Will the Press Force the Government to Finally Count the Number of Police Killings in America?

Media organizations like the "Washington Post" and the "Guardian" are trying to answer what should be a simple question: How many people are killed by the cops?

by Mike Pearl
Jun 3 2015, 12:00pm

Ferguson police protests. Photo by Flickr user Jamelle Bouie

On Saturday, the Washington Post published an article announcing that, as of Friday, American police had killed 385 people so far this year. On Monday, the Guardian published a similar feature titled "The Counted," which also included police killings that didn't involve firearms for a grand total of 467.

Whatever the number, it definitely seems like American cops have been killing people, many of them black men, at an alarmingly rapid rate. Eric Garner. Michael Brown. Tamir Rice. Walter Scott. The names are familiar to most of us by now—but hard data on police killings has always been hard to come by, because, incredibly, no government agency keeps track of those numbers.

On Tuesday morning, as if directly responding to the avalanche of reporting from major newspapers, Democratic senators Barbara Boxer and Cory Booker announced a proposal that would force all law enforcement agencies in the country to report every police killing to the Department of Justice. The senators also want info on race, age, and gender to be gathered, among other details.

The only question, of course, is why they aren't doing so already.

In an interview with VICE, Jon Swaine, the Guardian reporter who headed up that paper's project, applauded the Boxer-Booker proposal as a way to bring about transparency. "It doesn't seem likely to happen on its own," he said. Legislation does seem to be necessary, so we think it's a good step."

"We realized pretty early on that no one had collected the extent of the data we were trying to collect and that surprised [us] a bit," Steven Rich, database editor for investigations at the Washington Post, added in an email. "The biggest issue is that there's no simple way to do this."

The latest from VICE News: Video Released of Altercation Before Police Killing of Preacher Stuck in Flood Water

The two data sets have been mined for countless revelations already. More than 80 percent of the shooting victims, by the Post's count, had some kind of weapon. But put another way, 49 of the people shot to death this year by cops were completely unarmed (another 13, like Tamir Rice, were holding toys.) African Americans, the Post found, were killed at least three times as often as any other race after accounting for census numbers on population in any given area.

These tragedies will always spark outrage no matter what the data show, but reporting on police killings has historically made it impossible for the press to put them in context. That has been true whether claims are coming from protesters or a police department.

The existence of racial bias has been obvious, but without numbers, demonstrating its scale proved an incredible challenge.

Now we have those numbers, at least for this most violent of years. (The Post reported that the 385 people shot in the first five months of 2015—more than two per day—is more than double the rate reported by the feds over the previous decade.)

The Guardian's "The Counted," with its adjustable, filtered viewing modes, is an incredible resource as well. For instance, their numbers show that Oklahoma law enforcement has killed the most people per capita—22 in a state with a population under 4 million. We can also see that at least one person this year has been killed by cops in every state except North and South Dakota, Vermont, and Rhode Island. In addition to geography, the deaths can be arranged by factors like age, race, and cause of death.

Photo courtesy of the Guardian

Frequently, the Guardian offers a photo of each victim, arranged into a kind of grim calendar. The Washington Post, on the other hand, personalized their numbers by including the disjointed narratives of a few dozen incidents, superimposing events like the tragic final moments of a patient with mental illness next to stories of bona fide police heroism.

Since the US Department of Justice doesn't keep track of police killings, it relies on law enforcement agencies to report their own numbers. Before he stepped down, Attorney General Eric Holder told the press in January that this was "unacceptable." He complained that the Justice Department doesn't have "the ability right now to comprehensively track the number of incidents of either uses of force directed at police officers or uses of force by police."

Instead, we've had the FBI's Uniform Crime Reporting (UCR) program, which offers a small amount of help to data-hungry journalists. Their most useful publication has been the UCR's justifiable homicide numbers, which, again, are self-reported by cops around the country. These tend to be way behind—currently only counting up to 2013. To make matters worse, they include only those homicides deemed appropriate within the scope of the law, meaning a huge, and previously unknown, percentage of the picture has been left out. (And as the Wall Street Journal reported last year, even when it comes to so-called justifiable killings, police departments leave a healthy number out of their federal reports.)

Consequently, the FBI data revealed an average of about 400 deaths per year over the previous decade, according to the Post's calculations. Somewhat alarmingly, these new numbers reveal that there were around that many just in the first half of 2015. Part of the discrepancy stems from FBI's focus on "justified" deaths, but how big a part that plays is tricky to decipher.

"The trends that it purports to find and show are just completely meaningless," Swaine said of the UCR. "We don't know whether it's actually gone down when the figure has gone up, or vice versa."

For instance, the death of Walter Scott in April, an incident in which the officer shot the unarmed man repeatedly in the back, is unlikely to ever show up on a list of justifiable homicides given that the officer who pulled the trigger is being charged with murder. Thus, it most likely won't appear on the FBI list.

"The rate of participation [by local police with the UCR] has been falling steadily for years," said Jeffrey Fagan, a law professor at Columbia University in New York, before suggesting that the entire system of self-reporting is flawed.

"What are the incentives? They might be political," Fagan said. "A local police executive who wants the information to be publicized maybe? I don't know the incentive structures."

