COLUMBIA — If you’ve made a habit of speeding a little, not bothering with your turn signal or rolling through stop signs, your odds of getting stopped by the Columbia Police Department are worse if you happen to be driving through the First Ward.
The highest number of traffic stops by Columbia police occur in the First Ward, almost double the number in the next highest ward.
The First Ward is also home to the largest black population in Columbia, and every year since 2000, the Missouri attorney general's office has reported this fact: Too often, when Columbia police pull a car over, a black driver is behind the wheel.
To measure how much is too much, in its report, the state compares the proportion of black drivers involved in traffic stops with their representation in the community. It's a way to make communities aware of how much more often — or not — blacks are stopped by police, given their population in a town or city.
So, if blacks are 10 percent of the population but represent 20 percent of all traffic stops, they are said to be "overrepresented." In this example, they would be two times overrepresented, which is about the figure the attorney general has come up with each year for blacks in Columbia.
Not good news for either blacks or the police — or a community — but it gets worse.
A Missourian analysis found that blacks in Columbia were overrepresented even more than the disparity cited by the attorney general.
While the state reports blacks are two times overrepresented, the Missourian found that in some specific categories of traffic stops that disparity is as much as four times.
While there is as yet no solid explanation for why such a wide racial disparity persists in traffic policing, the figures present a clear obstacle to police and community relations, especially in the First Ward.
But while the results of the attorney general’s study seem to show an unequivocal bias against blacks, the response to the report from the police, the community and researchers has been a mixed bag. The debate over what to make of the numbers, or even whether anything should be made of them, has done more to muddle the issues surrounding racial disparities in policing than to clarify them.
A Missourian study analyzed the 2012 data used for this year’s state report on racial bias in traffic policing. Several different methodologies were used to look deeper into the numbers and help clarify what they say and, just as importantly, what they don’t. Several community members helped shape the debate surrounding the issue, while experts qualified the limitations and flaws within the state’s report. Finally, the police were asked to respond to the analysis and provided a clearer view of their approach to relations with the First Ward’s black community.
Missourian profiling analysis
According to data obtained from the Police Department, blacks were two times overrepresented in traffic stops in 2012. So while blacks made up 10 percent of the driving age population last year, they represented 22.5 percent of the traffic stops made around town.
Breaking down traffic stops into subcategories, though, reveals even greater disparities for blacks.
The Missourian looked at stops according to the pre-stop justification cited by an officer as the reason he or she pulled a car over. These justifications were categorized into moving, nonmoving and investigative. Moving violations include speeding, lane violations and failing to use a turn signal, among other classics of traffic law. Nonmoving violations can include violations such as expired tags or car equipment defects, such as driving with a burned-out tail light.
Investigative stops are a special category and do not necessarily mean the driver has committed any traffic violation. In these cases, the officer indicates the driver or car is being stopped in connection with a police investigation. That may mean the driver or car meets a certain description sought by the police in connection with a crime or the officer observes the driver under suspicious circumstances.
A single stop can have more than one justification cited.
Blacks are two times overrepresented in the moving and nonmoving categories, which is consistent with their overall disparity in the state report.
But in investigative stops, they are four times overrepresented.
That means black drivers represented more than 40 percent of investigative stops made last year, again as compared with being 10 percent of the population. Put another way, the percentage of blacks stopped by police under the category of "investigative" would be appropriate for a black population four times the size of Columbia's.
That striking disproportion is mirrored in the results of traffic stops for black drivers. Post-stop, blacks are again overrepresented two times in the number of citations and warnings issued. They are four times overrepresented in the number of arrests made.
The why is in question
There were 5,796 traffic stops made in the First Ward in 2012. The next highest ward, the Third, had nearly half that number at 3,088 stops made. According to the latest census data, the First Ward has 18,212 residents, 18 percent of whom are black, the highest among the wards.
The overlap of high black population and high levels of traffic enforcement has led to questions in the community about whether there is a relationship between those two facts.
