Enjoyed the episode? Want to listen later? Subscribe here, or anywhere you get podcasts:

…they randomly determined when officers would take the training… so it’s a really nice natural experiment. And they found that this one-day training program pretty dramatically reduced both complaints and use of force.

Jennifer Doleac

The killing of George Floyd has prompted a great deal of debate over whether the US should shrink its police departments. The research literature suggests that the presence of police officers does reduce crime, though they’re not cheap, and as is increasingly recognised, impose substantial harms on the populations they are meant to be protecting, especially communities of colour.

So maybe we ought to shift our focus to unconventional but effective approaches to crime prevention — approaches that would shrink the need for police or prisons and the human toll they bring with them.

Today’s guest, Jennifer Doleac — Associate Professor of Economics at Texas A&M University, and Director of the Justice Tech Lab — is an expert on empirical research into policing, law and incarceration. In this extensive interview, she highlights three alternative ways to effectively prevent crime: better street lighting, cognitive behavioral therapy, and lead abatement.

One of Jennifer’s papers used the switch into and out of daylight saving time as a ‘natural experiment’ to measure the effect of light levels on crime. One day the sun sets at 5pm; the next day it sets at 6pm. When that evening hour is dark instead of light, robberies during it roughly double.

The idea here is that if you try to rob someone in broad daylight, they might see you coming, and witnesses might later be able to identify you. You’re just more likely to get caught.

You might think: “Well, people will just commit crime in the morning instead”. But it looks like criminals aren’t early risers, and that doesn’t happen.

(Incidentally, a different experiment used the discontinuity in daylight savings time to quantify racial bias in police traffic stops.)

While we can’t keep the sun out all day, just installing more streetlights might be one of the easiest ways to cut crime, without having to hassle or punish anyone.

On her unusually rigorous podcast Probable Causation, Jennifer interviewed Aaron Chalfin, who studied what happened when very bright streetlights were randomly added to some public housing complexes but not others. His team found the lights reduced outside night-time crime by a massive 36%, even after taking account of possible displacement to other locations.

The second approach is cognitive behavioral therapy (CBT), in which you’re taught to slow down your decision-making and think through your assumptions before acting.

One randomised controlled trial looked at schools and juvenile detention facilities in Chicago, and compared kids randomly assigned to receive CBT with those who weren’t. They found the CBT course reduced rearrest rates by a third, and lowered the likelihood of a child returning to a juvenile detention facility by 20%.

Jennifer says the program isn’t that expensive, and its benefits are massive. Everyone would probably benefit from being able to talk through their problems and figure out why they make the decisions they do, but it might be especially helpful for people who’ve grown up with the trauma of violence in their lives.

A somewhat similar study of one-day ‘procedural justice’ training sessions for police officers in Chicago found they reduced civilian complaints against police by 10%.

Finally, Jennifer thinks that reducing lead levels might be the best buy of all in crime prevention.

There is really compelling evidence that lead not only increases crime, but also dramatically reduces educational outcomes.

In the US and other countries, there’s been a lengthy and mysterious drop in crime rates since the mid nineties, resulting in crime rates that are now just 25-50% of what they were in 1993.

That drop coincided with gasoline being deleaded. Before that, exhaust from cars would spread lead all over the place. While there’s no conclusive evidence that this huge drop in crime was due to kids growing up in a less polluted environment, there is compelling evidence that lead exposure does increase crime.

While average lead levels are much lower nowadays, some places still have shockingly high levels. Famously, Flint, Michigan still has major problems with lead in its water, but it’s far from the worst.

Jennifer believes that lead affects people’s brains in such a negative way that driving exposure down even further would be extremely cost-effective for its crime-reduction benefits alone, even setting aside broader benefits to people’s health.

In today’s conversation, Rob and Jennifer also cover, among many other things:

  • Misconduct, hiring practices and accountability among US police
  • Procedural justice training
  • Overrated policy ideas
  • Policies to try to reduce racial discrimination
  • The effects of DNA databases
  • Diversity in economics
  • The quality of social science research

Get this episode by subscribing to our podcast on the world’s most pressing problems and how to solve them: type 80,000 Hours into your podcasting app. Or read the transcript below.

