Freakonomics for Criminal Justice: Our Discussion with Economist Jennifer Doleac

Jennifer Doleac is a Stanford-educated economist, a former college economics professor, and the author of The Science of Second Chances:  A Revolution in Criminal Justice, published by Henry Holt and Company and available in print as of February 24, 2026.  

We spoke with Jennifer about her new book, which is billed as “Freakonomics for criminal justice,” and explores real-world experiments that suggest what policy changes and incentives offer the greatest hope of true criminal justice reform.  Doleac takes the reader through each stage of the criminal justice process, from arrest, to prison, and release, to identify data-driven solutions to make our society both safer and less punitive at the same time.  

Doleac is the Executive Vice President of Criminal Justice at Arnold Ventures, a philanthropy focusing on evidence-based public policy.  She is also a frequent contributor to The Washington Post, Time, and Bloomberg Opinion, and host of the Probable Causation podcast on law, economics, and crime.

Transcript

SPEAKER_01 0:30

This is the Just Pod.

SPEAKER_03 0:36

Welcome back to the Just Pod. From Philadelphia, I'm Justin Danilavitz.

SPEAKER_02 0:40

And from Washington, DC, I'm Janard Butler.

SPEAKER_03 0:44

Jennifer Doliac is a Stanford educated economist, a former college economics professor, and the author of The Science of Second Choices, A Revolution in Criminal Justice, published by Henry Halton Company and available as of February 24th, 2026. The Science of Second Choices is billed as free economics for criminal justice and explores real-world experiments that suggest what policy changes and incentives offer the greatest hope of true criminal justice reform. Doliac takes the reader through each stage of the criminal justice process from arrest to prison and release to identify data-driven solutions that she argues can make our society both safer and less punitive all at the same time. Jennifer Doliak is the executive vice president of criminal justice at Arnold Ventures, a philanthropy focusing on evidence-based public policy. She's a frequent contributor to the Washington Post, Time, and Blue Berg Opinion, and she hosts the Probable Causation podcast on law, economics, and crime. Jennifer Doliac, it's a privilege to have you. Welcome to the Just Part.

SPEAKER_00 2:03

Oh, thank you so much. I'm so happy to be here.

SPEAKER_03 2:07

The new book, your first book, uh is a terrific read. Um it is fascinating stuff. It is chock full of real-world experiments as as you describe them. Um maybe we can start off there by just sort of you explaining what is a real world experiment, why is there a need to look at real-world experiments? Um it's a topic that you come back to frequently in the book. Um and maybe you could just pick one of them to start us off with that that you find most interesting or fascinating that leads to the various conclusions that you draw in the book.

SPEAKER_00 2:46

Sure. Yeah, so I'm an economist by training, as you mentioned. Um, and economists are obsessed with distinguishing correlation from causation. That's, you know, we have we have our own theories to bring to the table sometimes, but I think often in policy conversations, our main contribution is really nailing down what's the causal effect of some policy or intervention. And so, you know, what makes this fun for us is that every once in a while you're able to run a real randomized experiment like you could in a lab. Um, I think a lot of economists feel like that's that's almost too easy, right? Like it's more fun if you have to work for it. Um, and so we're always looking for what we call natural experiments out in the real world. So, some policy change or some way that a program is implemented that sorts people or places into treatment and comparison groups as if at random, that gives us something akin to that lab experiment that we wish we had, um, and that enables us to measure what's the impact of that treatment, that program or policy on the people who are who are receiving it. Um, and so one example of this, there are a few different designs that show up over and over again. But one of my favorites is from one of my own studies on the effects of DNA databases that are used by law enforcement. And so these databases add DNA profiles of people convicted or arrested of certain offenses. Um, and and they're governed by state law in the United States. And so as of July 1st, 2019, say someone convicted of robbery in X state is required to have their DNA in the database. That means that anyone convicted of the same offense just the day before does not go in the database. And so that sets up a natural experiment where you have people convicted of the exact same crime with everything else about the criminal justice system going on as usual. But as of this implementation date, you have some people going into the database where if they'd gotten lucky and been and you know committed their crime or been convicted or whatever the trigger is just the day before, they wouldn't have been. And you can compare these sets of people over time, the people who were just barely in the database, the people just barely not in the database, and see what happens to them going forward. And because everything else is the same about them, if we see that their trajectories differ, if if what happens to them is different, then we can attribute that to the effect of being in the DNA database. And so in this case, what we what I find is um being added to the DNA database, which means that you're more likely to get caught if you re-offend, leads to a dramatically lower recidivism rate among among those folks that were added. So um it's winds up being evidence that people are deterred by an increased probability of getting caught.

SPEAKER_03 5:36

So interesting. Um another really great example that that you provide in the book um comes out of Suffolk County in in Massachusetts, right? And and as I understand it, the the natural experiment that occurs there is you you sort of have um a population of people who are arrested, um randomly assigned, uh given the circumstances of the situation in Suffolk County Nass. Um and so some of them get sent to ADA's assistant district attorneys who have a little bit of a lighter touch, they're a little bit more lenient. Some of them want to slam the book at the same, you know, otherwise similarly situated individuals. And uh you find from that comparison that the people who are given a soft landing, who are maybe not prosecuted, who are diverted out of the prisons or the criminal justice system, ultimately end up having a lower recidivism rate, um, which is fascinating and it's not what a lot of people might expect, right?

