Unlike Walter Scott’s
Horrific Killing,
Deaths By Police
Are Rarely Recorded
— They’re Not Even
Counted by the
Government
It’s not easy to watch the video of Walter Scott running
from officer Michael T. Slager, as Slager pumps not one,
not two, but eight shots into his back. It’s an unspeakable
horror. But the existence of that video also means that
Scott may be one of the few victims of police brutality
whose attacker could face justice. Slager has been fired,
and charged with murder.
with police, including killings, are not recorded on video.
In fact, they are not even tracked by the government. It
seems incomprehensible that with all the data that is
collected about Americans, there is no official count of
the number of people killed by the nation’s law enforcement
officers—let alone the reasons they were killed. D. Brian
Burghart, the editor and publisher of the alternative weekly,
Reno News & Review, is hard at work trying to rectify that.
His FatalEncounters.org is a project that uses crowd-sourcing
and media accounts to create a national database of people
who are killed through interactions with police.AlterNet’s executive editor, Don Hazen, interviewed Burghart via email about the shocking necessity for this project and what Burghart hopes to accomplish.Don Hazen: Why did you start Fatal Encounters?
D. Brian Burghart: I started the project because there is no adequate database by which citizens, researchers or law enforcement can track trends in officer-involved homicides across regions or time. Frankly, our government knows who I call on my cell phone, who I email, and probably even the contents of those calls and emails. The fact that it hasn’t kept track of the names of the people it kills and the circumstances under which it killed them is beyond my comprehension.
DH: Why is there no national database of homicides caused by law enforcement?
DBB: I don’t know the answer to that question, although I’ve thought about it for years. The 1994 Violent Crime Control and Law Enforcement Act mandated “the Attorney General shall, through appropriate means, acquire data about the use of excessive force by law enforcement officers,” and to publish an annual report, but it’s been ignored for 20 years and across administrations. In December of last year, Congress passed, and Obama signed the Death in Custody Reporting Act. It wasn’t very effective when it was last authorized 2002-2006, but maybe the political will to make it work has changed. We’ll see.
I’m kind of a cynic, but I think a lot of people will hope to sweep under the rug the discussion this country has been having about race when there’s a new president and a new attorney general.
DH: How many sources do you monitor? What can you share about your methodology?
DBB: We really only “monitor” the Internet. When we find a big source, for example, if a newspaper has done its city’s officer-involved homicides for the last decade, we’ll incorporate that. For a few states, there were central locations, like attorneys general websites, for the results of officer-involved homicide investigations, and we got those in. Sometimes reporters will send me their datasets. There are groups like Copwatch that keep an eye on this kind of stuff, but it’s pretty spotty. We get many of our current day-to-day links from KilledbyPolice on Facebook, which tracks officer-involved homicides on a daily basis.
Our general procedure goes like this: We have our development queue. This is just partial information we scraped off the web, including sources like the FBI. Anyone who wants to volunteer a few minutes can research an incident, and send it to fact-checking, using this web form.
From there, it goes to the main spreadsheet, where the new records get scraped off once a day or so and moved to the fact-check queue (the bottom spreadsheet) where I or another editor checks them against published accounts. These records are then moved back onto the main spreadsheet, given a unique identifier and uploaded into the database.
Sometimes people follow this procedure; sometimes they upload new incidents that aren’t in the development queue. If I receive a big dataset, I will build a separate spreadsheet and ask a volunteer to focus on it. We’ve done more than 2,000 public records requests so far.
DH: How many people have submitted new information? How reliable have you found these submissions to be?
DBB: More than 1,000, probably. One reason I can’t say precisely is because we allow people to post anonymously, otherwise, they wouldn’t submit. Many contribute because they’re afraid of the police so they don’t want their names attached. Accuracy, as opposed to reliability, has run the gamut. People aren’t generally professional researchers or journalists, so they’ll often get things wrong. We check against published accounts, but that’s no guarantee of accuracy either. Media bias, reporter experience, lots of things can affect how accurate the information is.
Our ultimate plan is to make public records requests for every law enforcement agency in the U.S., but those are also rife with errors, things like whether an address is a street or an avenue. It’s pretty amazing, really. As the submit form was refined, data got more and more accurate, but crowdsourcing is always going to be problematic. That’s why we have editors.
DH: Have you reached out to government agencies to help get data for Fatal Encounters? What was their response?
DBB: I’ve done many public records requests at the federal level and in several states. The responses have covered the gamut, from sending the information without comment or delay, to a delayed response to the request in such a way as to make the information useless, to outright refusal to obey public records laws and challenging me to sue for the information to which every American is entitled.
DH: Even if you could correctly identify every police homicide, how would this information help people? How would this data help us police the police?
DBB: I believe that being able to see outcomes over time and across regions will help Americans to identify agencies and policies that get better or worse outcomes in their interactions between police and the rest of us. We’ll be able to see trends like whether people of various races are killed at a rate that’s greater or less than the population. Also, comparing this data to other giant data sources, we’ll be able to research whether veterans are killed at a high rate. We’ll be able to see the effects of poverty on policing and officer-involved homicides. We’ll be able to see which communities kill mentally ill people, and at what rates. We’ll be able to see if communities where police kill more people have more people killing police. With this information, police will be able to look at crime rates and other factors to determine the policies that work in other communities to get the best outcomes for both police and citizens.
