Announcing In Context: Understanding Police Killings of Unarmed Civilians

Available on May 1, 2016

Pre-order now, at!

Since 2015, the media and the public have paid more attention than ever before to the use of deadly force by American police officers. That’s a great thing. Up until 2015, not many Americans had noticed that the data gathering done by the government on this subject was, to put it mildly, really bad.

In the United States, these incidents are reported on a voluntary basis for many of the important data points that ultimately comprise government statistics. American criminal justice system statistics about deaths related to law enforcement operations (of officers and of suspects/convicts) are compiled mainly by the Federal Bureau of Investigation, a national law enforcement agency with no real power to compel local, county, state and tribal police departments to actually report these numbers, and which provides few if any rigorous standards or definitions around what specifics are reported. Since 2010, The Bureau of Justice Statistics has been trialing an (absolutely excellent) open-source program to capture deaths in custody, but it has not yet been officially unveiledThe utterly excellent Bureau of Justice Statistics (, a component of the Office of Justice Programs in the U.S. Department of Justice, has a mission, "to collect, analyze, publish, and disseminate information on crime, criminal offenders, victims of crime, and the operation of justice systems at all levels of government. These data are critical to federal, state, and local policymakers in combating crime and ensuring that justice is both efficient and evenhanded." BJS gathers its primary data from the U.S. Bureau of the Census, the FBI's Uniform Crime Reporting program, and National Incident-Based Reporting System. The BJS Federal Justice Statistics Program collects data from other Federal agencies, including the Executive Office for U.S. Attorneys, the Administrative Office of the U.S. Courts, the U.S. Sentencing Commission, and the Federal Bureau of Prisons..

Now, this lack of useful data on police use of force hasn’t gone entirely un-noticed – it’s just that those who noticed it didn’t get any attention. They tended to be eggheads or nerds of one sort or another, and because crime was at all-time lows and life was generally pretty safe, there simply wasn’t the kind of driving force required to precipitate change at a macro societal level.

There were some exceptions. Some good reporters, some criminal justice professors and academics, and certain activists did notice this paucity of police use of force data, and they tried to do something about it. These efforts, though, were generally relegated to, “toil in obscurity” status. One such project was begun in 2012 by D. Brian Burghart, a journalist in Reno, Nevada, who began to count every killing of a citizen by police since 2000.

Burghart’s insight came to him serendipitously. He was perusing federal justice statistics and he came across a statistic he found odd and, frankly, unbelievable: according to the FBI statistics, police in Florida had never had an officer-involved justifiable homicide – at least in the years Burghart looked at. Knowing both cops and Floridians, Burghart sensed that this could not possibly be correct. The serendipity was that Florida was the only state that had never participated in this part of the reporting.

“From a journalist's ironic point of view,” Burghart told me, “it was the lack of transparency that caused the exposure. If I had picked almost any other state, I probably just would have said, ‘Oh, that's interesting,’ and continued to believe the stats put out by the FBI and the mainstream media and moved on.”

Instead, Burghart began to Google.

He soon came across many articles about police involved killings during the period in Florida, and realized that the FBI figures were wrong, and badly so. And that he could do something about it in the same way he just proved his point: through open source research and statistical analysis. Thus the Fatal Encounters project was born.

In America, a sea change began in 2012 and 2013. Empowered by the Internet and social media and an ability to gather, analyze and quickly and cheaply disseminate data, new groups began to bring attention to issues around police use of force and abuse of their authority. Cop Block, CopWatch and other groups determined to hold police accountable for their actions (and to highlight what they said was the hypocrisy of cops enforcing laws unfairly, or flouting laws themselves) began to use video and data to prove their point. Police militarization – the use by police of equipment originally made for the military but sold at steep discounts to local police departments – became a related subject of particular concern .

Burghart’s project, Fatal Encounters, became more important in 2014 when some high-profile killings by police of black men became, anecdotally, a “trend.” Newspapers began to dig in. When the QT gas station in Ferguson, MO was set ablaze by protesters angered by the shooting death by Officer Darren Wilson of Michael Brown, Fatal Encounters data ultimately provided fuel to what became a genuine grassroots movement that questioned how often police killed people – and more specifically, how often white police killed black people.