SPEAKER_00 6:44

Yeah, it's not what I expected going into that study. I think I was the real skeptic on the research team about um the potential upsides of this um, you know, this this theory of prosecution and the the uh hopes that we could reduce crime just by going easier on people. But the data in this case really convinced me that airing toward leniency, especially for first-time offenders, so that's really where we see the big, big benefits. If it's your first arraignment for, in that case, a nonviolent misdemeanor offense, then having that case dismissed in that initial arraignment hearing because you just happened to get lucky and got a more lenient prosecutor, you're dramatically less likely to reoffend. And so, so it really does seem like the, you know, the process of getting to that point, to that arraignment hearing wasn't no punishment, right? There are consequences for whatever it was you did. You were arrested, probably booked in jail, you're worried about this case, but then that's it. We didn't have to give you a criminal record and some sort of punishment. You course correct on your own, it turns out, for most for a really large share of those offenders. Um, and the same is true. There's another study looking at nonviolent felony defendants in Harris County, Texas, where I live now. Um, and uh what that study is looking at is the impact of a deferred adjudication for those first-time defendants and finds very similar results. If you're given the opportunity to avoid that first felony conviction, you're dramatically less likely, like on the order of 50% less likely to re-offend and wind up back in court in the future. So it really does, you know, helping people avoid that first criminal correct conviction really helps people um, again, just course correct on their own and find their way to a better equilibrium without uh burdening them with a criminal record.

SPEAKER_03 8:36

No doubt. And Jennifer, you know what it reminds me of is you you have these computer-generated tests nowadays, right? Like the SAT or ACT or what high school students typically take. And some of them that very first question can be very determinative. Because if you make a mistake on that first question, you are shunted onto a pathway that makes it super hard to get back onto the other branch of that decision tree and can have an enormous impact on your score at the end of the day. But yeah, what we're talking about is the the life score, the the resulting impact on one's life, right? So if you if you make a misstep with serious consequences, you're shunted onto a pathway that makes it very hard to get back. Um but the diversionary solution helps you kind of maintain some ability to remain on the on the good path and not completely destroy your future, right?

SPEAKER_00 9:36

Yeah, I love that. That's such a great comparison. Um and it's especially this is I think especially interesting in a context where, you know, uh in the policy space, we've been doing a lot of work over the past couple of decades to try to help people undo the effect of a criminal record later, you know, years later. And through policies like Ban the Box and Clean Slate, like we recognize that having a criminal record makes it harder to find a job, makes it harder to find housing. And so the hope, the hope is to try to help people kind of either clear their record or just find a way to undo that effect later. And it turns out that's really hard to do. Those policies just aren't effective. Um, and so being more judicious in how we're doling out the criminal record on the front end um does keep people from being diverted onto this worst path that then is very hard to kind of get back to a better equilibrium from.

SPEAKER_02 10:32

I wanted to ask, um there's a subtle title in the tension, a revolution in criminal justice. Now, revolution is a typical moral uprising, yet science is incremental and cautious and probabilistic. So why call this a revolution?

SPEAKER_00 10:54

Yeah. I think it's a fundamentally different way of thinking about this policy space than most folks who work in this space are used to doing. Um, I think this is a I think honestly, this is probably true of most policy spaces, but uh you you tend to have policymakers or decision makers who are rewarded for saying they have the solution to this complex problem, right? And they they're bringing the solution. And then their entire goal is just to implement that solution. And then everyone applauds and no one really has an incentive to figure out if that solution actually worked. Like we'd rather not know if it didn't work, right? Because that was the goal. The goal was to implement that solution. And it just turns out that's a really ineffective way to solve problems, especially in a space where these are really tricky problems to solve. Like if they were easy to solve, we would have done it already. The fact that we're still struggling with so many of these issues decade after decade is a sign that these this is just really difficult. There are a lot of complex factors, human behavior surprises us. Um, and so my view is that our goal should be to fail fast rather than pretending we're not gonna fail at all. And that's just a very different lens through which to see the policymaking process and um and the criminal justice process to large to a large extent. Um, and so so that's the lens I'm trying to bring here and the sort of different worldview uh to the to this policy space.

SPEAKER_02 12:20

Let me ask this um, because I think that what you've said and and you talking about the nuances in this is so important. And so let me press gently here. But let's say that if the evidence showed that a particular reform reduced recidivism overall, but slightly increased racial disparities in numerous contexts, how should policymakers weigh that trade-off? Is it safety first, is it equity first, or is it again something more nuanced?

SPEAKER_00 12:55

Yeah. So my view as an economist is um that, you know, I ask people what their objective is, and and then I help them maximize their allocation of resources to achieve that goal. And so I personally am agnostic about whether a policymaker's objective is purely public safety or purely reducing racial disparities or some combination of the two. Um, I really do view my role as um helping people understand trade-offs and what the empirical effect is of some intervention on on whatever outcomes they care about. Um and so I'm always very careful to say in in conversations about, say, how long sentences should be. We know uh from all of the evidence that if what we care about is reducing crime, long prison sentences are a waste of money. Now that's not the only thing we care about when we lock people up. We also care about retribution. You know, that is a reason that we we incarcerate people. I have nothing to say about that as an economist because the data can't tell us how much retribution or mercy someone deserves. But so I'm always careful to say up front, like if this is the outcome you're care you care about, then you should do this. If you care about this other outcome, you should do that. But I try very hard to um not inject my own preferences about what the out what the right outcome is to care about. Which I think is also a different lens from the way people talk about this, these issues.

SPEAKER_02 14:29

Yeah. Absolutely.

