Every year in the U.S., people die after police fire gunshots, tase and restrain them, or accidentally crash into their cars during pursuits.
But try to figure out how many police killings occur annually, and an unsettling answer emerges: There’s no official, accurate count.
At a time when police accountability, especially in the deaths of Black people, has become a major public issue, unreliable government numbers fail to portray the true scope, experts say.
“The federal government has done an abysmal job historically of collecting the data,” says David Klinger, PhD, a professor of criminology and criminal justice at the University of Missouri-St. Louis. Early in his career, Klinger had worked as a police officer in Los Angeles and in Redmond, WA, near Seattle. Much of his academic research has focused on police use of deadly force.
It wasn’t surprising, he says, when a recent study in The Lancet, a British medical journal, found that from 1980 to 2018, about 55% of deaths from police violence were wrongly classified as other causes of death in the U.S. National Vital Statistics System. This system tracks all death certificates and causes of death in the country.
“The misclassification of police violence in NVSS data is extensive,” the researchers from the Institute for Health Metrics and Evaluation (IHME) at the University of Washington wrote.
The University of Washington researchers compared the vital statistics system data to three non-governmental, open-source databases on police killings: Fatal Encounters, Mapping Police Violence, and The Counted. These three databases collect police killings from multiple sources, mainly media reports and public records requests.
Using these sources, the IHME team found that from 1980 to 2018, there were 30,800 deaths from police violence. The researchers estimated that more than half of these fatalities — about 17,000 — were listed under other causes of death in the National Vital Statistics System.
They also found that Black people were 3.5 times more likely than white people to die from police violence. Nearly 60% of deaths among Black people were misclassified, making this group the most under-reported of all races.
One of the open-source projects, The Counted, counted 1,146 police killings in 2015 (compared to 524 in the NVSS) and 1,093 in 2016.
The Guardian newspaper conducted the Counted investigative project after learning that the U.S. government had stopped counting police-related deaths because reporting from the nation’s more than 18,000 police departments yielded too little information to produce a credible number. Many departments had failed to take part and provide data.
The Counted not only tried to count every death, but offered snapshot descriptions of the killings, detailing a wide range of circumstances.
The overwhelming majority of people died by gunshot, such as James Bigley, a suicidal 20-year-old man in Oklahoma who fatally shot a woman trying to help him. Police chased Bigley, and one officer shot him after he got out of his car and pointed a gun at them.
In Minnesota, Dahir Adan, 22, stabbed 10 people at a mall before an off-duty officer shot him to death. Others listed in The Counted were killed during bank holdups, drugstore robberies, and carjackings.
In some cases, innocent bystanders have died. Police in Chicago accidentally shot Bettie Moore, a 55-year-old mother of five, after she tried to assist a neighbor with a domestic disturbance.
The Counted also found people who were hit by police cars during pursuits. In Arkansas, Sadine Dixon, 84, died after a deputy crashed into her car at an intersection while chasing two suspects.
In rare instances, officers have killed family members in domestic violence crises. Kyle Kurian, 25, an off-duty police officer in Long Beach, CA, killed his wife, Greta, 22, and then killed himself.
The project even tallied extreme outliers. For instance, Mary Knowlton, a 73-year-old woman in Florida, died after an officer mistakenly fired live ammunition at her during a volunteer citizen police academy “shoot/don’t shoot” drill.
Fatal Encounters, another open-source database, tracks U.S. police killings back to 2000. D. Brian Burghart, the former editor and publisher of the Reno News & Review, founded the project and has tallied 31,031 deaths from 2000 to 2021. It can be difficult to get information from police departments, he says, and government data remains incomplete.
Fatal Encounters captures and confirms considerably more data, he says, but he acknowledges the limitations. Compared to official national figures on highway safety, “It looks like we’re only catching half the chase-related deaths.”
But the national figures “are also missing a bunch that we have,” he says.
Many cases on the Fatal Encounters site include details. For example, in the death of Carlos Antonio Douglas, 41, on Sept. 20 of this year, Fatal Encounters provided a “media summary” of the incident in Florence, SC: “About 3:20 a.m., a deputy spotted a car driving 90 mph on Alligator Road. The deputy activated the blue lights and siren and tried to stop the vehicle. The deputy reportedly backed off the pursuit as the vehicle approached a red light at an intersection. Shortly after going through the intersection, police said the vehicle lost control, crossed the median, ran through a fence surrounding a retention pond and went into the water. Carlos Antonio Douglas was killed.”
But in other cases, Fatal Encounters notes that police have withheld the names, ages, and genders of the people killed and provided almost no details on circumstances surrounding the deaths.
While police might withhold names until relatives can be notified, Burghart doesn’t believe that’s always the case. Withholding facts can also help police departments avoid scrutiny, he says.
Also, police killings aren’t always reported when they occur, he says, “but only come out because of a lawsuit or a newspaper gets a public records request. Then that case will get reported.”
In one highly publicized case, the family of Ronald Greene, a 49-year-old Black man, filed a wrongful death lawsuit after the Louisiana State Police claimed that he had died in a car crash after a police chase in 2019. But police body camera footage showed a different scene: troopers using a stun gun and choking, punching, and dragging him.