"The police are where there is crime, but is that because the police are called there or because there is actually crime or something else?" First Ward Councilman Fred Schmidt said.
But while Schmidt said policing does have a geographic focus, he was hesitant to say that focus may disadvantage First Ward blacks more than others.
“I would look at profiling as a citywide issue,” he said. "We do have behavioral and cultural problems in the (Columbia) Police Department that impact certain areas. ... It's not like the police are not courteous (in the First Ward) and courteous in other areas. I feel that discourtesy is a citywide issue."
Darrell Foster, a member of the First Ward Ambassadors, has been outspoken about his belief in the presence of racial profiling in the Police Department.
Foster said that a majority of traffic stops occur in the First Ward because the Police Department has a negative perception of the people there. He said that often this perception is based on the fact that the First Ward has a high percentage of blacks.
Matt Akins, who in 2010 founded Columbia's Citizens for Justice, a citizen-run police oversight group, looks at profiling differently from both Foster and Schmidt. He said racial profiling isn't the problem; it's geographic profiling, meaning that the Police Department patrols certain neighborhoods that are perceived as high-crime areas more closely than others.
Foster, though, said he has witnessed something more than just hot-spot or geographic policing in the First Ward.
"There are a number of good officers in the Columbia Police Department," he said. "At the same time, there is profiling, instigating and outright setting people up to be victims of the system."
The Police Department will often make judgments, he said, based on where and when people are driving, especially in the First Ward. For example, residents who are driving home from a late work shift at 2 a.m. are sometimes pulled over on Worley Street near Douglass Park, simply because they happen to be driving late at night in a so-called high-crime area. Often, he said, these people aren’t doing anything illegal, but police are assuming they are.
Schmidt said though there are national data on racial profiling, there aren't enough local data to determine if Columbia police are racially profiling.
The Missouri attorney general's office releases a racial profiling report on vehicle stops each year that looks at police stops in cities and counties throughout the state. The methodology used was established in Missouri Senate Bill No. 1053, which was passed in 2000.
Each year the report calculates a "racial disparity index," which shows the likelihood that a certain racial or ethnic group will be stopped by the police as compared to that group's proportion of the local population.
For example, if the percentage of all stops made by the police that involved white drivers was 75 percent while the percentage of the local population who are white was only 50 percent, whites would be said to be overrepresented in traffic stops. Dividing those two proportions (.75 divided by .50 equals 1.5) produces the disparity index, with values of more than one indicating overrepresentation and under one indicating underrepresentation.
Using this methodology, the Missourian calculated a 2012 disparity index for blacks of 2.24, the highest the index has been since 2004. Since reporting began in 2000, the index for blacks has never been below 1.99.
For investigative stops, the disparity index for blacks last year was 4.13.
For arrests made as a result of stops, the index for blacks was 4.36.
A flawed report?
Stark as those figures may be, police departments have long voiced concerns with the accuracy of using this index to measure racial profiling. Some crime analysts find validity in those complaints.
Researchers and the police alike complain that the disparity index's implicit goal of matching the racial proportions of traffic stops exactly with those of the population is unrealistic. These numbers as they exist in the real world are too messy to align so neatly, they say.
"Mathematically, it is nearly impossible to have all of the racial groups equal to one," Columbia Police Department crime analyst Jerry East said.
East has met with community leaders from areas, including the First Ward, and made the case that these numbers don't indicate racial profiling is taking place, though the report is ostensibly designed to measure exactly that.
Another skeptic is Richard Rosenfeld, one of the researchers who compiles the state's annual profiling report and a curator’s professor at the University of Missouri-St. Louis’ Department of Criminology and Criminal Justice.
“The way the disparity index is calculated has drawn lots of criticism,” Rosenfeld said.