SPEAKER_03 14:30

Yeah, that's super interesting. Uh do you so I have to ask you about AI, artificial intelligence, and and the impact this might have on uh the economic analysis of criminal justice reform? Do you think we are, I mean, I could imagine that you could um sort of calculate the optimal prison sentence. It might be complicated, it might be very individual specific, but we we may be getting to a point where AI can do that. Um have you have you considered that in in your research, the application of AI and how it can help inform policy choices?

SPEAKER_00 15:10

Yeah, so there are, I mean, it's sort of similar to the idea of a risk assessment, right? And trying to figure out just whether, based on someone's, based on everything we know about a defendant, whether they they should be locked up, um, whether they should be detained pretrial, whether they're safe to release, what kinds of programs they should get while they're incarcerated. So we do try to crunch these numbers currently. Um, most of the tools we're using are not based on AI. They're based on more primitive, you know, more primitive algorithms. And so there's surely going to be a step up in the years ahead. And there's already a race from lots of high-tech vendors to introduce new tools in the system. Um, I think honestly, what people in the criminal justice space seem most excited about, and I'm curious if your conversations uh reveal the same thing, is just the potential of AI to make it easier for everyone to do their jobs better, to just like handle paperwork and make sense of evidence in criminal cases. And we have all of this digital evidence now from body cams and um all kinds of all kinds of different sources. And public defenders as well as prosecutors have to sift through all of that evidence and they have to do it quickly so that they can, you know, uh be able to handle a case and keep it moving. And that's just impossible given current technologies and current workforce shortages. And so in a in a context where, you know, throughout the criminal justice system, we don't have enough police, we don't have enough public defenders, we don't have enough prosecutors, we don't have enough 911 dispatchers, we don't have enough corrections officers, like at every stage of the criminal justice system, there are workforce shortages. Everyone I talk to is just really excited that AI might be able to make everyone who is there more efficient in order to help us just handle the workload that we currently have. But then I'm sure we will, you know, we'll we'll find some new uh new tools to help us make smarter decisions along the way. But I think, I think some of that, you know, less sexy stuff is gonna be the bigger game changer.

SPEAKER_03 17:09

Yeah, absolutely. I want to uh get back to the book for a second and ask you about um broken windows theory, which which you address in the book. Um the the term broken windows came from the Atlantic Monthly article in 1982 by two social scientists, James Q. Wilson and George L. Kelling. And the theory in essence is that a a community that has conditiona of decay, literally broken windows on factories, you know, uh burned out residences, weeds growing in lots, graffiti on the walls, um, there's likely to be a vicious cycle or a or a downward spiral from that for a number of reasons. Um I think in part people just sort of associate that with disorder and a and a view that people can get away with more. Um and the findings in your book I see as sort of uh a counter to to broken windows theory in a number of respects. So where there's an opportunity to to go go more harsh and to sort of sweep with a with a new broom and really throw the book. Instead, as in the Suffolk County natural experiment we discussed, it turns out that a bit more lenience, uh leniency can can achieve good effects. There's another example among many that comes from Harris County, uh, which you spoke about. And and that had to do with diversion, uh, where you had the opportunity after a failed ballot initiative um to sort of raise funds for a new jail, remarkably the initiative fails, right? It's it's what people would not expect to happen.

SPEAKER_00 19:01

Especially in Texas.

SPEAKER_03 19:02

Yeah. People are not keen on paying for this new jail. And so the reality is um there has to be a pressure valve to let inmates out, and they start diverting defendants who otherwise, you know, they would have housed in a new facility. Um so so again, it's sort of counter to the broken windows, it's a it's a bit of a lighter touch. Um do you agree with me that you're sort of reacting to broken windows in a very large sense, 30,000-foot view?

SPEAKER_00 19:35

Yeah, I think you know, as I was thinking through that chapter and really grappling, I was writing it as we were having all these conversations in a lot of big cities across the United States after COVID about what to do about disorder, right? And and it seemed clear that the reason disorder had increased in a lot of places was that we just stopped enforcing laws against trespassing or against shoplifting or whatever. There just weren't any consequences, right? And so I got this question all the time from policymakers. Does this mean broken windows works? And my my reaction was always, no, that's not really what it means. But this writing this right really forced me to kind of figure out like, where is it that I disagree? And I think, I think where I would diverge, at least from the popular conception of broken windows, um, is we can distinguish two things. There's there's the prop increasing the probability of getting caught, and there's increasing the punishment. And I think there's a lot of evidence at this point that people are deterred by increasing the probability of any consequences. So the swiftness and certainty of punishment. And there is much less evidence that people are deterred by the severity of punishment or the length of a sentence. And so to the extent that broken windows suggests we should not only, you know, arrest everyone who breaks the windows, but also throw the book at them, I would challenge people to break that into those two parts. I think it is worthwhile if we don't want broke if we don't want windows broken, we should arrest people who are breaking windows. But I don't think we have to put them in prison for two years, right? I think we should just have increase the likelihood that people face any consequences and that we'll see behavior change. Just having reliable, predictable consequences for bad behavior is a great way to change behavior. And that is just simply not the way most of our criminal justice system operates at this point, even for really serious offenses. I think across the country right now, the the number is it's about 60% of people of um reported homicides or homicides in general result in an arrest. And that means it's basically a coin flip. If you commit murder, you're going to get away with it. And that's for like the most serious offense where we really try, right? And then if you go down to less serious offenses, but still felonies, think about burglary or or larceny, um, the the clearance rate there is more like 10%. And so we were run running essentially a lottery where, you know, a very Unlucky few will get a really harsh punishment. And that's just simply the opposite of what decades of evidence from economics, but also feels like psychology have shown is the is a the best way to achieve behavior change. So I think I I am with the broken windows idea in terms of making sure there are that we take you know any crime seriously and have and increase the likelihood that people face consequences for it. But I I part with um at least those pushing for um harsh punishment for for those low-level offenses.

