It's hard to reiterate this point, which you've no doubt heard numerous times, often enough: There is no good data kept on officer-involved shootings. And, therefore, it's hard to know how often officers that kill people face criminal charges for doing so, in the line of duty or not. And, therefore, it's hard to know what to expect from the grand jury considering charges against police officer Darren Wilson. (An announcement on the grand jury's decision is expected tonight.)

Two weeks ago the FBI released one of the few sets of data made publicly available, which outlines the number of justifiable homicide incidents each year involving a law enforcement officer. Over at Wonkblog, Chris Ingraham contrasted that data -- 461 justifiable homicides by officers -- with the number of law enforcement officers that were killed in the line of duty: 27. The former is a recent high; the latter a recent low.

But that 461 figure doesn't tell us much. It's data from a self-selected group of law enforcement agencies, only including the sorts of deaths that don't result in any criminal charges (hence: "justifiable").

Trying to figure out how many people are killed by police each year, justified or not, often falls to local media outlets, as our Wesley Lowery noted earlier this year. There's a project from the Reno News & Review's Brian Burghart called Fatal Encounters that's built national maps of incidents it has uncovered. Over the weekend, the Salt Lake Tribune reported that killings by officers in Utah over the past five years has topped the number from gang members, child abuse or drug dealers.

That doesn't tell us much either. Is Utah representative? Is the Fatal Encounters map? More importantly, what happens in the aftermath of the killing? The Tribune addressed that question:

Nearly all of the fatal shootings by police have been deemed by county prosecutors to be justified. Only one — the 2012 shooting of Danielle Willard by West Valley City police — was deemed unjustified, and the subsequent criminal charge was thrown out last month by a judge.

Broad analysis of the aftermath of killings by police officers is necessarily spotty, given how spotty the data itself is. There has been some analysis of the question that's worth considering.

The "Police Integrity Lost" project from researchers at Bowling Green State University looked at a number of incidents from 2005 to 2011 involving police misconduct. (The project is funded, in part, by the Department of Justice.) Over that time period, the project uncovered 6,724 arrests of sworn law enforcement officers. The rate of arrest nationally, based on their data, was 0.72 officers arrested per 1,000 officers.

Of the number of arrests, 664 arrests involved a gun. (Nearly all of the justifiable homicides tracked by the FBI were caused by firearms.) Those 664 arrests involved 518 officers (some officers had multiple charges). Eighty-one of the charges were for murder or manslaughter. Conviction data was hard to come by, but at least two-thirds of those gun arrests for which data was available resulted in charges. Which is about as much of an answer as we can get.

It's a reminder that, even if we knew how many officers were arrested for their roles in killings, that doesn't tell us how many faced criminal charges.

There is some indication of how common convictions are in incidents involving police officer misbehavior. In 2010, the Cato Institute published statistical data from its police misconduct project, as the Post's Roberto Ferdman pointed out in August. That data, apparently the organization's most recently available, tracked far more incidents.

There were 4,861 incidents of misconduct tracked by Cato in 2010, including 1,575 that involved the use of excessive force. Those 4,800-plus incidents included 247 fatalities (that year alone). It's not clear how many of those charges resulted in criminal charges.

More interestingly, Cato also looked at misconduct incidents involving 11,000 officers between 2009 and 2010, finding that criminal charges resulted in about 3,200 cases, and just over 1,000 convictions stemmed from those charges. The number of officers sentenced to serve time for convictions was under 400 -- about 36 percent of those convicted, which is substantially lower than the incarceration for the broader public. Those incarcerations, we'll note, were also just over 3 percent of all the misconduct incidents.

From Cato. Compares rates on misconduct by law enforcement with rates from the Bureau of Justice Statistics for the broader population.
From Cato. Compares rates on misconduct by law enforcement with rates from the Bureau of Justice Statistics for the broader population.

How often do police officers who use deadly force against civilians face criminal charges for doing so? No idea. The best we have is guesses, none of them very good. As Ferguson has showed, even knowing that hard number would provide only some solace. For the families of those killed in incidents that seem questionable, only one data point is important.