Updated June 19, 2020, with data from 2015-2019

Data collected by the Washington Post on the use of lethal force by police officers since 2015 indicate that, relative to the proportion of the population, Blacks are over-represented among all those killed by police. As is evident in the figure below, (looking at the top blue bar) according to the US Census estimates, Blacks made up 12% of the population. However, from 2015 – 2019 they accounted for 26.4% of those that were killed by police under all circumstances. In other words, Blacks were the victims of the lethal use of force by police at nearly twice their rate in the general population. Whites make up the majority of victims of police use of lethal force (50.3%) from 2015 – 2019, BUT they also currently make up the majority of the population (61%). Asians make up about 5% of the US population but just 2% of the victims of the lethal use of force by police. Hispanics make up 18% of the US population and just over 18% of the victims of the use of lethal force by police. Native Americans make up 1% of the US population and 1.7% of the victims of the use of lethal force by police.

One way to think about it is over- or under-representation. All things equal, a socially-constructed racial group should be subject to lethal encounters with the police at a similar rate to their proportion in the population. The figure below shows the over- or under-representation of racial-ethnic groups from 2015 to 2019 comparing the proportion of victims of lethal police encounters to the proportion in the respective annual US population estimates. Data points that are above the 0% line indicate over-representation among victims of police lethal force and data points below 0% indicate under-representation. Since 2015, the earliest year of data collection from this particular source, blacks have been over-represented by around 15%, Hispanics, Native Americans, and Asians have been roughly similarly represented relative to their proportion in the US population, and whites have been under-represented by around 10%. (Note: At the time of writing, data from 2019 had a larger number of “unknown” racial/ethnic identities of victims, 20% compared to 3-10% in previous years.) The rate of over-representation for blacks is relatively stable from 2015-2018 and shows initial signs of an increasing trend in 2019.

Is that the right comparison? Shouldn’t we compare the percent of those killed by police to the percent of interactions with the police? Even more precise, shouldn’t we compare the percent of those killed by police to those encounters that were actually life-threatening to either the police officers or other people? There are several problems with such comparisons.

First, that data doesn’t exist. There are no national records of the nature of every encounter with police officers. Additionally, the definition of life-threatening encounters is subject to interpretation and video footage by bystanders or CCTV of many instances of the police use of lethal force has shown that what gets defined by officers as dangerous or life-threatening… is not. “Dangerous” is a subjective perception, even for police officers. Perceptions of danger are often tinted if not heavily skewed by explicit or implicit racist views that the black body is more dangerous. For more on the long history of race and policing see the book, Policing Black Bodies by Earl Smith and Angela Hattery, and From #BlackLivesMatter to Black Liberation by Keeanga-Yamahtta Taylor among many others. See here for scholars responding to the idea that blacks commit more crimes. See here for ideas on improving the research and need for more data. Policing does not occur in an objective social vacuum. The history of policing shows that it has been used to repress and control African American communities and other communities of color (see here and the books mentioned above).

Second, the assumption that police encounters and the occurrence of crime are highly correlated is problematic. Crime statistics are records of policing, not crimes. Records of reported crime are records of victims who contact the police. We cannot assume that existing data on crime is a full account of crime or criminal activity. For decades there was a deliberate effort guided by formal policies that placed a focus of policing on drugs, the so-called War on Drugs. These policies and policing disproportionally targeted people of color resulting in mass incarceration even though whites and blacks report similar rates of drug use (see here and here). Racial profiling has been documented in traffic stops (a.k.a. “driving while black”, stop and frisk enforcement in New York City, and more. Whites conduct disproportionate surveillance of blacks (see here and here). From Mapping Police Violence:

Policing happens in a social context. It is not objective. By comparing the proportions of racial-ethnic groups in the general population to the proportion of those killed by police, we take into account, without measuring directly, all the other social contexts that result in a disproportionate number of black victims of police lethal force. The social contexts of neighborhood segregation, income and wealth inequality, school resource inequality, school-to-prison pipeline processes, white supremacy, “color-blind” racism policy apathy, disproportionate poverty, voter suppression, labor force discrimination, and more.

