Episode Transcript

CARDIFF GARCIA: Jen. Hello. We’ve got our podcast voices on now. 

JENNIFER DOLEAC: We do. (LAUGHS)

CARDIFF: But you and I have spoken many times. 

JENNIFER: In normal voices. (CHUCKLES)

CARDIFF: I’ve interviewed you many times about your work. Now that this book is out, which seems like a compilation of so much of the work you’ve done previously, are you more nervous or excited for it to be out in the world?

Given that criminal studies is kind of a rough and tumble area of scholarship, isn’t it?

JENNIFER: It is. It turns out I have studied other more controversial things. 

CARDIFF: Oh, that’s helpful (CHUCKLES)

JENNIFER: Crime is lower on the list. So it’s actually like, this is fun. (LAUGHS)

CARDIFF: Wait, what is it below? What have you studied that it’s below? Just out of curiosity.

JENNIFER: Opioids and public health, actually, that’s the roughest and tumbliest.

CARDIFF: Because you’ve got — in the case of crime — economists, sociologists, scholars of criminal studies. What have you learned about how your findings and your assessment of others’ findings get absorbed in the rest of the field — not just economics of crime, but other fields that study crime?

JENNIFER: Yeah, I think most people are surprised that an economist would ever think about this topic. I meet someone for the first time and I say I’m an economist and I study crime and criminal justice policy, and they immediately think, okay, she must mean white-collar crime. And it’s like, no — homicide, actual real crimes, crimes you see on TV. (Not that white collar crime isn’t real crime.) But I think what economists bring to the table is an obsession with distinguishing correlation from causation. And that’s probably the source of most disagreement and upset in the field.

CARDIFF: Correlation versus causation — you introduce a new policy and you see these outcomes and you’re like, oh, that policy caused those outcomes. And sometimes, not necessarily. It might have been something else.

JENNIFER: Right. I think where this comes up most often is when people look at whether a reentry program is effective or if a program that’s in prison is effective. You’ll look at the people who participated in the program or completed the program and compare them to people who didn’t. It turns out, almost always, the people who completed the program do better later on than people who did not participate in the program.

My main concern with studies like that is usually that people have volunteered for the program or were carefully selected. A lot of reentry programs are very careful to heavily screen applicants based on how much they seem to want to change and only take the people who are most motivated. And it’s like, all right, if you’re comparing motivated people to unmotivated people, then maybe we’re not surprised they do better.

CARDIFF: It wasn’t the program.

JENNIFER: Exactly. It’s not the program.

CARDIFF:  It was the fact that people who already were inclined to improve, to change their trajectory in life, they ended up doing so, but they might’ve been able to do that anyways.

JENNIFER: Right.

CARDIFF: It wasn’t the program. They selected themselves into the program.

JENNIFER: Exactly. So what we really wanna know is: what was the value add of that program? That’s the big policy question. When we think about how we should be investing our public safety resources, our tax dollars in general, we wanna be investing in things that are actually going to move the needle, change the trajectory that somebody’s on.

And so those correlational studies are not giving us that.

So the three things I think economists are bringing to the table are: First, this obsession with causal inference — distinguishing correlation from causation.

But then the other two pieces are: Economists think a lot about how people respond to incentives. And so that’s really the theoretical lens we’re bringing to the table. And this assumption that all the decisions we make, all of us all day long, are really — we’re doing a complex calculus problem in our heads and trying to weigh the costs and benefits of things. 

And so changing the incentives we’re facing and the factors we’re considering will change behavior. And that’s just a very different way of thinking about human behavior than, say, what a sociologist would predict or political scientists would predict.

And then there’s this other piece about how should we spend our money? Where’s the best use of a marginal dollar? Do we put it toward prison? Do we put it toward police? Do we put it toward schools? If our goal is to reduce crime? And that’s the sort of thing that economists love to think about.

CARDIFF: Yeah. Do you think we’ve suitably upset all the other scholars of crime by now? (LAUGHS)

JENNIFER: Absolutely, I mean, we could probably do more, but— (LAUGHS)

CARDIFF: Let’s start talking now about what’s actually in your book. So before we started this chat, just to give the audience a little behind the scenes here, I was telling you that there’s so many studies in the book that we can’t go through how every single study was designed, ’cause then we’d never actually get into what the study actually found. 

So I just asked you to pick maybe your single favorite, most emblematic study for how this actually goes in researching these ideas. What is your single favorite, whether it’s one of yours or one that you cite in the book? Give us one study. Take us through it. Take us through how it’s designed and how it shares familiar features with a lot of the other studies that you look at. Because a lot of them are similar. They study different things, but the studies themselves are similar. 

JENNIFER: Yeah, it turns out in practice there are only a few research designs here that are commonly used in this kind of space. And the key here is that we want to approximate, as best we can, the randomized control trial that you might see in a laboratory. 

So that’s sort of, when we put our researcher hats on, what we’re hunting for — something in the real world that will get us as close to that treatment and control group as possible.

So we’re looking for something that introduces randomness into the process, so that people aren’t choosing to sign up for a program, or being carefully selected based on motivation or something like that. So one of the study designs that’s really popular in this space is thinking about the randomization of different cases to different decision makers.

So, for instance, we talk a lot about how, if we’re thinking about what the impact of prison is on people’s outcomes after prison or after their conviction and—

CARDIFF: The likelihood that they’ll commit crime again, for example.

JENNIFER: Right. We can’t randomize people to prison or not, right? That would be certainly illegal, probably unethical. A lot of people would be really uncomfortable with that. 

But it turns out that in practice we do randomize cases to different judges or cases to different prosecutors. And those individual decision makers, because they’re human, differ in what their best judgment is about what should happen to this person.

So we wind up, in practice, with — because we’ve got randomization to different decision makers who will naturally make different decisions — you have essentially randomization to the different treatments.

CARDIFF: So, for example, if you have two different prosecutors, one just by nature might be more likely to pursue aggressive charges and try to get somebody put in prison. Another prosecutor might be a little more lenient and just, again, by nature might be more likely to say, you know what, we’re not even gonna charge this person. We’re gonna let them go and I’m gonna focus on the tougher cases, or something like that. 

JENNIFER: Right. So I have this study with Amanda Agan and Anna Harvey where we look at this kind of randomization — a practice that essentially gives cases to different prosecutors as if it’s random — in Suffolk County, Massachusetts, where Boston is, to be able to study what the impact of that prosecution decision is.

So do you just drop the case, or do you pursue it? Prosecutors are doing this all day long. They have to do it really fast. They only have a couple minutes per case. There are so many — especially what we were looking at was nonviolent misdemeanor cases — there are a zillion of these that have to go through the court every day. They have to move fast.

