When they stop motorists in Vermont, cops don't just collect licenses and registrations. As of September 1, 2014, all police officers in the state must record the race of every driver they pull over.
The new mandate for "roadside stop" data collection is just one step in a national movement to halt discriminatory law enforcement. "Throughout the country, there's a crisis of legitimacy in policing," said Burlington Police Chief Brandon del Pozo. "Part of that crisis stems from a belief that the police are biased in the way they use their discretion. And traffic stops are something that police have immense discretion over."
Vermonters are still waiting to see what the new statewide data collection reveals. That's because while technically public, the information remains largely inaccessible. In 16 months, no one has compiled the raw data — a necessary first step before analysis. In fact, no one even knows if all of law enforcement is complying with the mandate to collect the data in the first place.
And as for merging the information from incompatible databases to allow a statewide look? Nobody is in charge of that.
The Vermont legislature mandated roadside stop data collection in 2014, as a result of prodding from the human-rights group Migrant Justice, among others. That was six years after a handful of police departments in Chittenden County began gathering it voluntarily, four years after the Vermont State Police started keeping track and two years after the legislature "encouraged" collection. Before Act 193, Vermont was one of only two U.S. states that did not collect race data associated with all traffic stops within its borders, according to del Pozo.
"It was alone with Mississippi in that regard," del Pozo said. "And given Mississippi's legacy, I don't think that's good company to be in."
Champlain Area NAACP president Mary Brown-Guillory said she doesn't need numbers to know there's bias in Vermont policing. Brown-Guillory, who is black, said she's been pulled over many times for no reason. Two local investigations support her claim: Vermont State Police data from 2010 and 2011 showed that black and Hispanic drivers are stopped, ticketed, searched and arrested significantly more than whites when they're pulled over — even though the rate at which minorities are found to have contraband is lower than the rate for whites. A 2012 report revealed similar discrepancies in the Burlington area.
Act 193 was supposed to make such data analyzable statewide, exposing police behavior in even the tiniest of Vermont's towns. "If we can't measure it, we can't manage it," said Mark Hughes, who is black and a cofounder of a Montpelier-based group, Justice For All, which aims to identify and dismantle institutionalized racism. Hughes wants to determine where bias exists so appropriate training can be directed at police within those departments.
Del Pozo said he agrees that citizens should not be singled out for their race. Data transparency cultivates trust, he said, and that's a crucial component of good policing. "When we need citizen cooperation for more serious crimes, part of the way we get it is by treating citizens well when we investigate them for little things like [traffic stops]," del Pozo said. In fact, the new chief just released data from 2014 on the Burlington Police Department website; by February, he plans to add prior years.
Preliminary assessment by Seven Days shows that white drivers made up 83 percent of the Burlington police stops in 2014, black drivers 7 percent, Asian drivers almost 4 percent and all other minorities less than 1 percent. Vermont driver's licenses do not include information about race or ethnicity, so officers are expected to make their own determinations. Not surprisingly, in nearly 5 percent of the stops, police record the race of the motorist as "unknown."
Census data from 2010, the most recent available, count the city's driving-age population as 91 percent white, 2.8 percent black and 3.2 percent Asian, among residents claiming only one race. The separate measure of ethnicity, which may overlap with race, counts 2.5 percent Hispanic residents.
Del Pozo said the apparently disproportionate stop rate for black drivers "definitely bears further investigation." That said, he approaches the discrepancy with caution: "When many of our social and economic problems unfairly fall along racial lines, it isn't reasonable to think that car stops by the police will be one of the few places where demographics match outcomes," he said by email. But widespread social inequities "can't just become a lazy excuse to just start stopping 'the usual suspects,'" he said in an interview. "That's when it becomes bias."
One other finding also caught del Pozo's attention: When stopped, white and black drivers received tickets and warnings at nearly the same rates — about 27 to 28 percent, respectively — while Asian drivers were 3 to 5 percent more likely to drive away with just a warning. However, among drivers of unknown race, a whopping 38 percent got tickets.
Del Pozo said he can't account for that discrepancy, which reflects activity before he joined the Burlington police force as chief. "We would have to do a case-by-case review of those stops in order to see if there was anything meaningful lurking beneath those unknowns," he said, adding that he doesn't suspect that the "unknown" category was used disingenuously to "soften" the stop data.
Because his department's analytical resources are limited, del Pozo said he's publishing the data with the hope that others with the time and interest will dissect it. His focus now is to lower the sheer number of "unknowns" recorded. "I'd like to ensure that every car stop has a race ascribed to [the driver]," he said.
That can be tricky in a short, often tense interaction outdoors under variable lighting. Plus, officers approach the public with varying degrees of exposure to diversity, and race can be complex to decipher, del Pozo said. He pointed out that most people would not identify him as Hispanic, for example, although with a Cuban father and Jewish mother, that's part of how he identifies himself.
"Since the state requires the collection of race data for car stops, it would be helpful for the state to facilitate this process by providing a person's self-declared race on his or her driver's license," he said. That prospect may be a lightning-rod issue, he acknowledged, but so is the correlation of race and car stops, "and the state has made the commitment to get to the bottom of [it]," he said.
