资源环境科技发展态势分析平台

There are myriad policies under discussion—banning chokeholds, mandating body cameras, even dismantling existing police departments. What does research tell us works to reduce police violence and improve public safety?

John S. Hollywood: The biggest case of abolishing a police department altogether is the City of Camden, New Jersey, which replaced it with a department under Camden County. They got positive results, including an overall crime reduction of more than a third.

It basically reestablished the department, who is doing the policing, how the organization works, and the strategy. And then the county and city governments worked with the community over several years to figure out how policing should work in Camden.

In terms of body-worn cameras, there's been a lot of variable returns, but they can be very effective in deterring attacks on officers and reducing the use of force. A main factor is whether or not officers need to keep their cameras on whenever they're doing law enforcement engagements with the community and that they let people know that they're being recorded from the second they come up on the scene.

As one officer put it, if everybody knows in advance they're on camera, they tend to behave a lot better. Whereas if an officer goes to turn on their camera in the midst of an escalating situation, that can make things worse.

Dulani Woods: Body cameras certainly have the potential to change behavior. But there are also lots of downstream effects. Police or agencies need policies for storing video. They have to know when they can share it or release it to the public and may have to redact faces. It's a bit of a double-edged sword. And then courts have difficulty receiving, displaying, and archiving that video in certain formats.

What about some of the policies pushed by reform advocates, such as Campaign Zero and its #8CantWait initiative?

Meagan Cahill: There are a number of policies that research shows are associated with fewer instances of officers using force, including banning the use of chokeholds and other methods of restraint. And there are eight policies that have been shown to be associated with fewer officer-involved shootings, so that's why there is a push now by activists to have police adopt these policies. Activists certainly are producing a lot of awareness, but I think we're going to need to push more on implementation, not just having policies on the books.

Bob Harrison: I don't know that police will need to be pushed too hard, either. A number of chiefs have already put out infographics to show they're complying with seven and are working on the eighth. Cultural change, however, is much more difficult. Working against it is the inordinate number of law enforcement agencies—more than 15,000 at the state, local, and federal level. We're not dealing with a monolith, but a bunch of small containers.

There's also a lot of discussion about how some social problems shouldn't necessarily be handled by police.

Cahill: There are a number of social challenges or issues that could be handled by agencies that are not the police.

For instance, homelessness can certainly be addressed by social service agencies without police involvement. Also, truancy. Prevention programs for early elementary school children can be very effective in terms of preventing future delinquency and criminality.

Police don't have to be excluded, but they shouldn't have to lead these efforts. The key is bringing others into the fold to address the multitude of problems facing a city.

Harrison: One of the largest mental health hospitals in the United States is the L.A. County Jails. Mental health issues, as well as drug use or drug abuse issues, are key preconditions to criminal behavior. But police are not trained or oriented to be mental health interventionists per se, other than through ad hoc training for crisis intervention or domestic violence.

So communities may be better served if mental health and social services professionals can deal with the preconditions rather than having police deal with the symptoms. I think that would clear away some of the issues that become the most contentious and difficult for the police.

We're hearing a lot about “defund the police.” What does that phrase mean, and how are police and policymakers reacting?

Woods: It's hard to find a definition that everybody would agree to. Some people have been interested in abolishing police agencies entirely. Others have talked about removing different functions, kind of like Bob was just discussing.

I've been surprised how much police executives in recent interviews have been willing to consider this. A number of them have been interested in handing off functions that historically weren't handled by police. Some have been willing to trade a certain number of officers or some funding in order to allow other parts of government to focus more on homelessness or mental health and allow the police to focus less on it.

Another charge is about systemic racism in policing. Define what that means to you and where you think we stand on the issue.

Hollywood: One of the things hampering the public debate is disagreement on that term.

When some police executives—current and former—are asked about systemic racism in interviews, they make clear that they're interpreting it to mean that officers are white supremacists or members of Klan or neo-Nazi organizations, or that they're actively enforcing laws intended to support racial suppression.

The other definition—which I think is more widely used by the critics of police— is that that there are major disparities in different public safety outcomes, and they need to be addressed. We can start with black Americans are shot and killed twice as often or more by police as whites. There are also disparities in terms of solving homicides and other major crimes for which African Americans are victims.

Certainly another thread is the history of policing, including the development of slave patrols and police participation in other forms of racial suppression. A few years ago, the president of the International Association of Chiefs of Police apologized on behalf of the whole profession for policing's historic involvement in racial suppression. At the same time, the president then went on to say that it is not fair to blame our officers today.

Cahill: To have police be willing to make changes that are necessary, we need them to understand that systemic racism doesn't mean that individuals are racist. It means they work within a system set up to perpetuate racist policies and approaches for centuries.

I would also argue that systemic racism carries throughout the entire criminal justice system. We can't just talk about police. We have to talk about what happens in courts, in corrections.

What else needs to be better understood to reform policing in the United States?

Hollywood: For five or six years we've assembled expert police panels for the Department of Justice and researchers to identify their science and technology needs. What's been most striking to hear is this: We need better strategies. We need better management. We need help with doing change management. We need help changing the culture. This has been the main story, not the classic science and technology needs.

Cahill: There are also huge gaps in data collection around a lot of different police activities, starting with police use of force. How much force is being used? How frequently? Against which types of civilians? What were the civilians doing? Having a much better handle on that would help us craft better policies and also examine policies already are in place to see if they are effective.

