Police admit data errors on shootings Department blames computer system for overstating incidents

'A level of inaccuracy'

Commissioner says audit will support claim of sharp drop

May 18, 1998|By Gerard Shields and Peter Hermann | Gerard Shields and Peter Hermann,SUN STAFF

Baltimore police acknowledge errors in calculating city shootings statistics, a finding that could undermine Police Commissioner Thomas C. Frazier's claim that shootings have dropped nearly 60 percent since his arrival.

Police blame their miscalculations on a 20-year-old computer system that they said mistakenly counted 1993 incidents

involving guns as shootings. The department acknowledged the errors after The Sun found discrepancies during a review of police statistics on shootings.

A police audit found that the department over-reported the number of shootings for two months in 1993 by 25 percent. The significance of the finding is that the department used those figures as a base to compare subsequent years.

But department administrators caution critics from concluding that shootings for all of 1993 were overstated until an audit can be completed. Frazier said he is confident that the audit will back his announcement that shootings dropped nearly 60 percent from 1993 to 1997.

"We code 600,000 reports a year and there will be data entry errors in that volume of reporting," Frazier said. "You take two months, multiply it by six and call me a liar and that is a flawed methodology."

But the department's concession of significant errors for the two months studied promises to fuel speculation over whether police administrators knew they were using faulty figures when announcing the decline in shootings to the public last year.

The culprit, the department said, is a computer system called the Management Information System, or MIS. It was established 20 years ago to organize annual city crime statistics for the FBI.

The FBI doesn't require cities to report shootings as a separate category. The bureau includes shootings in aggravated assaults. But many cities, including Baltimore, have begun tracking shootings as a way to gauge violence.

The problem, Baltimore police statisticians say, is that the MIS system wasn't designed to isolate shooting reports. In 1996, the department developed a reporting mechanism that retrieved shooting statistics more accurately at the district level.

The initial two-month police review shows that the department compared more accurate 1996 figures with faulty 1993 calculations that overstated the number of shootings, raising the disparity to nearly 60 percent.

"Anyone who deals with this MIS information accepts the fact that there is going to be a level of inaccuracy," said retired Lt. John Tewey, who collected 1996 shooting data as commander of the department's Violent Crimes Task Force.

Numbers a guide

"You use it as a guide," Tewey said of the MIS figures. "I don't think anybody is willing to go to the bank with it."

Questions over the reported decline in shootings surfaced last month after City Councilman Martin O'Malley accused Frazier of perpetuating a "massive hoax" on the city by fabricating police statistics to make Baltimore appear safer.

O'Malley compared police shooting reports for November 1993 and November 1996. He alleged that shootings were over-reported by 30 percent under Frazier's predecessor, Edward V. Woods. The department then intentionally under-reported shootings in 1996 by a similar 30 percent to show the 60 percent decline, O'Malley said.

O'Malley, a former prosecutor who chairs the council's Legislative Oversight Committee, has asked for additional shooting reports to study September 1993 and September 1996.

Figures reviewed

The Sun reviewed police shooting reports from November 1993 that showed the department over-counted shootings by 18 percent. The department's review for September and November shows that police over-reported shootings by 25 percent.

The department said its finding translates into a decline of 34 percent, about 25 percentage points lower than what Frazier proclaimed.

Police warn, however, that further comparison of shooting reports for the two years will likely result in the discovery of shootings that were mistakenly left off the MIS tally, errors that will make up for over-reporting.

Frazier points to city homicide statistics to show one cannot draw conclusions from two months of data. In May 1997, the city had 36 homicides. Multiplying that figure by 12 results in a yearly total of 432. Yet the number of homicides in the city last year was 310.

"The statistical logic is faulty," Frazier said.

No intent to deceive

During the past three years, police chiefs across the nation have been under increased pressure to show a drop in crime because of reductions made in big cities such as New York, New Orleans and Los Angeles. Two weeks ago, the Boca Raton, Fla., police chief resigned after investigators found that his department altered 400 crime reports in 1997 to make the city appear more safe.

Despite acknowledging that shootings statistics were over-reported for two months in 1993, Baltimore police deny intentionally misleading the public about city safety.

Baltimore Sun Articles
|
|
|
Please note the green-lined linked article text has been applied commercially without any involvement from our newsroom editors, reporters or any other editorial staff.