CCSU report examines racial bias in police traffic stops
Updated 5:57 pm, Thursday, November 9, 2017
In a wide-ranging report on traffic stops released Thursday, Central Connecticut State University researchers flagged a number of police departments in the state for disparities in enforcement that might indicate racial bias.
In “Traffic Stop Data Analysis and Findings, 2015-16,” researchers determined that State Police Troop B in North Canaan and local police departments in Berlin, Monroe, Newtown, Norwich, Ridgefield and Darien had “statistically significant” disparities in traffic stops of minorities compared with stops of whites.
The authors of the report — which also tracked overall town-by-town pull-over rates — said those agencies that were noted for racial disparities would be analyzed further. The numbers used in the study come from police data compiled from Oct. 1, 2015, through Sept. 30, 2016.
One way the study examined possible bias was by comparing stops during the day and those at night — when police are less likely to discern a driver’s race before pulling a vehicle over.
In Norwalk, the study indicated, Hispanic drivers were stopped during the day at a rate that was 2.3 times their rate of being stopped at night.
When researchers accounted for three years of data, they found additional towns and State Police Troops with “statistically significant” racial or ethnic disparities.
Ansonia, Groton Town, Madison, New Milford and State Police Troops C, G, H, and K made that broader list. Monroe and Norwich made both.
The study also looked vehicle-search success rates, also known as “hit rates,” that measure how often people are searched in a traffic stop against how often contraband is found.
White drivers who were searched after being stopped in Monroe were found with contraband 42.9 percent of the time, while black drivers were found with contraband only 8.3 percent of the time.
The study’s authors indicated those results might suggest more than one conclusion.
“Police officers make decisions to search in an effort to maximize their expectations of finding contraband,” the study’s text said. “The implication being that police will be more likely to search a group that has a higher probability of carrying contraband ... In turn, motorists from the targeted demography understand this aspect of police behavior and respond by lowering their rate of carrying contraband.”
Monroe Chief of Police John Salvatore said his department would look into the report’s conclusions.
“Speaking for my department, we hire good people,” Salvatore said. “We train them well and they’re out there doing what they think is the right thing and the proper thing. But that doesn’t mean that we’re not going to be looking at what is in this report.”
Salvatore said he has concerns about the report’s methods. He said it appeared that assessments of driving populations in the study did not factor in those passing through from other towns and cities.
Others took issue with the study, too.
In a written document dated Sept. 7, to the president of the Connecticut Police Chiefs Association, Stephen M. Cox, a professor in the criminology and criminal justice department at CCSU, pointed to problems with the report as part of a peer review of one portion concerning 2014-2015.
One point of contention for Cox was the way researchers determined the number of drivers in a given municipality.
“The use of population-based benchmarks and descriptive statistics to approximate towns’ driving populations has weaknesses and are not recommended to be used nor presented in these reports,” Cox wrote.
The report said Connecticut State Police likely exhibited the “largest and most persistent disparities” in the day/night stop tests. The agency also had very low “hit rates” across all minority groups.
Trooper Kelly Grant, a spokeswoman for Connecticut State Police, said the material in the report would be carefully reviewed.
“We look forward to continuing our years of cooperation ... to address any concerns raised by the results and to account for the many variables that are inherent in this research,” Grant said. “We trust that, as in years past, further analysis of the data in conjunction with the research team will provide a reasonable explanation for any apparently outlying results.”