And, and as you said, sometimes someone’s just a little bit more lenient. Someone’s a little bit harsher. They’ve seen some of these cases before and they just don’t like, you know, drug offenders, versus not. And so just by the luck of the draw, you as a defendant — you happen to get a lenient prosecutor. That means that you are now less likely to have your case pulled forward. You’re more likely to have your case just dropped in that initial arraignment hearing versus if you happen to be there the next day instead and gotten a different prosecutor — that prosecutor would’ve pursued your case.

And so basically, that sets up a nice natural experiment where, because the cases are randomly assigned to different prosecutors, that luck of the draw determines the likelihood that your case is pursued or not. And so that gives us really that experiment that we were looking for. It’s as if we flipped a coin to see if we pursue your case or not, but in a way that actually happens in real courtrooms every day.

And so in that study, we were able then to see, okay, what’s the impact of having your case dismissed in that initial arraignment hearing versus carried forward. And in practice, what we wound up finding was that it turns out if you get lucky — you get the lenient prosecutor — your case is dismissed, you are much less likely to re-offend in the future, which was a big surprise to me as a researcher.

CARDIFF: Why was it surprising?

JENNIFER: I am used to everything having trade-offs. Everything’s gonna have costs and benefits. And the question is just how much. Like how big is the cost versus how big is the benefit? And in this case, I figured if you don’t pursue the case, just mechanically some of those people are now not gonna be in jail, not gonna be on probation, and so they’re gonna go out and get into more trouble.

And the real question in my mind was, well, how big is that cost gonna be relative to how much time it would save everybody if we just dismissed some of these really low-level cases upfront? And then it turned out — there weren’t costs and benefits that were weighed against each other. It was all benefits. 

CARDIFF: Just benefits.

JENNIFER: Only benefits. We economists always say there’s no such thing as a free lunch —

CARDIFF: You found the free lunch!

JENNIFER: Turns out, there are a lot of free lunches in the criminal justice system. There are misaligned incentives all the time. We’re just making big mistakes all over the place, where if we just followed the data a little bit better, we could have less crime for less money. And those are big opportunities we should find.

CARDIFF: Yeah. I was also surprised by this because you had set up that part of the book as there’s always a trade-off. And specifically, for example, you were looking at the effects of having a DNA database where you essentially take the swab of a first-time offender — somebody who commits a crime for the first time or is arrested for the first time — you take a swab, now their DNA is in a database somewhere. And you find that it is actually beneficial to have this database for preventing future crime on behalf of the same person.

So that person’s gonna be less likely to commit a crime if they know that their DNA is somewhere, because the likelihood that they’ll actually get caught is higher.

JENNIFER: Mm-hmm.

CARDIFF: But you went through a whole list of possible trade-offs, and there were some there. 

JENNIFER: Right. And probably in most cases there are trade-offs. Maybe it costs us a little bit more money, but we get lower crime as a result. And so, okay. We just have to think about, is it worth it?

CARDIFF: People who don’t like the intrusion of privacy or surveillance technology.

JENNIFER: Yes. In that case.

CARDIFF: These are really big societal questions. But there are clear benefits and clear trade-offs. In the case of this chapter, you have — it’s literally titled Err Towards Leniency — so in the case of prosecutors deciding for a first-time misdemeanor person, let them go the first time. And see what happens. And it turns out that they’re less likely to commit a crime in the future than if they are prosecuted. 

That’s fascinating to me because it seems like there really are no trade-offs. 

But there’s a wrinkle in this too, which is that there still were costs on behalf of the person who committed the crime or allegedly committed the crime — getting arrested to get to that stage. And those things matter, being arrested and all that.

If you take all that away… people might hear this and say, “well, just don’t enforce the law. Just don’t enforce misdemeanors.” 

JENNIFER: Decriminalize.

CARDIFF: Yeah, decriminalize. That is not your conclusion. Give us a sense of the nuances there.

JENNIFER: Yeah, we heard that a lot when the study first came out. My colleagues and I were doing lots of interviews and hearing from lots of advocates saying, great, we’re so glad the study has proven that we should just decriminalize all nonviolent misdemeanors. It’s not worth it to arrest and punish anybody for trespassing, shoplifting, drug possession.

And we were very quick to say that is not what the study shows. Not having your case pursued, having your case dismissed at the first arraignment hearing, is not the same as no consequences. You are arrested, you’re probably booked in jail. You had to miss work for this court date. You’re probably really stressed out about it — especially for these first-time offenders who are the ones where we found all the benefits.

So it really is like giving those folks who don’t have a criminal record yet a second chance to avoid a first record. If this is your first run-in with the system, for a lot of the people, this is gonna be their rock-bottom moment. That’s maybe a wake-up call to say, “whoa, I need to stop hanging out with that guy I’m always getting drunk with,” or stop whatever it is. “I need to clean up my act.”

And it just turns out that in a lot of those cases, if we just get out of their way, they will course-correct on their own. Not everyone will, but a lot of people will. And then that helps us kind of triage a little bit, and we can focus then on the smaller group that remains and do something different for them.

CARDIFF: Yeah. One question I had — and I’m certain this is something that economists and that you in particular think about all the time — is when you have a finding like this as part of this randomized experiment, or just the randomized nature of how it worked in the real world, is if you change something, will this finding necessarily scale?

So if, for example, you know that prosecutors as a matter of policy are not going to charge you your first time you do something — or if it’s a misdemeanor, nonviolent misdemeanor — well, that sounds great. Now I’ll just go out there and it’s time to go to the mall and start sticking stuff in my pockets. That sounds great.

JENNIFER: Yep.

CARDIFF: So this finding is interesting, but it only works in the context of how the system is shaped. Now, if you reshape the system, will it still hold? How should we think about that?

JENNIFER: Yep. Yep. Absolutely. So, yeah, you can easily see how a policy of “everyone gets one free shoplifting charge” would be very bad. That is clearly going to lead to more shoplifting.

I think the way this is implemented in practice, and the actual policy change that the DA in Suffolk County implemented — that we were also, in that paper, able to study, which was very useful — established a presumption of non-prosecution, not a blanket policy that we do not prosecute these cases, but a presumption. So really just changing the default from, we’re always gonna prosecute these cases, to, we probably won’t.

And then it still is gonna vary across individual prosecutors and individual cases. And kind of depending on the specific case, a good reason to go ahead and prosecute someone might be that they’ve cycled through the system a few times, or they seem to have shown no remorse, or yeah — or a variety of other things. I’m sure we could all imagine.

But it really is kind of changing — it turns out in prosecutors’ offices in general, because there’s so many of these cases, prosecutors drop these low-level cases all the time. So if we’re talking about going from dropping like 30% of the cases to dropping 50% of the cases — not from zero to a hundred. And so it really is kind of this marginal shift.

But it was nice that in Suffolk County, where there was actually this policy change, that we were able to go in and say, okay, maybe in the courtroom, the individual defendant kind of knew they got lucky. They got this — they got a lenient DA, and they got their case dismissed, and they’d been sitting there all day and they saw everybody being convicted. And so they’re like, “Ooh, I dodged a bullet there. I better change.”