Rosenfeld draws a simple analogy to illustrate the problem with the state’s current methodology:
Along an interstate is a community with a black population of 100 percent. Now imagine a white driver, low on gas or otherwise just passing through, gets stopped by officers in the area. According to the state’s methodology, which uses the demographics of the residential population, we would expect the percentage of white drivers who get tickets in the area to be zero. It would be incorrect, though, to conclude the police are racially biased against whites if some white drivers are indeed stopped in the area. This is the problem: Drivers who get stopped passing through a community where they don’t live skew local statistics that may indicate racial overrepresentation.
This methodological flaw, called the denominator problem, is well-known by Rosenfeld and other social scientists, who say residential population is not the most accurate benchmark to compare with the racial breakdown of traffic stops. Instead, they say, the driving population of an area should be used, which takes into account passers-by.
Rosenfeld published a study in 2003 in which he and other researchers modeled the estimated driving population of 92 municipalities in Missouri. In the suburbs surrounding St. Louis, the difference in the populations was so stark in their estimates as to drop the disparity index more than 80 percent in one area, from a 12 disparity index for blacks to less than 2. In Columbia, the index dropped more modestly, about 6.1 percent, from 2.14 to 2.01.
Another problem in using residential population to estimate driving population is the assumption that the percentage of the population who are over the age of 16 and are licensed drivers is the same across all races. For example, if there are fewer licensees in minority communities, any racial disparity indicated by the residential population would actually be larger among the driving population.
In academic literature, this margin for error is accounted for with state data that indicate the race of licensed drivers in an area or through a statistical method known as Quasi-Induced Exposure. This method uses the race of drivers deemed not-at-fault in two car accidents as a statistically significant sample of the driving population. The Missourian attempted to obtain both of these data sets to better represent racial disparities in the 2012 traffic stop data. Neither was available in Missouri.
So with the methodological problems of the state’s racial profiling report, is there any reason to pay attention to the disparity index at all?
“I don’t think you should ignore the disparity index but bracket it with a lot of caution,” Rosenfeld said.
While the figures cannot be exact because of methodological problems, wide disparities are still relevant and can indicate a trend.
Using the discrepancy between indexes in Rosenfeld’s study (about 6 percent) as the best available indicator of an unknown margin of error, there is still good evidence blacks are overrepresented in Columbia police’s 2012 traffic stops (more than 300 percent in investigative stops and arrests made).
'A lead balloon'
Police Chief Ken Burton says he interprets the black skew in traffic stop statistics as a natural consequence of the department’s geographic policing strategy. The center of the city is the most intense hot spot of police activity, where there is a higher percentage of minority citizens, he said.
“The likelihood of stopping a person of color is multiplied by where (the police) are. The probability is going to go up, almost exponentially,” he said.
And where the police are is driven by calls for service made from the community.
Traffic stops aren't a good index to measure racial profiling, Burton said. In most cases, he says, the race of the driver isn’t known until an officer has already made the decision to pull a car over, usually after observing a traffic infraction from a distance. “But what the disparity index tells you is that the officer did know (the driver’s race), and that’s simply not true.”
Akins, for one, isn't sold on that argument.
"I definitely think race plays a role" in traffic stops, he said. "I can be behind a cop car and see the race of the cop in the car."
Burton did concede that in the special category of investigative stops, it's more likely that an officer will identify the driver’s race in the process of looking for persons meeting a certain description. In some cases, he said, those stops are generated by the community who call the police with a vague description of a suspicious person in their area.
But for Burton, these racial skews are all an unfortunate consequence of predictive policing, which he conceded "goes over like a lead balloon" in some cities.
“The data tells you where the crimes are likely to occur, and then the police concentrate there, and you’re caught up in the net," he said. "You’re not going to like it if you’re a law-abiding citizen.”
“I’m very sensitive to the fact that someone of African-American origin would say, ‘I think the police stop me just because I’m black,’ ” he said.
