Wednesday, November 01, 2006

Stop picking on the poor numbers

One more backhand at the bogus Dangerous Cities tale, then it's off to look for some live horses to beat.

Danger city USA
Get Charlotte out of the top 10 crime cities in the nation
If your house has been burglarized or your neighbor robbed at gunpoint, you don't need statistics to know there's a crime problem here. That unfortunate reality is why most won't be surprised that our city ranks among the 10 most dangerous big cities in America.

Sigh. No, Virginia, that's exactly why you do need statistics. They're what keep you from making dumb generalizations from the proverbial N=1. Should we infer from this that if your neighbors haven't been robbed at gunpoint, there's no crime problem?

Statistics are what help you sample properly, decide what you're going to measure and compare, and draw actual, credible inferences about what a crime problem might look like and where you do or don't have one. Whether you need hoked-up stats from a publicity-hungry "research" company, that's another question.

Pay attention, Mr. Mayor and members of the City Council. This is an important benchmark.

It is not an important benchmark. It's a made-up benchmark. Feel free to ignore it. The most sensible comment you've run about it was the police chief's "pretty meaningless." He was being generous.

Moral: Statistics are your friends. Treat 'em right, and you won't be the mercy of yahoos every time you turn around.

0 Comments:

Post a Comment

<< Home