But if it’s an actual policy, then maybe they don’t.

CARDIFF: You gotta be careful.

JENNIFER: You gotta be careful. And so being able to go from these proof-of-concept almost pilots — try it somewhere once, do the first study, we found that effect — and then the question is, what’s the actual policy we are suggesting that prosecutors implement?

And of course we got that question all the time after the study: what exactly is the line? Exactly which cases should we be dropping? And then we had to give the very unsatisfying answer a lot: well, we don’t quite know, but we’d love to help you learn in different contexts, basically where you’re drawing the line. It might be easy — imagine what a slightly different policy or slightly different guidance to your staff might be. And let’s implement that and let’s see what happens. And that’s basically the only way we’re gonna learn.

CARDIFF: You can tell I’m old, by the way, because I made a reference to going to the mall and stealing things. That’s the most teenager-in-the-1990s thing I’ve ever heard.

JENNIFER: (LAUGHS) Do malls still exist? Everyone just Googles “mall.”

CARDIFF: Yeah. “What is that?” Some 28-year-old listening to this right now is super confused.

Does this also apply for either more violent crimes or for felonies?

JENNIFER: Yes. So there’s another study that looks at the effect of, again, erring toward leniency, which means something a little different with more serious cases, but for nonviolent felony cases.

So the only studies we have right now are looking at nonviolent types of crimes. We need studies on violent cases, but we can look at felonies, which are crimes usually punishable by a year or more in prison. So more serious offenses — things like burglary, larceny.

And there’s a cool study in Harris County, Texas, where Houston is located— 

CARDIFF: Where you live.

JENNIFER: —my current home. Yes.

They looked at some cool policy changes that introduced sudden shocks to the likelihood that someone charged with a nonviolent felony would get what’s called a deferred adjudication, which means that if you go through a probationary period and you don’t do anything else wrong, you clean up your act, then the charges are dropped and you avoid having that felony record.

They found something very similar: people who got lucky and wound up getting a deferred adjudication because of this policy change were dramatically less likely to re-offend. And again, as in the Suffolk County example, the effect was really driven by first-time offenders. It was helping them avoid that first criminal record that had so much value.

They were also able to link this to employment and earnings and showed that people earned a lot more money. They got jobs.

CARDIFF: It wasn’t just about reducing committing crime in the future.

JENNIFER: It wasn’t just the criminal record. Mm-hmm. 

CARDIFF: That’s great. 

JENNIFER: Right.

So yeah, all of this together — this area is an example of a place where the research has really changed my prior about how this works. It just seems like you hear the idea and you think, that’s too good to be true. There’s no way that’s gonna work.

And then you have these studies — one after another — finding these big drops in recidivism on the order of like 50%. We’re not talking small effects. It’s like you’re half as likely to re-offend.

Most reentry programs struggle to find any impact on recidivism. So the fact that just helping people avoid that first record has such a big impact is really striking to me.

It feels like one place I wish we could move a little faster on the policy side.

CARDIFF: This is one where anecdotes can be very powerful and can, in some sense, overwhelm the data. Because if a policymaker hears, “Hey, this is great,” but they also hear about one really high-profile case of somebody who reoffended and did something horrible after being let go, politically that’s tough.

You have to sell that to people. Meanwhile, people are saying, “Yeah, but look at what this person did.” You let them go and then they burned down a church, or a house, or whatever.

I don’t want to say I’m despairing of people taking the data and implementing a policy that’s consistent with it.  But that’s one where, especially now where we’re in this attention economy and everything like that gets around. It seems like you have an uphill battle of convincing people on something like this.

JENNIFER: Totally.

I now work more on the policy side. I’m a researcher by training, but now I work at Arnold Ventures, which is a philanthropy that does a lot of policy work. And this is the big challenge. We have all these evidence-based solutions we want to help lawmakers scale, and they say, “Yeah, but what about this case?”

Even individual judges or parole officers know that if they let this person go, and they go out and do something bad, it’s their face that’s gonna be on the news. People will point to them and say, “Why didn’t you know?”

And the reality is, we just can’t prevent all crime unless we’re really just locking everyone up. If we really wanted to — I think the best example of this is the group that’s at highest risk of committing crime is young men. The best way to dramatically reduce crime, if we’re not thinking about costs and benefits, would just be to lock all the young men up until they turn 25, right?

There’s a reason we’re not doing that. That would be extremely costly to civil liberties. There are a variety of reasons that would be a terrible idea.

CARDIFF: Less podcasting competition, so I’m not totally out on it. But yeah, probably unfair. (LAUGHS)

JENNIFER: (LAUGHS) True, true.

And so we live in a society where we value people having their freedom unless they’re proven guilty and have a good reason to be behind bars. And so we’re just sort of always making these trade-offs.

But I think recognizing this sort of personal incentive — whether it’s the politician, whether it’s the judge, whether it’s the parole officer, making decisions about whether we want to give, whether we want to err toward leniency, let somebody out — they have their own personal incentive that they are responding to. And that one really terrible incident that winds up on the news could end their own career. And so that pushes us really across the board toward being harsher than would be optimal from a public safety perspective.

CARDIFF: I’m fascinated by the chapter on probation specifically because you mentioned a second ago probation in the context of somebody who got caught doing something, at least allegedly, but they can do this non-adjudication thing where if they comply with the terms of the probation, they just don’t get charged. It’s over. 

But there’s different kinds of probation, there’s probation for people who did get convicted, and if they do probation, they can avoid prison. There’s different ways to design it. And you have a whole chapter on this. 

Here’s a stat, by the way, that I didn’t know: In the country there’s 3 million people on probation. There’s 1.7 million people incarcerated. There’s almost twice as many people who are on probation as people who are actually in jail or in prison. And you write that almost 30% of people on probation ultimately fail to follow the requirements and end up back in prison, or they get penalized some other way for that.

JENNIFER: Yeah.

CARDIFF: What’s going on there? I mean, before we can even figure out how to do it better — what’s happening there? What is it about probation that makes it so hard to follow the terms? Not all of this is just people, you know, “okay, they’re criminals, so they can’t follow probation.” That’s not the case.

There’s something about the way it’s designed that’s frustrating people.

JENNIFER: Right. I mean the idea here — the way it should work — is that you’re released on probation, on community supervision. You’ve got a probation officer you have to check in with on a regular basis and a whole bunch of rules to comply with. Maybe you have a curfew. You have to be home by a certain time of night. You aren’t supposed to hang out with other people who have felony records. You’re not supposed to use drugs or alcohol.

CARDIFF: Can’t leave the jurisdiction, or something like that.

