Back in March, Zayd Atkinson was picking up trash outside his dorm at Naropa University in Boulder when a group of police officers confronted him, apparently refusing to believe that he lived there.
Atkinson, a Yoga Studies major who was picking up trash as part of his work-study doing groundskeeping around the dorms, says he feared for his life when an officer first drew a taser, and then a firearm. Atkinson says time slowed down and he went into fight-or-flight mode.
"The situation ultimately de-escalated when the manager of facilities ran up," says Atkinson, and confirmed that Atkinson did, indeed, live there.
"Zayd was stopped for being black, and this situation escalated because Zayd was black and the officer pulled his gun on him," says Siddhartha Rathod, an attorney with the firm Rathod Mohamedbhai LLC who is working with Atkinson.
Now, for the first time, the city has released a full year of data on the demographics of who gets stopped and arrested by police. It’s an effort that started in July 2017 after the consulting firm Hillard Heintze recommended it in an independent assessment of the Boulder Police Department. The data is open to the public here, and is summarized at the back of the department’s 2018 annual report in a series of colorful charts.
Even with all the nice charts though, it’s hard to understand what the data actually means.
At a quick glance, it seems that the police tend to stop people of different backgrounds roughly on par with the city’s resident demographics. For example, 90.7 percent of stops happened to white people, and 87.9 percent of Boulder residents are white. Similarly, 0.4 percent of stops happened to American Indian people; by comparison, 0.3 percent of Boulder residents are American Indian.
Where it gets more interesting is the part about when stops lead to arrests. When police chose to stop white, non-Hispanic people, 3.5 percent of those stops led to arrests. By contrast, 7.7 percent of the time that black people were stopped, they were arrested. In other words, stops lead to arrest twice as often for black people as they do for white people. (In all of this data, "stops" refers to "discretionary stops," meaning situations where officers chose to intervene, as opposed to, for example, responding to a 911 call).
"The data speaks the truth, to some extent, in my opinion," said Atkinson.
Nami Thompson, with the Boulder chapter of the NAACP and the Colorado Public Health Association, agrees the findings aren’t too surprising.
"One of the troubles we have is that we talk about 'Is this a racial issue? Is there a racial disparity?' as if the conversation just started today," says Thompson. "There are racial disparities and we have move forward from there."
But data like this is really complicated and Boulder is a good example of just how tricky it is to confidently answer the simple question, "Are police being racist, and in what ways?"
"We have work to do," says Julia Richman, Chief Innovation and Technology Officer with the City of Boulder. That much, she says, is clear. "The challenge of this data is that it doesn’t say what that work is."
For one, as Richman points out, working with small numbers can skew things. In the case of the arrest data mentioned above, that 7.7 percent actually refers to fewer than 30 people -- a number that, in other statistics situations, would have been tossed out. "A good rule of thumb is that an 'n' less than 30 is something you won’t include in analysis," says Richman, because "without trends, those small numbers could swing a lot." For example, one rowdy sporting event and a handful of arrests could drastically change the percentage, she says.
Another issue is that a "stop" includes everything from what Atkinson experienced -- a group of officers and an escalating confrontation -- to, on the other end of the spectrum, a single officer asking a homeless person if they’re doing all right. Furthermore, the stop data is anonymous, so there’s no telling how many stops actually all refer to one person.
Then there’s another big issue. Boulder’s daily population does not just include its residents. As the report points out, the influx of commuters, tourists, individuals experiencing homeless, and students means "large swings in population throughout the year, and even in a given day." And, it turns out, more than half of the time the police stopped people, they were stopping non-residents. Same goes for arrests and citations. So, lining up the demographics of people stopped versus people who live in Boulder, as I did above, isn’t showing the whole picture.
Amy Shoemaker, a data scientist with the Stanford Computational Policy Lab, says this is a classic problem -- the so-called "benchmarking problem."
"The most common thing that people do when they get stop data is they wanna look at stop rates, so 'What are the stop rates of different race groups relative to their share of the residential population?'" says Shoemaker, who works on the Stanford Open Policing Project, which tracks and analyzes traffic stops by law enforcement across the country. "But it’s really tricky to use those numbers to actually make any claims of discrimination or racial profiling because we don’t know what the correct basis of comparison is."
The people who are walking, driving and hiking around Boulder every day probably don’t have the same demographic breakdown as the people who actually live in that city. Same goes for drivers on the road, anywhere in the country.
For example, let’s look at the situation of Boulder officers stopping black people: 4.3 percent of stops happened to black people. Only 1.1 percent of Boulder residents are black. So, the stop rate is about four times what you’d expect if police were stopping residents -- and only residents -- randomly. On the other hand, the homeless are not counted among Boulder’s residents and 7.4 percent of them are black. If the police tend to stop homeless people more than other groups of people (like commuters, residents, or students), then the fact that 4.3 percent of stops happen to black people might not be a sign of racial discrimination among law enforcement, but instead just a reflection of who happens to be homeless.
So, based on just that tidbit of information, it’s not possible to say if or how the Boulder police are being racist, despite cases like Zayd Atkinson’s.
Shoemaker and her colleagues used a clever method to get around that issue, at least for data on people who were stopped while driving. As they wrote in a working paper earlier this year, there’s something called the "veil of darkness" test.
Basically, it should be harder for an officer to racially profile people if the officer can’t actually see them very well. So, instead of comparing the racial distribution of stops to residential demographics, the researchers compared stops before and after dark to see if there was a difference in racial breakdown.
"In aggregate, across the country, both in state patrol departments and in municipal police departments, we see that, indeed, there is a smaller portion of the stop population that is black when it is dark, which is evidence of racial profiling," Shoemaker says.
Likewise, the group didn’t just looking at the rates at which people of different races and ethnicities got searched after being stopped, they looked at how often those searches actually resulted in law enforcement finding contraband.
"Examining both the rate at which drivers are searched and the likelihood that searches turn up contraband, we find evidence that the bar for searching black and Hispanic drivers is lower than for searching whites," they write.
While the city hasn’t crunched the numbers on who gets searched, or applied the "veil of darkness" test, all that that data is now available online, and will be for years to come.
In the interim, Zayd Atkinson and his lawyer have decided to work with the city to improve its issues with racial profiling, rather than suing.
"I just think it’s really important to follow what you think is right with your heart," says Atkinson. "For me personally it’s about what we are going to do as not only a community but as humanity as a whole. How are we going to figure out how to become more interconnected with each other?"
Drawing on his yoga studies, he envisions something that involves meditation.
This story was produced by the Mountain West News Bureau, a collaboration between Wyoming Public Media, Boise State Public Radio in Idaho, KUER in Salt Lake City and KRCC and KUNC in Colorado.