Detroit police officials dispute FBI crime statistics released this week and say they have a more accurate count showing 863 fewer violent crimes were committed last year — but even with the lower figure, Detroit would still be ranked as the nation’s most violent big city.
Police officials say their numbers indicate there was a 5 percent drop in violent crime in Detroit from 2015 to 2016, as opposed to the 15.7 percent increase the FBI statistics show.
The flap over crime stats was addressed Thursday by the Detroit Board of Police Commissioners. If the FBI statistics are wrong, commissioners said, the police department should release the correct figures.
“There’s talk of Amazon and other companies coming here, but they’re not going to want to come here if they’re reading this is the most violent city in America,” commissioner Conrad Mallet said. “Somehow or another I think we need to create a record that this is what we believe the actual reporting to be.”
Board president Lisa Carter added that police officials should also release the correct figures to elected officials in Lansing.
But even by the police department’s count, Detroit would still hold the top spot as the nation’s most violent city. Detroit police figures show 12,842 violent crimes reported in 2016, as opposed to the 13,705 reported by the FBI.
According to Detroit’s data, there were 1,918 violent crimes per 100,000 residents in 2016 — slightly higher than the rate the
FBI reported for St. Louis: 1,913 per 100,000.
Assistant Chief Arnold Williams told the board the department’s numbers are available on the city’s website, adding the department would officially release its statistics at some point.
But Williams insisted the ranking isn’t as important as the public having the correct information.
“We can accept where we’ll fall as far as national rankings, but by having those incorrect numbers reported, it’s a slap in the face of every police officer in the city, and it’s a slap in the face to the city as well,” he said.
“If the correct stats show we’re No. 1, we take that, we own it and will work to get past it,” Williams said. “I don’t want to be No. 1, but if that’s what the numbers show, we’d have to accept it and come up with a pathway to get better — and if we can’t do that, you need to find new people who can.”
Williams said Detroit police officials tried to tell the FBI and Michigan State Police last year that their CRISNET reporting system was underreporting crimes.
“We’re not cooking the books; we were trying to tell them there were actually more crimes (in 2015) than were being reported,” Williams said.
But because data input into the CRISNET system immediately goes to the state police, which forwards the figures to the FBI, Williams said Detroit officials were told there wasn’t anything that could be done to amend the stats.
Williams added the FBI warns people not to use their statistics to compare communities — “but people ignore that,” he said.
The FBI says on its website: “Data users should not rank locales because there are many factors that cause the nature and type of crime to vary from place to place.”
Detroit Police Chief James Craig on Friday echoed Williams’ comments.
“I just want accurate data to be out there,” he said. “If the data shows we’re the highest in the country (for violent crime), then that’s what it shows. But going from a 15 percent increase in violent crime, which the FBI shows, to a 5 percent decrease, which the accurate numbers show, is kind of a big deal.”
Detroit police officials for years have complained about the CRISNET system. Detroit police in December purchased a new $9.1 million computer system, Superion, which officials say is more accurate.
Before switching to the new reporting system, Detroit police contracted with David Martin, a statistics expert at Wayne State University’s Center for Urban Studies, to pore through the CRISNET stats and fix errors. That’s how the department arrived at the lower 2016 crime stats, Craig said.
Deputy police chief David LeValley said the CRISNET system caused crimes to be over-reported or under-reported in some cases because the software doesn’t flag potential errors input by officers.
LeValley said the new system was tested for accuracy in July, and that there were only 17 errors out of the 7,000 crimes input into the system.