Thursday, May 4, 2006

Racial profiling and junk statistics

Here we go again with the phony racial profiling stories.

In the Daily Herald today, there's a top-of-the-front-page story on police stats released by DuPage County police chiefs showing that there is a slight variance is some communities between the number of traffic stops and racial composition.

Therefore, the suggestion is -- racial profiling!

Yet any first year college statistics student can tell you these stats -- and the hundreds of others collected under a bad state law -- are as significant as one of those polls on the 10 o'clock news.

The Manhattan Institute's Heather Mac Donald, who should have won about four Pulitzer Prizes by now but never will because she devastates liberal orthodoxy, has proven the folly of these statistical studies. Her groundbreaking, "The Myth of Racial Profiling" should be required reading for any lawmaker even thinking about passing racial profiling legislation based on junk statistics.

In short, Mac Donald shows that without including a variety of factors such as driving patterns, whether the stops were at night or day, differing crime statistics among racial groups, etc., the stats are utterly meaningless. Here's what she wrote in a recent op-ed in Massachusetts.

A reliable model of police discretion requires many more variables; the Northeastern team does not even hint at them. To now order Massachusetts officers to collect racial data, without developing a valid benchmark for that data, is senseless -- even more so given how minute the disparities measured by the recent study are. Cops will waste countless hours filling out forms that no one knows how to analyze, and they may think twice before stopping minority violators, lest they be accused of racism.

And she posits that these meaningless studies are not harmless. By the collection of the data and resulting distortive publicity, police officers eventually stop fewer minorities and some criminals who would have been deterred are not. Since minorities are disproportionately the victims of crime, ironically, these phony statistical studies are harming minorities most, Mac Donald says.

We have U.S. Senator Barack Obama and a bunch of state legislators to blame for this mess. While he was state senator, Obama sponsored this feel-good and do-bad legislation. It called for three years of data collection that will be released in a statewide report July 1. Get ready for a barrage of faulty analysis.

No comments:

Post a Comment