JENNIFER: Yeah. You can’t leave the county, things like that, without permission. And all these things. You have to find a job. So all of these things are really set up, ideally, as guideposts to kind of help you clean up your act, do things that are going to keep you out of trouble, avoid things that are likely to put you in high-risk situations. And that all sounds great.

And then if you can’t comply with those kinds of rules, the idea is that should be a flag. That’s a signal to the court that this isn’t going very well and you’re likely to commit another crime. So in theory, the rules should be things that are actually predictive of committing new offenses.

But then there are just so many stories out there of people who are like, “Yeah, but I have to leave from work to take a bus all the way across the city to meet with my probation officer at 1:00 PM, and then I lose my job, and I can’t.” And I opened this chapter with a story about a woman who had been pulled over for a DUI, got one of these probationary terms, and she wasn’t allowed to move into a different apartment. Then there was a mouse infestation in her house, and so she had to move, and then it just got caught up in —

CARDIFF: She wasn’t allowed to move.

JENNIFER: Crazy. Yeah. 

CARDIFF: You need permission from the judge to move or whatever.

JENNIFER: You read about it and it’s like, this is insane. There’s no way this is actually helping them create a structure or giving them a structure for a better, more law-abiding life. This is just a bunch of hoops you have to jump through that don’t really make any sense. So there are a ton of stories out there like that.

And then it becomes a question of: okay, lots of bad anecdotes, but I understand the theory. I understand why we want some rules. And we did have a system a long time ago — seventies and earlier — where if you’re on probation, basically you just send a postcard once a month to your probation officer. Most people probably want more structure than that too.

 So it’s like, okay, how is this actually working in practice? And are the rules that we’re giving people actually helping them build a life that is the life we want them building?

CARDIFF: Isn’t the answer, no?

JENNIFER: (LAUGHS)

CARDIFF: In the actual studies, it seems not. By the way, as you were going through some of the rules in a couple of these cases, I was sitting there thinking totally outside the criminal justice system: I would have trouble complying with all this stuff. You know what I mean?

JENNIFER: Much less if you maybe don’t have a car or don’t, or especially if you’re in these situations where it takes a lot of time to get around. You have much less control over various aspects of your life. Fulfilling a lot of these requirements is just really hard.

CARDIFF: Yeah. The finding specifically was that more tough supervisory requirements does not lead to better outcomes for these people.

JENNIFER: And it’s remarkable. Across the criminal justice system, one of the fun parts about this area to research for me as an economist was always, well, it’s really hard to randomize things in this space. You can’t randomly send people to prison. So you have to find really clever natural experiments, the stuff we were talking about before.

And this is an area where people have actually run RCTs on whether people should be on more intense or less intense supervision. They actually randomize people to high-intensity supervision, medium, or low. Which just seems crazy. How did they pull this off across all these different places?

So maybe that’s just the nerdy economist in me that finds that the most interesting piece of this whole conversation. But that means we have a bunch of RCTs in this space. And one study after another shows more intensive supervision produces zero public safety benefit. People rack up a bunch of technical violations because once you have more rules, you’re more likely to break some rules. And the problem with these technical violations is that they’re likely to send you back to prison — or to prison in the first place if you were just on probation.

So it costs a lot of money. It’s a hassle for the people on probation themselves. It costs money for the probation officer to supervise you. And then if you break a rule, you go to prison and now we’re paying to incarcerate this person. There are a ton of costs. And it’s like, okay, but if it makes us safer, there are benefits. But it does not make us safer! Just one study after another shows zero public safety benefits to go along with these costs.

So in general, right now, all of these studies are just saying: Do less. Do less. Whatever it is we’re doing is not productive.

CARDIFF: There’s one interesting exception to that that you note, which is people who struggle with substance abuse. And there you find that a specific kind of supervision is very helpful. People who have to check in constantly to make sure that they haven’t drunk or used drugs again that day. For them, if they know for sure what the terms of the supervision are. And it — I forget the acronym, it was swift, clear —

JENNIFER: Swift, certain, and fair.

CARDIFF: Swift, certain, and fair.

JENNIFER: Mm-hmm.

CARDIFF: And you cite this great anecdote from another scholar, the late Mark Kleiman, where he said the way these things are set up now is that you never know if the terms of supervision are going to actually be enforced. 

And then it’s almost like a kid whose parents tell the kid, “Hey, don’t steal cookies out of the cookie jar.” And the kid starts stealing outta the cookie jar, but they let them get away with it every time, until one day they suddenly say, “By the way, you’re grounded for five years.” And it’s like, wait, what? You know what I mean?

Because people who struggle with these substances tend to be very present-oriented. They’re not thinking distantly into the future. It helps to be more on top of them, but to give them very clear guidance. If you break this, yes, you might go back to jail, but also we’re here to help you out. So just check in with us constantly and we’ll see.

JENNIFER: Right, right. One of the big themes that comes up across the crime and criminal justice policy literature is that what matters most for deterring crime and changing behavior is increasing the probability that people get caught quickly. It’s not the severity of the consequence.

In general, that’s basically the opposite of what we do now. We have very low likelihood that you’re going to get caught if you commit a crime, but then we send you to prison for a really long time for the rare people who do get caught. 

And so this is an example of: you’ve got a bunch of probation conditions, and in practice, for instance, the rule that you’re not supposed to hang out with other people with felony records. How on earth are we going to enforce that? Are we going to follow you around and check the background of all your friends? Of course not. That’s completely impractical to enforce in any kind of consistent way.

But the thing about drug and alcohol use that is often banned for probationers, especially those who have a history of that problem — that’s something we can check. There are easy drug and alcohol tests that we can require people to take. So there are these programs that require people to come into the jail once or twice a day, blow into a breathalyzer, and if you have been drinking, you immediately are in jail for the night. If you’re not, you go on your way, go to work, come back tomorrow.

And it turns out that something like 99% of people who come in and take these tests are sober. They just stop drinking. Some don’t. Some have a real problem and need treatment. Then you can triage and help those people differently. But it turns out that just the shift in the incentives, as you said — because people are more present-minded — they’re not thinking that far ahead. It’s about increasing the probability of getting caught rather than increasing the punishment.

Perhaps this is more salient for people who have a drug and alcohol problem. Maybe they’re more present-minded than most. But it’s kind of true of all of us. We all weight the present more than the future. And turns out — there are a lot of studies  that look at people who are just in prison in general, people with criminal records in general, (who) do discount the future more than the average person. So in general, the people whose behavior we’re trying to change are especially not thinking about the long term. The more we can find ways to move up the benefits of behaving well to the short term, the better off we’re all going to be.

CARDIFF: The marginally higher present-mindedness of people who tend toward crime.

JENNIFER: Mm-hmm.

CARDIFF: I’m guessing that’s also related to something else you note throughout the book, which is that crime tends to be a young person’s game. Young people are way disproportionately more likely to commit crime, and that’s also one reason why long prison sentences don’t make sense, that it’s more helpful to make sure that people know that there’s a higher likelihood that they’ll get caught and serve some time than it is to just put them in jail for a very long time.