The coroner’s report ruled that Greene had died accidentally after crashing his car into a tree and made no mention of the struggle with police. The FBI ordered the autopsy to be re-examined and rejected the car crash as the cause of death. Instead, the new forensic review pointed to other factors, among them: police hitting Greene in the head and restraining him at length, as well as cocaine in his system.
Coroners and Medical Examiners
Often, reports from medical examiners and coroners make no mention that police violence contributed to deaths, according to the IHME researchers. In fact, they believe that the death certification system plays a major role in the undercount.
In police-related deaths, a medical examiner or coroner must fill out the cause of death on the death certificate. “However, only some cities have forensic pathologists to act as the coroner,” the researchers wrote. “In small, rural counties, the coroner can be a physician with no forensic training, the sheriff, or a mortician.”
In some cases, police might not tell examiners of their role in the death. If the death certificate does not mention that the person was killed by police, the case could be misclassified.
Also, the researchers wrote, “Many medical examiners and coroners work for or are embedded within police departments,” creating “substantial conflicts of interest” that could discourage them from listing police violence as a cause of death. Furthermore, according to the researchers, one national survey found that 22% of medical examiners have reported being pressured by others, such as coroners, to change the cause of death on a certificate.
The researchers suggested several actions: Improved training and clearer instructions on how to document police violence on death certificates; allowing forensic pathologists to work independently from law enforcement; and awarding pathologists whistleblower protections under the law.
If there’s not adequate information after a death, Burghart says, public suspicion and anger may grow.
Lack of transparency clouds the contentious public debate over police violence. Some charge that aggressive policing and systemic racism lead to a disproportionate number of police killings among people of color.
“Today, U.S. police are heavily militarized, and fatal police violence disproportionately affects Black, Indigenous and Hispanic people,” the IHME researchers wrote in The Lancet. “Police are more likely to shoot Black civilians than white civilians given the same levels of criminal activity, even when the civilian is unarmed.”
Others, including Klinger, counter that the vast majority of police shootings are justified and that unjustified police shootings are not only uncommon, but typically not motivated by bias. In his research on police use of force, he has reviewed hundreds of investigative case files and large data sets, he says.
Regarding police shootings, he says, “The vast majority of the people that are killed by the police have a weapon and are using it in an aggressive fashion, either against the police or an innocent third party, a crime victim, their spouse whom they are upset with, their girlfriend, whatever the case might be.”
As a former police officer, he says that cops often try to hold fire.
“I’ve seen officers, myself included, time and time again — when they had lawful warrant to shoot — not shoot,” he says. “The number of encounters where police officers could legally shoot somebody far exceeds the number of cases where officers actually pull the trigger.”
“Officers are trained to manage encounters in ways that will reduce the number of times that they would need to shoot and to shoot the fewest number of rounds that they would need to in order to resolve a situation,” he says. “Now, police officers sometimes don’t abide by their training and sometimes they rush into things and create shootings that could have been avoided. But the police department doesn’t want the officers to do that. What they want the officers to do is slow stuff down, de-escalate.”
Even if a victim was unarmed, that doesn’t automatically mean that the shooting was unjustified, according to Klinger. He cited one case in which an unarmed woman was killed while driving her boyfriend, who was actively engaged in a shootout with police. And some unarmed people are capable of overpowering officers and killing them with their own guns, he says.
However, Klinger limited his comments to police shootings, not cases involving other forms of police violence, such as chokeholds or, as with George Floyd, from an officer kneeling on his neck.
Samuel Sinyangwe, a data scientist who founded Mapping Police Violence, one of the open-source databases used in the Lancet study, believes that bias drives much of policing. In an October online presentation titled “Using Data to Fight Police Violence,” he said, “Our society is structured by race and racism, and policing is no exception.”
He launched his site in 2015, inspired to do so after a white police officer fatally shot Michael Brown, an 18-year-old African American, in Ferguson, MO, igniting protests.
To date, Sinyangwe has gathered information on 13,147 police departments and 2,878 sheriff’s departments across the country. He has done much research, or, as he has said, “a deeper dive into the data for explanatory factors, what explains police violence.”
Within the Columbus, OH, police force, for example, he was able to drill down to the level of individual police officers. He obtained use-of-force information from 2001 to 2020, including officers’ names, personnel files, how many times each officer used force and against whom, and disciplinary records.
“There were officers that never used force. Most officers used force a handful of times,” he said.
But at the extreme end, there were officers who he believes could have been guilty of misconduct. “You see officers who used force over 100 times. There’s a set of officers who use force at dramatically higher rates than the average officer,” even by national standards, he said.
“Officers that are much more likely to use force against other people or shoot people or engage in misconduct…those officers tend to spread those behaviors throughout the organization,” Sinyangwe said.
When highly violent officers supervise or train others or work the same patrols, they can influence others to engage in similar behaviors, he said.
Police departments can use data to predict the spread of misconduct, particularly among those exposed to violent officers, he said. Then departments could develop interventions or take steps to remove excessively violent members.
Better Data Needed
Despite differing viewpoints on police violence, there’s mutual agreement on the need for high-quality data to guide an issue that has gripped and divided the country.
“There’s a lot that you can do when you have the data,” Sinyangwe said during his online presentation.
“I think that this is a vitally important topic,” Klinger says. “My hope would be that we could have honest dialogue rooted in really good, sound analyses.”