Disturbing History - The False Confession

Episode Date: October 28, 2026

You'd never confess to a crime you didn't commit, right? Out of the first 375 people in the United States cleared by DNA evidence after a conviction, 102 of them had.In this episode of Disturbing History, Brian looks at how an innocent person ends up in an interrogation room saying the words, who built the room, and what it did to a long list of real people whose names belong in the record.The episode opens with Peter Reilly, an 18-year-old in Canaan, Connecticut, who came home from a church meeting in September 1973 and found his mother dead. After roughly 25 hours of questioning and a polygraph he was told he'd failed, Reilly began telling state police that it really looked like he'd done it.The tape of that interrogation survived, and so did the witness statements that put him miles away, which sat undisclosed in a prosecutor's file for four years. From there the story goes back to the 1936 Supreme Court case Brown v. Mississippi, where three Black tenant farmers were whipped into confessing and a deputy testified it wasn't too much for a Negro, and forward to the Chicago polygraph examiner John Reid, whose nine-step technique replaced the third degree and still dominates American police training. Reid's own signature case, the 1955 confession of Darrel Parker in Lincoln, Nebraska, turned out to be false when a career criminal's deathbed admission came to light decades later. The episode covers Miranda v. Arizona, the 1969 Frazier v. Cupp decision that made it legal for police to lie about evidence in all 50 states, and the research of psychologist Saul Kassin, including the 1996 experiment in which 100 percent of innocent college students signed a confession under the right conditions, and the 2004 study showing that innocent suspects waive their Miranda rights far more often than guilty ones do.Then the cases.The Central Park Five, questioned for up to 30 hours and told they could go home. The Norfolk Four, where a detective later convicted of extortion kept adding sailors to a rape and murder that DNA tied to one man acting alone. The Beatrice Six in Nebraska, where a sheriff's psychologist taught suspects that repressed memories of a murder could surface in dreams, and five of them pleaded guilty to a crime committed by a man who'd been excluded years earlier by a blood-typing error.Earl Washington Jr., with an IQ around 69, who came within nine days of execution in Virginia on a confession that got the victim's race, height, and number of wounds wrong. Christopher Ochoa and Richard Danziger in Austin, Jeffrey Deskovic in Peekskill, 14-year-old Michael Crowe in Escondido, Juan Rivera in Waukegan, Marty Tankleff on Long Island, Damon Thibodeaux on Louisiana's death row, Huwe Burton in the Bronx, the Dixmoor Five, the West Memphis Three, Brendan Dassey, and the torture victims of Chicago's Area Two under Commander Jon Burge.The episode closes with what has changed since: the Illinois recording law sponsored in 2003 by a state senator named Barack Obama, the 2017 decision by one of the country's largest police training firms to stop teaching the Reid Technique, the juvenile deception bans now in ten states, the British PEACE model built after the Guildford Four and Birmingham Six, and the fact that in every state in this country, police may still lie to an adult about the evidence.It ends back in Canaan, where the murder of Barbara Gibbons has been open for 53 years and her son remains the only person ever tried for it.Email BrianGet Brian's Books Leave Us A VoicemailVisit Our WebsiteHave a forgotten historical mystery, disturbing event, unsolved crime, or hidden conspiracy you think deserves investigation?Send your suggestions to [e-mail] History is a dark history podcast exploring unsolved mysteries, secret societies, historical conspiracies, lost civilizations, and the shadowy stories buried beneath the surface of the past.Follow the show and enable automatic downloads so you never miss a deep dive into history’s most unsettling secrets.Because sometimes the truth is darker than fiction.

Some stories were never meant to be told. Others were buried on purpose. This podcast digs them all up. Disturbing history peels back the layers of the past to uncover the strange, the sinister, and the stories that were never supposed to survive. From shadowy presidential secrets to government experiments that sound more like fiction than fact, this is history they hoped you'd forget. I'm Brian, investigator, author, and your guide through the dark corner.

of our collective memory. Each week I'll narrate some of the most chilling and little-known tales from history that will make you question everything you thought you knew. And here's the twist. Sometimes the history is disturbing to us. And sometimes, we have to disturb history itself,

just to get to the truth. If you like your facts with the side of fear, if you're not afraid to pull at threads, others leave alone. You're in the right place. History isn't just written by the victors. victors. Sometimes, it's rewritten by the disturbed. You'd never confess to a crime you didn't commit, right? Most people answer that one fast. You know what you did and what you didn't do, and no

detective in a small room is going to talk you out of your own life. I spent 16 years in law enforcement, and I would have told you the same thing for most of that time. The people who confessed to murders they didn't commit, I figured, were a special category. Drunks. Kids with some wrong upstairs, people who wanted the attention. Then the DNA testing started coming back. Out of the first 375 people in this country cleared by DNA evidence after a conviction, more than a hundred of them had confessed. 29 percent. Not accused of confessing, not rumored to have confessed. They sat in a room with police officers and said they did it in detail, sometimes on videotape, sometimes with a

court reporter. And then the biology came back years later and said somebody else did. Those are only the cases where there happened to be semen or blood or skin under a fingernail that somebody saved. Most crimes don't leave that kind of evidence. Most confessions are never tested against anything. This episode is about how a person ends up saying the words. Some of these cases involve the murder of children and sexual violence and the details are part of the record. so listener discretion is strongly advised. The first case is one of an 18-year-old in Connecticut named Peter Riley, and the reason to start with him is that the state police recorded what they did to him,

and the tape survived to tell the tale. On the night of September 28, 1973, Peter Riley was 18 years old and living with his mother in a small house in Canaan, a town in the northwest corner of Connecticut. He went to a meeting at the Methodist Church, that evening. When he came home a little after 9.30, he found his mother, Barbara Gibbons, on the floor of her bedroom. Her throat had been cut. Both of her legs were broken. He called for help, and the first people he reached were neighbors, and then the state police came. The state police

took Peter to the barracks that night. The next day, they drove him to Hartford for a polygraph. He didn't have a lawyer, and nobody in his family was with him, because his mother was his family. He'd been awake for most of a day by the time the examination started. The examiner, a sergeant named Timothy Kelly, told Peter that the machine showed he was lying. Kelly told him the polygraph doesn't make mistakes. Peter said he didn't remember killing his mother. Kelly and the other officers told him that was possible, that people sometimes block out things like this, and they worked on that idea for hours.