Because some people actually age out of their “doing crime” period.

JENNIFER: A lot of people do a lot of dumb, reckless stuff when they’re 16, 17, 18. And by the time they’re 40, they’re married with kids in a job. And I think we probably all have friends we could think of who grew up a lot. (CHUCKLES)

CARDIFF: Or as somebody in my forties now, I’m way more tired than I used to be. Leaving aside the issue of crime in the book and everything — I don’t have the impetus to go out there and start causing trouble. I’d rather read a book. I’d rather read The Science of Second Chances. I’m on my couch. (CHUCKLES)

JENNIFER: I’m in bed by 8:00pm, so I can’t commit crime anymore. This is a hypothesis I haven’t seen people test. Is it actually an impulse problem or is it just THAT we’re more tired? (CHUCKLES)

CARDIFF: This is interesting because in the very next part of the book, you give an example of something where a little bit of ambiguity actually might be helpful. So we went from “Okay, people in this specific situation where they struggle with substances, if they’re given this clear, certain guidance, it’s really helpful to them.”

But then you talk about what happens when people are already in prison, and the possibility of parole, where you go before a parole board and you might get let go early, but there’s no guarantees. And you compare that to a fixed sentence where it’s just: you’re in prison for five years with no chance of getting out. 

It’s very interesting to me that in this case, the possibility of parole has very powerful effects on how people behave in prison. Not surprising, I should say! But still.

JENNIFER: Yeah, it’s based on the same idea basically. People are not thinking super far ahead and so you really need to provide short-term incentives and consequences for good behavior. And so in this scenario, what happened is that during the 90s, a lot of states in the United States started moving toward what are called “truth in sentencing” policies. So it used to be that you’re sentenced to four years in prison, but you have the possibility of parole after two years if you really engage in rehabilitative programming and exhibit good behavior and all of that. And then at some point people were like, “Wait, you said four years. Why is he out in two? That’s not fair.” And both the left and right agreed that this was a terrible system. 

And so we reformed lots of states. This is something that we battle at Arnold Ventures, as there are a lot of states that are still trying to implement truth and sentencing policies. And we’ve learned, it turns out, they’re really bad because they take away the shorter run incentive to engage in rehabilitation when there isn’t the possibility of an early release.

So you would think you’d have plenty of incentive to invest in your own rehabilitation because once you get out, you’re gonna want a job, you’re gonna want to build your life, so you might as well start preparing for it now. But as we’ve been discussing, this is a group that is not very good at thinking very far ahead.

And so providing the shorter term incentive that, if you take that GED class and engage with substance use treatment and stay outta trouble, don’t fight anybody, then you could get out six months earlier, that really works. And then the next question might be, “Well, okay, they’re engaging in all this programming, but does it actually help?” And the answer is yes. 

So we have these nice studies because we’ve tried these truth and sentencing policies in different places, economists have been able to study the impacts of them. Basically what you see is that once people no longer have the incentive to engage in this programming because they’re not gonna be able to get out earlier, they stop engaging in the programming and then once they do get out, recidivism is higher. 

And so all that programming actually did seem to work. Maybe it’s just the having to practice being good, whatever it was, the incentive to be on good behavior and do the right thing was really effective.

And when you take that short-run incentive away, people don’t do it anymore. And then when they do eventually get out and most of these folks are gonna get out eventually, we’re all worse off.

CARDIFF: The point about incentives is so powerful and it is one of the single most economisty things about this approach. I will say, I get why people are queasy with the idea that a parole board of strangers gets to decide this. It’s so much power over somebody’s life to hand over to strangers.

And depending on the specific group of strangers, some people are gonna get a lenient board and say, “Yeah, great. You can get out now.” And some are gonna get a harsher one. There is an arbitrariness to that, a randomness to that, I guess, that just feels wrong. It feels off. So there, there is a clear trade off, I think. Because some people are gonna get screwed in a situation like that. 

JENNIFER: Yes. But we do that already throughout the criminal justice system. We do it with the judges at the front end. We do it with the probation officers, the prosecutors; there’s just randomness throughout the system. And that’s just the consequence of something where it’s not obvious what the right answer is. So we trust professionals to use their best judgment. 

Your best judgment might be different from my best judgment. And that’s just natural. And that’s where we kind of have to land. We have, in general, I think in this country, we swing back and forth between giving people more and less discretion and constantly trying to battle this fundamental trade off. 

One thing that’s really interesting to me about this conversation is that parole boards — the idea here is that people do change while they’re locked up and the parole boards have the opportunity to then respond to that.

So the flip side of someone being sentenced to four years with the possibility of parole after two is that you also have the possibility of staying four. So if the parole board is seeing you’re not doing so well and you’re still fighting a lot of people and you’re getting into a lot of trouble while you’re incarcerated, they’re not gonna let you out after two. 

They’re gonna keep you the full four because you’re still a public safety threat, and that’s a nice option to have. I think the other interesting aspect of this conversation is we could do much better job than we currently do of giving parole boards the information they need to make good decisions.

It turns out right now, this is just not a policy area we’ve invested a lot of time into thinking about how can we help them make really smart decisions. One group that they don’t talk to very much, just because of the way that the system’s designed right now and the way information flows, is the corrections officers who are in the facilities with everybody. They have a lot of information about who needs to be there and who doesn’t.

So my team is thinking a bit about how can we facilitate the communication between corrections officers who are doing the hard work of trying to keep everybody safe. They have a lot of information about these individuals. How do they get that information to the parole boards so that they can help keep the community safe?

CARDIFF: We’re deep into the book now and deep into the conversation. Now, I’m gonna bring up the thing that, if I’m not mistaken, might have given you the biggest headache of all these topics. I’m not totally sure. There might have been other headaches I’m not aware of. 

JENNIFER: (LAUGHS) Yeah.

CARDIFF: But the back and forth on the research about something called “Ban the Box” really has been playing out publicly and quite heatedly for, I don’t know, more than a decade?

JENNIFER: Yeah. More than a decade.

CARDIFF: Let’s set this up. Ban the Box is a policy by which, I think it’s state-by-state, if I’m not mistaken. 

JENNIFER: Yup. And sometimes cities.

CARDIFF: And sometimes cities. So if a state passes a Ban the Box policy, it means that employers are not allowed to ask prospective job seekers, people who wanna work for their companies, whether or not they have a criminal record. That’s the so-called box. 

So you Ban the Box. And the intention behind this policy seems like a good one, which is we want people who have gone through the prison system. We want them to be able to get back on their feet. We want them to have a second chance. We want them to get work. And if employers know that they have that criminal background, then they’re gonna be less likely to hire them.