Other departments may already be posting data they've collected. But absent a central clearinghouse for it, there's no way of knowing without reaching out to more than 100 individual agencies, departments, offices and constables. Some say time and technical hurdles prevent them from posting information about their roadside stops. Winooski Police Chief Richard Hebert said he's willing to share the raw data, but he feels it would be meaningless without analysis.
Allen Gilbert, executive director of the Vermont chapter of the American Civil Liberties Union, said many assumed that Act 193 would play out differently — that once the data were collected, they would be compiled and crunched. He's since realized that the wording in the law is too vague. The legislature directed police to work with the Vermont Criminal Justice Training Council "with the goals of collecting uniform data, adopting uniform storage methods and periods, and ensuring that data can be analyzed. Roadside stop data, as well as reports and analysis of roadside stop data, shall be public."
Rick Gauthier is eager to know exactly what that means. He's the executive director of the VCJTC, which operates the Vermont Police Academy and oversees training of law-enforcement officers statewide. "Absent every agency reporting to some central location, there isn't any way to ensure that every agency is in compliance," Gauthier said.
He said the board of the Vermont Incident-Based Reporting System, which collects criminal justice data for the Federal Bureau of Investigation, recommended that departments submit quarterly reports to the council. But that advice isn't binding. Gauthier said he also offered to post a collection of links to web pages where departments publish their own data, but so far none has taken him up on that.
Some question whether the Vermont Criminal Justice Training Council is the right entity to compile the data. Hughes feels it's too close to law enforcement. "By a long stretch, they are not the most impartial organization to aggregate this information and make it available," he said.
Brown-Guillory believes the information is flawed to begin with because police officers themselves are providing it. "You can put that the person had a broken headlight or didn't have their license," she said. "You can skew the data."
"There are a lot of different ways to analyze ... a set of data," del Pozo said. "And the worry is that the methods you choose and the way you present it and the conclusions you draw, rather than being dispassionate, will be colored by your own bias about police work, positive or negative."
Differing reports on the same data demonstrate del Pozo's point. In 2011, the Vermont State Police asked Jack McDevitt and Chad Posick of the Institute on Race and Justice at Northeastern University to analyze their stop data. The researchers found "few areas in which racial and ethnic disparities exist." Three years later, University of Vermont professor Stephanie Seguino reanalyzed the same data, and took McDevitt and Posick to task.
"By grouping all minorities together, [McDevitt and Posick] miss the differences in outcomes of Asians as compared to blacks and Hispanics, thus diluting their ability to identify racial disparities," she wrote. In other words, by analyzing each race separately, Seguino found "wide disparities" in the treatment of black and Hispanic drivers compared with whites, while Asians are actually stopped and searched less.
And then there's context. Before releasing Burlington's data, del Pozo is trying to add a note when external circumstances, such as probable cause or reasonable suspicion, prompted a stop. Del Pozo said it's consuming staff hours to prepare the data for release — adding those notes, removing duplicative or erroneous records, consolidating multiple violations from a single traffic stop into one record. Smaller departments, he noted, may not have the personnel to manage the work.
Yet another hurdle to aggregating the data statewide: Departments use one of two different database systems to record incidents. Once each agency's data are ready, the fields will have to be aligned in order to collect them in one place — basically, a single spreadsheet for the tens of thousands of roadside stops that occur every year. Clearing that technical hurdle without statewide oversight will be a challenge, and Gauthier said law-enforcement officials are concerned about the price of ongoing data analysis.
Hughes said that shouldn't be a factor. "I found myself somewhat incensed when I was thinking about the fact that there is a cost associated with my safety," he said. "It seemed to me there might be a few line items that might be a little less important."
Finally, Act 193 failed to place a deadline on the data's statewide aggregation or publication. Law-enforcement agencies can collect it, but they never have to furnish anything until they're asked.
Suzi Wizowaty, a former House representative who now heads the nonprofit Vermonters for Criminal Justice Reform, said her organization has been working with Rep. Bill Lippert (D-Hinesburg) to introduce a bill that would add a deadline for data reporting and put someone in charge of meeting it.
"What we are asking is to make the data available on a public website so that anybody can analyze it," Wizowaty said. Calling cost concerns a "red herring," she recommended putting "dis-identified data" out there to "let people who have the skills analyze it." That's exactly what del Pozo is doing in Burlington.
Sen. Dick Sears (D-Bennington), who sponsored Act 193, also said he expected to be looking at the whole state's data by now. He anticipates tightening up the law's language in the 2016 legislative session to make sure it's compiled.
Traffic stops are just a small part of the criminal-justice system, but Hughes and del Pozo agree that they represent a crucial step toward repairing public trust in law enforcement.
Still, Hughes is anxious to move on. The narrow focus on this one piece of the problem serves as a distraction from the broader, more fundamental issues, he said.
"We have to get past this discussion of whether or not [racism is] there," Hughes said. "And get past all the accusations and all the denials and the delays and the requests for empirical data and delays of release thereof. And have a discussion in Vermont that black lives matter."