Close Mobile Menu

How Often Do Cops Kill Citizens? Given “Scandalous” Data Gaps, Nobody Knows

January 23, 2015
by Erik Neumann

Franklin Zimring calls it “scandalous.”

The UC Berkeley law professor—one of the nation’s leading criminal justice experts—is referring to what he discovered when he set out to analyze four decades worth of FBI data on police and citizen killings. Incidents in which citizens killed on-duty police officers had been meticulously recorded. But when police killed citizens? Those incidents were recorded haphazardly, if at all.

In fact, the data was so spotty that he had to resort to finding cases on Wikipedia.

“I knew there would be some problems, but how big the problems are and how much killing by police had become an orphan statistic that nobody accumulated or cared about was really quite striking,” says Zimring, director of the Center for Studies in Criminal Justice, whose analysis with doctoral candidate Brittany Arsiniega appears this month in the Ohio State Journal of Criminal Law.

Nor are the Berkeley academics the only ones frustrated by a dearth of data on police use of deadly force. It has triggered heated debate over the accuracy of the claim—widely circulated on social media—that every 28 hours, a black person is killed by police, security guards or self-appointed vigilantes. It has inspired investigative reporters and even some private citizens to launch various crowd-sourced reporting projects in an effort to document officer killings on their own.

And it has helped prompt Congress to approve a long-languishing bill requiring better reporting, which President Obama signed into law last month. Even so, some researchers say the new Death in Custody Reporting Act doesn’t go far enough, and this week Attorney General Eric Holder called for more stringent reporting of police killings of citizens as well as the killings of cops in the line of duty.

Having accurate data could provide valuable strategies for police reform and training, increasing safety for officers and citizens alike.

Concern about police use of deadly force has heightened nationwide after the officer killings of unarmed citizens such as Michael Brown in Ferguson, Missouri; Eric Garner in New York City; and Cleveland’s Tamir Rice, a 12-year-old who was carrying a toy pistol.

The currently available data, spotty as it is, comes from the FBI’s Uniform Crime Reporting Program (UCR), the national repository for crime statistics, which has been in existence since 1930. The previous requirement to collect stats on police killings directed city police departments, county sheriffs offices and other agencies to report such incidents to the FBI.

But that law expired in 2006. Reporting became voluntary—and many agencies stopped doing it.

The information that does exist in the UCR is problematic too. All incidents recorded in the police use of deadly force category are referred to as “justifiable homicides,” further defined as “the killing of a felon by a officer in the line of duty,” or “the killing of a felon during the commission of a felony, by a private citizen.” But such language simplifies complicated events. An example: A citizen is pulled over by an officer for a traffic stop. Both parties are nervous. The driver reaches for his wallet and the officer shoots him. “He’s not a felon,” Zimring says.

Or consider Tamir Rice; 12-year-olds are unlikely to be felons.

The Berkeley researchers concluded that between 400 and 500 times a year, an officer kills a citizen in what police consider justifiable homicide of a felon—usually under threat from an armed citizen.

But there is no reliable way of knowing what that number would be if included a comprehensive tally of all incidents—including, for example, those not considered justified.

As Zimring and Arsiniega shifted their research from the UCR to Wikipedia, they perused countless news reports from around the country to document the details of each one: the weapons involved, whether the cause was a robbery or domestic dispute, and the victim’s ethnicity, among other factors. They are still aghast that there should be a 25 percent margin of error in their findings—and that their best data came from error-prone Wikipedia.

The gaps in the data fuel the obsession of Nevada journalist D. Brian Burghart. In early 2014, he was driving home from work when he passed a crime scene cordoned off by multiple police cars. “I knew instinctively that a cop had died there or killed someone,” he says. At home on his computer, Burghart tried to find out how often killings by police occurred, but he got nowhere. “I’ve been an investigative reporter for 20 years. I couldn’t believe that.”

“”I’m con­vinced to my core: The lack of such a data­base is in­ten­tion­al,” Burghart con­tends. “No gov­ern­ment—not the fed­er­al gov­ern­ment, and not the thou­sands of mu­ni­cip­al­it­ies that give their po­lice forces li­cense to use deadly force—wants you to know how many people it kills and why.”