Here is a specific example of how the level of policing does not correlate to the level of crime but rather the rate of the black population. In 2019, Lance Hannon and Aaron Siegel examined the police service areas of Philadelphia. In those areas where blacks make up less than 50% of the population, the correlation between police frisking people during traffic stops and the violent crime rate of the area is positive. Meaning that police are more attuned to the high rate of violent crime in some areas and are subsequently more likely to frisk people. We can see how that logic is reasonable for officers to engage. See that relationship illustrated below in the figure from their article in the sociology journal Contexts.

However, in police service areas of Philadelphia that are more than 50% black, the relationship between frisking and the rate of violent crime disappears. Traffic stops in these areas result in police frisking the drivers no matter the rate of violent crime. This relationship is evident in the figure below.


As the authors state, “One potential explanation for these trends is that police officers may not perceive differences in levels of dangerousness among Black neighborhoods in Philadelphia. As is evident in our graphic, majority Black areas tend to have higher average violent crime rates than other spaces. Yet it is also true that many predominately Black places have violent crime rates significantly below the city’s average. Consider the case on the far left in Panel B. It has a violent crime rate that is only one third of the city’s average, but a frisk rate that is three times that of the comparable low-crime places in Panel A (places that are not predominately Black). It is possible that low-crime Black areas end up being painted with the same brush as high-crime Black areas due to neighborhood-level racial stereotypes.”

The incidents that drive the protests and organization of Black Lives Matter are largely focused on the police use of lethal force on unarmed blacks. See some of the stories behind the data and headlines here. The figure below shows the total annual number of unarmed people of all races, genders, and ethnicities killed by police from 2015 to 2019. Since 2015, the number of unarmed people killed by police has clearly declined.

We can look at the distribution of race and ethnicity among those that were killed by police while unarmed in a couple of different ways. The first is to look at all those who were unarmed when killed and, within that group of victims, examine the racial and ethnic distribution. The figure below does this. Each bar represents 100% of the people who were unarmed when killed by police that particular year. Each bar is then proportionally divided by race-ethnicity. In 2015, 34% of those who were unarmed when killed by police were white, 40% black, and 20% Hispanic. The percentage of whites among this group of unarmed victims increases to nearly 49% by 2018 while the portion of blacks decreases to 36% and Hispanics to just under 15%. Again, 2019 data should be approached with caution due to the dramatic increase in victims in the database with “unknown” race or ethnicity. Just as with the ratio of all those killed to the proportion of the US population, while whites make up the highest percentage of those that were unarmed when killed by police, blacks remain overrepresented in the group relative to their proportion with the US population. Whites and Hispanics remain underrepresented among the victims. I focus on these three racial-ethnic groups because combined they account for nearly all those killed by police while unarmed.

The other way to examine the distribution of race and ethnicity in relation to the police use of lethal force on those that were unarmed at the time is by looking at the proportion of all those that were killed by police by race and then look at the proportion of each race that was unarmed when killed. The figure below does just that for whites, blacks, and Hispanics. Each line represents the percentage of victims of that particular race-ethnicity that were unarmed when killed. So, in 2015, about 15% of blacks that were killed by police were unarmed. About 11% of whites killed by police were unarmed and a bit more than 6% of Hispanics killed by police were unarmed. In 2016 and subsequent years, there is a visible decline in the percentage of victims of each race-ethnicity that were unarmed when killed by police. 2019 appears to show near equal percentages of each group being unarmed when killed by police. Counts rather than percentages show that in 2019, 9 blacks, 19 whites, and 6 Hispanics were unarmed when killed by police. In 2015, 38 blacks, 32 whites, and 19 Hispanics were unarmed when killed by police.

