Ilook closely at investigative stories written by The Post. Investigation is a key competency of The Post and should remain so, not just because of the Watergate legacy but because this is one of the prime functions of newspapers — holding government and institutions accountable. I hear this from readers all the time; they want fact-based, hard-hitting investigations that keep people honest.

One recent investigative story was Cheryl W. Thompson’s front-page article on Feb. 19, “The trick to D.C.’s homicide closure rate.” It had no factual inaccuracies — all the statistics in it are true — but in its language and tone, it seemed to tell a story more of gotcha than of scandal. I don’t think D.C. Police Chief Cathy L. Lanier used trickery in her reporting of statistics to mislead the public about the solving of homicide cases.

Rather than suggesting that Lanier was fudging numbers, I think the story would have worked far better as a straightforward explanation of how the Metropolitan Police Department, other major police departments and the FBI keep homicide statistics — and of some of the pitfalls in that method.

The story led to an unusual editor’s note from Executive Editor Marcus Brauchli that appeared on Page A2on Feb. 26 and is now appended to the online version of the story. Lanier’s op-edrebutting the story also appeared in the Feb. 26 newspaper, on the Local Opinions page in the Metro section.

Lanier, as you might imagine, was hopping mad after the original story; I know because she and her senior aides met with me after meeting with top editors. They made several valid complaints, but I’m going to stick to the statistics for this column because they go to the heart of this story.

The problem with homicide statistics is that there isn’t a perfect way to express them.

Let’s say there are 100 homicides in 2011 and the police make arrests in 60 of them during the calendar year. That’s a closure rate of 60 percent. But let’s say 10 of those killings occurred in December, and there were no arrests in those cases until January and February 2012. You might think that the police would go back and revise the 2011 closure rate to 70 percent.

But that’s not how the FBI or major police departments around the country calculate it. Closing those 10 cases from 2011 early in 2012 puts them in the 2012 totals for closed homicides. For that matter, a homicide from 1989 that is solved in 2012 is also put in this year’s closure rate.

Now, let’s say that, in 2012, another 100 homicides are committed, and the police again make arrests in 60 of those cases during the year. They get to add to that latter number the 10 cases from late 2011 that were solved in early 2012, resulting in a closure rate of 70 percent.

And if the number of homicides is steadily declining over time and the number of solved cases is going up from prior years, the closure rate goes up.

This is precisely what is happening in the District.

In fact, one of Lanier’s priorities since she became police chief in 2007 has been the solving of cold cases — those older than three years. This is why, for 2011, the homicide closure rate hit 94 percent: There were 108 homicides, and the closure rate included about 60 solved cases from 2011 — the “in-year rate” — plus nearly 40 others from prior years.

In a sense, then, the official homicide closure rate is more of a rolling rate. Lanier usually puts the in-year statistic in her annual reports as well, but as more of an asterisk. The Post in its story called the in-year rate the “true” rate.

I guess you could say that, but, as Lanier points out, police should be given credit and incentive for going back and solving cold cases. The families of the murder victims certainly think so, as Lanier told me.

I think the MPD should be very clear in its annual reports about the rolling rate and the in-year rate; that way, no one can suggest any manipulation. And reporters will probably need to explain the two rates to readers every year. But is that so onerous? I don’t think so.

Patrick B. Pexton can be reached at 202-334-7582 or at updates, read the omblog at