on numbers of people shot and killed by police in a year, and how many interactions the police had with the public over the same time frame. Phil was right to point out the limitations of those numbers and how there is a wider relationship between police and minority communities that is not conveyed in those figures. I completely agree, as I did in the episode. The purpose of my comment was more to point out the scale of the challenge facing movements (both within and outside of policing) trying to reduce the numbers of unarmed Black people shot and killed by police. This is not to say the work isn’t vital – of course it is! The point was to appreciate the reality of the challenge posed. This post outlines the sources of my numbers, and explains the limitations of those data.
The police-involved shootings data come from the Washington Post (WaPo) here. The WaPo data include the race of the deceased, the circumstances of the shooting, and whether the person was armed. The primary limitation of this data set (as they recognize in their ReadMe.md) is that WaPo is “documenting only those shootings in which a police officer, in the line of duty, shot and killed a civilian — the circumstances that most closely parallel the 2014 killing of Michael Brown in Ferguson, Mo., which began the protest movement culminating in Black Lives Matter and an increased focus on police accountability nationwide.” This data set does not include deaths in police custody, shootings by off-duty officers, deaths by reasons other than shooting, or incidents where police shot but did not kill the suspect.
While the WaPo data have limitations, it is generally recognized that efforts to collect these data through private and nonprofit organizations have done better to reflect the reality of the situation than officially recorded statistics. The Bureau of Justice Statistics (BJS) has maintained data related to Arrest-Related Deaths (ARD) since 2003. However, in 2010, BJS identified an underreporting by states and moved to a data collection regime more aligned to the WaPo approach, identifying cases using open source methods and validating by following-up with law enforcement agencies, and medical examiner and coroner offices. The ARD program should include more cases because it includes death by any cause in police custody; however the WaPo data are also trusted by many people because it is independent of government data collection. That said, of all the available data sources I recognize that it is limited in scope, as is clear from this table from a Bureau of Justice Statistics technical report published in 2016.
The most recent complete year of data from the WaPo database is 2019, documenting 999 fatal shootings by police in that year. By race/ethnicity, 403 were white, 250 were Black, and 163 Hispanic. The remainder were Asian, native American, other or unknown.
I (arbitrarily) considered here lethal, potentially lethal or perceived to be potentially lethal as someone in possession of a gun or "toy weapon", "knife", "samurai sword", "sword", "Taser", "ax", "machete", "nail gun", or "bean-bag gun". Excluded from this list, but open to discussion are baseball bats, hatchets, and hammers.
Of the 250 Black deaths by police shooting in 2019, two-thirds (168, 67%)of the people were recorded as being in possession a firearm. From the remaining 82, when we exclude the list of potentially lethal weapons (or things that could be perceived to be lethal) from the list in quotation marks in the paragraph above, we are left with 43 shooting deaths of Black people who were not armed with a lethal (or potentially or perceptually lethal) weapon. The race-specific percentage of these non-potentially lethal encounters ending is a deadly police shooting is 17.2 percent for Black fatalities (22% for Hispanic fatalities and 20% for white fatalities).
I took the number of public-police contacts from the most recent Bureau of Justice Statistics report on this topic, Contacts Between Police and the Public, 2015, available here. Table 7 details results of the Police-Public Contact Survey (PPCS) which asked people the number of times they experienced various types of contact with police during the past 12 months, including reporting crime, being stopped on the street, or being in a traffic stop. Summing the total number of contacts within the past 12 months results in a total in excess of 75.7 million.
The number of unarmed fatal shootings of Black citizens occurs at a rate of a bit less than one in every million police-community contacts. However, I also recognize that comment in the middle of a dynamic interview with Phil is incomplete in that my denominator was all police-community contacts. 9.4 million of these contacts were with Black citizens. Therefore the ratio of fatal police shootings for every million police interactions when race/ethnicity is incorporated in both numerator and denominator is about 4.56 per million interactions for just Back citizens (43/9.4m), 3.75 for Hispanic people, and 1.63 for whites.
I should also add an additional caveat that these data are the most recent available, but they are not comparing equivalent years. Furthermore, the police-community contact information comes from a survey – the Bureau of Justice Statistics’ 2015 Police-Public Contact Survey. The PPCS is administered to a nationally representative sample of about 71,000 people whereas the WaPo data is more representative of a population measure. As such, further caution is required in interpretation. A useful additional report is the one into Police Use of Nonfatal force, 2002-11. See table 9 at https://www.bjs.gov/content/pub/pdf/punf0211.pdf
Note: This page was updated August 2020 in numerous places in the light of a revised dataset from the Washington Post and a revision to the list of potentially lethal non-gun items.