These are crucial questions that we as law enforcement officers and administrators and researchers, at the local, county, state and federal level, have failed to answer in a manner that people can understand, believe and, most important, trust.

By early 2015, two major newspapers – the Guardian and the Washington Post – had begun their own projects to count some of these killings. Each placed upon their projects a slightly different spin. Both the Post and the Guardian based their initial count on Fatal Encounters’ data, but then quickly moved on to their own gathering and reporting systems. Their methodologies varied – the Post stayed focused on police shooting deaths, while the Guardian opened its scope to all kinds of killing by police, regardless of method and regardless, at times, of whether the death actually involved police, per-se. The Guardian statistics, for example, included deaths that occurred in federal prisons, off-duty personal disputes, and even traffic accidents in which a police vehicle was involved. This broad methodology upped the anecdotal total of “police killings” and was included in The Guardian’s reported totals. But The Counted project was broadly popular, and very important culturally. It made more accessible, through beautiful graphical visualization the statistics about police killings in a way that was highly successful in focusing the attention of unprecedented numbers of Americans on the questions about how they are policed.

As 2015 progressed, more database projects sprang up, also based on Fatal Encounters, and run by many groups. They became better at counting the high-level number of people killed by police, but they were leaving out facts, such as why the decedent was killed. There were often no details, or only the ones that served a narrative, but other key data were absent. There was no context. And these flawed or incomplete stories were serving as the basis for statistical analyses that seem ill-conceived.

Anecdotal narratives and bad math isn’t new in this arena – in October of 2014, watchdog journalist ProPublica wrote that young black males stood a 21-times greater chance of being killed by police than young white ones. They based this statistic on a flawed mathematical formula that treated shootings as random events as likely to affect anyone, then doing division to correct for the population difference between black males and white males.

In a similar thesis, the Guardian began to track a “rate” of killings of blacks versus whites, showing that blacks were killed by police at a rate (per million in the US population) that was approximately 4 times higher than whites.

The reporting was missing some important points. As a police officer myself, I was aware that many incidents in which police use force begin with a civilian threatening another person or the officer. As someone who runs a company that provides data analytics to law enforcement, I recognized that the data in the news stories was incomplete.

Let me be clear at the outset; my objective was by no means to defend the officers indiscriminately or to exonerate the guilty. Rather, the media narrative I was seeing was flawed because it wasn’t based on robust data or rigorous process. It wasn’t analysis.

As an illustrative example, consider this story: Filimoni Raiyawa died while battling officers in San Francisco in July, 2015. The Guardian said that Raiyawa had, “been involved in a traffic collision, fought with another at the scene, had run away and then allegedly fought with and injured two officers before becoming unresponsive.” This account may be factually true as far as it goes, but here is what was left out of that highly selective summation.

A deeper investigation into the records and reports surrounding the case would add to this narrative the following; that Raiwaya, who was 6 feet tall and weighed 265 pounds, was the caregiver to 96-year-old Solomon Cohen, whom Raiyawa had just beaten to death half an hour earlier. Soon after, Raiyawa rear-ended a BMW, pushing it across a street and into a parked car. When its driver emerged, Raiyawa began making disturbing statements about God and “God’s will,” and chased the driver around the block. When police responded, Raiyawa beat a female officer several times about the head and shoulders until she was barely conscious on the ground. Raiyawa then threw the officer’s standing partner by the wrist.

Witnesses saw two female police officers trying and failing to subdue Raiyawa. One said, “Nothing was going to get him to stop moving." Raiyawa continued to fight officers and finally collapsed after he tried to break into a restaurant.

This sounds more like a case of tremendous restraint on the part of the officers being beaten who did not use their guns. When Raiyawa died, his death was due to existing health issues aggravated by his decision to strain himself by murdering one person, attempting to murder another, and attacking others. That is a lot of context to leave out and hard to justify as police murder.

In the middle of 2015, I began a project to track killings by police of unarmed civilians. I decided that it should not have any text or prose because any attempt to describe an incident was inherently editorial. I wanted to ask questions to which the answer would be “Yes,” “No,” “Unknown,” or “Not applicable.”