The interrogation ran on and off for roughly 25 hours. The reader asked at one point whether it was possible he'd done it and didn't know. He started saying things like, It really looks like I did it. He described a struggle he said he couldn't actually picture. He signed a statement. That tape is the reason the case became famous. It's not a confession in the way you'd imagine one.

It's a boy being talked, by degrees, into doubting his own memory until the officer's version is the only one left in the room. A jury convicted him of manslaughter in April of 1974. The people of Canaan had mostly never believed he did it, and they raised money for his defense. The playwright Arthur Miller, who lived nearby, got involved, and so did the novelist William Steyron and the director Mike Nichols. They hired investigators.

In 1976, a judge granted Peter a new trial on the basis of new evidence. Before that trial could happen, the state's attorney who'd prosecuted him, John Bianchi died of a heart attack. The man who took over the office went through Bianchi's files. In them, he found statements from witnesses, including an auxiliary state trooper that put Peter miles away at the time his mother was killed. Those statements had never been turned over to the defense.

In November of 1977, the charges were dismissed. A special one-man grand jury appointed by the governor later cleared him entirely. The murder of Barbara Gibbons has never been solved. Nobody hit Peter Riley. Nobody threatened to hurt him. The state of Connecticut didn't need to, because what happened in that room was a technique. And the technique has a name and a sales history.

But before the technique, there was the beating. In March of 1934, a white planner named Raymond Stewart was murdered in Kemper County, Mississippi. deputies arrested three black tenant farmers, Ed Brown, Henry Shields, and Arthur Ellington. A deputy named Cliff Dyle and a group of white men took Ellington to the home of the dead man, hanged him from a tree limb, let him down, hanged him again, tied him to the tree and whipped him. He didn't confess. They let him go. A day or two later, Dyle arrested him again, drove him across the state line into Alabama, and whipped him until he agreed to whatever statement the deputy dictated. Brown and shields were stripped, laid over chairs, and beaten with a leather

strap with buckles on it until they confessed too. At trial, Guyle was asked on the stand whether he'd whipped Ellington. He said he had, and he said, it wasn't too much for a Negro. The jury convicted all three men and sentenced them to hang. In 1936, the United States Supreme Court threw those convictions out in a case called Brown v. Mississippi. Chief Justice Charles Evans Hughes wrote that the rack and the torture chamber couldn't be substituted for the witness stand. That decision is the first time the federal court said plainly that a confession beaten out of a person can't be used against him in a state court. That ruling changed the business. Police departments that had relied on what was called

the third degree had to find something that worked without leaving marks. A Chicago police officer named John Reed had been working on exactly that. that. Reed was a polygraph examiner, and in 1947, he left the department and opened his own firm. He paired up with a Northwestern Law Professor named Fred Inbow, and in 1962, the two of them published a manual called Criminal Interrogation and Confessions. It's still in print. It's been through five editions. The company John Reed founded has trained, by its own count, hundreds of thousands of police officers, federal agents, and corporate investigators. And for most of the second half of the 20th century, if you learned how to interrogate a suspect in this country,

you learned the read technique, or something derived from it. The method has nine steps, and the first one is the whole philosophy in a sentence. You tell the suspect you already know he did it. Before that, there's supposed to be a non-accusatory interview, where the investigator asks open questions and watches for signs of deception, things like fidgeting, breaking eye contact, crossing arms, vague answers. If the investigator decides the person is lying, the interview ends and the interrogation starts. And the interrogation is built on the premise

that the person in the chair is guilty, and the only remaining question is how to get him to say so. That decision is made in a conversation that might last 15 minutes. The manual teaches investigators to read posture, eye contact, the length of a pause before an answer, whether the subject's answers are too detailed, or not detailed enough, and to treat the wrong pattern as evidence of a guilty mind.

Then the door closes, and the investigator who walks back in is no longer gathering information. He's gathering a statement. From there, the steps are about controlling the conversation. You cut off denials before they're finished. You don't let the suspects say, I didn't do it, because every time he says it out loud, it gets easier to keep saying. You offer themes, stories that make the crime understandable. She provoked you. It was an accident. Anybody would have snapped. You contrast two versions of the suspect, the cold-blooded one and the decent man who made a mistake, and you invite him to pick. You move your chair closer. You watch for the moment the person stops arguing and goes quiet. And when it comes, you ask a question that assumes guilt and offers two answers, both of which are a confession. Did you plan this, or did it just happen? Researchers who study this boil the technique down to two words.

Maximization and minimization. Maximization is the stick. You exaggerate the strength of the evidence. You say the witnesses have already talked. You say the lab results are back. Minimization is the carrot. You downplay the seriousness. You suggest the courts will understand.

You imply that this will all go easier if he just gets it off his chest. Read and associates will tell you, And they've told courts and reporters for years that the technique is only meant to be used on people the investigator reasonably believes are guilty. They'll tell you that the manual warns against promising leniency and against interrogating juveniles or people with intellectual disabilities the same way you'd interrogate an adult. They'll tell you that when a false confession comes out of a room where a retrained officer was working, the officer departed from the training. I'll give them this much. The manual does say those things. The part I have trouble with, as somebody who sat on the other side of that table,

is that the whole structure depends on that first judgment. The interrogation starts because the investigator decided, from a conversation, that the person is lying. And the research on whether trained investigators can actually tell when somebody's lying from body language and eye contact is not kind. Study after study puts police accuracy at detecting lies right around the level of a coin flip. and the training makes officers more confident without making them more accurate. If the gate is a coin flip, everything downstream of the gate is being applied to a lot of innocent people. In December of 1955, in Lincoln, Nebraska, a city forester named Daryl Parker came home for lunch

and found his wife Nancy dead in their bedroom. She'd been raped and strangled. The Lincoln police brought in John Reed personally to polygraph the husband. Reed questioned Parker for most of a day. By the end of it, Parker had confessed. He recanted almost immediately and said Reed had worn him down. But the confession went to the jury, and the jury convicted him of murder in 1956. He got life.

