Wednesday, September 22, 2010

Crime Data

It is crucial that the data behind crime rates be developed on consistent, transparent benchmarks and data gathering policy be accessible, but, according to this article from the St. Louis Post Dispatch, that may not be the case, and that is troubling, very troubling.

An excerpt.

“ST. LOUIS • Many residents of the Central West End neighborhood awoke Feb. 21 to find a big crime scene: dozens of car windows smashed overnight by thieves.

"When I came outside, the whole doggone neighborhood was out" sweeping up glass bits and filling out forms for a police officer, recalled Dan Thomas, of the 5000 block of Washington Place.

“Police said 52 cars were broken into there and on nearby blocks.

“A few years ago, the city might have counted 52 larcenies in its official statistics. But that day, the police tallied just three larcenies — one per block.

“It may sound fishy to the uninitiated, but that's how the FBI wants local police to count certain property crimes that happen close together — as single incidents. The big cop shop next door, the St. Louis County Police Department, has done it this way for years.

“St. Louis police Chief Dan Isom acknowledged last week that the department has only recently changed its crime counting methodology to adhere to federal guidelines.

“But while the chief has frequently touted a dramatic drop in the city's crime rate, he's never before publicly disclosed any change to the way his department counts crimes.

“Isom's monthly presentations to the Police Board feature crime report graphs that almost always point downward.

“Earlier this month, he told a luncheon crowd of 1,000 business leaders that crime was down 25 percent over three years.

“A local expert on the city's crime trends said the department should have been more upfront.

"When a change is made, the public is owed an explanation," said Rick Rosenfeld, a criminologist at the University of Missouri-St. Louis. "Especially when you begin to count multiple victimizations as one."

“In an interview, Isom said he did not know how much of the apparent drop in crime was really only on paper.

“He said he was "fairly confident" that crime really was in steep decline.

“The change has been an "ongoing process of reviewing our statistics, not one big revelation," he said. "As you find little problems, when you do report that to the public?"

“Isom said the department also realized it had been under-counting assaults by counting them by incident, instead of by victim. A shooting in which five people were targets now turns into five aggravated assaults instead of one, he said.

“The city police have had few outside eyes on recent years' data. For years, they hired a St. Louis University dean to review their crime data but discontinued the audits after 2007.

“The Post-Dispatch has been unable to analyze recent crime trends because, starting in 2008, the department blocked reporters' access to some data that had been made available in previous years.

“Isom noted that the Missouri Highway Patrol, which compiles crime statistics statewide, gave the department a 92 percent compliance rating.

"I'm very confident that what we're doing is appropriate," he said.


“Experts in policing said the questions about St. Louis' crime numbers provide more proof that so-called "Uniform Crime Statistics" aren't very uniform — or meaningful — despite the importance that people place on them.”