Soon after, Burghart started the website Fatal Encounters, which relies on crowd-sourced news reports and public records to collect accounts of killings. It currently lists about 4,000 accounts of police killings since January of 2000.

Since 2011, three other citizen representatives have started their own crowd-sourced record-keeping on police killings to help fill the current information vacuum. They include: Jim Fisher, a former FBI agent-turned crime writer; Kyle Wagner, a journalist for the Gawker Media-owned website Deadspin; and an anonymous Facebook moderator who uses the page name “Killed By Police.” Most of these projects look the same: long open-source spreadsheets or databases. Citizens are asked to enter incidents and provide links to news accounts, and the entries are later fact-checked by the site’s moderators.

Burghart calls the FBI’s Uniform Crime Reporting Program “window dressing” to make it look like the government is collecting reliable numbers.

At the bottom of the UCR is a paragraph that reads: “The FBI’s Uniform Crime Reporting (UCR) Program serves as the national repository for the collection of crime statistics and its primary objective is to generate reliable information for use in law enforcement administration, operation, and management. Participation by law enforcement agencies in the program is voluntary.” Burghart notes that agencies don’t have extra time and money to afford to do things voluntarily.

“The biggest thing I’ve taken away from this project is something I’ll never be able to prove, but I’m convinced to my core: The lack of such a database is intentional,” he wrote. “No government—not the federal government, and not the thousands of municipalities that give their police forces license to use deadly force—wants you to know how many people it kills and why.”

Representatives from the UCR would not respond directly to questions about their data collection methods, but they did release a statement: “In a given year, more than 18,000 agencies contribute data to the FBI; however because of computer problems, changes in records management systems, personnel shortages, or a number of other reasons, some agencies cannot provide data for publication.”

Many researchers insist that accessing complete data about killings between police and citizens would allow officials to help prevent future conflicts between citizens and officers. It would allow law enforcement to emulate best practices, and it would allow a community to monitor whether its police force has a problem.

“One of the big reasons that things like Ferguson happened was [citizens] were afraid because they feel like they have a target on their back. They’re going to assume the worst if they don’t have information,” he says.

“It’s not a question of whether you care about citizen lives or police lives,” Zimring says, “you care about both.”

Even with their limited data, Zimring and Arsiniega’s study reached a number of findings:

  • Between 1976 and 2012, killings of both police and citizens declined (police officer deaths by 69 percent and citizens by approximately 31 percent).
  • A byproduct of improvements in protecting police against violence: While the ratio of citizens killed by police, to police killed, was about 3 to 1, it now is nearly 8 to 1.
  • Although police killings of whites are drastically higher than police killings of African-Americans, killings of African-Americans are proportionally about three times higher than whites. 
  • Nearly all officer deaths are caused by guns rather than other weapons, which leads the study to observe that there are “reasons to suggest that police use of deadly force is not (necessary) when citizens brandish knives, blunt instruments or use personal force.”

The best step forward, Zimring contends, would be the creation of a rigorous, audited reporting system that requires police departments to document incidents of officers using deadly force.

The new Death in Custody Reporting Act requires states and federal law enforcement agencies to report deaths of any person by police who is “detained, under arrest, or is in the process of being arrested,” or “en route to be incarcerated.” It mandates that this information be reported with a “description of the circumstances surrounding the death” to the Department of Justice, the agency that oversees the FBI. Included in the law is the ability of the DOJ to enact financial penalties on states that don’t comply with record keeping.

The FBI is also reevaluating how its Unified Crime Reporting Program collects data. An FBI representative wrote, “The FBI is reviewing its process for compiling data on justifiable homicides by law enforcement.” The agency plans to issue a report to the UCR advisory board in two months.

Still, Zimring is skeptical about the efficacy of these new plans, noting the complexities of this issue that have led to bad reporting over the years. The new law doesn’t solve the current problem of tasking states with reporting on what is actually a municipal issue: deaths occurring in local police departments. Likewise, he says, police departments have a vested interest in looking good, so there’s little incentive for them to report bad behavior.

Ultimately, he says, it’s vital to create a way to audit police departments collecting this data.  “We’re gong to have to very carefully craft a layered response to these problems. And I think it’s going to start happening. I don’t think the issue is going to go away.”

Share this article