Reed used the Parker case for years as a teaching example of how the technique works. Parker was paroled in 1970. In 1988, a career criminal named Wesley Peary died in prison and his lawyers were then free to reveal something he'd told them under attorney-client privilege. Peary had confessed to killing Nancy Parker. He'd given them details. Nebraska pardoned Daryl Parker in 1991, and in 2012, the state's attorney general stood up at a press conference, apologized to him, and announced a settlement of $500,000. Parker was 80 years old by then. So the signature case of the man whose name is on the technique was a false-conflict.

confession. The company disputes that characterization and has said that Parker's statement contained details only the killer would have known. The state of Nebraska decided otherwise. The Supreme Court had a chance to put limits on all of this, and it did, up to a point. In 1966, the court decided Miranda v. Arizona, and if you've watched any American television, you can recite the warning. What people forget is why the court wrote it. Justice Earl Warren's opinion spends pages quoting from police interrogation manuals, including the one inbaugh and Reed had published four years earlier. The court laid out the tactics, the isolation, the false confidence, the themes, the two-choice questions, and said the atmosphere those tactics

created was inherently coercive. Miranda was the court's answer. You tell the person he can stay silent and have a lawyer, and if he gives that up and talks anyway, the pressure in the room is his problem. Three years later, the court decided a case called Frazier versus Cup out of Oregon, where police had falsely told a suspect that his cousin had already confessed. The court said the lie was relevant, but not enough by itself to make the confession involuntary. That one sentence has been the law ever since. In every state in this country, a police officer may lie to you about the evidence. He may tell you your fingerprints are on the knife when there are no fingerprints. He may tell you a witness picked you out of a lineup that never happened. He may tell you your

friend in the next room just gave you up. He may tell you the DNA came back when nothing was ever sent to a lab. I did this work for a long time and I never thought much about it. It was legal, it was taught, and it worked. The question nobody asked in the training room was what it works on. Miranda itself turned out to be a weaker shield than the court expected, and the reason is a strange one. innocent people wave it. In 2004, Saul Kasson and a colleague named Rebecca Norwick ran an experiment where some participants committed a mock theft and others didn't. And then, all of them were questioned by a detective who read them their rights.

Stay tuned for more disturbing history. We'll be back after these messages. 36% of the guilty ones agreed to talk. 81% of the innocent ones did. When the innocent participants were asked afterward, why they'd given up the right to stay silent, the answer came back the same way over and over. They hadn't done anything, so they had nothing to hide. So why wouldn't they talk? Casson calls it the phenomenology of innocence. The person who most needs the protection is the one most likely to

hand it back, because he's sure the facts will take care of him once he explains. In 1996, two psychologists Saul Casson and Catherine Kiesel published an experiment that's become the most cited study in the field. They brought college students into a lab for what they were told was a reaction time test. Each student typed letters that another person read aloud. Before starting, they were warned not to touch the alt key because it would crash the computer and ruin the data.

About a minute in, the computer crashed. It was rigged to. The experimenter came in upset and accused the student of hitting the forbidden key. Every one of those students was innocent. None of them had touched it. 69% of them signed a written confession saying they had. When the experimenters added a second person who claimed to have seen the student hit the key, a false witness, and sped up the typing so the student felt less sure of what his hands had done,

the rate of signed confessions went to 100%. More than that, in the high-pressure condition, 65% of the students later told a stranger in the hall, someone they thought wasn't part of the study, that they'd done it. They'd come to believe it. And about a third of those made up details. I hit it with the side of my hand. I think my pinky slipped. Later researchers pushed the same design further.

In a 2005 study, a team led by Melissa Rusano that included Casson, set up a mock crime in which participants had a chance to cheat and then were questioned about it. The ones who'd actually cheated were more likely to confess when the investigator minimized when he suggested the whole thing was no big deal and everybody did it. The innocent ones were vulnerable to the same move. Minimization raised the false confession rate right alongside the true one, and the people hearing the confessions afterward couldn't tell which were which. The stakes in that lab were nothing. Nobody was going to prison for crashing a computer. If you can get 100% of innocent

college students to sign in those conditions with a fake witness and a little time pressure, The question of what 16 hours and a murder charge can do starts to answer itself. Casson had already, a decade earlier, worked out a way of sorting false confessions into three kinds, and the field still uses it. The first kind is volunteered. Nobody pressured the person. He walked in and offered it. After the Lindberg baby was kidnapped in 1932, something like 200 people confessed to the crime.

After the murder of Elizabeth Short in Los Angeles in 1947, the case the newspapers called The Black Dahlia, dozens of people came forward to say they'd done it, and police had to spend time clearing each one. In 2006, a schoolteacher named John Mark Carr was arrested in Thailand after telling investigators he'd been with John Bonae Ramsey when she died. He was flown back to Colorado in a business class seat, and his DNA didn't match anything at the scene. The charges were dropped within two weeks.

weeks. Voluntary confessors tend to get caught because police are skeptical of them. Detectives know about the Lindbergh 200. A man who walks in off the street gets tested against the holdback facts, the details the department never released, and he usually fails. The most expensive voluntary false confessor in American history was probably Henry Lee Lucas. In the early 1980s, Lucas, a drifter already convicted of killing his mother and in custody in Texas began confessing to murders. Hundreds of them. Police departments from all over the country flew into interviewing. A Texas Rangers task force cleared more than 200 open homicides on his say-so. In 1985, the Dallas Times Herald published work record and receipt evidence showing Lucas had been

in other states, sometimes a thousand miles away on the dates of killings he'd claimed. And the Texas Attorney General's office issued a report concluding that most of the confessions were impossible and that investigators had fed him the details. Lucas himself later said he'd confessed to everything they brought him because it got him better treatment, trips out of jail, and strawberry milkshakes. One of those murders put him on death row. In 1998, Governor George W. Bush commuted that sentence, the only one he commuted in office, because the evidence showed Lucas had been in Florida when the woman in Texas was killed. The second kind of false confession is the compliant one. The person knows he didn't do it.