Producer: Keiran Harris.
Audio mastering: Ben Cordell.
Transcriptions: Zakee Ulhaq.


The biggest problems with US policing and the criminal justice system today

I think the problems that are certainly at the forefront of everyone’s minds right now are around the unnecessary escalation of incidents between community members and law enforcement. So maybe someone calls the police for help but, you know, when they get to the scene, they wind up making an arrest that could have been avoided. Or they use force that could have been avoided. And, in general, I think people would prefer less of that if it’s possible. The other problem is related to racial bias in policing. And this isn’t a problem that’s specific to policing or the criminal justice system. We know that racial bias is a problem throughout society in the United States and lots of other countries too. And so this is a problem that is everywhere, and we should not be surprised that it’s also a problem in the criminal justice system.

So I’d say those are the two big problems in the policing space that we’re kind of trying to figure out how to solve. And then there are myriad ways that we might solve them. And so they kind of range from trying to hire better people, trying to train them better, trying to put better incentives in place. You know, there’s lots of discussion about the role of unions and how hard it is to discipline or fire a cop that repeatedly exhibits bad behavior. There’s a lot. There’s a lot we don’t know about how to do a lot of policing better. Even defining like what is a good cop? What exactly is it that we want police to be doing with their time? It’s actually a really difficult question. And certainly one where the evidence doesn’t have much to say to guide us.

Procedural justice training

So there’s a new study that just came out in PNAS that has caught my attention that I find myself talking about a lot: measuring the effectiveness of procedural justice training for police officers. So basically, it’s like a one day training, this program they put together. So not super extensive. But basically really just helping police officers and pushing them to make the process of interactions with citizens feel more fair, and to leave the citizens with the feeling that they were heard and understood and they trust the person who’s making the decision. That they kind of had taken everything into account fairly. That they trust the process, basically. In general, trainings aren’t necessarily easy to measure the effectiveness of because it could be a very selected sample who takes the training, but in this case, they needed to train all 8000 plus officers in Chicago with this training.

They couldn’t do it all at once, so they basically had to do 25 officers at a time, once a month for many years. And they randomly assigned officers to particular times of when they took the training. So it wasn’t just like, you know, the most senior people first, or volunteers first, or something like that. And so what that leaves us with, is it’s a really nice natural experiment where you can compare the behavior of officers who took the training earlier to officers who took the training later and then measure in that way what the impact of the training is on behaviors and the behaviors that they were looking at were things like complaints received from citizens, and use of force by officers. And they found that this one day training program that officers just took once pretty dramatically reduced both complaints and use of force significantly.

And that was on average across all officers. And I have a hunch that if they were to zero in on just the officers that had used more force in the past, maybe they would see a bigger impact for those officers. Because you’re going to have a lot of zeros in there that kind of water it down. And it also seems like the sort of thing that repeated training could be helpful for. So I think there’s no more experimentation to be done there. But I think it’s a really nice example. Both of a program that police departments should be trying, but also that staggered rollout design that they used just because of capacity constraints, right? They couldn’t train everyone at once. That’s going to be a problem a lot of police departments are going to face in any new training program. So using a similar staggered out design where they just randomize who goes first, who goes second, who goes third, I think, is a really nice model for implementing new training programs in a way that then allows us to measure whether they’re effective. So I really liked that study. And I think that their approach is promising.

The impact of body cams on officer behaviour

We have all these randomized controlled trials now, and basically, on net, they found no effects. So in some places they seem somewhat helpful in terms of reducing complaints and reducing officer use of force. In other places, they actually make things worse and force goes up.

The first big study like this in the United States was done in Washington DC, and they just found no effect on anything. And so there are a couple of possible reasons for this, right? Let’s say one is, especially in the current context, every cop assumes they might be on camera all the time anyway, because everyone’s got cell phones and there are cameras everywhere, businesses and whatever else. And so they were already treated, to the extent of having their behavior filmed.

