Who’s most likely to commit crimes in Wichita? Police say math will tell them

Wichita police are using a different way to identify people most likely to commit a crime. And those on the top offender list have no way of knowing they’re on there.

Police adopted the top offender model, which uses a mathematical equation to determine a person’s likelihood of committing a crime, in March. The predictive policing method has been used by some other law enforcement agencies for years.

The program launched as city officials worked out a settlement over a multi-year lawsuit that called the Wichita Police Department’s gang list unconstitutional.

“There is a lot of subjective human input into that (gang) list,” Chief Joseph Sullivan said during a Wichita Police Department town hall meeting on June 8, the first time police spoke to the public about the top offender model.

“This completely eliminates all of that. I cannot give somebody a high score in any way shape or form. It’s simply based on the math and nothing else. I can’t influence it.”

The mathematical formula is not based on a person’s arrests or convictions. Instead, it assigns 35% for a person’s criminal history based on Wichita police cases, 35% from intel police receive and 30% for a personal connection in Wichita Police Department cases to other people who also might be committing crimes.

Victims in a case could be included, but with less weight than someone who was arrested, police said.

The top offenders list is being called a new gang list by some. It’s been criticized as a way for cases to constantly be made against people on the list if they are under surveillance or live in areas with greater police presence.

“It seems they just broadened their gang list to include the whole community … that’s too invasive,” said Aujanae Bennett, president of Northeast Millair Neighborhood Association, who attended the town hall meeting. “Why would you retain such information, why would you create such a list if you weren’t going to watch these people?”

Police also told the roughly 25 people in attendance at the town hall meeting that they would soon launch a second program called risk terrain modeling, which uses a mathematical equation to figure out areas where crime is most likely to happen based on neighborhood factors, such as vacant buildings and liquor stores.

Who’s on the list?

People have no way of knowing if they are on the top offender list. Police would not say what personal information is included in the list.

Wichita Deputy Chief Paul Duff told those at the town hall that people make the top offender list only if they commit a crime.

After the meeting though, when asked if a victim could be included in the list, Capt. Aaron Moses said a victim could be included based on connection and criminal history but that would have “less influence in the score calculation of the criminal history than the role status of (a) suspect or arrestee.”

A decay rate in the formula ensures people who don’t commit more crimes won’t stay on the list forever, police said. The longest would be five to seven years for a homicide and “just a few months” for minor crimes, Wichita police crime analyst Geoffrey Vail said.

ACLU staff attorney Kunyu Ching, however, said it’s easy to keep people involved in police cases in overpoliced areas, or if police are following a certain person, making it a “skewed sample.”

“I also think that in general there is a heavy potential here for discrimination against protected classes or demographics,” she said. “I think it’s common knowledge these sorts of algorithms can produce bias, including racial bias. I think it’s also well known that Black and brown communities and low-income communities are policed far more heavily than others and that combined, these circumstances can snowball and become sort of a self-enforcing feedback loop that results in overpolicing certain areas or demographics.”

Police are now using the top offender model only to find people who might be involved in a shooting. It could be expanded to include other crimes, police said. Every week, officers receive an updated list with the Top 12 offenders.

When asked what happens to people on that list and if they’ll be under surveillance more, Moses said, “No. Not necessarily.”

Moses said police could use that list to “prioritize individuals … when they are involved in less severe incidents that otherwise wouldn’t garner much attention.”

“If we have a hotspot for crime, our analysts can use this mathematical equation to identify people who are believed to be involved in that activity or most likely to be involved in that type of activity in the near future so that we can provide that information to our patrol officers to have some data-driven enforcement information,” Moses said.

Top offender lists, gang lists have both led to lawsuits

Programs like the top offender model and Wichita’s gang list have led to lawsuits.

A program similar to the top offender model in Pasco County, Florida, led to ongoing federal lawsuits after a Tampa Bay Times investigation found that people on the list had been subjected to more “monitoring and harassment by sheriff’s deputies.”

Wichita police’s gang list was challenged in court in 2021 by Kansas Appleseed and the ACLU of Kansas lawyers on behalf of Progeny, a nonprofit juvenile justice organization. People could have ended up on the list just by being seen with known gang members.

In April, Wichita reached a settlement in its gang list case. A federal judge is expected to sign off on the agreement on Aug. 23.

Here is what that settlement entails:

• Wichita paying $550,000 in court and legal fees

• Wichita paying $75,000 annually for three years to a judge-appointed special master to oversee how the Wichita Police Department uses the list

• A mechanism for people to see if they are on the list and the ability to appeal being on the list

Matthew Guariglia, a senior policy analyst at civil liberties and data security organization Electronic Frontier Foundation, said “absolutely yes,” he thinks the top offender model is Wichita’s new gang list.

“Mathematics provides a veneer of legitimacy,” he said.

Sullivan says this way removes any biases.

Ching said that without more information she couldn’t say if the top offender model is the new gang list.

“But I would certainly hope that the city and WPD specifically would have learned something from the gang list lawsuit,” she said. “I think policing could be a lot more effective with a good amount of transparency and honesty with the public.”