When reached for comment, the FBI pointed VICE to comments by Director James B. Comey who, like Holder, has publicly criticized the lack of data. He recently said that during the Ferguson protests last year, he asked his staff for a report on the number of police killings, but "they couldn't give it to me, and it wasn't their fault."

FBI spokesman Stephen G. Fischer wrote in a follow-up email, "We actually have several initiatives underway to improve the collection of crime data. One is to encourage agencies to submit data to the National Incident Based Reporting System (NIBRS). Another is a proposal before our Advisory Policy Board to expand data collection efforts to include nonfatal shootings by law enforcement in the line of duty.

"But remember," Fischer added, "provision of any data to the UCR program is done on a voluntary basis."

So two major newspapers are now offering up the data analysis the federal government wasn't able—or willing—to generate on its own. The figures highlight the incredible discrepancy between the number of deaths that have been going unreported in the UCR's justifiable homicide report and the tiny number of cops who face criminal prosecutions. The Post notes that less than 1 percent of the 385 fatal shootings in 2015 have resulted in criminal charges for the officers.

That's a total of three.

The Guardian and Washington Post are far from the first to attempt such analysis, however.

"It should be noted that various groups have attempted to track aspects of this in the past and at present," Rich, the Post's database editor told me, without naming names. "They collect varying levels of data, and do various levels of fact-checking." The organizations Fatal Encounters and Killed by Police are credited by both the Guardian and the Washington Post as having taken a crack at this mess. Deadspin launched an effort of its own last year, citing the work of blogger Jim Fisher. There's also a list of law enforcement killings compiled by Wikipedians.

Swaine differed with some of the methodology of Killed by Police in particular. They included car accidents, for instance, which the Guardian ruled out. "If you're stopped for a traffic stop, drive off to escape, and crash into a wall, we don't think that's really fair to include," he said.

But despite the obvious power of these reports, they aren't exactly panaceas. The highly controversial death of Baltimore's Freddie Gray, for example, was left off the Washington Post's list because he wasn't shot.

"It's immensely difficult to define what a 'police killing' is," according to Rich. "With guns it's obvious. With Tasers it can be, too. But it's very hard to make determinations on things like in-custody deaths without a coroner ruling the incident as a homicide. We thought we could collect the most data accurately if we didn't have to get into making determinations." (It's worth noting that the Baltimore medical examiner did, in fact, rule that Gray's death was a homicide right before local prosecutor Marilyn Mosby charged six cops over his death.)

Swaine said the Guardian's list includes any case where it "was clear that police actions had killed this person," which excludes the car accident mentioned above. "Self-inflicted deaths" in the presence of police don't count in the Guardian's list, nor apparently do they count in the Washington Post's version, but this is one of the trickiest areas of all.

"There are at least a few out there where it's unclear if police fired or if the citizen shot him or herself. We left those out unless there was complete certainty that a police officer fired the shot," Rich said.

The Guardian has a similar policy. For instance, the death of Bruce Lee Steward in Colton, Oregon, was officially ruled a suicide by cop, and Swaine pointed out that it "might have been missed or dismissed by other counts." However, he said, "the police shot him, so we decided to put him in our database."

In the end, both lists come with disclaimers about not being comprehensive. "Anyone claiming to be definitive or completely comprehensive is either not telling the truth or has effectively FOIA'd [Freedom of Information Act requested] 18,000 state and local police departments for the data," Rich said.

For his part, Fagan, the policing expert at Columbia, was less than euphoric about the new data sets, even if he sees value in them.

"They've got lots of fancy graphics," Fagan told VICE of the two newspapers' features, before later adding, "They both duplicate the Wikipedia page, and the Fatal Encounters website. Fatal Encounters is really interesting because it gives you a spreadsheet."

Fagan conceded that the Guardian's count was interesting for the simple reason that it includes non-gun deaths, but actually suspects that the Post's shootings-only figures might be more reliable.

"Inherently, the one with non-firearm deaths is [useful]," he said. "But things can happen in non-gun homicides. I'd be more confident from a social-science perspective in the shootings-only database."

Either way, this kind of rigorous numbers-gathering has undeniable long-term potential. The LA Times, for instance, has been keeping careful data on Los Angeles-area police killings as part of their ongoing "Homicide Report" series. In November, an activist scribbled some statistics on a sidewalk saying, "LAPD killed 1 person per week since 2000. 82 percent were black or brown." The folks at Homicide Report were able to check that intel against the cold hard facts, and found that the graffiti was inaccurate, if not absurdly so.

But it takes time and money to keep something like this going. Homicide Report has been in operation since 2007 and undergone vast transformations over that span. These new reports, which will require constant updates from every corner of the US, seem even tougher to maintain, and we already know the Washington Post's list isn't long for this world.

"We intend to continue collecting data through the end of the year," Rich told me.

Maybe by 2016 the federal government will have taken over the job. But even if they haven't, at least there's a framework being laid down by reputable sources for what it means to keep tabs on police violence in America. It's left to the rest of us to figure out what to do about it.

Follow Mike Pearl on Twitter.

Big data
The Guardian
Vice Blog
police shootings
eric holder
being shot
The Washington Post
Cops shooting people
Small Data
Data on police killing people
killing citizens
Steven Rich
how many people are getting killed by police
cop shootings
getting shot