SPEAKER_02 22:39

I want to dive into that a little more philosophically. At its core, broken windows assumes disorder is predictive of danger. But in the cities that Justin discussed, that you discussed in your book, and in cities throughout America, um if we're being honest with ourselves, the cities where you see factories with broken windows and graffiti and weeds growing in yards and um vacant homes, things of that nature, it's typically associated with racial segregation. Those areas are typically predominated by minorities, and what we label subsequently then as a disorder can ultimately be culturally coded. So how do we prevent empiricism from laundering subjective judgments about whose behavior feels threatening?

SPEAKER_00 23:48

Yeah, I think it's a great question. I think um even in cities like San Francisco, right, which are not particularly diverse, um, you have you know encampments and public drug use that were making people feel very unsafe. So I think this happens, I hear you, it happens more in more segregated places, but I think especially post-COVID, we saw that it happens everywhere to some extent. Um I and I think the the real conversation we were, we were having as a country then was whether it's okay to think of disorder as being crime, right? I think that certainly the conversations I was having with lots of people in the criminal justice space is like you have a lot of folks who are saying, well, disorder isn't, you're not, you're not actually in danger. This is not actually correlated with murder, right? And San Francisco is a great example. You have a lot of disorder, a lot of traffic accidents, but very few murders. And so San Francisco was really never an unsafe place from in terms of violence, but people felt unsafe. They didn't like walking around their neighborhood anymore. And I think what my takeaway is from these more recent conversations is that it is okay to value order in and of itself, right? That there's something of a social contract that we all feel like we've signed, if especially if we live in an urban area, there everyone just needs to abide by some basic norms if we're all going to be able to live here in this, this these close quarters together. And so people being upset about disorder in and of itself, I think was a lot of the policy conversation. And then figuring out what to do about that, right? I think in your point that, you know, the broken windows philosophy was really based on this idea that disorder begets violence. I don't think we have any evidence of that at this point. Um, you know, the the I'm thinking even of um, you know, there's a one of many papers showing that putting more police on the streets reduces violent crime. Also looks at, you know, you you see changes in just low-level arrests and you're putting more police on the streets, and sometimes that increases low-level arrests. That's not what is changing the violent crime rate. Um, and so, so there have been people who've been able to kind of distinguish these two things and really demonstrate that these low-level arrests are really not the thing that are keeping us safe. That doesn't mean that they're not warranted in and of themselves. And so I do think it's important, I think it's an important conversation to have as a society. How much do we value order versus disorder? And which of these rules do we think are important to enforce? And then if, you know, once we decide which rules those are, we should enforce them if we want to just discourage that behavior. But we should not think of that as being our strategy to reduce homicide. If we want to reduce homicide, we're gonna have to do something different over here and not not focus on the um public drug use.

SPEAKER_02 26:47

Is that I No, I think that I, you know, it's it's it's a hard question. And I I I I I think that that is a a very uh very thorough answer. And I you you rely so heavily on science, which is very appreciated because it's a novel approach to, you know, I mean this is political science, but it's a novel approach to a problem that has been um I think picked apart and whiteboarded and tortured in so many different ways, except the way that you're doing it. And and so I applaud you in that and and and I I I I'm curious if we follow the science wherever it leads us, yeah, years from now, even when it becomes politically inconvenient, yeah, do you believe incarceration rates would fall dramatically, or do you believe that some communities may see increases in targeted enforcement as a result?

SPEAKER_00 27:57

Um I genuinely believe that focusing more on increasing the probability of getting caught is our ticket out of mass incarceration. Like I do genuinely, based on all the evidence we have, think that that is the the type of intervention that will deter and prevent crime and that will start a really beneficial cascade through the system where we have fewer and fewer people cycling back through.

SPEAKER_02 28:26

And I agree. And I must let me let me say I agree a thousand percent. I think that a lot of people are, I think a lot of people are are, I think a lot of criminals specifically, both of us being you know former prosecutors, yeah. Um there you you do see an ethos of I don't believe that I'm gonna get caught, so I'm gonna do whatever I want. So I absolutely agree that it's not about the idea of deterrence, it's about the belief that I am actually going to get caught. So let me not engage in this action whatsoever.

SPEAKER_00 28:53

I need to go, yeah, it's time to clean up my act.

SPEAKER_02 28:55

It's the it's the label on the smoker's box. It's the label that tells you you're going to get cancer and die. Do not smoke this. We don't have that at gun shops. We don't have that when gangbangers are training youngsters to become drug dealers. So I agree with you a thousand percent.