Harrison: There's so little out there that is consistent. The Washington Post keeps a database on officer-involved shootings that has general reliability, but not any statistical or data reliability. The FBI gathers information on law enforcement officers assaulted or killed, but without the depth necessary to really get into what are the preconditions. Quite often people are grasping, and policy decisions get made based on judgment and intuition without having to be data-informed.

Woods: We need the information at all levels, too. Local data would allow police leaders and managers to make better decisions and compare and contrast themselves and their policies to other agencies. Having that information at state or national levels would let researchers find larger patterns to guide policymakers as they modify laws.

Over the last several decades, the police have been moving from paper records to electronic record systems. So there is a large collection of records out there that could be turned into data to support decisionmaking. But a lot of it is in incompatible systems. So that's another need: record keeping systems that enable data sharing.

What about facial recognition? Amazon just cut off police from using its facial recognition technology. What's your view of how that can or should be used?

Woods: I think facial recognition will have a role one day; it just automates what we're doing in person. But until it can do it reliably and we're confident that it's not going to introduce bias or codify biases that are already there, it's probably appropriate for law enforcement and tech companies to take a step back.

Hollywood: And how are agencies actually going to use facial recognition and other “surveillance technologies?”

One of the big ways I've seen agencies and companies get into trouble is when they just collect huge amounts of data without any sense of what they're going to do with it beyond something that will involve catching bad guys. And we don't have much cybersecurity protection on it, or policies controlling access to it.

In the United Kingdom a company took huge numbers of photos of people, I believe, walking to the Kings Cross Station and through a mall to get to it. And they didn't really ever do anything with it. It just sat there in their data set.

Woods: Another thing we need are standards to decide when it's appropriate to use not just facial recognition but any of these automated analytic technologies. How are we ensuring that the algorithms that they're based on are fair and unbiased? How can we manage the algorithm itself? Is it transparent enough for an outside organization to evaluate?

Harrison: Inevitably most crime is investigated retrospectively, so that gives police the impetus to gather all sorts of data, because you never know what might be useful after the fact.

The use of stingrays to scrape cell data is something that police just started doing with very little policy guidance. They could see who's on the phone and what number they have and all sorts of intrusions into personal privacy. They did so many things to try to serve a noble purpose, but without thinking about all the potential violations to personal privacy.

RAND has done work around the country, not only in New York but also Minneapolis, Chicago, Cincinnati, Los Angeles. Are there lessons out of that?

Cahill: We have a grant from the National Institute of Justice to look at sustained levels of violence in communities, including Minneapolis and Durham, North Carolina. We're looking at the role of police in levels of violence and how they help or hinder efforts to reduce violence. It is too early for any firm conclusions, but it certainly is a very interesting time to be doing this work in Minneapolis.

Woods: We looked at Chicago's Strategic Decision Support Centers. These bring a number of technologies together, including cameras and ShotSpotter, which is a system of microphones that can locate where a gunshot might have occurred. Then, in a district monitoring center, analysts can observe the cameras and use them for supporting police in the field. We found that system did contribute to reducing crime in the areas where they were implemented.

Hollywood: One of the things we saw was the importance of process in addition to technology. They had daily meetings where analysts were displaying information so that commanders could look at what was going on in their district every 24 hours. And making decisions to respond to what had just happened was something that was very valuable. Because otherwise using CompStat—computerized statistics—they were doing preparations based on crime trends over the previous month or up to a quarter.

Harrison: The closer you can get data in real time to the people in the field and the people who direct policing, the more effectively the police can address issues where problems exist. We're still ramping up from CompStat toward a real-time crime data dashboard that you can query crime issues such as what are auto thefts like by time of day or day of the week? Where do people want police visibility? Where do you have the most complaints from the public?

Cahill: In our research on NYPD's neighborhood policing, we see a change in focus for CompStat. They're bringing in a lot of different measures: How many community contacts do you have? What are the problem-solving efforts that you're making? A number of agencies now are using CompStat like that to measure their relationship with the community.

Do you have any advice for policymakers, police, or protestors?

Harrison: First, people are demanding change. This gives the police a natural opportunity to start a conversation with those who perhaps they have been holding at a distance.

Second, police departments often convey what they consider important through data collection. So if they conduct quantitative and qualitative community confidence surveys, as well as data on the frequency of contacts, the non-enforcement contacts, and things like that, those will help build a bond with the community. And with less lag in information, that can transform from a reliance on random patrols to give officers greater focus and direction to do things that both build confidence and ensure public safety.

Woods: Police now have an opportunity to sit down with policymakers, to say: “This is what you have asked us to do. These are the resources that you've asked us to do it with. Let's talk about how you want us to get from where we are to where you want to be.” That might mean handing off homelessness responses and acute mental health events. They may have to fight to get a seat at the table, but they have to if they're to reshape what's on their plate.

Harrison: They're going to have to not only be at the table, they're going to have to demonstrate visible leadership by sincerely listening and having empathy for people who have been at a distance from them historically into present day.

Cahill: To me, transparency is the key word. We need to get through the barriers with police unions to reveal police disciplinary actions. And we need police agencies to collect and share more data on their own use of force, including officer-involved shootings. That all needs to get aired out before we can make any progress.

Hollywood: This really is an opportunity to systematically improve the way policing is done.

Imagine if we had under 10,000 homicides a year in the United States. Imagine if we had half the number of people who were killed in officer-involved shootings, or other uses of force. Imagine 25 to 50 fewer officers killed per year in the line of duty.

These are stretch goals. But this is a unique opportunity to start addressing some of the major disparities that we have.