WPD did not tell the ACLU about the new program, she said.

Bennett, the neighborhood president who called the program the “new gang list,” said people should know if they are on the list and when they get off.

“I’d like to see what is the racial profiling of that list? Is it 50% Black, 50% white? We know that’s not the case, but I’d like to know … because that would help us hold them accountable for overpolicing,” she said. “If they want that system they need to make it work for everybody,” she said. “If they can’t do that, they don’t need to do it at all. I really don’t think they should do it at all but if you are going to do it, it needs to have some parameters.”

Data has shown that minorities have been disproportionately affected by both the gang list and traffic stops.

The city’s gang list, which is still in use, has more than 5,200 people on it, with 60% of them Black, 25% Hispanic and 6% white. Roughly 3,500 people will be removed from the list as part of the settlement.

Citation data also shows that Wichita police have given minorities disproportionately more tickets . Black people make up 11% of the Wichita population, but they received 18% of all traffic tickets between April 19, 2021 and December 31, 2022, according to Wichita police data. Hispanic people were twice as likely as white people to receive a non-moving violation, the data shows, and Black people were nearly 3.5 times as likely.

Modeling used elsewhere

The top offender model program Wichita police implemented is not new to law enforcement.

Nearly 200 law enforcement agencies across the country use predictive policing techniques, according to the Electronic Frontier Foundation .

That number is probably low. The foundation uses crowd sourcing to compile its information, which means Wichita’s new program hasn’t been recorded yet.

Wichita based its top offender model on a program the Milwaukee Police Department started in 2016. Wichita uses a similar mathematical equation.

Milwaukee used it as a way to curb a rise in teens involved in motor vehicle theft and robberies, but later expanded it to include gun violence.

A 2018 Milwaukee police presentation to other law enforcement officials talked about “targeted police contacts” and “preemptive targeted contacts” of people on the list.

“So the people we put on this list, we want to focus on, they’re the ones causing problems,” a law enforcement official said during the two-hour-plus presentation. “Let’s provide officers with locations of … who they associate with, current criminal activities.”

That slide in the presentation included a card with mock information for a person on the list that consisted of things like their school or where they work, any gang affiliation, a picture of them, address and any intelligence on them.

Milwaukee police crime analyst Garrett Knuth said during the 2018 meeting that officers could use the list to be proactive.

“There’s surveillance, so sitting outside an active offender’s house, you may be able to observe some stolen vehicles,” he said. “Be creative, neighborhood canvass. Don’t knock on the … offender’s door, knock on the neighbor’s door and ask if they have a known auto thief in the neighborhood. Maybe you can get some intelligence that way. Home visits. You have to be careful on these, because some parents just get upset with the police. If we keep knocking on doors we can burn bridges. But, at the same time, you want to do home visits to reach out to the youth … for two parts: one is, ‘hey we have our eye on you,’ but also, two, to offer resources to them.

MPD’s media relations department said the program “played a big role” in cutting down on vehicle thefts during the Kia/Hyundai theft trend.

“In 2022, our most recent year of completed data, 52% of (people on the list) were criminally involved in at least 1 vehicle theft, robbery, or firearm offense while they were published on the (list),” the agency said in an email. “This shows us that the Program does a good job of correctly identifying the key offenders.”

The department did not respond to questions about whether people on the list are notified or if the program has faced any lawsuits.

The Pasco County Sheriff’s Office in Florida touted drops in property crime under its modeling program since it started in 2011.

“But Pasco’s drop in property crimes was similar to the decline in the seven-largest nearby police jurisdictions,” according to the Tampa Bay Times . “Over the same time period, violent crime increased only in Pasco.”

Former Pasco County Sheriff’s Cpl. Royce Rodgers told The Tampa Bay Times that they would find violations for people on the list when they or their family wouldn’t cooperate.

“We would literally go out there and take a tape measure and measure the grass if somebody didn’t want to cooperate with us,” he told The Tampa Bay Times.

The sheriff’s office sent the Times a 30-plus page statement defending its program, saying other departments used similar techniques and accused the paper of labeling “basic law enforcement functions” as harassment.

The implementation of the program in Pasco County has led to its ongoing federal lawsuit.

Give it a year ‘to see what happens’

The town hall meeting was held in Council Member Brandon Johnson’s district. Police have held other town hall meetings in other districts and have more planned.

Johnson, who advocated for reforming the gang list even before being elected in 2017, said people should have due process of knowing if they are on the top offender list and appealing whether they should be.

“I think everyone is right to have their own suspicions about it,” he said. “But I think we also need to give it a healthy one-year period to see what happens from it and if it is similar (to the gang list), we have work to do.”

Johnson said, after a year’s time, he’d like to see demographic and location data to see if the list is targeting certain groups or areas.

He said people need to hold police accountable if this is misused.

“Anybody who was concerned about that with the original gang (list) has a right to be concerned about that with this,” he said. “The proof is going to be in the pudding. If we are actually targeting people and following them … let me know, I will address it … People have to hold us accountable … I don’t tell anyone that they shouldn’t be concerned, because we’ve seen things in the past that we said weren’t going to happen and they did.”