Supplemental: Upbeat statistics/horrible logic!

MONDAY, JANUARY 4, 2016

The way our discourse works: A letter in today's New York Times grabbed our attention with some upbeat statistics.

The letter came from Wllliam Bratton, New York City's police commissioner. Bratton was responding to claims last week by Raymond Kelly, his predecessor.

We don't know if Bratton's letter settles Kelly's claims. That said, we were struck by the statistics shown below, which appear near the end of his letter.

Just to be clear, Bratton is discussing shooting incidents by citizens, not by police officers. He describes an extremely large drop in such events in the past twenty-five years:

BRATTON (1/5/16): Shootings are not a Uniform Crime Report reporting category, but a subset of aggravated assaults. The N.Y.P.D chose to track shootings separately from aggravated assaults when I was police commissioner the first time in 1994. We did this because the city had a severe shooting problem at the time.

To establish a baseline, we counted shootings in 1993. There were more than 5,200 that year, or about 100 a week, as opposed to about 1,130 in 2015. To make the 1993 count, and to count going forward, we established the N.Y.P.D. definition of a shooting, which has been in use ever since, including during all of Raymond W. Kelly’s years as police commissioner.

We used the shooting data to focus on shootings as never before, and by 1998, the N.Y.P.D. had pushed down shootings by more than 3,500 incidents a year.

To state the obvious, Bratton is describing a very large drop in shooting incidents. He seems to attribute the drop to good police work, without considering other possible factors.

As a general matter, Bratton's statistics correspond to the drop in murders in New York City. According to the leading authority, that number has dropped from 2,245 murders in 1990 to 328 in 2014, the lowest number dating back through 1960. (According to the leading authority, no numbers are available from 1937 through 1959.)

We're always amazed by statistics like these, especially by the limited role they seem to play in our national discourse. Given what is being counted, 328 remains a large number. But the nation's large drops in crime statistics seems to gain amazingly little purchase, as we noted in this recent post about Chicago's homicide rate.

That's right—Chicago! In 2014, Chicago had its fewest homicides since 1965. No wonder we call it Chi-raq!

Relatively speaking, Bratton's statistics are remarkably good. Today, they appear on the Times letters page. Such statistics seem to play little role in our front-page journalism. Almost surely, most Americans have no idea that these drops in crime have occurred.

Encouraging statistics may play little role in our discourse, but there's always plenty of bad logic around. That's especially true when the horrible logic supports a tribal narrative.

In this morning's award-winning post, we said we'd show you more bad work from last week's Washington Post. We've decided to postpone that service for a few days. It's too depressing to beat up on the tribe in separate-admission doubleheaders.

That said, we were thinking of an argument about the shooting of Tamir Rice, a presentation which first appeared in a piece by Wonkblog's Christopher Ingraham. For those who want to read ahead, you can just click here.

Elementary logic heads for the moon when preferred narrative comes into play. Needless to say, James Downie borrowed Ingraham's logic then added something resembling a bungle in this subsequent column, which appeared in the hard-copy Post.

We'll return to those analysis pieces in a few days. In the meantime, we're always struck by statistics like Bratton's—by how little dust they raise.