So maybe we can start with the barriers and then later move into what actually works and what you've learned about that.

So this was my first real experience as a professor engaging in the policy conversation.

It was the first paper I had that like really kind of got a lot of policy attention and was part of an active policy conversation where the evidence was really relevant or, you know, people should have viewed it as relevant.

And so, so what I found and, you know, it was my paper and then there's another really, you know, high profile paper that came out about the same time.

And both of our papers found that what happens when, so Ban the Box removes that question, that box that you have to check saying if you have a criminal record, and the idea is this should help people get their foot in the door.

Now, anecdotally, employers just throw away your application if you check that box.

That's, of course, incredibly unfair and demoralizing to people who have a criminal record and have done the work and now would be great employees.

So I don't know if you could share a little bit about that.

Yeah.

Like, that should deter crime, but how do we know if it works? Um, and so the, the natural experiment I wound up using in, in, um, a couple of my, my studies on this was that the, the date of...

So these ex- these databases are expanded over time in any given state.

So they might start with people convicted of murder and rape, and then you add people convicted of robbery, and then you add people convicted of burglary, and then you go on down the list, and they get bigger and bigger.

But that, in each of those individual expansion dates creates its own natural experiment.

So you wind up with someone who's convicted of robbery on, you know, May 30th, and they don't go in the database, but then the law changes on June 1st, and someone convicted of robbery June 1st or 2nd does go in the database.

Everything else about them is the same, everything about the criminal justice system, their sentence, the, uh, the community they're gonna be released back into.

Yeah

But she says for too long, we've been focused on the wrong levers to deter crime, that we've been too focused on longer sentences when it's actually more effective if we increase the probability of being caught.

In practice, we know that most people who are committing crime are not thinking very far ahead, and so increasing the length of a sentence from one year to two years or five years or 10 years, none of that's gonna change behavior if people aren't thinking past this week.

But increasing the probability of getting caught, that does change behavior.

That does register.

So really, what we've learned over time is that the most effective way to change behavior and reduce crime is increasing the swiftness and certainty of consequences, not so much what the consequences are.

Doleac says one way to increase the probability of getting caught is by putting people who are at high risk of committing crimes into law enforcement DNA databases, like the one we heard about earlier, CODIS, the database that matched the DNA of Ashley Spence's attacker.

And so the idea here is that DNA databases are constantly looking for matches between these two things, between the unsolved crimes and the people who we know are, you know, have been convicted of or arrested for crimes in the past.

And, uh, looking for leads in the case that police might not have, um, have come up with themselves in, in some other way.

Hmm.

Um, but that's not always possible.

We can't, for instance, we can't randomly assign people to go to prison or not.

Um, and, and we shouldn't, [clears throat] right? And so the question then becomes: how do we find a good control group as if we were in a lab? And that's where economists are really pros at this.

So I'm an economist by training.

This is our bread and butter, is finding natural experiments that give us something really close to that ideal randomized experiment.

And so for instance, in, in the case of, you know, if we wanna know what the impact of going to prison is, we can't randomize who goes to prison, but we do very fr- frequently randomize different cases to different judges, and it turns out that different judges have different preferences and different professional opinions about who should go to prison and who shouldn't.

Mm-hmm.

Why err on the side of leniency? How is that gonna help for first-timers?

Yeah.

It does sound counterintuitive, and I kind of, so I like, I, in the book I kind of have these, these concepts almost side by side because I think it does make people realize, like wait, these ideas aren't necessarily...

The left didn't get it all right or the right didn't get it all right.

Like, there are, this is kind of complicated, right? And this is a reason we need some research to see what works in practice.

So, um, so the idea here is, and this has been a big public debate over the past, I don't know, five to 10 years, especially with the rise of progressive prosecutors in a lot of communities.

A lot of people saying, you know, the, the criminal justice system actually has a, a criminogenic effect.

What are they doing that we can learn from?

Tell us what we found in Kansas.

So this is, um, a really cool study done by a researcher named Ryan Sakoda, and he found, uh, a situation where there were, um, the sentencing reforms in, in Kansas put into place.

It- the, the, the status quo was that if you are released from prison, you are immediately put on post-release supervision, and this was after, this was during a context when, you know, truth in sentencing laws had been in effect, so most people were not eligible for release on parole anymore.

And so I think a lot of politicians looked at that, and they're like, "Well, just re- immediately releasing someone into the community from, you know, complete structure to zero structure, that's gonna set someone up for failure, so we wanna give them some sort of supervision to help with the transition." So they added post-release supervision terms for anyone coming out of prison.

Um, and anyone who was initially sentenced to probation but then had a technical violation that sent them to prison, had their probation revoked, they also, when they were re-released from prison, were required to be on this post-release supervision term.

So just kind of like, you know, there's this su- post-release supervision that really is this additional sentence, essentially, [laughs] uh, that people have to, have to endure and, and complete successfully.

Um, and so there was a policy change in Kansas that removed that post-release, release supervision term for that group of people who were originally sentenced to probation.

Yeah, go ahead

When you were a graduate student and then when you moved on into your academic appointments, what were some, because it does sound like like econ as a whole is like such a broad odd field, what were some of your kind of personal areas of interest and inquiry?

Yeah.

When I first went to graduate school, I think I assumed I would study education or the economics of education or higher education, probably because that's where I'd spent most of my life, right? That's the system I knew best.

I'd been a student for a really long time.

And so it really wasn't until it was later in my graduate training where I and all of my colleagues were hunting around for these natural experiments that we could apply our new toolkit to.

And I read a newspaper article about law enforcement DNA databases.

And the point of the article was that, you know, it seems really arbitrary that in some states, you know, states are determining the rules about which groups of people convicted or arrested for certain crimes are required to provide DNA to the database.

I mean, only in very ideal