Kajieme Powell’s shooting death last week by St. Louis police, which was caught on video, raised the same questions as those following Michael Brown’s death in nearby Ferguson just weeks earlier: How often do police use force? How often is it excessive force? And how often does that excessive force result in a civilian's death?
Short answer: It’s impossible to know.
While a handful of groups, from the libertarian Cato Institute in D.C. to an app startup run by teenagers in Decatur, Georgia, track incidents of police misconduct, no federal authority comprehensively and reliably documents the use of force by police officers across the country. As FiveThirtyEight points out, the FBI tracks “justifiable police homicides”—around 400 per year since 2008—but their Supplementary Homicide Report is "beset by systemic problems," not the least being that it relies on voluntary information.
The Justice Department notes that when it comes to use of force, "there is no single, accepted definition among the researchers, analysts, or the police," but cites a definition from the International Association of Chiefs of Police’s National Police Use of Force Database project report from 2001 on the use of force from the previous decade: “the amount of effort required by police to compel compliance by an unwilling subject.” Of course, as that report later states, "the test of whether a particular use of force incident is, or is not, excessive can be determined by the administrative and/or civil outcomes of the incident investigation or complaint.” That's especially problematic given that in most states, as The New Republic's Yishai Schwartz reports, "as long as there is a modicum of evidence and reasonable plausibility in support of a self-defense claim, a court must accept the claim and acquit the accused."
It's no surprise, then, that the IACP study found that of the 8,082 citizen complaints about police use of force that were reported to the project from 1991-2000—out of a total 150,026 incidents—only 750 resulted in the complaint being upheld after investigation.
The available data since that study is equally paltry. The federal Bureau of Justice Statistics' Police-Public Contact Survey, a supplement to the National Crime Victimization Survey, is distributed only once every three years to a random sampling of the population either by phone or in person (the last one, in 2011, was read to 62,280 people, of whom 49,246 responded). A cumulative report looking at 2002, 2005, and 2008 found that there was no statistical change in the percentage of police-public contacts that resulted in use or threat of force: It hovered around 1.5 percent. Then, in 2011, it rose to 2.5 percent.
But that doesn’t necessarily mean that the use or threat of force went up because the survey was redesigned in 2011. According to BJS statistician Lynn Langton, 15 percent of those surveyed received the old design, and only their responses were used to calculate that police contact escalated to threat or use of force 2.5 percent of the time. Langton wrote in an email that the "relatively large standard error" caused by the smaller sample size meant it could range by +/-1.2 percent. The redesign has also delayed until 2015 the data collection scheduled for this year.
In 2002 and 2005, the survey didn’t include any questions about being shot by an officer. The only way to report it was when the questioner asked, “Did the police officer(s) … Use or threaten to use any other force against you?” That question comes early in the survey, and it’s possible that at that point, the respondent wouldn’t necessarily realize this was the only opportunity to volunteer that information. On the 2008 and 2011 surveys, the only opportunity to report being shot by an officer was when asked about any injuries caused by interacting with the police. Being shot still wasn’t listed as a type of force threatened or used. Then again, some of the people who would answer "yes" to that question are dead.