Most fundamentally, we’ll know the numbers of officer-involved homicides in our own communities, so we’ll know whether our own communities have particular problems, and whether we as citizens need to change things.
DH: What, in your opinion, would make a homicide by police “justified”?
DBB: Every human being has the right to defend themselves. I’ve looked at thousands of these things now, and most are clearly self-defense. I just don’t think there should be any officer-involved homicides that don’t get a full public scrutiny. These are public employees paid with tax dollars. This is how we are supposed to manage our government personnel in this country.
DH: Looking at your site, it looks like homicides skyrocketed in recent years. Why is that?
DBB: While I believe police homicides are increasing, we don’t have the data to prove that yet. Our data makes it look like they skyrocketed in recent years, but that’s just because we haven’t begun the systematic day-by-day searches beginning in 2000, yet. Part of the apparent increase is because of the growth of the Internet, which means we have access to more information each year progressing from 2000. Part of that is because in the early 2000s, digital memory was expensive, so media outlets routinely purged their archives. Also, because of the crowdsourced aspect, people tend to remember the more recent homicides, so they are more likely to report those.
We have 12 states we believe are “complete”: Montana, South Dakota, Oregon, New York, Nevada, New Hampshire, Rhode Island, Connecticut, Massachusetts, Vermont and Maine. (Florida is within a week or so of completion.) I say “complete” because no matter how complete we think it is, there are homicides that are not reported either in the media or in public records requests, and they’ll crop up from time to time. For example, Las Vegas police reported a homicide to the Wall Street Journal that they did not report to me. I don’t know which it was, because the WSJ just did raw numbers, but we’ll eventually get it.
DH: Of the data you have collected so far, are there any trends you can extrapolate from it? Is there anything that has shocked you? Is there one specific area of the country where police homicides are most prevalent?
DBB: I always hesitate to try to extrapolate except in states for which we have “complete” data. This project, the public face of it anyway, has only been going since March. We estimate we’ve only collected 30 percent of the data. Here’s what can be said, though. People of color are killed at greater rates than they exist in the population. Mentally ill people are a high percentage, maybe 25-30 percent, of the people killed by police, particularly when you consider that drug abuse is considered a mental illness. Most people, around 96 percent, of people killed by police are men.
It sure looks like law enforcement in the western states kill more people per capita than eastern ones, but as Florida progresses, I may have to reassess that statement.
Some other trends that become apparent aren’t even in the data. One thing that becomes obvious is how lazy the mainstream media is. For example, the race of victims and police are often not reported by media. The media gutlessly lets police withhold names of people they’ve killed on the thinnest of rationales. Media rarely get photos of victims from families, so often the only publicly available image is a mug shot, which of course, works to support the narrative that the person killed is a career criminal.
The official dispositions of officer-involved homicides are rarely reported. Again, it seems as though the media just assumes if an officer killed somebody, it’s justified, but that creates a de facto collusion to tell the law enforcement story, but not another side.
DH: What does it say about the state of law enforcement in America that a project like yours is even necessary?
DBB: Honestly, I don’t think it says that much about law enforcement, but more about government. These numbers are tracked in most developed countries. My feeling is, just as with other data, if the government collected this data, law enforcement would use it to modify policies and procedures.
Law enforcement doesn’t like to kill people, and some of my reporting in the Reno News & Review has shown that officers almost always have severe psychological and emotional trauma when they kill somebody in the line of duty.
DH: What do you need from people to take Fatal Encounters further? How can people reading this article help?
DBB: Like everything, it comes down to time and money. The government ignored the need to comprehensively collect this information for a very long time, and there’s a lot of data to collect. We have an army of volunteers out there, helping with data entry, helping with a redesign of the site, doing visualizations, doing research with the data. People can donate money to the 501(c)3 on the website. People can research and submit information here.
DH: Anything else you think readers should know about the database and your work?
DBB: I think the biggest thing that people should know is that we’ve only just begun. Assuming we continue at the rate we’ve gone for the last year, as of April 1, we expect to require 100 more weeks to complete the “media reports” portion of the database. The public records portion is being done concurrently, but it’s much more expensive and time consuming. The more data that goes into the database, the more accurate it is, and the more we can do with it.
It won’t be very long before we can compare what we have to other big datasets, like the U.S. Census data, to see the racial makeup or socio-economic characteristics of areas where people are killed. We’ll be able to do overlays with GPS coordinates to see how every law enforcement agency is performing in this particular area. We’ll be able to check names against Veteran’s Administration roles to see rates at which veterans are killed by police (and I’m pretty sure we’re going to find out it’s much higher than the general population rates).
In less than two years, we’ll be able to see what agencies get the best outcomes using what policies, and when the best policies are in place, we’ll have fewer police killed by criminals and fewer individuals killed by police, which is the whole purpose of this thing.