He says he did anyway because at that moment, in that room, confessing looks like the fastest way out. He's been there 12 hours. He's been told the evidence is overwhelming, and the only thing left to decide is whether the prosecutor sees him as a monster or as a man who made a mistake. He's been told, or allowed to believe, that he can go home if he just signs. He thinks the lie will get sorted out later because he's innocent and the truth will come out. most of the cases I'm going to tell you about are this kind. The third kind is the one that happened to Peter Riley

and to the students in Casson's lab who told the stranger in the hallway they'd done it. Internalized. The person starts out certain, is told over and over that the evidence proves otherwise, is offered an explanation for why he doesn't remember, and eventually comes to believe, at least for a while, that he must have done it. The people this happens to tend to be young, tired, trusting of authority, and already a little unsure of themselves. It doesn't take much once they accept the premise that their own memory might be wrong.

Across the two big databases of American exonerations, the numbers on who falsely confesses line up the same way every time. In the National Registry of Exonerations, about 12% of all exonerated people had confessed. Among exoneries who were under 18 at the time of the crime, it's about a third. Among those who were 14 or 15, more than half. Among exoneries under 14, nearly 8 out of 10. Among people with a documented mental illness or intellectual disability, roughly 7 out of 10. The Innocence Project's own figures on DNA cases show that about half of the people who falsely confessed were 21 or younger when they were questioned.

That's who ends up confessing, and the cases bear it out one after another. In the spring of 1989, a 28-year-old investment banker named Trisha Miley went for a run in Central Park after work and was found hours later near the 102nd Street transverse, raped, beaten, and so badly injured that she wasn't expected to live. The same night, a group of teenagers had been in the park, and some of them had assaulted, and robbed other people. Police rounded up a number of boys from Harlem over the following day. Five of them ended up charged with the rape. Kevin Richardson and Raymond Santana were 14. Antron Lecrae and Yousaf Salam were 15. Corey Wise was 16. Wise wasn't even on the police list. He went down to the precinct to keep his friend Salam company and ended up questioned himself.

The boys were held in question for somewhere between 14 and 13. 30 hours each, depending on which one you're talking about. Some of their parents were present for parts of it and absent for others. Santana's father was there, then left to get food and was kept out. Wise, who had a hearing impairment and a documented learning disability, was questioned without any adult of his own. Detectives told each boy that the others had already put him at the scene, which at the start wasn't true. They told them they were witnesses, that they could go home once they'd explained what the others had done. Four of them gave videotaped statements. Salam gave an oral statement and refused to sign anything. On the tapes, each boy minimizes his own role and describes

the others doing the worst of it. The statements don't agree with each other about where in the park the attack happened, who held her down, who hit her, or what was used to hit her. None of them matched the physical evidence, which included semen recovered from the victim that didn't belong to any of the five. The prosecution knew that before trial. The theory they took to the jury was that there must have been a sixth participant who got away. Two juries in 1990 convicted all five. The four younger boys got five to ten years in juvenile facilities. Wise, as a 16-year-old, was sentenced as an adult, and he served nearly 13 years in state prisons, including Rikers and Attica. In prison, Wise crossed paths with a man named Matthias Reyes. Reyes was serving a life

sentence for a string of rapes and a murder committed in the months after the Central Park attack. In 2002, Reyes told a corrections officer that he'd raped the jogger, alone, and the DNA matched him. He described the crime in detail that fit the physical evidence. He had, it turned out, attacked another woman in the park two days before Maley, and that crime had been reported. to the same precinct. The Manhattan District Attorney, Robert Morgenthau, reinvestigated and asked the court to throw out all five convictions. In December of 2002, a judge did.

The city fought the civil case for more than a decade and finally settled in 2014 for $41 million. Every one of those boys had recanted within weeks of the tapes. All of them said the same thing about why they'd talked. They were told they could go home. Eight years later and 300 miles south, in July of 1997, a Navy wife named Michelle Moore Bosco was found raped and stabbed to death in her apartment in Norfolk, Virginia. Her husband came home from sea and found her.

The investigation went to a Norfolk detective named Robert Glenn Ford. In 2010, in an unrelated matter, Ford was convicted in federal court of extortion, and of lying to federal agents. He'd been taking money from defendants in exchange for favorable treatment. He was sentenced to 12.5 years. The detective who got four confessions in the Moore Bosco case

was, a decade later, a convicted felon. The first man he brought in was Daniel Williams, a sailor who lived across the hall. Williams' own wife was dying of cancer at the time. Ford questioned him for roughly 11 hours overnight. He told him he'd failed the polygraph. He told him a witness had identified him. Williams eventually confessed,

and in his first version, he said he'd beaten her to death with a shoe. Then the autopsy came back, and it showed she'd been stabbed and strangled. Williams's confession changed to match. His DNA didn't match the semen at the scene. Rather than treat that as a problem with the confession, Ford treated it as evidence of a second man. He brought in Williams' roommate, Joseph Dick, another sailor. Dick confessed after hours of questioning and put himself and Williams at the scene. Dick's DNA didn't match either. So there had to be a third. Dick was pressed to name more people, and he named Eric Wilson, and then Derek Tice, and the circle kept growing until seven men had been charged with the rape and murder of a woman in a small apartment that showed no sign of more than

one attacker. The medical examiner had found no evidence of a gang assault. None of their DNA matched. In 1999, a man named Omar Ballard, already in prison for raping another woman in the same neighborhood, wrote a letter to a friend saying he'd killed Michelle Morbosco and that the men charged with it were innocent. He said he'd acted alone. His DNA matched. He later pleaded guilty and described the crime in a way that fit every physical detail. The prosecution's response was to add Ballard to the conspiracy.

He'd done it with the sailors, they said, even though Ballard said he'd never met them, and the sailors said they'd never met him. Charges against three of the men were eventually dropped. Williams and Dick pleaded guilty to avoid the death penalty. Tice went to trial and was convicted. Wilson was convicted of rape and acquitted of murder. It took until 2009 for Virginia's governor to grant conditional pardons that got the three men still inside released.