Mapping Police Violence puts it this way (using their similar but not exactly the same data):

In sum, the number of unarmed people killed by police declined after 2015. Arguably, with a well-trained police force, this number should be zero. While whites constitute both the highest number and percentage of those killed by police and those unarmed when killed by police, they also make up a majority of the population in the US (~60% non-Hispanic white in 2019). Blacks are disproportionately impacted by the use of lethal force by the police relative to the general population. Blacks continue to make up a disproportionate number of all those killed by police and the number of those that were unarmed when killed by police. If we look at the victims of police lethal force by race in 2019, a similar proportion of the whites, blacks, and Hispanics killed by police were unarmed when killed.

While this is not the only issue that The Movement for Black Lives and Black Lives Matter focuses on, it is key. Their efforts (organizing, protesting, advocating, etc.) have driven some policymakers and some police departments to take note and take (some but not enough) action to reduce the use of lethal force. While the data indicates that the problem of police lethal force clearly remains racialized, there are signs that incidents of lethal force by police are decreasing for those that are unarmed.

Teach well, it matters.

Follow Sociology Toolbox on Twitter and Facebook

A video by sociologist, Rashawn Ray, from the Brookings Institute.
How can we improve police accountability in the US?

Other excellent resources on the topic:

Click on the images below to be redirected to web pages or articles.

Search a specific police department on Mapping Police Violence‘s tool:

. . .

Here is an attempt at documenting all the unarmed people of color killed by police from 1999-2014

Screen Shot 2016-01-28 at 9.10.41 PM

Also, see some of my other related posts on this topic (click on the titles below to go to the full post):

Screen Shot 2016-08-12 at 11.53.12 AM
Screen Shot 2016-08-12 at 11.53.46 AM
Screen Shot 2016-08-12 at 11.54.46 AM
Screen Shot 2016-01-20 at 12.35.11 PM
Screen Shot 2016-01-20 at 12.36.33 PM
Screen Shot 2016-01-20 at 12.37.09 PM
Screen Shot 2016-01-20 at 12.37.50 PM
Screen Shot 2016-01-20 at 12.38.30 PM
Screen Shot 2016-01-20 at 12.39.15 PM

 . . .

Click on the image below to go to the full story and polling data

. . .

A recent study arguing that the militarization of the police helps neither them nor the community’s view of the police. Read more here

A recent article on the predictors of Black Lives Matter protests. Click on the image to link to the full article.

. . .

“A review of the shootings of unarmed people shows that officers were reported to be under physical attack in about 40 percent of the cases. The remaining 60 percent involved a variety of circumstances, including individuals’ making provocative movements or verbal threats (31 percent) or fleeing, or being shot unintentionally or in undetermined circumstances, according to a review of news reports and video of the incidents. The news accounts cited in the Post database are typically summaries based on information provided by police at the time of each event.”

. . .

. . .

Screen Shot 2016-10-02 at 1.41.32 PM

“Instead of making officers more accountable and transparent to the public, body cameras may be making officers and departments more powerful than they were before.  …  First, many officers are (either earnestly or conveniently) forgetting to activate their cameras when they’re supposed to. Take the case of Terrence Sterling, an unarmed 31-year-old black man who was fatally shot this month by local police officers in Washington, D.C., after his motorcycle crashed into their car. Contrary to District of Columbia policy, no officer at the scene activated their body camera until after the shooting. The city released footage of Sterling’s final moments this week—but that video begins more than a minute after shots were fired. …. The third threat is that many states have introduced or passed new laws that restrict public access to footage while preserving police access.”

. . .

The federal government will soon require police departments and law enforcement agencies to report information on the deaths of any citizens when interacting with police – during a traffic stop, after an arrest, in jail, etc. The data will be gathered quarterly and will hopefully address any gaps in the data that others, like the Guardian above, have tried to fill.

Screen Shot 2016-08-12 at 11.49.14 AM

Screen Shot 2016-07-13 at 9.47.00 PM

. . .

Also from last year, an article exploring the weakness in the FBI data on civilians killed by police. Click on the image to be linked to the full article.

Screen Shot 2016-07-13 at 10.19.31 PM

. . .

“With increasing use of police body cameras come new tests of transparency and trust. This half-hour documentary looks at the consequences for law enforcement and communities, from the rollout to the courtroom.”

. . .