Connecticut’s deep data dive on racial profiling

Rep. Minnie Gonzalez questions researchers Ken Barone, Matthew B. Ross and James Fazzalaro.


Rep. Minnie Gonzalez questions researchers Ken Barone, Matthew B. Ross and James Fazzalaro.

The first advanced analysis of police traffic stops in Connecticut found significant racial disparities in three municipal departments and two State Police troops, with red flags raised in another dozen departments, researchers said Tuesday.

The analysis presented to legislators was based on applying seven statistical and descriptive measures to 620,000 traffic stops by all but one of the state’s 103 law enforcement agencies in the year ending Sept. 30, 2014.

“The statistical disparity provides evidence in support of the claim that certain officers in the state are engaged in racial profiling during daylight hours when motorist race and ethnicity is visible,” the analysts concluded.

Groton Town, Granby, Waterbury and state police troops in Hartford and Tolland had significant disparities, with evidence of potential bias found in Wethersfield, Hamden, Manchester, New Britain, Stratford, Waterbury and East Hartford.

Further analysis will explore if a handful of officers in those dozen departments are responsible for stopping blacks and Latinos in numbers far higher than a series of benchmarks, including racial makeup of an area’s estimated driving population.

Researchers said another five departments had “disparity threshold levels” that exceeded the norms and were worthy of further study: Meriden, New Haven, Newington, Norwich and Windsor.

Only Stamford police failed to provide data, denying the researchers information on another 25,000 stops, a problem blamed on technical issues that have since been resolved, said Andrew Clark, the director of the Institute for Municipal and Regional Policy.

The institute, based at Central Connecticut State University, analyzed the enhanced data collected under revisions passed in 2013 to a 1999 law, the Alvin W. Penn Racial Profiling Prohibition Act, that had produced inconsistent data and little useful analysis.

As recently as 2011, only 27 of 103 police agencies had bothered to submit traffic stop data. Under the new effort, the institute provided police agencies with software that allowed it to mine traffic data and analyze it in standard formats.

The project was funded with a grant from the National Highway Traffic Administration.

The study puts Connecticut ahead of other states in analyzing racial profiling, an issue that has come to prominence locally after a racial-profiling scandal in East Haven and nationally in the aftermath of a police shooting in Ferguson, Mo.

The institute applied seven benchmarks to the data, including the “veil of darkness” standard, which compares daylight stops to those after dark, when a police officer is presumably less apt to know the race of a motorist until the car is stopped and approached.

The study also compared the percentage of minorities stopped in individual communities to the statewide averages and to local driving-age populations. Twenty-nine communities stopped significantly higher percentages of minorities than the statewide averages: 13.5 percent, blacks; and 11.7 percent, Hispanics.

The analytical tools were chosen in consultation with minority groups and law enforcement officers, who all offered a cautious endorsement of the general approach taken by the institute.

Kenneth Barone, a policy specialist for the institute, called the study the “most comprehensive approach” in the U.S.

“The release of this report is evidence that Connecticut is well positioned to lead the nation in addressing the issue of racial profiling and increasing trust between the public and law enforcement,” Barone said.


The findings were presented Tuesday to the General Assembly’s Public Safety and Security Committee. The report and underlying data are available online.

To read the report or download raw data on which it is based, go to

Michael P. Lawlor, the governor’s criminal justice advisor, said the data collection system provides an unprecedented degree of oversight into a police officer’s most frequent measurable interaction with the public: The traffic stop.

Chief Douglas Fuchs of Redding, who represented police chiefs in developing the study parameters, said the chiefs and State Police are committed to honestly examining traffic-stop data for evidence of racial profiling.

“Law enforcement should be stopping a virtual mirror image of who is actually driving on our roadways,” Fuchs said. “We have not fought this any step of the way, nor would we fight this.”

But determining exactly who is driving on a community’s roads at any given time can be complex, reflecting local driving population, commuters, shoppers and others.

Fuchs said some chiefs are suspicious of one measure — the estimated driving population — that is supposed to account for out-of-town commuters. He would like to see those estimates compared with observations made from roadsides.

“We need to take four or five of the communities that have been called out in this report and let’s put boots on the ground,” he said.

Andrew Matthews, the president of the Connecticut State Police Union, had no ready analysis of the data that showed minorities drove 37.5 percent of all vehicles stopped by Troop H in the Hartford area and 15.2 percent of those stopped by troopers assigned to Troop C off I-84 in Tolland.

Matthews said the 230,000 traffic stops by State Police yielded only a dozen complaints.

But he said any evidence of racial profiling should be examined and acted on.

Some communities showed evidence of bias by one measure, but not others. Traffic stops by the highlighted departments and state police troops scored high on a matrix of several measures.

Werner Oyanandel, the executive director of the Latino and Puerto Rican Affairs Commission, said researchers on his staff believe that the institute took a conservative approach to analyzing the data, taking care to note that racial disparities are not always evidence of discrimination.

“So, the problem might be larger than what is reported,” he said.

But he said the transparency of the data is a tool for both communities and police leaders to assess whether racial profiling is a local problem.

“I am confident that the chiefs are going to be looking at their towns,” Oyanandel said. “I’m sure this will be very helpful.”

The data allows researchers and local police brass to examine how many stops were attributable to individual officers.

Sen. Gary Winfield, D-New Haven, said the data provides a series of metrics that allow for a constructive community-police conversation, rather than broad attacks by critics and a defensive posture by police.

“It is never helpful when people say, ‘The police are…’”

Winfield did not fill in the blank.

“It shuts down the conversation. This is not about going after police,” Winfield said. “This is about rooting out any bad police officers.”