It took until 2016 for a federal... judge to vacate the convictions of Williams and Dick, finding that no reasonable juror could have convicted them in light of what was now known. In March of 2017, Governor Terry McCalliffe granted all four absolute pardons. Joseph Dick's confession is the one I think about most from that case. Dick was, by every account including his own, a young man who was easily led. He eventually came to believe he'd been there. He testified against Derek Tice. He described the crime in a courtroom as though he remembered it. Years later, after Ballard, he said he didn't know anymore what he'd actually remembered and what the detective had given him. The case that shows

internalized confession spreading through a whole group happened in Nebraska in the same state that had apologized to Daryl Parker. In February of 1985, a 68-year-old widow named Helen Wilson was raped and murdered in her apartment in Beatrice, a town in the southeast part of the state. The crime went unsolved for four years. In 1989, a former Beatrice police officer named Bertet Searcy who'd left the force to farm and had told the Wilson family he'd solve it, joined the Gage County Sheriff's Office as a deputy and took up the case. Circe built his case around a man named Joseph White and a woman named Ada Joanne Taylor, who'd lived in the area in 1985 and had since moved away. Stay tuned for more disturbing history.

be back after these messages. He arrested them in Alabama and North Carolina. Then he added Thomas Winslow, James Dean, Kathy Gonzalez, and Deborah Sheldon, six people in all. None of them had any physical connection to the apartment. What makes this different from the other cases is what happened after the arrests. The sheriff's office had a psychologist named Gawain Price, who was also a reserve deputy. Price met with the suspects. He explained to several of them that it was possible to commit a crime and repress the memory of it and that the memories could come back through dreams. James Dean started dreaming about the murder. He reported his dreams to investigators, and the dreams got more detailed, and the details got closer to what the investigators knew about

the scene. Sheldon dreamed, too. Taylor came to believe that she'd held a pillow over Helen Wilson's face. She described it in court. Five of the six pleaded guilty or no contest and agreed to testify. Joseph White was the only one who refused. He went to trial in 1989, and the jury heard Taylor and Dean and Sheldon describe his part in the murder, and he was convicted and sentenced to life. White never stopped asking for DNA testing. In 2008, the state finally tested the semen and blood from the scene, and it belonged to a man named Bruce Allen Smith, who'd been a suspect early on in 1985, had been living in Beatrice at the time. and had died in Oklahoma in 1992.

Smith had been excluded back in 85 by a blood typing error. All six were cleared. Nebraska pardoned them in 2009. Joseph White, who'd served 19 years for refusing to say he'd done something he hadn't, went back to Alabama and took a job at a steel mill and was killed in an industrial accident there in 2011 before the civil case was decided. In 2016, a federal jury awarded the six,

including White's estate, $28 million against Gage County, finding that Cursi and the others had recklessly manufactured the case. The county appealed as far as it could. The Supreme Court declined to hear it in 2019. Gage County, population about 22,000, raised its property taxes to the legal maximum to pay the judgment. Some confessions are false because the person was broken down. Some are false because the person was never capable of understanding.

what was being asked of him. In June of 1982, a 19-year-old mother named Rebecca Lynn Williams was raped and stabbed 38 times in her apartment in Culpepper, Virginia. She lived long enough to tell her husband that the attacker was a black man she didn't know, acting alone. The case went cold. Eleven months later, in a neighboring county, a farmhand named Earl Washington Jr. was arrested for breaking into an elderly neighbor's house and hitting her with

chair. Washington was 22 years old and had an intelligence score of around 69, which put him in the range of mild intellectual disability. He'd been in special education and could not read above an early grade school level. He was, as people who knew him put it, eager to please and inclined to agree with whoever he was talking to. Over two days of questioning, Washington confessed to five separate crimes. Investigators dropped four of them because his confessions couldn't be reconciled with the facts. The fifth was the Williams murder. In that confession, he said the victim was black. She was white. He said she was short. She was five foot eight. He said he stabbed her two or three times. There were 38 wounds. He said nobody else was in the apartment.

Her two small children were there. When officers drove him through Culpepper and asked him to point out the apartment. He couldn't, and they had to show it to him. When he was asked whether he'd raped her, he said no, and then the question was repeated, and he said yes. The statement that went to the jury was typed by the police and read back to him, and he signed it with his name, which was about the limit of what he could write. He was convicted in 1984 and sentenced to death. In 1985, Earl Washington came within nine days of execution, a fellow prisoner who'd been doing legal work on his own behalf, noticed that Washington had no lawyer and no pending appeal and got word to outside attorneys. They filed the papers that stopped the clock. It took until 1993 for

DNA testing to be done on the seaman from the scene, and the results, while not conclusive by the standards of that day, pointed away from Washington. Governor Douglas Wilder commuted his sentence to life in January of 1994 on his last day in office rather than pardon him, and Washington Washington stayed in prison for six more years. In 2000, Governor Jim Gilmore ordered new testing with newer methods, and the results excluded Washington entirely. He was pardoned for the murder in October of that year and walked out in February of 2001. The DNA eventually matched a man named Kenneth Tinsley, who was already serving time for another rape and who pleaded guilty to killing Rebecca Williams in 2007. In a civil trial, a federal jury found that the state police

investigator who'd taken Washington's confession a man named Curtis Reese Wilmore had fabricated it and awarded Washington two and a quarter million dollars against Wilmore's estate Earl Washington spent 17 years in prison nine and a half of them on death row on the strength of a statement that got nearly every checkable fact wrong in October of 1988 a 20 year old named Nancy DePriest was raped and shot in the back of the head while opening a pizza hut in Austin, Texas. Two young men who worked at a different pizza hut across town, Christopher Ochoa and Richard Danziger went by the restaurant a couple of weeks later.

