Miltonette Craig began researching racial profiling by police after she witnessed similar inequity while working in the court system. Craig had obtained her law degree in Georgia and returned home to Miami where she was employed as a law librarian in a prosecutor’s office and later as a judicial assistant.

“I could see the disparities in who was getting arrested and who ended up having to take plea deals because they couldn’t afford attorneys,” Craig said. “That really piqued my interest. And when I decided to start studying criminology and criminal justice, I ventured to the criminal justice side, so what we do after crime happens, and policing is always on the forefront of current events. Definitely, the events in Ferguson, Missouri, gave me a little bit of tunnel vision. And so that’s a large contributing factor to why I focus on racial bias and policing.”

Craig earned a Ph.D. from Florida State University and arrived in 2018 at Illinois State, where she is an assistant professor in the Department of Criminal Justice Sciences. In January, Craig was recognized as a 2020 Emerging Scholar by the journal Diverse: Issues in Higher Education.

In the following Q&A, Craig talks about her research, specifically her dissertation. She examined vehicle stop data reported by Missouri law enforcement agencies from 2001–2007 and the police departments’ responses to those findings. Craig also spoke about an article she published with her Florida State colleagues in 2017 in the Journal of Criminal Justice titled “Thinking Fast, Not Slow: How Cognitive Biases May Contribute to Racial Disparities in the Use of Force in Police-Citizen Encounters.”

The interview has been edited for brevity and clarity.

According to your dissertation research, Missouri police departments stopped black drivers 45 percent more often than white drivers and 39 percent more often than Hispanic drivers. At the same time, police found contraband in the vehicles of white and Hispanic drivers more often than on black drivers. What does this tell us?

If our benchmark numbers are proper, it’s that the resources being used to stop and search are not being done efficiently. So it’s probably biased influence that these drivers we have in mind are the typical suspect. But clearly, you’re finding contraband on them in lesser numbers, so you need to change your approach and, maybe, you need to be searching more white drivers if your goal is to get contraband off of the street.

We’ve seen a lot of departments where their patterns look that way. It’s almost nonsensical, because you’re searching and stopping, and you’re doing this for safety and law enforcement and all of that. But the outcome isn’t helping, right? So, you’re not getting (contraband) off the street, as you said you were. Then that’s when it looks targeted. Because of the inability of just raw numbers to explain things, we can’t for certain say this is profiling. But a lot of studies say if you have that contraband hit rate inequality, then it’s probably going to be something based on bias and discrimination. We just can’t officially label it that unless an officer says, “Oh, yeah, that’s why I do it.” And it’s rare that they would do something like that.

In your dissertation you mentioned that the police seem defensive in their written responses. Did you buy their explanations?

I think a lot of them were actually legitimate. There’s a huge issue with racial profiling research called the denominator problem. In order to calculate whether different groups are being under- or over-represented, you have to do the calculation for the number of people stopped in reference to their presence in the population.

However, what population count do you use? If you’re just looking at the census for a particular area, it’s not truly reflective of the drive-in population, because we drive to places where we don’t live. Some of the agencies mentioned, “We have a large mall here. We have the community college here. So we get more out-of-jurisdiction traffic.” But it’s hard to determine the driving population as well, because it’s not the same every single day. You would have to have someone at every intersection in your jurisdiction counting. “OK, this is how many white drivers there were. This is how many black drivers there were.” And that’s clearly not feasible or a good use of time and resources.

Those are valid concerns, but the overall tone of their responses, like the defensiveness and other things, showed unconscious biases. It wasn’t just about, “We don’t have the proper benchmark.” There would be things like, “Well, we are over-stopping Hispanic drivers in this area, because a lot of them are here illegally, and they can’t read English and understand the traffic signs.” It’s like, oh, I didn’t expect you to put that down in an official government document like that. So some of the things were objectionable, and it was just trying to see what method they use to explain why the numbers look the way they do.

Do you find that these kinds of reports detailing racial profiling by police are useful in marshaling resources to pressure police to change their policies?

I think that’s a major shortcoming of the Missouri data collection initiative. So it’s mandated by law that you have to collect this information, report it to the attorney general. But that’s been going on since 2000, and they have not used the data in any significant way. Are they using it to inform policy? Are they using it to investigate particular departments that seem more problematic than others? The data is just sitting there.

I think it should be used in that way. Let’s look at these patterns. Let’s try to have focus groups and figure out why the patterns happen. What are we doing that might be unconsciously contributing to these disparities? Should we look more at trying to implement driver safety programs that would decrease the number of stops we have to make? They’re not using the data at all. And so that is very problematic.

