The Journal Sentinel’s Attack Dog

On Sunday, the Milwaukee Journal Sentinel ran a mildly interesting watchdog story telling us that FBI audits of local police department crime data are “shallow” and infrequent. Fairly ho-hum stuff for Sunday’s front page, but the story, you see, was also part of the paper’s continuing effort to justify its misleading attack on Chief Ed Flynn and the Milwaukee Police Department for supposedly manipulating crime data. Sunday’s story let us know the FBI audit is far too small scale to uncover data screw-ups by Milwaukee police. Hence the need for ever more JS watchdog stories — all written by Ben Poston —  attacking the department.

The newspaper’s original watchdog story, as a column I wrote in June noted, “could have been a snoozer story about bureaucratic errors in computer coding which resulted in a minor, three percent difference in how violent crime was recorded.” Instead, the JS led readers to believe something nefarious was afoot, running quotes from experts calling the mistakes “suspicious” and proof of a “failure of leadership,” a likely sign of deliberate misreporting, “playing fast and loose” with data and “cheating the public.”

As one expert quoted by the newspaper put it, “These are not accidents. Somebody knows what’s going on. Somebody understands the implications of reporting it this way… The question is: Who is making those decisions?” Gee, I guess that would be Chief Flynn.

The story was filled with logical and evidentiary holes, all itemized in my column. The newspaper’s major finding was that 500 aggravated assaults had been misclassified as simple assaults, thus lowering the violent crime rate by three percent in 2011.  So Flynn’s team scoured through their data, reviewing 34,000 “assault-type reports” from 2006 through 2012 which included disorderly conduct, simple battery, family offenses and other incidents, some of which might have been more accurately coded as aggravated assaults. The point here is obvious: if computer coding mistakes are being made, it’s quite likely the errors could go both ways, to upgrade non-felonies as violent crimes as well to wrongly downgrade aggravated assaults as simple assaults.

Sure enough, the department review of six years found that while 5,307 crimes were under-reported (for example, calling an aggravated assault a simple assault), 1,176 crimes were over-reported as more serious than they actually were. The adjusted totals for 2011, the key year — and the basis for the entire Journal Sentinel attack — showed crime actually declined by more than the department originally reported.

The review also found that computer errors were actually higher under  Flynn’s predecessor, Chief Nan Haggerty, further undercutting the newspaper’s suggestion that Flynn was cooking the books.

The results, presented at a Common Council meeting provoked outrage from aldermen, most of which the newspaper of record chose not to report. As Alderman Jim Bohl put it, the newspaper’s “smoking gun has just turned into a pop gun.” Bohl called on the newspaper to admit the problems with its reporting. “If you are a responsible publication then you report it.” His quotes never made it into the JS story on the meeting.

Alderman Joe Davis noted that he hadn’t heard from one member of the police department complaining that he or she was instructed to misclassify data. The newspaper didn’t report this.

Nine alderman published a statement accusing the newspaper of misleading the public. The newspaper did report this, but did not link readers to the statement, didn’t publish it in its entirety and didn’t quote some of the strongest charges, including these:

“It should be noted,” the aldermen declared, “that before issuing this statement many of us pleaded with the reporters and editors at the Journal Sentinel to correct their own mistakes. They refused.”

The paper “went beyond facts, and it used poor methodology to get there. In statistical jargon, it ‘failed to set a baseline for comparison’ when it asserted that crime was up, not down. This was immediately obvious to anyone with any statistical background who reviewed its coverage. This was pointed out to the paper repeatedly — but it stuck with its story.”

“We will not speculate as to why the paper of record… has spent weeks misleading the public. But the facts are that it has and the public should know.

‘That misleading information — whether deliberate or sloppy — needs to be corrected by someone.”

The JS ignored all this and focused on one sentence, where aldermen charged the newspaper told the public that the mistakes in data “were done on purpose in order to fool the public…” The JS story jumped on this, with a quote from editor Martin Kaiser that such a claim was never made in its original story. In fact, as I’ve noted, the story was studded with quotes from experts suggesting the computer errors were done to cheat the public.

If there was any doubt the newspaper intended to continue beating up on the department, Sunday’s story on FBI audits answered that question. The story twice reminded us the JS previously found 500 aggravated assaults that were wrongly called simply assaults and twice told us the internal review by Flynn’s team showed more than 5,300 aggravated assaults were underreported since 2006. The paper made no attempt to balance that off by noting the department’s review showed it had also made errors that over reported violent crime, and that the errors had declined since Flynn became chief and that violent crime had actually declined more than his department had reported.

The Sunday story also told us the department’s 20 percent error rate is 10 times higher than “the national standard.” But nowhere in the story does the newspaper tell us how many police departments in the nation actually meet this standard. In fact, its story has quotes from experts saying that computer data problems are widespread in police departments and that they need to do “far better audits” of data.

If this problem is so widespread, it completely undercuts the original watchdog series, which suggested Milwaukee’s department stood out nationally for its sloppy data.

Yet the newspaper continues to flog Milwaukee mercilessly. Today, it ran a front page, top-of-the-fold story telling us that Edward Claughton, the consultant chosen by the city to audit the police department’s data, has publicly supported the department against the newspaper’s stories.

The first question that comes to mind: if this is a nationally-known analyst of data who questioned the JS analysis, why wasn’t he ever quoted in the newspaper’s stories assailing Flynn for alleged data manipulation? It confirms my suspicion the newspaper went shopping for experts to criticize Milwaukee, and ignored any who might have supported Flynn.

Indeed, today’s story shopped for the one alderman, Bob Donovan, who can be depended on to support the newspaper’s thesis that Milwaukee is on a campaign to mislead the public about police data. Not one of the nine alderman who criticized the newspaper is quoted in today’s story. The story also sought a quote from police union leader Mike Crivello, who is a constant critic of Flynn.

As to whether Claughton and his PRI Management Group can do an objective job of auditing the department’s data, my guess is that his views would be shared by most other experts in the field. Indeed, the newspaper’s Sunday story told us that data errors are widespread in police departments.

Claughton told the newspaper he is a certified auditor who must be “100 percent” objective and must follow strict auditing standards. And Claughton surely knows the newspaper will jump all over him for the smallest error. Under the circumstances, it’s hard to imagine him producing a biased report.

But I’m guessing the Fire and Police Commission will be pushed to hire some other expert. If it doesn’t the JS will keep attacking.

The issue at this point has nothing to do with accurate crime reporting and everything to do with the JS justifying its consistently biased coverage of this issue. It’s a truly sad situation. As the aldermen’s statement noted about the newspaper, “it is our sincere hope that… the public is not forced into a crisis of confidence in regards to this important institution. We all want to live in a city with a great, responsible and trusted newspaper.”

These sentences, needless to say, also went unreported by the Journal Sentinel.