UVM Professor Stephanie Seguino. Photo credit: Ian Thomas Jansen-Lonnquist
Vermont Business Magazine A new University of Vermont study reveals notable disparities in how police officers from 29 departments across Vermont treat drivers by race. The research finds racial disparities in traffic stops, searches, arrests and outcomes – which vary by police agency. It is the first study of statewide traffic policing and race, covering Vermont’s largest police departments, and follows a 2014 state law requiring police to collect race data.
Blacks and Hispanics ‘oversearched,’ despite less contraband
At the state level, Black and Hispanic drivers were searched approximately three to four times the rate that White drivers were searched.
Despite lower search rates, White and Asian drivers were more likely to be caught with serious contraband leading to citations or arrests.
Among drivers stopped, Black and Hispanic drivers were more likely to receive tickets than White drivers.
Given their shares of the driving population, Black and Hispanic drivers were stopped more than expected, while Asian and White drivers were stopped at rates below their shares of the driving population.
Read the full study for data on individual departments: http://www.uvm.edu/giee/pdfs/SeguinoBrooks_PoliceRace_2017.pdf
Exploring ‘police discretion’
“Search and hit rates are among the best available indicators of racial disparities in policing,” says UVM economist Stephanie Seguino.
“Given that Black and Hispanic drivers are searched more, but found with less contraband, it suggests police use a lower threshold of evidence for these searches,” adds Seguino, a researcher at UVM’s Gund Institute and College of Arts and Sciences. “It also points to potential inefficiencies in policing.”
The researchers recommend improvements in data collection and bias training to address the disparities, especially in areas with elevated stop and search rates.
• Ticket rates: At the state level, Black and Hispanic drivers are more likely to receive a
citation once stopped than are White or Asian drivers.
• Arrest rates: At the state level, the Black arrest rate is almost double the White arrest
rate. At the agency level, disparities differ. For example, at the high end, Black
drivers stopped by Rutland police are 2.6 times more likely to be arrested than White
drivers, subsequent to a discretionary stop (excluding arrests on warrant), and in
Williston, 2.3 times more likely.
• Search and “hit” rates: At the state level, Black drivers are four times more likely to be
searched, subsequent to a stop, than White drivers. Hispanics also experience
elevated search rates compared to Whites; they are almost three times more likely to
be searched. Asian drivers are less likely to be searched than White (or Black and
Hispanic) drivers. In contrast to these search rates, Black and Hispanic drivers are
less likely than White or Asian drivers to be found with contraband that leads to a
citation or an arrest. Officers would appear to have a lower threshold of evidence for
searching vehicles with Black and Hispanic drivers. This suggests a problem of oversearching
of Black and Hispanic drivers as compared to a possible under-searching
of White and Asian drivers. Variations exist at the agency level. However, only a few
agencies have sufficient data to make statistically reliable inferences on racial
differences in hit rates. Among those that do (Burlington, Rutland, Vermont State
Police), hit rates of Black drivers are lower than of White drivers. Hits that result in
arrests—indicative of more serious contraband—occur also at a lower rate for Black
drivers than White drivers for all of three of these agencies as well as Williston.
• Stop rates: Black and Hispanic drivers are stopped at a higher rate than their share of
the population while White and Asian drivers are stopped at rates that are below
their population shares. Stop rate disparities are often subject to criticism because
researchers typically lack precise measures of the driving population. We have sought
to overcome that by using accident data on the race of not-at-fault drivers. Also,
most of our indicators of racial disparities are based on post-stop outcomes, which
do not rely on estimates of the driving population.
• Male drivers are more likely to be stopped than female drivers, regardless of
race/ethnicity. But the racial disparities in male shares of stops are notably large. At
the agency level, for example, in Middlebury, among Black drivers stopped, 88% are
male, while among White drivers stopped, 62% are male. Overall, Black and
Hispanic males comprise a larger share of stopped drivers in their racial/ethnic
group than do White males, suggesting a possibility that Black and Hispanic males, in
particular, are targets of heightened police scrutiny.
• Officer stop rates of Black drivers: Twelve agencies provided traffic stop data by officer,
allowing us to calculate within-agency disparities in stop rates. The results indicate
that the disparity in Black/White stop rates at the agency level cannot, in general, be
attributed to the behavior of just a few officers. The data indicate that this behavior
is common to many officers, perhaps suggesting more pervasive cultural norms
within agencies that contribute to disparities. Of note, in Brandon Police
Department, 67% of officers stop Black drivers at a rate that is 50% greater than
their share of the population. A sizeable share of officers in Bennington, Manchester,
Middlebury, and Winooski also stop Black drivers at rates higher than expected,
given population shares. In addition to disparities in stop rates by race by officer, we
also found evidence of variation by officer in the completeness of their legally
required data reporting.
• Data quality: Missing data is a concern. Some agencies were not able to respond to
our request for data. Moreover, many agencies have high rates of missing data in key
categories. For example, in St. Albans Police Department, race was not recorded in
29% of stops, and in Addison County Sheriff Department, race was missing in 17%
of traffic stop incident reports. Missing data undermines efforts to accurately assess
the degree of racial disparities in traffic policing.
Largest study of its kind, but more data needed
The study is the largest to date on Vermont police and race, and the first to compare data across multiple departments.
It covers the largest 29 of Vermont’s 70 police agencies – including 24 municipalities, 3 county sheriff’s departments and Vermont State Police – which provide law enforcement for the majority of the state’s population.
“Many agencies had high rates of missing data in key categories,” says co-author Nancy Brooks of Cornell University. “More work is needed to improve the quality of the data collected. That’s key to assessing and improving racial disparities in policing.”
While approximately 30 US states routinely collect police traffic data by race, some states have not analyzed it – and even fewer make the information publicly available.
The study, “Driving While Black and Brown in Vermont,” follows Seguino’s and Brooks’ previous work on individual police departments in Burlington, South Burlington, and the Vermont State Police.
“When it comes to police and race, the data suggests that Vermont may not be as different from other states as some might think,” Seguino says.
Agencies included in the analysis:
Addison County Sheriff
Grand Isle County Sheriff
Rutland County Sheriff
University of Vermont (UVM)
Vermont State Police (VSP)
Source: UVM 1.9.2017