Another possibility is the hypothesis about the officer’s decision-making in that moment is wrong. It’s not necessarily that they’re like, “Well, I know I don’t need to use force here, but like, I want to anyway, and nobody’s gonna notice”. It might be that most force is used because police are legitimately afraid and that could be a training issue, right? Like maybe they shouldn’t be afraid. And there are other ways we could help them figure out how to deal with that situation. But if they’re using force because they’re legitimately afraid for their lives, then putting a camera on them isn’t going to change what they do. They think they’re doing the right thing in the moment. The other possibility is that we now have all this footage, but because of the fact that we’ve been having this conversation about how we can’t fire anybody, there really isn’t much accountability.

Police effect on crime

If you increase the number of police officers by 1%, we can generally expect somewhere between a 0.1 and a 1% reduction in crime. So that’s a pretty big range, but it’s negative. And it’s pretty big. I think the way to think about this is just in general, hiring more police officers is one of the most cost-effective ways to reduce crime. And if we were just looking at the cost of officers and the social cost of crime, then our cities are massively underpoliced, not overpoliced. Like it would be cost-effective to add more officers. And it’s really only once you add kind of the additional social costs that we haven’t done a very good job quantifying them yet, that you get to a point where it’s like, “Oh, maybe that’s not right and we should be cutting police forces”, but that’s fuzzier in all our minds.

I think that the general consensus is that (the large impact of more police is) mostly a deterrent effect. So it’s not necessarily like catching people after the fact and locking them up, although there is a crime reduction effect from incapacitation as well. Like if you put someone in prison, they can’t commit more crime out in the street. But most of the effect, I think, in most studies that are able to figure out what the channel is, it’s coming from deterrence. So, you know, there are all of these randomized controlled trials of like hotspot policing, where we put a cop on this corner, but not on this other corner over here. And we randomized where we put them. And those studies routinely find a big drop in crime on the street corner where you put the cop.

I think a lot of crime is just like crime of opportunity. You know, it’s not carefully premeditated. It’s like, “Oh, I see a laptop left on that park bench. That person doesn’t seem to be around. Maybe I’ll grab it”. And if there’s a police officer standing right there, you don’t do it.

The effects of DNA databases

If your profile’s in the database, it can be frequently compared with crime scene evidence. And so the reason we might expect a deterrent effect is that you know you’ll be more likely to get caught if you offend again. So basically if the database matches you with crime scene evidence, then that’s a lead that police might not have had otherwise. So it’ll identify you as a suspect in a crime if you might not otherwise have been a suspect. So we see this big deterrent effect, but the DNA database itself could actually lead to an increase in likelihood that you get caught. I think that’s the whole point. And so if you just sort of look at the data we might’ve expected, if it had no effect on recidivism, we would expect an increase in the likelihood of being reconvicted, right, because you’re just more likely to get caught.

And so the fact that, especially in the US context, where I had to just look at the net of those two things, I still see a significant negative effect means that that’s an underestimate of the benefit. So I find that in the United States, being required to submit a DNA sample to the database reduced the likelihood of a new conviction within five years by about 17% for serious violent offenders. And that was statistically significant. It reduced the likelihood of another conviction by 6% for serious property offenders.

Articles, books, and other media discussed in the show

Jennifer’s work

Peer-Reviewed Publications

Working Papers

Published papers by other people

Everything else

Related episodes

About the show

The 80,000 Hours Podcast features unusually in-depth conversations about the world's most pressing problems and how you can use your career to solve them. We invite guests pursuing a wide range of career paths — from academics and activists to entrepreneurs and policymakers — to analyse the case for and against working on different issues and which approaches are best for solving them.

The 80,000 Hours Podcast is produced and edited by Keiran Harris. Get in touch with feedback or guest suggestions by emailing [email protected].

What should I listen to first?

We've carefully selected 10 episodes we think it could make sense to listen to first, on a separate podcast feed:

Check out 'Effective Altruism: An Introduction'

Subscribe here, or anywhere you get podcasts:

If you're new, see the podcast homepage for ideas on where to start, or browse our full episode archive.