Utah State University’s new “predictive policing” model might at first sound like the plot of “Minority Report.”
But the school says it’s far less dystopian than the 2002 sci-fi thriller, and much more about predicting needs like which football games might require more officers on patrol or figuring out where students are reporting the highest levels of concerns and why.
“It’s been phenomenal,” said Blair Barfuss, the recent police chief at USU. “We started having answers that I never anticipated seeing. And it’s all data-driven.”
Barfuss and Erik Christensen, an executive officer overseeing the university police department, unveiled the modeling last week during a campus public safety conference that brought together administrators from colleges across the state to learn how to improve security for students.
In recent years, Utah State has been the focus of criticism, particularly about how its police department and other campus offices — as well as Logan police — have responded to sexual assaults. The U.S. Department of Justice investigated the northern Utah school and reported in 2020 that it found “significant failures” in how the school handled complaints.
In December 2021, Utah State student Kaytriauna Flint filed a lawsuit alleging the school was continuing to protect its football players against claims of sexual assault, including in her case, even after the DOJ report. She referred to recordings of then-USU Police Chief Earl Morris and football coach Blake Anderson making derogatory comments about victims. Morris resigned, Anderson apologized and Flint has since settled her case.
That history is part of why the school brought in Barfuss to overhaul the department, though he left a few months into his position to deal with family medical concerns. Christensen, who was previously an investigator in the police department, has taken over studying how the organization can improve.
Both Barfuss and Christensen said colleges have the resources to do data studies and should be leading the way in “predictive” policing. “We have blindspots on our campus,” Christensen said. “We need to dig down into the numbers to address the issues.”
Here’s a look at three things USU is studying with its new model.
1. Studying crime hotspots on campus
When classes started this past fall, Barfuss said, USU’s police department started receiving a high volume of calls from students and faculty reporting that they’d smelled marijuana.
“It got so highly reported that we had community leaders, church leaders addressing our administration wanting answers,” Barfuss said.
Barfuss laughed as he said reports of marijuana use didn’t seem too out of the blue for a college campus. But it became more interesting, he said, when the department decided to map the reports.
The department expected to see hotspots on the 98 student housing buildings on campus — the most on-campus housing of any university in the state.
But none of the reports came from there. Most often, the odor was reported around classroom and research buildings on the north side of campus, with a few additional reports from the south end.
After a few phone calls, Barfuss and Christensen figured out that, as an agriculture-focused school, there’s a research lab at USU growing cannabis to study its uses. Researchers harvested the plant at the same time classes were starting. Then they composted it and put it in flower beds around the north side of the campus.
They also learned from the grounds crew that there are families of skunks that live on both the north and south side of campus that likely led to some of the reports, as well.
Digging into the data, Christensen said, allowed the department to find the real cause of the spike in reports. And they told the campus about what was happening — rather than just brushing it off on what they were biased to believe was the reason, he added.
The department intends to use hotspot mapping to look at more serious issues in the future, including assaults. If there are certain places where assaults are more common — such as the dorms — they’ll look at talking to students there about prevention and reporting.
They would also study whether there are areas without enough lighting, for instance, or whether students aren’t aware that they can request a police escort. This could also help to determine if there is a repeat offender, which was an issue with the Torrey Green and the fraternity assault cases.
2. How to police big events, such as football games — and save money in the process
Christensen made his own algorithm to study large campus events, starting with football games.
The equation looks at attendance of past games for six years by competitors, the number of offenses that were reported to police, the time a game started and even the weather to determine the likelihood of what kind of crimes might happen at a future similar game.
At games with higher temperatures, for instance, there were more medical calls for attendees experiencing heat fatigue. So Christensen used the model to staff more medical officers for upcoming games where the temperature would be in the same range.
At games against certain competitors, there tended to be more alcohol violations, which required more officers to make arrests. The games against Brigham Young University were the lowest for that, said Barfuss with a chuckle, noting that BYU’s Honor Code prohibits students from drinking. But those games, which had stronger rivalries between fans, tended to have more disputes called in.
Christensen used the algorithm to predict what might happen at the Air Force Academy game last fall. It predicted 1.5 medical 911 calls. There were two during the game, and staff had planned accordingly.
The data modeling also showed more reports were called into police at the end of the first quarter and middle of the second quarter of a game. So Barfuss and Christensen started to change how many officers they had on shift to match that — rather than having police staff sit around at the start of the game with nothing to do, which costs the university money in overtime costs.
It’s referred to as surge staffing, which is also used by FedEx and UPS, Christensen said. For five homes games this past year, USU saved about $10,000 using this model to better staff games based on need.
The modeling can also be used to look at events like the Howl, the large Halloween party held annually at USU. In years past, there has been a spike in assault and groping reports at the event.
USU has 3,500 hours of special events each year, Christensen noted.
He can also use the data to determine daily staffing; most calls come in between 2 p.m. to 6 p.m. This is particularly helpful, Barfuss said, as police departments — especially those on college campuses — struggle to be able to afford enough officers.
On 4.6% of days, Christensen found, callers were waiting an average of almost two hours for help (with priority given to more serious cases, which had shorter times). But by shifting staff, those delays can be addressed. And the data can help show the university’s administration the need to fund more officers.
3. Planning for the likelihood of certain crimes, like drug arrests
Like with football games, the data analysis by Christensen can take the number of crimes for years in the past and predict their growth or decline. For this year, it predicted 3,658 incidents, which is a drop from the 4,406 last year.
Christensen started examining types of crime. He studied drug arrests, building a model for how many USU might expect for 2022. After ruling out outlier years, like 2020 where there were fewer kids on campus because of the COVID-19 pandemic, the equation predicted there would likely be 16.5 drug arrests by USU police last year.
By the start of December last year, they had 12, Christensen said, and he was worried the analysis was off. He emphasized that the department wouldn’t try to get more arrests just to meet the prediction, but he did keep more staff on that month to meet the expected demand in arrests.
Then, four more calls came in at the end of the semester from kids partying as they celebrated graduation. The department ended with 16 drug arrests for the year, he said: “Now that’s a good model.”