SPEAKER_00 29:12

Yeah. But to the kind of meat of your question, are there ways that people are people going to find ways to implement these laws that further different ends? I think this is where human behavior does not, you humans don't always do exactly what we hope they'll do based on the law we wrote, right? And so this is in some ways the spirit underlying a lot of the book that I've written is just we have to be trying new things if we want to find our way to a new and better equilibrium, but we have to do it in a way that we're keeping our eye on the outcomes we're trying to change and make sure that whatever it is we're trying is actually working. Um, and so I would say, I mean, you know, a related conversation I have a lot about increasing the probability of getting caught is one real, you know, a lot of the best ways to do that have some or some variation on surveillance, better surveillance, right? More cameras everywhere, DNA databases, drones as first responders, like all of these new high-tech tools. And so we have to be really thoughtful about what are the data safeguards we're putting in place to make sure we don't have misuse of that information, et cetera. Um, and so that can go to part of your question if we want to make sure that the data and the technologies are not being abused. But I think, I think the other piece is just to get really clear on like what are the outcomes we think are important here. And if we value both public safety and fairness, or, you know, minimizing racial disparities or however, whatever the outcome is that you you wanna you wanna put on the table, we just need to be keeping our eye eyes on both of those balls. And um and I think too often it's just amazing to me how many times I'll have conversations with folks in the policy world who are rooting for some policy. And I'll say, that's great. What is the outcome you're trying to change? Just want to make sure like we know, we'll know it, we'll know, we'll know how it worked, if it worked or not, right? Like what's the outcome you'd want to measure to see if it changes or if it worked? And they just sort of like their eyes glaze over and they just kind of look at me blankly, right? And it's just, I think a lot of times we just don't pay enough attention to what is it that we're hoping that this thing we're doing is going to achieve? And what are some unintended consequences we might be worried about that we should also be measuring and tracking? And I think if we had more of those conversations up front, not only would we be able to tell if our interventions are having the intended consequences we'd hope for, but we'd also know if they're having having unintended consequences. So I guess my main answer is just if we're worried about an increase in racial disparities or worried about some people abusing some laws in certain ways, let's put those outcomes on the table so that we can all track them.

SPEAKER_03 31:57

Jennifer, when you talk about the probability of being caught as a key driver and the analysis, it makes me think about um a great additional natural experiment that you talk about in your book, um, and that is SCF, the concept of swift certain and fair sanctions, particularly in the case of um offenders with a drug abuse history. And so and this is also interesting because it's sort of an exception to a purely lenient approach, right? It's not entirely about leniency, because in this instance, with swift search and fair sanctions, uh there are there are absolutely consequences for a violation of probation, I'd say. Right. But what I thought was so interesting is the the difference between sort of an unpredictable consequence from violating supervision conditions, um, even under intense supervision. That's one of the takeaways is that the intensity of the supervision is not necessarily the be-all and end all.

SPEAKER_00 33:19

Yeah, so the the community supervision space is one that's really challenging because we do have just remarkably randomized trial after randomized trial, like actual RCTs out in the world where we randomly assign people to different levels of supervision. And if you're at a higher level of supervision, you rack up technical violations and wind up back in prison, but it doesn't reduce new crime in any way. It just doesn't actually improve public safety. And so this signals to me that, you know, either the rules that we're asking people to follow are not good signals of risk, right? And so if you break this rule, it doesn't actually the idea is that should be a signal that you're a high risk person and we should put you in prison because community supervision isn't going to work. It's either, they're either uncorrelated with that risk or we're just not enforcing them very well. Um, but either way, what what we're doing right now is not working. And so let's just stop most of it, is my read of the of the research here. Um, except in this case of uh Swift Certain Fair sanctions for alcohol and drug use. And I think what makes this a really interesting space to experiment is that, you know, from a lot of the rules that we we ask probationers and prolees to to abide by, it's not really practical to enforce them in a predictable and consistent way. So, you know, one rule people often have to follow is not to associate with other people with felony convictions. Well, what are we gonna do? Follow people around and check the records of every person they talk to on the street? Like that's crazy. Like we're not actually gonna do that. And so again, this is a space where people presume surely figure out really fast, well, I have to follow this rule, but it's not actually, I'm not actually gonna get in trouble if I break it until maybe one day they get unlucky and then they're thrown back in prison. Um, but with substance use, we can do frequent drug and alcohol tests and see if you've been using. And so there have been a few different versions of this that have been um rolled out in ways that we can study them. One is a program called Hope that was developed in Hawaii. Another is 24-7 Sobriety that was originally designed in South Dakota for DUI defendants. And the idea in both cases is that you have um very frequent random drug tests. Or in the 24-7 sobriety example, you have just you go in every morning and evening and breathe into a breathalyzer. And if you blow hot, you immediately are spending the night in jail. And so that's the idea here is you just have like frequent tests. And if you, if you fail, you're immediately incarcerated, but for a very brief period of time. And it turns out people learn really quickly that there are consistent consequences and they just they just stop drinking and they stop drug use. And um, I think in the 24-7 sobriety context, it's something like 99.9% of people who are enrolled in the program have have negative tests, right? They just stop using. And, you know, one one quip I've heard is that had him explain this is it's just it's easier to refuse the first drink than it is to refuse the third drink, right? So if you can stop people from even having the first drink, that's really helpful. Other places that have tried setting the bar at, you know, point oh, you know, BAC of 0.04 instead of zero, it's not as successful for exactly this reason. Like you really need to say like zero tolerance, no drinking, and then people can follow that rule. Um, and this is really counter to, I think, the way or the theory of change that the public health community often offers, which is the only way to address substance use disorders is through treatment. And so we should be mandating treatment, but inset, you know, changing incentives isn't gonna do anything. We have this example where changing incentives actually does does a lot for a lot of people, doesn't completely solve the problem, and but it does help us identify really quickly who this is not working for, who are gonna need more than this. And it also another way of looking at it is it could just be providing the incentive people really need to go engage in that treatment that was available all along. But now they're realizing oh, like the system is actually like it's gonna hold me accountable for this. And so I really need to figure out how to um get my life together. But regardless, like these programs seem really effective. Um, and it's really cool to see them rolling out across different states.