When you talk about bias in these cases, are you talking about implicit or explicit bias?

Most often it is implicit. I don’t think a large majority of officers go out with the intent to discriminate against minorities. But because of the messages that we receive in the media, regular television shows, and the news, we always see this typical subject—they call it the “symbolic assailant.” And if you have it in your mind that this person is the likely offender, that’s the automatic place that you go to. So even if you thought you were acting in good faith, you’re still being influenced by the stereotypes. Even with training, you’re still susceptible to it. So there needs to be more focus on that: How do we unpaint that picture so that we can be more fair in how the laws are applied?

Since then, have you done any other research involving police stops and racial bias?

The next step using that data is comparing before and after. So I would be looking at not only the tone of the responses, but the language being used, “post-racial America” after Barack Obama’s election and before. I would be looking to see if language is different. Is it more defensive? Do they feel more obligated later on to paint themselves as fair? One of my mentors said, “You should probably continue to look at that and look at the post-Trump administration to see if there are fluctuations in it.”

I definitely like to look at the qualitative end of things, because they give us more of the why, as opposed to the what and the how. But at some point in the future, I do want to conduct focus groups with officers just to see, “How did you get to the point where you are now? Do you believe that you operate with bias? If so, do you believe it’s justified? And why is that?”

Police officer standing next to car pulled over

In the other study, you examine how cognitive biases may contribute to racial disparities and the use of force by police. Please define cognitive bias for me.

Cognitive biases can be neutral, positive, or negative. We have a limited amount of time to process information and limited means to do so. And in order for us to get things done in an efficient manner, we have to make these shortcuts. So from our experience, this symbol means X, and from our experience, this symbol means Y. But in that shortcutting process, we don’t have the full picture. It’s going to eliminate important parts of that image or person or a place. Cognitively, you just don’t have the full picture and you’re making an assessment without all of the pertinent information.

In the study, you talk not only about the cognitive biases of the police, but also about the cognitive biases of the people they are encountering. How would the cognitive bias work with somebody who is being approached by police?

A lot of people do not have firsthand experience with law enforcement. A lot of it is from vicarious experience. So you heard it in the news, or your best friend was pulled over. Because historically, the police force and the black community haven’t meshed very well—they have their history in slave patrols. And generation after generation, there’s not going to be a good relay of facts about the police department.

If that’s what you have been exposed to, from stories with friends and families that police officers aren’t on our side, or the only time you see them in your neighborhood is when they’re coming here to arrest someone, not to respond to a call for service, you’re going to then develop a negative attitude as well. And you’re in this encounter, and maybe the officer is doing everything they’re supposed to be doing and not showing any bias. But because you are coming into the situation feeling as though you’re being targeted, or as though you’re going to be treated unfairly, then things can escalate.

Both parties are there thinking the worst of the other. That’s where we see the catastrophic outcome.

In the study about the cognitive biases, what was your and your colleagues’ conclusion?

What we concluded with is trying to almost map out how these decisions are made cognitively and try to get people to think about thinking. We don’t put a lot of attention on the thought process. We just make decisions and hope that they are effective and efficient. When you can interpret multiple issues that may influence this event, then you’re better able to expect them and manage them when they happen.

So yes, citizens need to know either through some kind of programming or information session, that, yes, there is friction between these two groups. But the ultimate goal is to try to keep you safe, your community together, so on and so forth. You have to, at some point, give the police department a chance, if they are doing what they’re supposed to be doing. Call them out for the illegal, unjust things.

Is there evidence-based training that could help with cognitive bias?

Some of the training programs now that police departments are using do talk about the unconscious bias issue, how not to stereotype people as symbolic assailants, and things of that nature. But the effectiveness of the programming is what still needs to be investigated on a large scale.

Do you think the police forces, in general, will change?

I would like to be optimistic about it. I think a large contributing factor is our political climate. When it is as divisive as it is now, I don’t even see what the incentive would be for a police department to say, “Let’s do better.” Because if no one is going to take away their funding or punish them or fire them, then what is the push other than a moral obligation to do better? I feel as if we get to a place where we’re more cohesive as a nation, then maybe there will be a heavier burden placed on these authoritative institutions like police departments.

But now, I’m largely pessimistic about it happening within the next couple of years. It probably will be quite awhile before we can see broad equity in how interactions happen between officers and the public.