And so we want to take away that information, essentially, and see if that helps. This has now been tried in many places. There have been many studies on this, including your own studies on this. Why don’t we do this? Why don’t you sort of take the listener through what’s been found and then we’ll back up into some of the specific back and forth and why this has been such a controversial topic.

JENNIFER: Sure. These policies are based on this very human impulse: “Gosh, we wish employers didn’t care about the criminal record. We’ll just tell them they can’t ask” —

CARDIFF: Seems fair. 

JENNIFER: Seems reasonable. 

CARDIFF: Because these people, they’ve gone through the system.

JENNIFER: Absolutely. We want people to have a second chance. We wish employers agreed with us, and maybe they even do, but they’re trying to decide who to hire. And what the studies have found, to the first question, is just does it actually help people with criminal records get a job? Seems like the answer there is unfortunately no.  

There are a couple reasons that might be the case. One might be that it turns out doing a formal background check on someone, which under these policies you can eventually do — it’s usually after you’ve made a conditional hiring decision. So employers get to eventually run someone’s background check, but the idea is they’ve done an interview with the personal already, gotten to know them a little bit. Maybe the person’s explained their criminal record already to them so they have a more complete picture. They have more information than they would’ve had otherwise. 

It could be that they just Google the person and see if they’ve got an arrest record online. A lot of this information is available for purchase, and so it could be that they get the information some other way.

It could also be that they don’t do that, but they go through the process. But then when they do run the background check, they see they have a record, then they don’t hire them. And then that feels especially bad because now you’ve just wasted this person’s time. They’ve given them hope. 

CARDIFF: I mean, they’ve gotten a callback. They’ve gotten interviewed. You’re saying sometimes it happens at the end of the process, not at the beginning.

JENNIFER: And then at the very end they’re like, “Sorry man, you’ve got a record. This isn’t gonna work out.” And I’ve heard from folks who run nonprofits that this does happen a fair amount and it’s really dispiriting for people who’d rather just know upfront.

If you’re not gonna hire someone with a record, just tell me. Don’t waste my time. So anyways, we do have research now that’s showing that people aren’t getting jobs, but then even more concerning, and this was what first really interested me about this policy, there’s a real potential of unintended consequences here.

So if employers care for whatever reason about whether someone has a criminal record, they don’t wanna hire someone with criminal record, even if that feels really unfair, even if that for society makes it harder to reduce recidivism, et cetera, that is the concern the employer has and the incentive they’re responding to. Now that they can’t ask upfront if you have a criminal record, they’re in most cases gonna try to guess. 

That’s gonna be the human thing to do. They’re not just going to say, “Well gosh, I’ll just treat everyone as equals.” That’s what we would love for them to do. But in reality, what we all do all day long is try to kind of make our best guess based of the information we still have.

And in practice, because there is such large racial disparities in the United States,  between people who have records and people who don’t, it turns out that employers will discriminate, especially on race and young black men without college degrees, the group that is most likely to be helped by this policy, if it works ’cause they’re the group that’s most likely have a criminal record, actually we see employment for that group fall after Ban the Box goes into effect. 

So what seems to be happening now is that Ban the Box is making it harder for young black men who do not have a criminal record to get a job interview and get a job because employers are assuming they probably have a record and then they just interview somebody else.

And so we’re not helping people with records. We’re making it harder for a group of people who don’t have records, but have other challenges in the labor market, making it harder for them to find jobs. So on net, this is just a big mess.

CARDIFF: So it’s interesting and I wanna make sure that we distinguish between two different things for the listener. One thing that also is pretty clear in the book is that there is a big race gap in terms of hiring no matter what. It already exists. Ban the Box, no Ban the Box, there’s racism in the labor market. 

JENNIFER: Yes.

CARDIFF: That’s there. One thing you’ve found, and at this point there’s just an abundance of studies showing it, is that something like Ban the Box actually tends to increase the racial hiring gap. I guess economists refer to this as statistical discrimination. Where [employers] say, “Okay, there’s a higher share of Black men, let’s say, who are convicted of crimes, higher share than white men.” And so what they end up doing, if they don’t have the information, is they end up discriminating on the basis of that context. 

Statistical discrimination is something I’d been familiar with for a very long time, but I thought of it as a theory. And it was an interesting theory, but that this is something where you’d really want to test it in the real world to see if it actually is a thing.

And the abundance of Ban the Box policies [and subsequent studies] seem to kind of confirm that, yeah, companies do this. It’s a real thing.

JENNIFER: And people do it. The reason I think it’s a really useful concept is if we’re trying to reduce discrimination, you have to think, “Okay, what is the source of the discrimination to begin with? What is the problem we’re really trying to solve?” And economists think of two types of discrimination. 

One is animus. It’s just, “I don’t like people who’ve broken the law. There’s nothing you can tell me about them. I just don’t like rule breakers. That’s just it. Don’t wanna associate with them.”

Statistical discrimination is about information. “I am really worried about whether someone is gonna show up on time every day and stay out of trouble and be a good employee. That’s really hard to observe from a job application or even a first interview, so I have to kind of go on other information that I can see that is correlated with those things I actually care about. So I don’t actually care about the criminal record, but I view a criminal record as basically a negative credential. I have reason to believe it is associated with these other bad things I’m trying to avoid.” 

And so what’s nice about statistical discrimination is if we wanna reduce it, if we’re able to give people more information about the stuff they actually care about, they’ll rely on that thing they were using as a proxy less.

And so that’s true for the criminal record. It’s true for race. There’s other previous evidence that even just making it easier to check people’s backgrounds — when the internet became a thing in the 90s. It wasn’t always easy to check someone’s criminal record as an employer — and once it became easier, you saw the racial gap in hiring fell dramatically. 

So it seems like what people were doing, employers were doing, was using race as a proxy for whether someone had a criminal record. Once they were actually able to just check the criminal record, they didn’t care about race anymore —

CARDIFF: Well, they did, but they cared less than they did before. 

JENNIFER: Less. Less. They cared less about race than they did before. And so adding the box was actually a solution or partial solution to that earlier problem. So taking it away was a step backwards. 

CARDIFF: That seems the lesson of this chapter, which is add information. Don’t take it away.

JENNIFER: That’s my main takeaway from a lot of this. And we see this in other spaces too. There’s interesting research around drug testing enabling. There are a lot of reasons people don’t like letting employers do drug tests of employees. Turns out when you don’t let them do that, they try to guess who might be using.

And then you see increases in racial gaps in hiring. Same thing for credit check bans. On average, people who are Black have lower credit scores. So there’s a lot of concern that allowing employers to check someone’s credit score before hiring them is adding to these racial gaps in employment.

It turns out that when you ban the credit check, that actually increases the racial gap. It’s the other way around. It’s because employers, again, are gonna use race as a proxy when they can’t ask. So in general, I think the lesson here is the way information flows in these types of hiring decisions is just really important and information’s really valuable.