A manager thought they were acting strange and called it in. Ochoa was 22, a high school honor student with no record. Austin police questioned him across two days. The lead interrogator was a sergeant named Hector Polanco. Ochoa later testified and the city of Austin never seriously disputed, that Polanco told him he'd get the death penalty, that he showed him where on his arm the needle would go, that he threw a chair, and that he told him Danziger had already given him up. Ochoa was told the only way to avoid execution was to confess and testify against his friend. He did both. He pleaded guilty, got life, and took the stand against Danziger in 1990. His testimony put the gun in Danziger's hand. Danziger was convicted of the rape and sentenced

to life. In 1991, Danziger was attacked in prison by another inmate who mistook him for someone else. He was kicked in the head repeatedly. Part of his skull had to be removed. He suffered permanent brain damage and will need care for the rest of his life. Starting in 1996, a man named Akeem Joseph Marino, serving life in Texas for other crimes, began writing letters. He wrote to the Austin police. He wrote to the district attorney. He wrote to Governor George W. Bush. In each letter, he said he'd raped and killed Nancy DePriest alone, that he'd found religion and that two innocent men were in prison for it. He told them where he'd hidden the bank bag and the handcuffs from the restaurant. The police had already recovered those items from

his mother's house years earlier in an unrelated search. Nobody acted on the letters for two years. The Wisconsin Innocence Project eventually took Ochoa's case, and in 2000, DNA testing matched Marino and excluded both Ochoa and Danziger. They were released in January of 2001. Austin paid Ochoa $5.3 million, and Danziger $9 million. Christopher Ochoa went to law school at the University of Wisconsin and passed the bar. He's a practicing attorney. The story of Jeffrey Descovic is the one of the one. I'd point to if somebody asked me how a confession can beat DNA at trial.

In November of 1989, a 15-year-old named Angela Correa was raped and strangled in a wooded area in Peekskill, New York, north of the city. Descovic was 16, a classmate who'd been absent from school on the day she disappeared, and who seemed to police to be taking an unusual interest in the case. He went to her wake. He cried more than the other kids. He offered theories to detect Over about two months, police spoke with him repeatedly. Then, in January of 1990, they drove into a polygraph examiner in Brewster, 40 minutes away, with no parent and no lawyer. The session ran roughly six hours.

He wasn't fed. He was told he'd failed the test. Near the end, he gave a statement, and by the time the detectives were done, he was on the floor under the table in a fetal position, sobbing. Before trial, the seaman recovered from the victim was tested. It wasn't his. The Westchester prosecutor went forward anyway. The theory presented to the jury was that Angela Correa, 15 years old,

must have had consensual sex with some unknown boyfriend shortly before Descovic killed her. The jury convicted him in December of 1990. He got 15 years to life. He lost every appeal. In 2000, his federal petition was thrown out because his lawyer had filed it four days late. The court had told the lawyer the deadline was a few days later than it actually was. The Second Circuit said it was sorry, but the rules were the rules. In 2006, a new Westchester District Attorney, Janet DeFiori, agreed to run the old DNA profile through the state database.

It matched a man named Stephen Cunningham, who was serving time for strangling another woman in Peek, a few years after Correa. Cunningham confessed. Descovic was released in September of 2006 after nearly 16 years. He'd been 17 at sentencing. When he walked out, he was 32. The Michael Crow case in Escondito, California,

is the one where you can hear the technique doing it to work on a child because the Escondido police recorded all of it. In January of 1998, 12-year-old Stephanie Crow was found stabbed to death on the floor of her bedroom. Her parents, her grandmother, and her two siblings had all been in the house overnight. Nothing was missing. The doors were locked. Detectives decided quickly that the killer was inside the family, and they settled on her 14-year-old brother, Michael.

They separated him from his parents and placed him in a facility for abused children, where he wasn't allowed to see them. Over the next several days, they interrogated him for more than ten. hours in total across multiple sessions. They hooked him to a device called a computer voice stress analyzer, told him it was a lie detector, and told him it showed he was lying. They told him they'd found Stephanie's blood in his room, which was false. They told him there were two Michaels, the good one and the bad one, and that the good Michael should help them understand what the bad one had done. They told him that if he confessed, he'd get help instead of prison. Michael denied it for hours.

Then he started saying he didn't remember, but maybe he'd done it. Then he said, he must have. He said, I'm not sure how I did it. I just know I did it. He never described the crime in a way that matched the scene, because he didn't know what the scene looked like. Detectives then brought in two of his friends, Joshua Treadway and Aaron Houser, both 15.

Treadway was questioned for roughly 22 hours in two sessions, one of them running past three in the morning, and eventually gave an elaborate, detailed confession in which the three boys planned the murder and he served as lookout. Hauser never confessed, but gave what the detectives got him to call a hypothetical description of how the murder might have gone. All three were charged. What the detectives had set aside was a man named Richard Toot. Tuett was a schizophrenic transient who'd been in the Crow's neighborhood that night, knocking on doors and looking in windows, asking for a girl named Trey.

Cray. Several neighbors had called the police about him. Officers had stopped him, talked to him, taking his sweatshirt, and let him go. A year later, as the boys were headed for trial, the prosecution had Tewitt's sweatshirt retested. Three spots of Stephanie Crow's blood were on it. The charges against the boys were dismissed in January of 1999. Tewitt was eventually tried, convicted of voluntary manslaughter in 2004, had that conviction overturned on appeal, and was acquitted at a retrial in 2013. Nobody is in prison for the murder of Stephanie Crow. The city of Escondido and the city of Oceanside paid the three families $7.5 million. In 2012, a California judge took the unusual step of declaring Michael Crow, Treadway, and

Hauser, factually innocent, a finding that erases the arrest from the record. The Ninth Circuit, in the civil case, described the interrogation tapes as something that could shock the conscience. Michael Crowe was 14. He'd found out his sister was dead one morning, and by that night, he was in a room being told a machine said he killed her. In Waukegan, Illinois, in August of 1992, an 11-year-old babysitter named Holly Staker was raped and stabbed

in the apartment where she was watching two small children. A 19-year-old named Juan Rivera, who had a history of special education and emotional problems, was on home electronic monitoring for a burglary at the time. The monitor showed him at home. A jailhouse informant put his name forward anyway. Lake County detectives questioned him over four days. On the fourth night, after an all-night session,

Rivera gave a statement so inconsistent with the facts that the detectives decided it couldn't be used. They went back in. By early morning, he'd given a second, cleaner one. Between the two sessions, guards at the jail found him in his cell incoherent, pulling out his hair and banging his head against the wall. A nurse described him as being in a psychotic state. He signed the second confession a few hours after that.