Burton also contends that if there is indeed a problem with racial profiling, it's specific to individual officers. The state's report, however, measures racial profiling as a departmentwide problem, and Burton says that implies all officers are prejudiced rather than just those individuals who may be racially profiling.
Regardless, he says the report he sees every year just doesn’t give him enough information about any specific problem to act on it.
“That’s one of the flaws with the attorney general’s data: It doesn’t go far enough," he said.
According to Rosenfeld, one small change that would go a long way toward more helpful data would be asking drivers stopped for traffic violations whether they live in the area where they were stopped and including that information in the annual statistics.
“I want to hear a chorus of voices. Let’s do it right,” he said.
Digging deeper: root problems
Karen Aroesty of the Missouri/Southern Illinois Anti-Defamation League agrees. In 2010, she founded the Round Table on Bias-Based Policing to discuss a better way to analyze racial disparities in policing. The group comprises a wide range of members from law enforcement including Burton, who is an active member; Missouri Immigrant and Refugee Advocates; the American Civil Liberties Union; the National Association for the Advancement of Colored People; and various academics, among others.
"I worry about the profiling statistics," Aroesty said. "How do we get the communities to really dig deeper (into issues of racial profiling)? It's spectacularly inefficient."
Aroesty's group plans to introduce a bill for the 2014 legislative session in Missouri that will address the various methodological problems with the profiling report in its current form and propose ways to communicate the results in a way that promotes constructive conversation around the topic.
"We want the data to be effective as an educational tool and not break down a community," she said.
Attorney General Chris Koster has not yet declared formal support for the group's purpose, though a representative from the office has participated in the ongoing discussions, Aroesty said.
A representative of the attorney general's office released the following statement in response to questions about whether any plans to improve the annual report were being considered:
"The (state's) reporting method was designed more than 10 years ago and is the product of the law passed by the General Assembly in 2000. Certainly, we are willing to consider improvements in collecting information, keeping in mind the important value of year-to-year comparisons and historical data."
Despite these flaws, Aroesty said, "If these stats do anything, they keep highlighting that there is a problem."
Geography of race and policing
To supplement the attorney general’s report and explore the connection between the racial makeup of a community with traffic policing, the Missourian conducted its own geographic statistical analysis. It was designed to test whether race played an important part in predicting where traffic stops occurred in 2012.
Spatial lag regression models, a geo-statistical analysis technique, were used to model the relationship between the number of traffic stops in an area and other local variables. These included:
- racial makeup
- total population
- calls for service made to the Columbia Police Department
- traffic volume, estimated using the Missouri Department of Transportation's average daily traffic survey.
Using the Missourian's models, the number of calls for service in an area was the most consistent predictor of greater numbers of traffic stops in an area. The racial makeup of an area was not a statistically significant predictor across models.
Still the best models only explained 48 percent of the difference in the number of traffic stops between areas of Columbia. That means the other 52 percent could be explained by either chance or another variable or combination of variables not factored into the models or considered as part of the study. These could include bias, whether racial, geographic or otherwise.
While these results seem to support the police’s claim that calls for service drive high concentrations of traffic policing, they aren’t the final word on whether racial makeup of a neighborhood matters.
But even if race does not cause traffic stops, perceptions of racial bias might still be an effect of traffic policing in Columbia.
Regardless of why police are pulling someone over — whether bias plays a role — intense focus on a single community can create a sense of unfairness.
Mistrust in community
What all of this means for how the community perceives bias by Columbia police is unclear.
Schmidt pointed to a report released by the city of Columbia in March 2012 as one effect. The report showed that approval ratings for the Columbia Police Department had dipped during a seven-year period. In 2005, 81 percent of people said they were satisfied with the Police Department, compared to 69 percent in 2012.
"There's a mistrust between police and the people in the city," Schmidt said. "You can tell (from the report) that people don't trust the police."