And we should always be airing toward more information rather than less. And I’m sure there are examples where that’s not always not gonna move us in the right direction, but I think it’s a good general rule of thumb. And in this particular context, it’s just super important.

CARDIFF: There’s an anecdote in the book about when you went to present to Congress on these findings. And you said that like you’d brought studies to a knife fight. This is something that you concluded later as you were reflecting on it because you showed these studies, but one of the other guests who disagreed with you, one of the other people who were testifying to Congress, started calling you all kinds of names and saying, “You don’t have any context for this. You’re trying to destroy people who have gone through the prison system already.” Stuff like that. 

It’s kind of interesting to me that this seems to have become quite personal. That this one did not stay in the “we disagree on the methodology of how to conduct the studies” and things like that. This one became something bigger than that.

JENNIFER: Yeah. This study, I think, was the first one that I wrote that actually really put me into these policy conversations. And it turns out that the policy/political conversation is often very different from the conversation you have in an academic seminar. And so it would — the goal was just to win the fight however you could.

I think about this policy area a lot because I think it was a space where a lot of economists especially were very concerned about this policy when we first heard about it. It’s like, this could really backfire, but we’re researchers. And so, that’s a theory. Don’t wanna go out there and say we shouldn’t do it because it might do this.

We wanna wait for the evidence. And there are other reasonable people who think it’s gonna help, so who are we to say? But then by the time the research was actually written, there was a really established Ban the Box lobby that still exists, that’s still out there passing these laws.

And at that point you have a lot of people whose careers depend on these policies being passed. Not on solving the problem, but on the specific policy solution. And that was then a very difficult thing to be fighting. And so it’s just — as someone who cares a lot about policy and thinks a lot, even more directly about policy now than I did when I was writing the study, it becomes a tricky balance between thinking about when you feel like you know enough to raise some red flags.

There is a real risk, too, if you wait too long. It’s not just that there are some people whose lives could be harmed. In general, we’re always trying stuff that may or may not work, but it’s really the political aspect of having these lobbies that emerge whose livelihoods depend on not changing their mind basically.

I don’t know the answer to that one, but it’s a real struggle.

CARDIFF: One example of adding information that does seem to work is if somebody coming out of prison has a kind of credential that shows that they have done work either in prison or something like that, it’s almost like proof of concept.

Like, “Yes, I’m gonna be transparent with you that I have been convicted of a crime in the past, but towards the end of my time in prison, I also completed this program and here’s a credential from a judge.”

That does seem pretty powerful. That does seem to work. Did I read that right?

JENNIFER: Yeah. So they go by different names in different places, but rehabilitation certificates or employability certificates, basically, if you go in, you can go to a judge and kind of say, “Look, I’ve done all this work.”

And try to convince them, “I’ve been rehabilitated.” And they can give you the certificate that then you can take to an employer. And there’ve been these nice studies, just like the studies showing that employers do routinely discriminate against people with criminal records. It turns out that if you have the criminal record plus the certificate, basically completely wipes out the effect of the criminal record.

You’re called back at equal rates to someone who had no record at all, when the person with the record but no certificate is still struggling. This policy seems to be doing what ban the box was hoping to do. Get the person’s foot in the door with the call.

These studies are all looking at callbacks so far. We’re hoping there’ll be some studies of these policies (which) are now in effect in a lot of states. 

CARDIFF: People actually getting hired.

JENNIFER: Actually see if people get hired. Actually see what happens to recidivism, things like that. But at least this first stage, we’re seeing that people are getting their foot in the door and having a chance to compete, which is great.

CARDIFF: I wanna do a bit of a lightning round on a couple of things. Something like that might help people get a job ultimately, you’re studying it, but it seems promising. 

Turns out having a job itself, though — not necessarily as effective as you’d hope at reducing recidivism. So people still seem to commit crimes at similar rates whether they have a job or not, after they get out of prison. Borderline shocking to me. I didn’t know that and I thought I’d been following your work pretty closely, but when I came across that in the book, I was like, well, that, number one, sucks, but number two, I was surprised by that.

JENNIFER: I think most economists are surprised by this, and it’s taken a long time for those of us in this field to be like, “Yeah, I think this is actually where we are. This is unfortunate.” And I think the way I think about it at this point is a job is probably necessary but not sufficient.

It’s hard to imagine being able to build a stable, crime-free life without a job. So you probably do need a job. The source of this conclusion is we’ve got some nice randomized trials of like transitional jobs programs where someone gets a job for six months, the hope is to help them learn some job skills, get a reference.

CARDIFF: Some money in their pocket.

JENNIFER: Money in their pocket. So we see those people show up. They work during this period, and then once the six months ends, they’re right back down with the control group in terms of whether or not they have a job that doesn’t actually work at transitioning them into the labor market.

And while they’re engaged in the transitional job, they’re actually not any less likely to re-offend. Maybe if the study’s found any effect, it’s super small. And so that’s super disappointing, ’cause it does feel like — there’s one theory here that if you have to show up at 9:00 AM every morning to work, you’re not gonna be out late. You’re not gonna be causing trouble with your friends. You’re now surrounded by a whole bunch of other people who are also trying to change their lives and are hardworking. 

That should be a good influence, and there are a bunch of reasons to think, plus the money issue, that this should be helpful and keep you out of trouble, and it doesn’t.

But then we do have other studies showing if you happen to get lucky and get released from prison during a really hot labor market in your county, there’s a housing boom and everyone’s hiring for construction workers or something, then you are less likely to re-offend. And so jobs are helpful there, but if we just give you a job, it’s not helpful.

And so it’s a bit of a puzzle. But in —

CARDIFF: Economic conditions, which are associated with more jobs. Those are helpful. 

JENNIFER: Yep. 

CARDIFF:  But not necessarily just having a job in isolation.

JENNIFER: So for the longest time, I did wonder if it’s maybe the friends and family of the person who’s released in prison who are better off, during the hot job market.

Regardless, we have not figured out. There isn’t real clear evidence at this point that giving someone a job is actually useful at reducing recidivism, which then puts me in the place thinking — maybe I haven’t completely given up on the idea that a job is necessary, but not sufficient.

That’s the not sufficient piece. So there’s still other stuff that’s a problem and maybe it’s mental health issues, maybe it’s substance use, maybe it’s not having the network be able to find a job. I don’t know. But there’s clearly other stuff.

CARDIFF: Quality of the job or something. Like in a hot labor market, jobs—

JENNIFER: Pay well.

CARDIFF: —pay more, higher raises, better benefits. More flexibility, stuff like that. And we don’t have a lot of hot labor markets to pull from in the last quarter century.

JENNIFER: Right. And that’s really hard for going and telling a policymaker, “If you could just make your local labor market really hot, that would be very helpful.” And they’re like, “Please get outta my office.” (LAUGHS)

CARDIFF: “Please leave. Thanks.” (CHUCKLES). 

