There are a number of factors that can lead a suspect to confess to crimes they did not commit. WGBH's Morning Edition Host Joe Mathieu spoke with Northeastern University law professor and WGBH News legal analyst Daniel Medwed about false confessions and why they happen. The transcript below has been edited for clarity.

Joe Mathieu: So we're talking about false confessions — admitting guilt to a crime that a suspect did not commit. Which seems completely bizarre to most people on the surface, but how often is this happening?

Daniel Medwed: It happens surprisingly often. I saw one study of 250 innocence cases, where 40 cases were identified as having involved false confessions — where the suspect falsely claimed to be involved in a crime he didn't commit. More than half of those 40 cases involved either juveniles or folks with mental or cognitive deficits. And this is really important, because when juries hear about confession evidence, they tend to overvalue it. I saw another study that indicated that when there's any confession evidence at all in a trial, juries vote to convict 73 percent of the time, regardless of the strength of the other evidence against the person.

Mathieu: So why mostly juvenile or mentally challenged suspects? Is this about coercion?

Medwed: Well, to some extent, folks who are young or who have mental challenges can be easily swayed and led in a particular direction. But there are many other factors as well that relate to police interrogation tactics. So for one thing, we all know about the Miranda case. You have a right to remain silent [and] a right to an attorney before the police can interrogate you while you're in custody. But get this: more than 80 percent of suspects waive their Miranda rights and decide to go it alone with the police.

Mathieu: 80 percent waive their right to remain silent?

Medwed: Yeah. I think most of these folks think that they can do better than the police. Or they feel that if they were to assert their Miranda warnings, that somehow they would increase the suspicion and they would become the target.

Mathieu: Is there a case in Massachusetts where this has happened?

Medwed: Yes, there have been several. But the one that comes to mind involves a man named Victor Rosario. There was a house fire in Lowell in the early 80s — eight people died, five of them were kids. The police suspected an arson, and Victor Rosario immediately emerged as the chief suspect. He was interrogated for, get this, seven and a half hours at a time when he was suffering from extreme alcohol withdrawal ... and he rendered a bunch of incoherent, quasi-confessions that were used against him. There was also a problem with the translation, he was a non-native English speaker. He spent 32 years in prison for those crimes [and] was exonerated in 2017.

Mathieu: That was a big story. Is there any way to stop this then from taking place so often?

Medwed: Well, there are a couple different ways to stop it. For one thing, we could put firm limits on the duration of these police interrogations. Victor Rosario was interrogated for seven and a half hours. I've seen another study that said that about 75 percent of false confessions occur after interrogations of more than six hours. So why not just set firm time limits on this?

Second, we could set rules regarding how the police interrogate. So the police are not allowed to engage in physical torture — that was outlawed back in 1936 — but they can utilize interrogation techniques that, in my view, constitute psychological torture. So for instance, they can lie. They can lie about the nature of the evidence against the suspect. They can say that your fingerprints were present at the crime scene when they weren't. Courts call this a ruse; police are allowed to use ruses. I call them lies; they're allowed to lie to you. So maybe we could restrict some of the techniques that the police are allowed to use.

And here's a very simple solution. We could just record all of these police interrogations from beginning to end, so we could determine whether or not they have the addition of a false confession, and juries themselves could evaluate whether it's legitimate or not.

Mathieu: They don't have to record them now?

Medwed: In Massachusetts, the Supreme Judicial Court has indicated that should be the practice here. And about 25 states, either by legislation or judicial fiat, record interrogations. But again, that's only about half the states. It's something that should be done across the nation.