SPEAKER_03 37:37

Yeah, it sounds really promising. On a different note, maybe we could just talk about uh the prison system, prison sentences, and how it's done differently in different parts of the world. I found this really interesting in the book. The discussion of really nice sort of college dorm-like uh facilities in Norway, uh better health outcomes in Sweden, um, the impact of another natural experiment, inmates kind of randomly assigned to newer facilities in Colombia, where as a result it was tracked and found that there was a lower rate of recidivism. What do we have to learn from practices in other parts of the world in terms of our prison system here in the United States?

SPEAKER_00 38:23

Yeah, so it's I think thinking about how to reform prisons, especially in the US, is um is especially challenging because we don't have many studies yet. This is changing, but we don't have many studies yet that are exploiting these kinds of natural experiments to figure out what works inside a prison. Like how should we be um either redesigning prison units or what kinds of different programs should we be offering? But we do have a lot of those studies going on outside the US. And so this was a great opportunity to really tour the world, I guess, and and see, you know, what are other countries trying that maybe we could uh take pieces of and try in the United States. And of course, there are lots of differences across different countries. Not automatically true that just transplanting a Norwegian prison into the US is gonna make us like Norway, but um can give us some ideas. Um, so yeah, so so in Norway, there, you know, I think Norwegian prisons are often held up as a model of what rehabilitation could look like, right? They're they're very quiet, they're often out in nature somewhere, people can walk around and um and maybe they leave to go to work and it's just it's very tranquil, uh, which is the opposite of I think how people would describe American prisons. Um, and studies have shown there that if you know, if you happen to get uh your cases randomly assigned to a more harsh judge and you're sent to prison in Norway versus not, um, you're actually better off in terms of you are more likely to um stay out of prison if you jury recidivism falls, you're more likely to be employed, those sorts of things. Um, a closer cousin is open prisons that are in Italy. So this is uh Italian prison systems are closer to the US model, but they do have a version of this where they have open prisons where, again, you can walk around, you aren't just confined to your cell or your unit. Um, some people can leave and go to work during the day. Inmates um run restaurants and do all kinds of stuff like this. And again, using a similar natural experiment where because of random chance, some people wind up in that open prison versus a traditional closed prison. The folks in the open prison do much better in terms of lower recidivism. Uh, same in this Colombian prison, where random chance, they're sent to the newer facility, they have more space for programming, higher staff to inmate ratios, so they're safer and just less crowded. And so you're not spending as much time potentially getting into arguments with other people around you. Uh, recidivism falls for those folks too. So we are learning a lot about what's going on in and what other countries are trying. And honestly, at this point, uh I think my main focus and and a lot of what my team at Arnold Ventures is is thinking about is how can we spur more innovation in the US prison system so that we can start getting into the habit of trying new stuff ourselves and doing it in a way that we can learn from, right? And so the the only way we're gonna, I think, you know, prisons are are have been have been run the same way for many, many, many decades. Um, and we are in a bit of a crisis situation right now where there are massive workforce shortages. Uh corrections, a lot of, a lot of prisons are so understaffed, and then they're really violent because they're understaffed. And then people quit because they don't want to go to the work to to that that job anymore. And then it just, it's a vicious cycle. And so we really we have to do this, is something's got to give here, and we've got to be trying new stuff. So we're talking to a lot of corrections leaders all the time about what that might be.

SPEAKER_03 42:04

Are you finding a receptive audience to that? Because you know, one of the things that that strikes me is the criminal justice system is just so slow to evolve. And here we have all of these fantastic ideas from you and and other of your colleagues. Um but by the time we adopt any one of them, it it just seems like we'll be in a in a in a different entirely different environment. But what what are corrections officials saying to you about their openness to making changes?

SPEAKER_00 42:37

Yeah, I mean, this is certainly, I'm sure there's there's lots of variation across the system. The system's very big, right? And every state's different and every facility is run by a different person. But I think there is pretty widespread agreement at this point across both corrections officers as well as the families of people who are incarcerated that this system is not working. It's at a breaking point. And so I think it is leading to a lot of um individual facilities. It's leading to a lot of states being willing to just try something radically different because it's just so clear that the status quo is unsustainable. So, for example, we are funding a few randomized trials of Scandinavian-style prison units in South Carolina, Missouri, Pennsylvania, uh, taking that, you know, really that Norwegian model and um, but implementing it in a way where eligible inmates are randomly assigned to be in that unit or a standard unit. And we'll track over time what the impact on those inmates is. What's been fascinating is that most of, I mean, people seem seem to like it. People are optimistic. We're hoping it reduces violence in the prison as well as um improving recidivism. But the reason people are most excited about it is that the guards love working in those units. And so Pennsylvania actually just expanded their pilot to a bunch of other facilities because. They, like everyone else, are really struggling with the workforce shortage. They cannot hire and retain enough corrections officers. Um, and the guards love working in these units. And so, and it's just it makes it a more meaningful job. They feel like their job is to actually, you know, help the people, their clients succeed. They're helping to rehabilitate and help put them on a better path. And that's just more meaningful at the end of the day than just standing around and trying to like, you know, not get attacked and not let anyone else um uh be hurt. And so, so that is so now we're trying to line up these pilots in a way that will be actually able to rigorously measure the impact on the staff because it seems it seems that that's the most interesting outcome to policymakers. So I really, I mean, this is a space where I think necessity is the mother of invention. And um, another another example of this that I I love is um, I don't know if you guys, you guys have seen the Netflix documentary or docuseries Unlocked, a jail experiment. Um, so there is uh there's one in Georgia and they just released a second season, which blows my mind in Arizona, where um sheriffs in in jails there are like, this is, you know, we're understaffed, people are locked on 23 hours a day. That just makes conditions worse. They're just, you know, riled up when they do, we do release them. And then we have to lock everybody down again. And this is just a terrible cycle. And so they try they're trying something where they just unlock the doors, take the guards out, and say, you all have to figure out how to run a community and take care of each other. And um, so I think just the fact that stuff like that is happening is a sign that everyone recognizes we need to be trying some new things here.