JENNIFER: “We’ve thought about that. Yeah.” (CHUCKLES)

CARDIFF: “We have to hope that the Federal Reserve does its job.” “Oh, you’re telling me a good economy is good. Thank you.”

JENNIFER: Thank you. (CHUCKLES) Insight from the economist. (CHUCKLES)

CARDIFF: “Amazing. I didn’t didn’t have a 21-year-old junior staffer intern who could tell me that.” (LAUGHS)

A few things I’m just gonna mention, and I will essentially tease the book if people want the details of it because I have something important to ask you as we close. 

Something else that seems not to work: something called wrap-around services, which is essentially a flood the zone thing where somebody gets out of prison and you offer them almost everything: counseling, mental health services—

JENNIFER: Job training, all the things.

CARDIFF: The whole thing. It doesn’t work!

JENNIFER: Yep.

CARDIFF: What does seem to work: giving people money, especially right after they get out of prison so they don’t get desperate. That does seem quite promising. I don’t know how definitive —

JENNIFER: Or like key moments where they’re at risk of becoming homeless or something, it’s like, “Great, let’s pay your rent for the month.”

So, and then you’re less likely to be arrested. 

CARDIFF: Don’t cut people off from Medicaid or healthcare generally.

JENNIFER: Mental health care is something I think a lot about right now. 

CARDIFF: That’s helpful for people. It’s also helpful, by the way, for their kids and the likelihood that the children of people who commit crimes will themselves grow up to commit crime, that kind of thing.

JENNIFER: That’s public benefits in general. It’s thinking about SNAP benefits and stuff like that. Nutrition for young kids is, turns out, really protective for lots of different reasons that we could have a longer conversation about. 

CARDIFF: Big section that, again, I’m gonna leave as an incentive for people to buy the book, which they absolutely have to. Big section on how to help kids, and then also teens and young adults, which I thought was really powerful as well

JENNIFER: There is this punchline. Helping people change their trajectory they’re on. And pulling people out of the system is not impossible. I hope that is a takeaway from the book. There’s a lot we could be doing to change them, to put someone on a better path, but man, is it much easier if we stop them from ever being pulled into the system to begin with.

So doing some of that preventive stuff for kids and young adults would be a really beneficial policy move. 

CARDIFF: I was thinking throughout the book: what are some of the through lines besides helping people get a second chance at life? The truth is that your approach is not one that lends itself to through lines. It’s very, as I think you refer to it as an incremental approach, like do these studies see what works, and then put in place policies that are consistent with the data.

But it’s not like this overarching philosophy. It’s not this big structural theory for how to do criminal justice better or anything like that. And I guess what I’m wondering is does that make your job harder as a salesperson for this? You see what I mean? 

JENNIFER: Yeah.

CARDIFF: It’s easy to have a slogan, it’s easy to be like, “This is our North Star.”

You don’t really have one of those. Essentially your theory is: there’s all this data we have and yeah, one new policy in isolation may not make a huge difference, except for the people specifically affected, but societally it may not. But if you do all these things at once, that would have a big effect.

JENNIFER: Well, in some of these small incremental changes we discussed before, like the erring towards leniency for first time offenders, a 50 percent reduction in recidivism is not a tiny change. 

And so, some of these small changes do have big effects. But yeah, I guess I think of the overarching theme here as being, “We should be approaching this policy space, and really all policy spaces, with more of a scientific mindset and a scientific approach, rather than with the kind of big idea about what is going to be effective.”

So thinking of everything as being a process of trial and error and focusing on getting really clear on: what is the outcome we’re actually trying to change? What does success look like? 

And keeping our eye on that ball. Our goal is to solve the problem. Our goal is not our preferred policy solution.

And so really just shifting everything to thinking about how can we — we have to go out and just try stuff. I think a lot of people hear me talk about things about whether or not something’s evidence-based and they say I just wanna slow us down. I’m just trying to put sand in the gears or whatever and slow the process of moving forward.

And my response is always, “the only way we’re gonna learn what works is if we try stuff.” We have to be going out there and trying stuff, but we have to do it in a way where we stay humble. We don’t become wed to solutions until we know if they’re effective and we’re implementing things in a way that we can actually learn from.

And that’s not impossible. There are also a lot of people just don’t know about how economists think about studying the impacts of things. That there are these natural experiments we can use. We don’t have to just be running randomized trials. And so there is a way that we can be learning as we go, and it’s really the “we have to learn as we go approach” that I think is the new lens, the different worldview, I guess I’m trying to bring to the conversation.

CARDIFF: What have you learned about persuasion in doing this work?

JENNIFER: A colleague of mine, Kevin Ring, who leads our criminal justice advocacy team, is constantly reminding me that lawmakers do not care about research. (CHUCKLES) And so we have—

CARDIFF: (LAUGHS) That’s a bummer. “Thanks, I just spent two decades doing research, so I really appreciate it.” (LAUGHS)

JENNIFER: Yes, “Thank you.” So we have a nice agreement that’s like, “Okay, I get to sign off on something that’s evidence-based, and you sell it however you have to sell it, as long as you don’t cite bad studies. If you’re gonna cite a study, at least cite a good one.” 

But yeah, so stories, individual people’s, there’s so much of policy making that is about individual relationships, just getting to know and trust people — in terms of lawmakers, in terms of advocates. And so really, the other line I have learned from my colleague Kevin is that “you learn from people you love.”

And so just really trying to get to know individuals and see them as people that share your concerns and goals is very helpful. It’s probably not surprising to any of the other social scientists out there, but the economists are like, ‘this is fascinating.” (LAUGHS)

It’s also interesting because I’ve realized it’s the same issue on the research side that we’re really trying to solve.

It’s like if we want more and better research, if we want people to go out there, researchers and practitioners, to work together to study the impact of some new policy. And my policy colleagues will say, “Well, why aren’t these studies being done? Maybe the researchers don’t know that’s an interesting question.”

It’s like, yes, of course they know. But they’re trying to cold call the police departments and no one’s returning their calls. If we actually introduced these people and they got to know each other and trust each other a bit, then they might wanna work together. 

So it just turns out that so many of the problems we’re trying to solve from all the way from research, to policy, to implementation, are just very human problems where we just need everyone interacting in the real world. Which is fun, which means we get to just throw some parties and make sure everybody gets to meet each other.

CARDIFF: Perfect place to close. Jen, your book is available for pre-order everywhere. Who’s the publisher?

JENNIFER: Holt/McMillan.

CARDIFF: Ok, so go to the Holt/McMillan site, or go to Amazon, or go to Barnes and Noble, or wherever you want. Go to your indie local bookstore.

JENNIFER: Yes. That would be fantastic.

CARDIFF: Go order it. It’s fantastic. Thanks for doing this.

JENNIFER: Thank you, Cardiff. I really appreciate it.