How to actually stop police brutality, according to science

An aerial view shows people gathering to pay tribute at a makeshift memorial in honor of George Floyd, on June 3, 2020 in Minneapolis, Minnesota.

(Image credit: CHANDAN KHANNA/AFP via Getty Images)

Cities across the U.S. have been rocked by nightly protests against police brutality following the May 25 killing of a Black Minneapolis man named George Floyd by a White police officer.

And as videos proliferate of police arresting or tear gassing seemingly peaceful protestors, the issues raised by the protestors seem more insurmountable than ever. But researchers and activists say that solutions are no mystery: Evidence-based changes to policy around policing can reduce deaths at the hands of the police. These steps alone can't end racism overnight or erase the myriad inequalities in American society, but they can save lives.

There is no comprehensive government clearinghouse for data on police killings or police use of force. After the killing of Michael Brown in Ferguson, Missouri, in 2014, several private and nonprofit groups began keeping their own databases. These include Mapping Police Violence, an effort led by data scientist and activist Samuel Sinyangwe, Fatal Encounters, a catalog by journalist D. Brian Burghart, and efforts like the Washington Post's Fatal Force database.

Thanks to databases like these, it's clear that Black people are killed at a disproportionate rate by police officers, making up 24% of deaths despite being only 13% of the population, according to Mapping Police Violence. But the databases rely on media reports of deaths, not police department, city, state or government data, for the simple reason that many police departments are not forthcoming with this information.

Even when the government does keep data, it's incomplete and often held on laughably out-of-date technology. In the summer 2019, Delehanty embarked on an effort to get raw data from the Federal Bureau of Investigation's Uniform Crime Reporting Database. The email provided by the FBI for researchers to request data bounced back. The phone number for researchers led to a phone tree that automatically hung up after Delehanty picked the academic option. He finally reached a person by using the field office's media line, only to learn that the only way to get the data was by mail, on a CD (compact disk). After a few weeks of waiting, the CD arrived and Delehanty dug out a computer that still had a CD-ROM drive. The data was in an old, rarely-seen format (a fixed-width delimited text file) without the necessary file that would automatically define the data columns. It took days to define the columns by hand, Delehanty said.

Sometimes, incompetent data management by the government means that information just doesn't exist. Edward Lawson, Jr., now a data analytics researcher for the state government of South Carolina, once tried to find out from the Defense Logistics Agency, part of the Department of Defense, how much military equipment was being sent to police departments around the country. He learned that prior to mid-2014, the agency had simply been updating each quarter's information in the same document, erasing and rewriting whatever inventory had been transferred the previous quarter.

Police department data should be accessible through the Freedom of Information Act (FOIA), which allows citizens to request records from public agencies. But FOIA requests often come up empty, in many cases because police decide they simply do not want their department's data scrutinized. On Twitter, one data scientist who used to work on police use-of-force research wrote that some departments are forthcoming. Others ignore requests, deny them summarily or ask for enormous fees — such as a deposit of $1 million — to release records.

For decades, police departments have been gradually adopting more and more gear from the U.S. military. Departments get this gear in a variety of ways, but one common route is the 1033 program, which provides free surplus military gear to departments for the cost of shipping. Some of this gear is innocuous, Delahanty told Live Science — filing cabinets, gloves, binoculars and other run-of-the-mill supplies that departments would otherwise have to buy on their own. But departments have also received equipment such as grenade launchers, bayonets and mine-resistant ambush-protected vehicles (MRAPS), which are military trucks designed to take blows from improvised explosive devices in Afghanistan and Iraq.

Both Delehanty and Lawson have found that police departments with more military equipment from the 1033 program kill more people. In a paper published in Political Research Quarterly in 2018, Lawson and his colleagues found that in all 50 states between 2014 and 2016, the number of police-involved deaths rose with militarization, as measured by the value of the equipment sent to a department via the 1033 program, even after controlling for factors such as population numbers, poverty, race and violent crime. In 2017, Delehanty and colleagues reported in the journal Research & Politics that in four states where they had records (Connecticut, Maine, Nevada and New Hampshire), military equipment via the 1033 program was linked with more killings by police. In a given year, a department with no 1033 requisitions could expect 0.287 killings of suspects, on average, Delehanty found; those with the max expenditure could expect 0.656 killings, more than twice as many.

It's likely that departments with a militaristic, us-versus-them mindset seek out more military equipment, Lawson said. But Delehanty's findings hint that the cycle can feed on itself, with more military equipment encouraging a more violent force. By comparing departments over time, he and his colleagues found that the annual change in military equipment could predict a department's suspect deaths in the next year. A department with no new equipment in a year could expect 0.068 fewer suspect deaths in the following 365 days. A department with the most new requisitions could expect 0.188 more deaths. The researchers even found a similar increase in police killings of dogs, suggesting that cops weren't necessarily gearing up for big, casualty-heavy raids with their requisitions. They were simply becoming more violent in general.