I chose to inquire into police killings of unarmed civilians because I feared that other projects doing so were simply creating body counts with the intent of them recording higher and higher. It was obvious we should be counting. What was not obvious was, if we sought to use the data to affect outcomes, to inform training and policy and debate, what should we be counting?

When we look at what should be counted, my main goal is to differentiate killings by police that are police behaving the way one would expect police reasonably to behave – that is, standing between civilians in danger of losing their lives and those who would take the civilians’ lives from them. In short, when officers behave as they did in the Raiyawa case above.

No one complains when a police officer uses deadly force to end a hijacking in progress or a school shooting or an armed robbery. What no one likes is when it appears that police are using their authority and deadly force abusively, taking the lives from people without justification.

I decided that the only way to tell the difference between these kinds of events is to ask as many questions as you can about how the police got involved, what happened when the police got involved, and what specific actions lead to the use by the police of deadly force against a person who was, presumably, unarmed and therefore not dangerous.

The answers are revealing. They show that more than 70% of the unarmed civilians killed by police in 2015 were in the middle of committing crimes such as robbery, carjacking, assault, or serious destruction of property or burglary. Unarmed people killed by police in 2015 had already injured and assaulted civilians in 26 cases, and murdered other civilians in 2 cases – this despite the fact that the decedent was not armed.

“Unarmed,” it turns out, does not mean “not dangerous.” The Washington Post found in its review of the almost-1000 people shot to death by police in 2015 that, in 74% of cases, the person killed was in the middle of a deadly attack against the officer or another person, and in 16% of cases, the person was making threatening movements or gestures or ignoring police orders to stop. That makes nine in ten. Our numbers generally agree, but they go further.

I am a law enforcement expert, and more specifically, an expert on police data and statistics. I wanted to provide some insights as to trends and other patterns that we recognized in the data that might be useful in asking the next question you ask after you figure out whether you want to count something and what it is that you want to count, which is: What do we do with what we’ve counted? My project set out to answer these kinds of questions. I sought to look back at a cohort with the benefit of lack of political pressure, lack of media sensationalism, and the fact that I am not personally involved in any of these cases.

This book is co-authored by Ben Singleton, a 10-year police detective; and by Ed Flosi, a 27-year police veteran who currently operates his own company that provides use of force and defensive tactics training for law enforcement agencies and provides consultation and expert witness services to defense attorneys and prosecutors in cases where officers were alleged to have used excessive force. The book has been reviewed by several other officers, administrators, statisticians and editors. We wanted our work to be as useful and immediately applicable as possible, because the intent of this book is to make these results, these statistics, actionable. We want to change the debate from one of blame to one of action to review and update policies and procedures.

Body counts make for great headlines. And if you are in the business of selling newspapers, it’s a fact that reporting on the 1,100 people who were killed in 2015 by the police will sell a lot of newspapers and drive significantly online traffic. What it will not do is change anything about how the nearly 18,000 law enforcement agencies in the United States operate or use force.

America is a very violent place, and its police must address violence not on a statistical level, but on an individual, face-to-face, in-person, at-bad-breath distance. While we study numbers on killings, it is crucial to remember that there is nothing theoretical about police work in America.

It is important to consider context. While I am of course not saying that any death is acceptable – I absolutely do not say that – it is important to keep in perspective the numbers of deaths we are talking about that appear to be unjustified. As we will see in this book, the overwhelming majority of cases in which police killed a civilian were not just justified, but they were absolutely necessary for public safety or the preservation of other human life.

And the fact that relatively few unjustified killings occur each year is not a result of police being without fault, or being above reproach, but rather it is that deaths are so thoroughly investigated, and the evidence so widely available and transparent that, simply put, if you are looking for malfeasance in a killing, you’re hardly ever going to find it.

While deaths at the hands of police are very sensational, they are probably the worst place to look for police malfeasance. Later in this book I will discuss where to look for malfeasance – hint: it’s the much larger volume of cases where people don’t die. But, to begin, I would like to give a summary of the 153 incidents during 2015 in which an unarmed civilian lost his or her life after an encounter with an American police offer.

Pre-order now, at!

Get Notified Of New Posts
(We won't ever sell, rent, or lend your email; nor will we ever spam you)

Announcing In Context - March 30, 2016 -