Schmidt said that distrust means some people who feel they've been subjected to racial profiling might not even come forward with complaints. Often, he will receive emails or Facebook messages from family members or friends who tell him they know someone who might have been mistreated by the police. But, those people rarely, if ever, come forward and speak to him directly. And that makes it hard for him to go to the police with a complaint.
Akins thinks young people, especially, should be educated about the best way to make a complaint about problems with the police. Young people in the First Ward are more likely to complain to the police directly during an encounter than talk to their council representative after the fact, he said. He wondered how many people under 30 even know who their council representative is.
Even when young people do file a complaint with the Police Department, they don’t expect the police to take their complaint seriously, he said. So, to avoid seemingly inevitable disappointment, people who might have been mistreated might avoid filing a complaint in the first place.
'Starsky and Hutch' model
Despite these obstacles, there is evidence the Police Department has attempted to repair relations with the First Ward community, specifically in Douglass Park, which the department holds up as an example of community policing.
Burton praises the efforts of two police officers, James Meyer and Jamie Dowler, assigned to patrol Douglass Park last summer after a spate of criminal incidents in and around the park last spring.
Burton says despite their focus on serious, ongoing problems in the park, the two officers were given special instructions:
“I don’t want you down here enforcing typical law enforcement reaction to violations of the law. I want you to create relationships with law-abiding members of the community.”
According to Burton, that meant not writing tickets for small violations against the letter of the law. He said the program was a huge success, and the two officers — who gained the nicknames "Starsky" and "Hutch" — became accepted as members of a community. Park-goers even threw one officer a birthday party. After a general easing of criminality in the park last summer, Burton praised the new model of community policing.
Darrell Foster said the officers have indeed established good relationships with park-goers.
"It is a learning process for everyone involved. They (the officers) have been fair and consistent. They haven't come in and pinpointed anyone to harass or just arrest. They are not perfect," he said. "But, they have the attitude that they want to work at it."
Akins agrees with Foster. He said that Meyer and Dowler have created rapport with the people in Douglass Park by not trying to take a forceful, approach. The less-authoritarian approach, Akins said, has been much more effective for getting people to cooperate with police. He'd like to see officers take a similar approach to patrolling other areas in the First Ward.
But Burton is skeptical about whether such an approach can be applied to traffic policing. “The mission is different," he said. "The mission of a traffic officer is to correct errant driving behavior.”
That mission doesn’t allow for letting observed infractions slide.
But still it’s unclear whether the police’s focus on the First Ward, even if because of higher numbers of criminal infractions, brings about racial bias, meaning that blacks are subject to more general police scrutiny than whites by virtue of the communities where they live.
That bias, in turn, can cause a reaction in the community against the police that presumes every officer is harboring private bigotries.
Burton cited better communication as key to breaking that destructive cycle.
“The more educated the public is about how we do our business, the better off we’re going to be,” he said. “It has to be that routine contact that we establish with those communities where the law-abiding citizens don’t fear the police and we’re dealing just with the criminal element that we all know is there.”
Changing a culture
But what to finally make of the state's annual report?
"As much a matter of statistics, this entire thing is a matter of culture and changing culture," Aroesty said.
As part of addressing that culture, Aroesty leads annual training for officers designed to help them confront racial profiling they see or implicit biases they may hold.
The training uses some powerful examples: One program invites officers to the Holocaust Museum and Learning Center in St. Louis where they discuss how the police in Nazi Germany were used to subvert justice on the basis of racial prejudice.
The bill Aroesty hopes to push through the legislature will specifically ask for more required training for officers. Now, the state requires just three hours of racial sensitivity training every three years.
But Aroesty cautions that the responsibility for confronting racial biases in policing does not belong solely to the police.
"Community folks have to be at the table meaningfully," she said. "Otherwise, how do we address these issues?"
But until those relationships are established, it’s clear the state’s annual racial profiling reports — flawed though they may be — will continue to be a significant obstacle to both the First Ward and the police in their efforts to learn to trust one another.
Supervising editor is Katherine Reed.