After studying four years of traffic-stop data in Burlington, University of Vermont economics professor Stephanie Seguino has concluded that there’s evidence of police officers targeting black drivers.
Seguino, who conducted the analysis with Cornell professor Nancy Brooks based on data recorded by officers from 2012 through 2015, presented findings to the city council Monday night.
Among them: Black drivers were more likely than their white counterparts to get pulled over in Burlington, and, once stopped, they were less likely to get off with just a warning. White drivers got warnings 69.4 percent of the time, whereas black drivers got warnings only 61.6 percent of the time.
Black drivers were more likely to be searched after being stopped: 3.3 percent of the time, compared to 1.1 percent of the time for white drivers.
But searches of black motorists turned up contraband only 46.2 percent of the time, whereas searches of white drivers produced illegal items 63.5 percent of the time. That could mean white drivers were getting under-searched or black drivers were getting over-searched, Seguino suggested. The difference in arrest rates between races, she noted, was not statistically significant.
Seguino drew attention to the disparity in search rates, while Police Chief Brandon del Pozo downplayed it, noting that the total number of searches was small. Just 52 black drivers were searched during the time period.
Seguino’s analysis also revealed what she characterized as “quite a substantial disparity in stop rates by officer.” On the high end, 15 or 16 percent of an individual officer’s stops involved black drivers; on the low end, black drivers made up less than 2 percent of an officer’s stops.
Unlike other states, Vermont doesn’t ask people to disclose their race on their driver’s licenses. But in 2014, the state began requiring police officers to record this information during traffic stops. Seven Days reported in January that it was unclear whether police departments were complying with the new law, in part because no one was charged with making the data publicly available.
The Vermont Senate is currently reviewing a bill the House passed in March that would require departments to submit their data to the Crime Research Group of Vermont, which would maintain a central database open to the public.
During her presentation to the council, Seguino noted that there were significant gaps in the data collected by Burlington police officers, which, she said, made it difficult to complete a thorough analysis. But compared to some other departments, Burlington is actually ahead of the curve in that it’s been collecting traffic stop data for years, which del Pozo has pledged to publish online every quarter.
Acknowledging that some of Seguino’s findings were “troubling,” he emphasized the importance of making this data transparent — both within the police department and to the public.
Got something to say?
Send a letter to the editor and we'll publish your feedback in print!
“Troubling” doesn’t even come close…
I am not the least bit surprised. While Vermont purports itself to be a liberal bastion, the law enforcement community clearly has a racial bias. Extremely troubling, but not surprised.
Seems to me University of Vermont economics professor Stephanie Seguino and Cornell professor Nancy Brooks is trying to start some racialism here in Vermont..especially by our police..I’ve seen a lot more whites and other race get stop then blacks..so don’t even go there, white, black, purple or orange..if you break the law you will be stopped..you will be questioned..Seems to me these so called professors can be doing something a lot better then trying to make a ant hill into a mountain..
Glad that Dr. Seguino is pushing this issue – would like to see real attention to the disparities in stop rates by officer. That means there’s some “bad apples” and really getting serious about that would be important.
Every time…EVERY time I have been to Burlington in the last 7 years I have seen a person of color pulled over by a white cop…every time.
The search data is troubling but as to initial traffic stops is it possible this is based on human beings breaking the law, whatever their color? And as to tickets vs warnings, whether they cooperated with and respected law enforcement from the get-go or argued with & challenged law enforcement?
For example, did this study control for the fact that Burlington has seen significant spike in diversity from the new African immigrant refugees? Some of whom, through no fault of their own other than growing up in Africa and not the US, may not be familiar with American traffic rules & thus have broken the law more frequently? In that case, traffic stops are valuable educational tools to help immigrants understand the law. Most new Americans are outstanding people who are contributing to Burlington but it could have easily skewed the results. Nothing in this reporting suggests the professors controlled for this fact.
Also, as Burlington has gotten grittier over the last 10 years, with more homelessness and more drug crime, the fact of out-of-state drug dealers cannot be ignored. There are plenty of white people involved (including Vermonters who host the dealers). However, anyone associated with law enforcement will tell you there is disproportionate number of dealers coming from the Bronx, Hartford, Springfield, Boston, etc. and good number are African-American. Whether it’s lack of job & educational opportunities or whatever drives people of whatever race into crime (including white people like Bush & Cheney who commit treason by lying America into war for private profit of their companies and friends), we cannot ignore that unfortunately some of the crime in Burlington and VT is driven by out-of-state African Americans. I hope the study examined this fact and the journalist will follow-up with the professors.
These professors are definitely race baiting. The issue of searches is telling. They make their case on a 2.2 percent difference. With only 52 searches of black drivers, only 4 searches less means whites are searched more often. Even in sample sizes of thousands, margins are 4-5% at a minimum.
This is not to say that BPD doesn’t have problems or that all these statistics are wrongly interpreted. Just be careful to look at who is presenting statistics and why.
While this in interesting, with no controls reported for things like:
number of years driving, number of years living in the USA, Country of Origin (as one person already pointed out, if you grew up driving where the traffic laws are very different then the odds of getting stopped go up)
there is no information about warnings to indicate if a person was warned the first time and ticketed for failure to address a safety issue (head light out) at a later date.
Perhaps information about that nature of the traffic stops…. (distracted driving, speeding, defective equipment, failure to obey traffic signals, failure to yield to pedestrian traffic)
Additionally, before accusing officers of having a racial bias… consider where individual officers patrol. If officers often patrol residential areas that has a higher concentration of a given ethic group one would reasonably expect all interactions with the public to reflect this.
no information was reported about the same driver / vehicle stopped multiple times… (a family car with a headlight out that is driven often as multiple family members work various shifts) could easily result in the same vehicle or driver being stopped before the repair could be made.
I would like access to the data used for this (individuals names and officer names obfuscated or replaced with unique identifiers so individuals and officers can be tracked but not identified) so see what other trends can be found.
I’m curious about bias for day of the month… to see if traffic stops happen more often near the beginning/end of the month… and if searches happen more often on week days / week ends / nights / days… there may also be some interesting information about where traffic stops happen more often.
Yeah I wouldn’t be surprised if this were the case in Burlington and the rest of the US, but you can’t publish results if they aren’t statistically significant…