After George Floyd was killed under the knee of a Minneapolis police officer, reporters from The Associated Press wanted to know how many other people died after encounters in which law enforcement used not firearms but other forms of force that are not supposed to be fatal.
The U.S. government should be tracking these non-shooting deaths, but poor implementation and inconsistent reporting by local law enforcement agencies means no one really knows their scope.
A team of journalists led by the AP has spent three years reporting on deaths following “less-lethal violence.” For that investigation, conducted in collaboration with the Howard Centers for Investigative Journalism and FRONTLINE (PBS), reporters created a new database that provides the most complete accounting yet of these cases and new opportunities to understand patterns in police work .
The study identified 1,036 deaths in the past decade resulting from encounters involving less-lethal force. Some cases are known. Others have not been publicly reported. The total is undoubtedly an undervalue; deaths can be difficult to verify, partly due to deliberate withholding of information.
More than 800 of the more than 17,000 law enforcement agencies in the US had at least one documented fatality. The country’s twenty largest cities were responsible for 16% of deaths.
To be eligible, the cases had to meet the following criteria. Verifying each case required primary source documentation, usually data from government agencies. News reports or lawsuit allegations alone were not enough to substantiate a case.
WHAT: Encounters that involved at least one form of force, coercion, or a weapon less lethal than handcuffs.
Holding someone face down in what is known as a prone position and Tasers were the most common forms of violence. Complaining with fists or knees, takedowns, and devices to restrain people’s legs were also common. To a lesser extent, chokeholds, pepper spray, spit hoods, dog bites and beanbag rounds fired from a shotgun were involved. Reporters ruled out deaths from gun and car accidents following police chases.
Inclusion does not always mean excessive violence. In about half of the cases, the medical examiner or coroner concluded that law enforcement caused or contributed to the death.
Fewer than ten deaths were ruled suicides. Such cases were only recorded when officers used significant force, such as a Taser to shock someone so they would stop cutting themselves up.
WHEN: Meetings that took place from January 1, 2012 to December 31, 2021.
WHO: Officers who have contact with the general public, often on patrol.
Most of the cases involved local police officers or sheriff’s deputies, but some officers were state or university police. Deaths involving only jailers, prison guards or federal agents, such as those from the Border Patrol, were excluded. When others assisted police—whether private security guards, civilians, paramedics, firefighters, prison guards, or federal agents—the force they exerted was not counted. The exception was when medical personnel administered sedatives, sometimes at the encouragement of police.
TRUE: Deaths often occurred on site or shortly afterwards in hospital. The most common location for a meeting was in or near the deceased’s home.
Fatalities in prisons or police station cells were excluded unless a person died shortly after violence during an arrest, or the violence continued in prison with the involvement of arresting officers. Deaths in prison were excluded.
To document deaths, reporters from AP and the Howard Centers for Investigative Journalism at Arizona State University and the University of Maryland made about 7,000 requests to public agencies in all 50 states and the District of Columbia. The resulting repository of documents ran to more than 200,000 pages and contained hundreds of hours of body camera footage. Although some records were released free of charge, the effort to collect them cost more than $39,000 in fees.
In designing the strategies for extracting and interpreting all this data, AP reporters consulted experts in law enforcement, public health, forensic pathology and other fields.
Reporters first filed requests to obtain data from government agencies that could collect the names of people who died after encounters with police. These requests went to the attorney general or state police as well as to the chief medical examiner or the health department. Where a state had centralized the data, reporters searched tens of thousands of names to identify those who died after less-lethal violence, rather than in shootings, car chases or while behind bars.
Many states do not track these deaths or consider them public information. In several states, reporters went county by county, calling coroners and emailing requests for lists of possible cases.
Reporters identified additional potential cases to review based on tips from sources, news stories, lawsuits and databases that investigators and others have shared to track deaths. No death was recorded based on this information alone.
After obtaining a name, reporters filed records requests with law enforcement agencies, prosecutors and medical examiners, seeking incident reports, autopsy reports, internal investigations and video. They also read court documents and depositions, and in some cases contacted families for data. AP also contacted each law enforcement agency involved in a death with a questionnaire about use-of-force policies and training.
In total, about 270 of the 1,036 cases identified by AP did not appear in any of the three known, public databases that include non-shooting deaths.
The study also looked at coroners and medical examiners, whose opinions about how and why someone died influenced investigators and prosecutors. Based mainly on death certificates or autopsy reports, but also police reports, investigations by local prosecutors, court cases or government databases, reporters collected the official cause and manner of death in 951 cases.
Often, agencies have released documents but blacked out information or blurred videos. Officers routinely used vague language like “struggle” or “taken into custody” in their reports, glossing over more serious violence documented in other evidence.
Laws in some states — notably Alabama, Iowa, Nebraska and Pennsylvania — restrict public access to government records, so reporters could not always determine whether a death met the investigation’s criteria. Elsewhere, data was lost to time, flood or fire. Other requests for documents failed without resolution or were rejected outright. And high costs for records — including an estimate of $10,000 for video of a death in Nevada — also limited access.
Because access to documents varied widely, comparisons risk being misleading. Based on the strict criteria that reporters used, the investigation was able to document only one death in Philadelphia, a city of 1.6 million people — the same number as Philadelphia, Mississippi, with 7,000 residents.
In addition to the 1,036 deaths, there were about 100 others that reporters had reason to believe would be cases based on news reports or lawsuit allegations. These cases are not currently in the database because reporters have not yet independently confirmed the details, sometimes because agencies declined to release information.
AP continues to seek information on deaths that may meet the criteria described above and are not included in the database. Send information about the case and any documents you have to investigative@ap.org.
___
This story is part of an ongoing investigation led by The Associated Press in collaboration with the Howard Center for Investigative Journalism programs and FRONTLINE (PBS). The research includes the interactive story Lethal Restraint, the database and the documentary ‘Documenting Police Use Of Force’, premiering April 30 on PBS.
___
The Associated Press receives support from the Public Welfare Foundation for criminal justice-focused reporting. This story was also supported by Columbia University’s Ira A. Lipman Center for Journalism and Civil and Human Rights in partnership with Arnold Ventures. The AP is solely responsible for all content.
___
Contact AP’s global investigative team at Investigative@ap.org or https://www.ap.org/tips/