SPEAKER_03 45:41

It is super interesting and be really fascinating to see the outcome of all of that, um, including the pilot programs in in Pennsylvania, which is where I'm based. So I will be eagerly watching that. Um, before we start wrapping up, I'll tell you one thing that really surprised me in the book is the ban the box initiatives, BTB, and the comment that you make that they actually have the opposite effect of what they're intended to do. Um they, you know, you expect that by um not requiring potential employees to check the box and whether they have a prior criminal conviction, that that would be a helpful thing. Uh that you're giving them a clean slate and a fresh start. But apparently it's not so. And I was hoping you could just talk about that before we wrap up.

SPEAKER_00 46:31

Yeah, this was research I did um uh a while back now. It was really my first foray into the policy uh making process and my first interaction with policy advocates. And it was as a you know junior faculty member. It's like, oh, we're not in academia anymore here. Um, but yeah, the the you know, the goal of these programs, these well-meaning policies, is to help people with criminal records get their foot in the door. I think another goal is to help reduce racial disparities in employment. I think there are a lot of people out there who think uh, you know, employers are barred from considering race um in hiring and being able to consider a criminal record is almost like a loophole because um because of the large racial disparities in the United States in terms of who has a criminal record. And so employers are uh saying they're discriminating based on a criminal record, but really they're just trying to, you know, discriminate in favor of white people. And so, so this racial justice element was, I think, a big motivation for these policies. But uh at the very least, we're hoping that they they improve employment for people with records. The problem with these policies is that you know, there's clearly some underlying reluctance that employers have to hire someone with a criminal record. And these policies don't directly address that in any way. They don't change the incentives employers are facing in their hiring decisions or, you know, who they need in terms of a good employee. They're just telling them they can't ask at the front end if someone has a record or not. And so when someone has an economics background like me, hears about a policy like this, our minds immediately jump to the possibility of statistical discrimination, which is when when you can't see something that you really care about, you basically take, you look for other things that are correlated with that underlying characteristic, and then you discriminate based on that. And so in this case, if employers really care about the criminal record and they can't ask anymore, then because of the racial disparities that, you know, are part of the motivation for this policy, they then can discriminate based on race instead. And and they might, in particular, discriminate against young black men with limited education, sort of the group that is most likely to have a criminal record, who should be helped if the policy is working, but would be hurt if employers are now just guessing that this whole group is likely to have a record and they're just not gonna bother to interview them. And so I have a study looking at the rollout of ban the box policies across the country and what the impact of employment on employment is for different groups. There's um about the same time, some colleagues of mine ran a ran a really cool field experiment sending out resumes or job applications for fiction uh fictional applicants before and after Ban the Box went into effect in New Jersey and New York City to measure what happens to racial disparities in employment. And it turns out Ban the Box dramatically increased racial disparities. So in my paper, we found that it reduced employment for young black men without a college degree by about 5%, which is a big effect for a group that already struggles in the labor market. And then there's subsequent research showing that Ban the Box also is not improving employment for people with criminal records, which I think we'd all sort of assumed. Well, at least some people will be helped, but it's clearly we've got this other group that's really hurt because they now can't signal their clean record to employers. Um yeah, so this is a this is a space where, I mean, this evidence has been out for over a decade now, and ban the box policies are still being introduced in state houses. Uh, they're still in effect. I think it's it's an example of a place where the policy became the goal rather than solving the problem being the goal. And by the time we had the evidence, unfortunately, there was a really strong ban the box lobby that whose whose livelihoods depended on um continuing to push for this specific legislation. And that's really unfortunate.

SPEAKER_03 50:27

Interesting. Best intentions going awry. Yeah.

SPEAKER_00 50:31

And it's a place I just want to emphasize like I think it's a really important problem. Like, I really wish we could channel this energy toward trying different solutions that might have more of an impact. It's just really makes me really sad that like everyone agrees it's a really important problem to solve, and we've gotten stuck on this one policy that isn't working and is actually harming some people. Like, this is exactly why we need to be um yeah, uh eager to fail fast rather than pretending we're not going to fail.

SPEAKER_03 51:00

Absolutely. Um, so I'm I'm curious, um, with all of the great policy insights in your book, um, with limited resources, um, what would you prioritize if you could pick one or two, given the limits of time and investment, where can we as a society get the biggest bang for our buck today?