Stay tuned for more disturbing history. We'll be back after these messages. He was convicted in 1993. The conviction was overturned on appeal. He was convicted again in 1998. In 2005, DNA testing on the semen excluded him, and the Illinois courts granted him a third trial. At that trial, in 2009, the Lake County State's Attorney's Office argued that the 11-year-old

victim had been sexually active with somebody else, that the seaman came from that unknown person, and that Rivera had raped her without leaving any and then killed her. The jury convicted him a third time. In December of 2011, the Illinois Appellate Court reversed, and it didn't send the case back. The judges wrote that the state's evidence was so weak and the DNA so decisive that no reasonable jury could have convicted, and they barred a fourth trial. Rivera walked out in January of 2012 after nearly 20 years. In 2015, Lake County paid him $20 million, the largest wrongful conviction settlement for a single person in American history at that time. The DNA from the

Staker case was later matched to evidence from a different Lake County murder, which has raised questions about yet another conviction there. The man who killed Holly Staker has never been charged. A Long Island case shows the lie about a co-defendant taken about as far as it can go, because the co-defendant was the suspect's dying father. Marty Tancliff was 17 in September of 1988 when he woke up in his family's house in Beltaire on Long Island and found his mother, day, and found his mother and his father unconscious and bleeding. He called 911. The Suffolk County detective who took the case,

James McCready, decided within hours that Marty had done it. During the interrogation, McCreedy left the room, came back, and told Marty he'd just gotten off the phone with the hospital that his father had come out of the coma and said Marty was the one who attacked him. That was a lie. Seymour Tankcliffe never regained consciousness.

He died a month later. Marty, who'd been told by the one authority figure in his life that his father had named him, said that maybe he'd blacked out. McCready wrote up a statement. Marty refused to sign it, and when the family's lawyer got there a short time later, he repudiated it entirely. The jury heard it anyway. He got 50 years to life.

His father's business partner, a man named Jerry Steuerman, owed Seymour Tancliff a large sum of money, had been at the house playing cards of the nightbook. before and within days of the attack faked his own death and fled to California under an assumed name. The detectives never treated him as a serious suspect. It took a private investigator, working for free, almost 20 years to develop the witnesses that got the conviction vacated in 2007. Marty Tankcliffe served 17 years. In 2020, he was admitted to the New York bar. Damon Tibodeau was 22 in July of 1996 when his 14-year-old

old step-cousin, Crystal Champagne, disappeared from a Jefferson Parish apartment complex outside New Orleans. Her body was found the next evening by the levy, strangled with a cord. Tibado had helped search for her. Detectives brought him in, and over nine hours, with no sleep and no lawyer, he confessed to raping and killing her. He described a rape. The autopsy found no evidence of any sexual assault. He described the cord as something it wasn't. He described things about the scene that were wrong and left out things that were right. He recanted within hours. He was convicted anyway and sentenced to death. He spent 15 years on death row at Angola, most of it in solitary confinement. In 2007, the Jefferson Parish District Attorney, Paul

Connick, agreed to a joint reinvestigation with the Innocence Project, which is rare. The DNA on the cord wasn't his. In September of 2012, the conviction was vacated and he was released the same day. He became a long-haul trucker and died in 2021 at 47 of complications from the coronavirus. Pui Burton was 16 in January of 1989 when he came home from school in the Bronx and found his mother, Caziah, stabbed to death. Detectives questioned him without a parent and told him he'd be treated as a juvenile and go home if he admitted it. He confessed. Within days, a man who rented the downstairs apartment, Emmanuel Green, was found with the Burton's missing car, with a prior record of violence, and had told a girlfriend things that implicated him. Green was killed in an

unrelated robbery before the case came to trial. Prosecutors went forward against Burton on the confession alone, and he was convicted in 1991. Burton was paroled in 2009, having served nearly 20 years. In 2019, the Bronx District Attorney's Conviction Integrity Unit agreed to vacate the conviction. Part of what persuaded them was a review of other cases handled by the same three detectives, which turned up a pattern of confessions taken from teenagers in the same way. Saul Kasson, the psychologist from the Alt-key study, testified as an expert on how Burton's confession had been obtained. He'd been coached through a rehearsal before the recorded version was taken. The Dixmore case, south of Chicago, is the one where the lie about a co-defendant was simply

passed down a line. In November of 1991, a 14-year-old named Cateresa Matthews disappeared on her way home from school in Dixmore, Illinois. Her body was found three weeks later in a field. She'd been raped and shot in the mouth. The case sat for almost a year, and then Illinois State Police investigators picked up a 15-year-old named Jonathan Barr and questioned him, And then his friend Robert Taylor, also 15, and then James Harden, Robert Lee Veal, and Shane Sharp, no relation to me that I'm aware of, the oldest of whom was 17. Veal, Sharp, and Taylor each confessed, after long sessions, and after being told the others had already named them. Each confession put all five at the scene and had them take turns.

The seaman recovered from the girl was tested before trial. It belonged to one man, and that man was. wasn't any of the five. Prosecutors went forward anyway. Veal and Sharp took plea deals and testified against the other three. All five were convicted, with sentences running up to 80 years. In 2011, the DNA was run through the state database and matched a man named Willie Randolph, who'd been 33 at the time of the murder, had lived nearby, and had a long record of violent crime. He'd been out of prison on parole for a few weeks when Catarisa Matthews was killed. All five convictions were vacated that November. Robert Taylor and Jonathan Barr had been locked up

since they were 15. In 2014, the Illinois State Police settled with the five for $40 million, and Cook County later paid a settlement of its own. In West Memphis, Arkansas, in 1993, three eight-year-old boys were found murdered in a drainage ditch. A 17-year-old named Jesse Miss Kelly Jr., with an intelligence score around 70, was questioned for about 12 hours. The police recorded 46 minutes of it. In the recorded portion, he placed the crime in the morning when the boys were in school and had to be corrected by the detective. He got the manner of death wrong. He described details that didn't match the scene. His confession was used to arrest Damien Eccles and Jason Baldwin, and the three of them spent 18 years in prison, Eccles on

Death Row, until they were released in 2011 under a plea arrangement that let them maintain their innocence. If you line up those cases, the pattern hardly needs explaining. Teenagers, people with intellectual disabilities, long sessions, mostly overnight, no parent, no lawyer. A lie about the evidence. A lie about what a co-defendant said. A lie about the polygraph. The offer, spoken or implied, that confessing means going home. And in almost every one, A confession that was wrong about facts the real killer couldn't have gotten wrong, which the jury heard about and convicted on anyway. Juries believe confessions, and that's what makes the rest of it dangerous.