Training is often cited as a way to reduce racial biases among police officers and encourage de-escalation. Some training methods have evidence to back them up. For example, training in procedural justice, which focuses on fairness, was shown in one randomized experiment to reduce police officers' likelihood of ending encounters with arrests or using force, according to a 2018 study published in the journal Criminology & Public Policy. But training is a nebulous concept with little oversight, and departments don't necessarily turn to evidence-based programs. In 2017, for example, Fox 9 reported that the St. Paul Police Department's "main attraction" in its annual equity training was watching the children's movie "Zootopia." There are also questions about the efficacy of methods like implicit bias training, in part because of a lack of standards for these training methods and in part because the lessons may not translate to stressful circumstances, as The Atlantic reported in 2017.

1. Requiring officers to de-escalate before using force;
2. Using guidelines defining the types of force that can be used to respond to specific situations;
3. Restricting or banning chokeholds and strangleholds;
4. Requiring a verbal warning before using deadly force;
5. Prohibiting officers from shooting at moving vehicles except in extreme circumstances;
6. Requiring officers to exhaust other options before resorting to deadly force;
7. Establishing a duty by officers to intervene if one of their colleagues is using excessive force;
8. Requiring officers to report all uses of force or attempted use of force.

Police union contracts are also associated with police violence, mainly because contracts can be written to make it very difficult to fire or discipline officers for misconduct. A 2018 thesis by Oxford University graduate student Abdul Rad found that U.S. cities with more police protections had higher rates of police abuse, even when controlling for variables like racial demographics and crime rates.

In some cases, drastic measures are needed. After a major police-corruption scandal in 2010 and persistently high crime in Camden, New Jersey, the city decided to entirely disband its police department in 2013, starting a new department with an emphasis on community policing. This means policies are put in place to reduce the use of force and increase accountability, as well as to build trust between police and the community, according to CityLab.

Research shows that more comprehensive investments in communities pay dividends. A 2017 study published in the journal American Sociological Review found that across more than 250 cities, a greater number of nonprofit organizations was linked to declines in crime. For every 10 organizations in a city of 100,000 people, the murder rate dropped 9% and the violent crime rate went down 6%, the researchers reported. A similar study published in the journal Urban Affairs Review this year used Denver as an example and found that a higher density of nonprofits in an area was linked to lower crime, even after controlling for demographics and other factors.

Police are often the first to respond when someone with mental health issues is in crisis, and some evidence suggests that as many as 1 in 4 of those killed by police are mentally ill. , So some police departments have launched Crisis Intervention Teams (CITs), which consist of specially trained officers who work to get mentally ill individuals into treatment rather than escalating into a potentially violent encounter. The first of these programs, in Memphis, started in the late 1980s after police killed a mentally ill Black man who was cutting himself with a butcher knife. Research on CITs is challenging because different departments commit different levels of training and effort into their programs; but some evidence suggests that having a CIT in place results in less frequent use of SWAT (special weapons and tactics) teams, according to a 2008 review.

Some places have gone even further, taking police out of the mental-health response equation. In Eugene, Oregon, a 911 diversion program called CAHOOTS redirects calls about mental health-related issues to a team of mental health professionals and medics, rather than to police. Program staff also reach out to the city's homeless, giving out supplies and referring people to medical care, according to The Bend Bulletin.

Independent and civilian oversight of police departments can go a long way toward reducing bad behavior. A 2015 study in the journal Police Quarterly found that departments that had citizen complaint investigations reviewed by an outside citizen agency were more likely to find that the complaints had merit, rather than dismissing them without consequence. (The study also found that Black complainants were more likely than those of other races to have their complaints dismissed.)

Another form of oversight involves citizen watchdogs. This aspect of policing the police has grown organically with the rise of video-enabled smartphones. Interestingly, research suggests that wearing body cams does not reduce police violence directly. But knowing that police officers could be wearing body cams but are choosing not to activate them might alter the public's opinion of police behavior, said Fabian Neuner, a political scientist at Arizona State University.

Meanwhile, having bystander video of police killing suspects appears to be altering the conversation on police racial bias and brutality. After the Michael Brown Jr. shooting in 2014, Neuner and his colleagues Hakeem Jefferson and Josh Pasek found a wide gap between Black people and White people in America as to whether the officer should have been charged and even over basic facts of the case. For example, 91% of Black Americans in the study thought the officer probably or definitely should have been charged, compared with 42% of White Americans. Meanwhile, 23% of White respondents thought Brown had a weapon, compared with 4.4% of Black respondents.

Stephanie Pappas is a contributing writer for Live Science, covering topics ranging from geoscience to archaeology to the human brain and behavior. She was previously a senior writer for Live Science but is now a freelancer based in Denver, Colorado, and regularly contributes to Scientific American and The Monitor, the monthly magazine of the American Psychological Association. Stephanie received a bachelor's degree in psychology from the University of South Carolina and a graduate certificate in science communication from the University of California, Santa Cruz.

How to actually stop police brutality, according to science

1. Track the problem

2. Demilitarize

3. Change police culture

4. Invest in alternatives

5. Instill oversight