New Haven is getting safer, according to University and city data. Yet as the 2011-’12 class year begins, a controversial ranking as the “fourth most dangerous city” has angered New Haven residents, and elicited retaliatory statements from Yale and New Haven officials.

The Federal Bureau of Investigation released its annual Uniform Crime Report on May 23 — which compiles the violent crime statistics reported by municipal police departments over the previous year — setting off a rush of media comparisons and rankings. Many news websites published stories using this data to rank New Haven as the fourth most dangerous city behind Flint, Mich., Detroit, Mich. and St. Louis, Mo.

[ydn-legacy-photo-inline id=”4093″ ]

But City Hall Spokesman Adam Joseph told the News that these websites’ statistical analysis obfuscated the truth about Elm City crime, and Yale administrators are still reaching out to the campus community to correct the perceived inaccuracies. Yale Police Department Chief Ronnell Higgins wrote an Aug. 25 email to students condemning the rankings.

[ydn-legacy-photo-inline id=”4090″ ]

“You should be aware that there have been dubious statistical ‘rankings’ circulating about the crime rates in New Haven relative to other cities,” Higgins wrote. “One claim this summer received much play in the media, even though it is not a valid comparison.”

Websites for Business Insider, Michigan’s WNEM-TV, news aggregator Newser, and investor-focused 24/7 Wall Street posted stories on May 23 that used raw totals of “violent crime” — including murders, rapes, robberies, and arsons — from the FBI’s crime report, and then divided that sum by the report’s cited municipal populations, achieving a measure of violent crime per capita, where New Haven placed fourth in the nation.

Despite the negative press from the FBI report, its data actually showed that New Haven had successfully reduced total violent crime by nearly 10 percent in one year — from 2,195 incidents in 2009 to last year’s 1,978. University dated confirms this trend: Yale saw its lowest crime rates in 20 years in 2010, according to Higgins.

Yet Mark Abraham, executive director of local nonprofit Data Haven, wrote in the New Haven Register in June that the media analysis used incorrect population numbers — the FBI cited neither 2000 nor 2010 census data, recording New Haven as home to 124,856 people while the newest figures are over 129,000.

Additionally, Abraham faulted the rankings as failing to recognize the difference between the relatively small area of New Haven’s metropolitan area and somewhere like Jacksonville, Fl., an area that is legally defined as more than 42 times the size of the Elm City. If the actual boundaries of the greater New Haven area are taken into account, then the Elm City ranks near the middle of the 350 major U.S. cities, Abraham wrote.

“As a municipality, New Haven is highly unusual: despite being home to 80,000 jobs, it is so small in land area that most of its working residents are employed in other towns, and vice versa,” Abraham wrote. “When comparing places, good researchers define a city not as a municipality, but as the “place” where, by standard methodology, the majority of people live, work and shop.”

Although this analysis did not account for actual population versus residential population (New Haven swells by as much as 28,000 people every day during work hours), most local media outlets ran with the story, including the News, NBC Connecticut and WTNH.

Compounding the city’s image problem, the week the rankings were released, New Haven had seven shootings in 48 hours. Still, some New Haven residents recognized the statistical problems with the rankings.

“The feedback I hear from neighbors, is that they are really more angry at those parts of the media who’ve sensationalized things,” Michael Morand ’87 DIV ’93, former associate University vice president for New Haven and state affairs, told the News.

As of Aug. 25, New Haven has seen 23 murders in 2011, only one less than 2010’s final tally.

Danny Serna contributed reporting.