Casson and others have run studies where mock jurors are shown a confession, told explicitly by the judge that it was coerced and should be disregarded, and the conviction rate still goes up. Confession evidence is more persuasive than eyewitness testimony, more persuasive than character evidence, and in the Descovic and Rivera cases, it was more persuasive than DNA. People think the way I used to think. Nobody would say they did it if they didn't. And confessions contaminate everything around them. The Innocence Project found that in more than half of its false confession cases, the confession contained details that hadn't been made public.

That sounds like proof of guilt until you ask where the details came from. In case after case, the answer is that the detective supplied them. Sometimes on purpose, through leading questions and photographs and crime scene walkthroughs, sometimes without meaning to, because a man who's been told what the evidence shows for 12 hours knows what the evidence shows. The Brendan Dassey interrogation in Wisconsin is the clearest example on tape. Dassey, 16, with an intelligent score in the low 70s, is ask what happened to Teresa Hallbach's head, and he guesses, and he guesses again.

and the detective finally says, all right, I'm just going to come out and ask you, who shot her in the head? And Dassey says, his uncle did. Then he asks whether he'll be back at school in time for sixth period. A federal magistrate threw out Dassey's conviction in 2016. A three-judge appellate panel agreed. The full Seventh Circuit then reheard the case and reinstated the conviction by a vote of four to three, with the majority saying the state courts hadn't been unreasonable in finding the confession voluntary. The Supreme Court declined to take it. Brendan Dassey is still in prison and isn't

eligible for parole until 2048. Once a confession exists, forensic analysts know about it when they examine the evidence. Eyewitnesses who were unsure become sure. Alibi witnesses start to doubt themselves. Prosecutors, faced with DNA that doesn't match, construct an unknown sixth man or an unknown boyfriend rather than question the statement. The confession stops being one piece of evidence and becomes the frame everything else gets fitted into. Some things have changed. In 2001, the Chicago Tribune published a series documenting 247 murder cases in Cook County since 1991, in which a confession had been thrown out, or the defendant acquitted, or the charges dropped. Chicago had a particular history. A police commander named

John Burge and the detectives under him at Area 2 had spent the 1970s and 80s extracting confessions from black suspects with electric shock, suffocation with plastic bags, and mock executions. And dozens of those convictions have since been thrown out. The pattern at Area 2 came to light largely because of one man, Andrew Wilson, who was convicted of killing two Chicago police officers in 1982, and who arrived at the jail after his interrogation with burns on his chest and thighs that a jail doctor documented and refused to ignore. Wilson's civil suit and the lawyers who kept pulling at it for 20 years produced the list of other men who described the same black box with the hand crank, the same typewriter cover held over the face.

Burge was never convicted of the torture because the statute of limitations had run. He was convicted in 2010 of lying about it under oath and served three and a half years. The city has paid out well over $100 million in settlements, and in 2015, established a reparations fund for the survivors. In 2003, Illinois became the first state to pass a law requiring that homicide interrogations be recorded from start to finish. The bill's sponsor in the state Senate was Barack Obama. Today, a majority of states require recording in at least some categories of cases, and a lot of departments do it voluntarily, even where they don't have to. Recording doesn't prevent a false confession. The Crow and Dassey interrogations were recorded in full. It does mean a judge and

jury can see how the statement was built, instead of being handed the finished product and told it was freely given. In 2017, Wicklander Zulowski, one of the largest police training firms in the country, and a company that had taught the reed technique for more than 30 years, announced it would stop teaching it. The firm said the confrontational approach carried too high a risk of producing false confessions. Reed and Associates said its competitor had misrepresented the method and continues to train it. In 2021, Illinois became the first state to make a confession from a minor inadmissible if police obtained it by knowingly lying about evidence or leniency. Oregon followed the same year. Ten states have now passed some version of that law. In every one of those states,

and in all the others, the police may still lie to an adult. The British went a different direction 30 years ago, and they did it because of their own false confessions. In 1974, bombs and two pubs in Guilford killed five people, and four young people, including a 20-year-old from Belfast named Jerry Conlin, confessed after days of questioning that Conlin said involved beatings and a threat that his mother would be shot. The confessions contradicted each other and the physical evidence. All four were convicted and spent 15 years in prison before the Court of Appeal quashed the convictions in 1989 after it came out that detective's notes of the interviews had been rewritten.

The Birmingham Six convicted the same year on confessions obtained the same way, were cleared in 1991. Out of those cases, England and Wales rebuilt their interviewing from the ground up, required every suspect interview to be recorded, and in 1992 adopted a method called Peace that's built on open questions and letting the subject talk without confrontation, accusation, or lying about evidence. Officers there aren't permitted to tell a suspect the evidence shows something it doesn't. In this country, the Reed technique is still the dominant method taught. Some departments and some federal agencies have moved toward the British model, and a federal research group formed after the Iraq war to study interrogation of terrorism suspects, came back saying rapport-based questioning

produces more accurate information than confrontation does. But the room most American suspects walk into tonight is the room John Reed designed. Peter Riley is in his early 70s now. After the charges were dismissed, he stayed in Connecticut and worked as a car salesman, among other things, and has spoken over the years to law students and reporters about what happened in that polygraph room. The tape of his interrogation is used in training by people who study false confessions. The witness statements that cleared him sat in John Bianchi's files from 1973 until Bianchi died in

1977. Nobody has ever explained why they were never turned over, and Bianchi isn't here to answer for it. The state police never developed another suspect in the murder of Barbara Gibbons. The one-man grand jury that cleared Peter in 1977 heard testimony about other people who'd been in the area that night and found no basis to charge anyone. The file is a judge. The file is a Canaan has been open for 53 years, and the Connecticut State Police still list the case as unsolved. Her son was the only person ever tried for killing her. He'd been awake for most of a day, and he was told by a sergeant that a machine had proven he was lying, and after 25 hours, he said that it really looked like he'd done it.