SPEAKER_00 51:23

Yeah, well, this is a conversation I and others on my team are having every day with lawmakers across the country, especially at the state level. And so what we're we're really pushing based on the evidence in the book and a variety of other evidence is to shift attention from trying to fight crime with long sentences to trying to fight crime by investing in investigations. So increase the probability that people will get caught, increase clearance rates. This is a so we there are a few different ways you could do this. You could expand DNA databases, as I, as we talked about earlier. You could um give police more resources to buy technology that will help them um identify suspects more quickly. Um, it turns out most detectives don't get any sort of training when they take the job. It's more of a promotion based on seniority. So we could, you know, we're working with an organization to design a detective school and we're gonna try to roll that out as a randomized trial and see what the impact is on investigations and solving crime. So I think this is a place where I think it's largely driven by what's easy for legislators and just in just, you know, the tools at their disposal. If they're worried about crime, it's pretty straightforward to change sentencing guidelines, right? That's something they can just write into law. And now everyone convicted of this crime is spending more time in prison. But it just turns out that's really not a cost-effective way to achieve the goals that they they have. And it might not be quite as straightforward to use their legislative tools to um to improve police investigations, but it is something where there's a lot of bipartisan support and where we just have a lot of evidence that that's gonna have have a much bigger impact. And so, so directing resources to that area to solve more crimes faster is my number one recommendation to policymakers. If I had a magic wand, I would uh uh, because this is much more expensive, but I had a magic wand, I would uh push for lead abatement, take lead out of the environment. Um, because lead exposure for young kids has really big impacts on violent crime later and is a harder problem to solve, just technically. It's just there's a logistical challenge to getting rid of all the lead, but um that's something I wish we we invested in more as a country.

SPEAKER_03 53:47

We'll have to save that for the next podcast.

SPEAKER_02 53:51

We will. I'm I'm I'm gonna I'm gonna wrap up with kind of some lightning round style questions um that I'd love to get your answers uh for our our listeners. If a 22-year-old with a felony conviction reads the science of second chances, what do you hope they feel? Reassured, empowered, challenged. What do you hope?

SPEAKER_00 54:15

Yeah, empowered and hopeful. Ultimately, I am extremely optimistic about this policy space. And I hope anyone who reads the book feels the same way, but especially someone who has been caught up on the system already, there are a lot of things that we can do to help them build a better life.

SPEAKER_02 54:31

If a prosecutor reads this book, what do you hope that they think or rethink?

SPEAKER_00 54:36

Yeah. Same, same. And and I'll add on if someone in that like the practitioner or decision-making role, recognizing that they a lot of the the things they take for granted as just the way you do things or you know, the way that they they've been trained, or even just based on their best experience, might not be the best way to do things, and being willing to just try new stuff, right? And recognizing that's not a sign of weakness or a sign that that that you know uncertainty is not a sign that they don't know what they're doing. It's a recognition that we're just we're all fallible and these are really hard problems to solve. And so, so an interest in trying new things but learning along the way.

SPEAKER_02 55:22

Finally, when historians look back, will they say that the revolution, this revolution that you are a part of in criminal justice, began with moral awakening or with data clarity? And which is more durable?

SPEAKER_00 55:37

Oh my gosh. Um I mean, surely both, right? I think I think there is um this my book follows a number of other books that really tried to impress upon people the scale of the problem. Um, and those are really important. I think of the new Jim Crow in particular.

SPEAKER_02 55:58

Outstanding book.

SPEAKER_00 55:59

I agree. Yes, wonderful book, very important, clearly. And also, I think those the books that that really focus on describing the scale of the problem, the the limits of those books uh tend to be that they, you know, they don't they don't present potential solutions, right? I think a lot of people read them and they're just like, oh man, this is terrible. We're just gonna have to burn it all down. And I hate feeling that way. It makes me feel really powerless to just see a big problem that just feels insurmountable, right? And what I'm really trying to do is showcase all the ways, all the tools at our disposal to figure out how to solve these problems and to demonstrate that these are not that that we do have some agency here and there these these solutions are within reach. And so um I'm hoping maybe it's a one-two punch.

SPEAKER_02 56:52

Well, I I must say, um, especially with the references that you've made and how how versed you are in in so many different aspects of uh this arena, that it is wonderful to hear from somebody who isn't just watching a carnage and watching the tyranny, but who's actually coming up with solutions. Because I agree with you, we are at a time in our history where um this quote unquote idea of wokeness somehow is attached to learning something that most of us already knew was happening all the time. And so when it's presented to us starkly, we're like, oh my god, this is terrible, and we stop the end. And as you said, presenting solutions is the next step. And um, I sure I'm sure I share share Justin's sentiments when I say um you've presented a lot of ideas, a lot of hope, a lot of optimism, um, a lot of sobriety as it relates to this issue and this topic. And um I I I couldn't agree with um so many of your positions more. And I I I think you're doing God's work, and I I you know applaud the work that you're doing and and just ask for you to keep fighting the fight and you know keep educating us all on what we can do better.

SPEAKER_00 58:25

Thank you. I really appreciate that.

SPEAKER_03 58:27

Jennifer, it's been wonderful to have you on the book again. Now, February 24th, 2026 is The Science of Second Choices, A Revolution in Criminal Justice by Jennifer Doliac. It's a really wonderful read. Thank you so much for being with us today.

SPEAKER_00 58:43

Thank you. Thank you for having me.

SPEAKER_02 58:45

Thank you. It's been a pleasure.

SPEAKER_03 58:50

This episode of The Just Pod was edited and produced by Keosue. Our original theme music was composed and mixed by Janar Butler.

SPEAKER_02 59:00

If you would like to join us on a future episode of The Just Pod, or if you have questions or topics for us to address on a future episode, please send them to our producer, Keoso, at k yo.suh at americanbar.org.

SPEAKER_03 59:19

This has been another episode of the Just Pod. Thanks so much for joining us. From Philadelphia, I'm Justin and Daniel Evans.

SPEAKER_02 59:26

And from Washington, D.C., I am Jannard Butler. Please join us again the next time.

SPEAKER_03 59:32

The Just Pod is a production of the American Bar Association's criminal justice section.