So why do so many suspects confess to crimes they didn’t commit?
By SAMUEL GROSS and MAURICE POSSLEY
This article first appeared in the Marshall Project on June 12, 2016.
"You have the right to remain silent."
If you’ve ever watched any of the tens of thousands of hours of television devoted to crime dramas, you know the first warning given to suspects who are arrested and questioned. And the second: “Anything you say can and will be used against you.” The Miranda warnings—named for Miranda v. Arizona, the 1966 Supreme Court decision that required them—celebrated their 50th anniversary on June 13. In that period, they have become so ubiquitous that it’s easy to forget their origin and purpose.
Miranda was the culmination of 30 years of Supreme Court cases that were designed to protect criminal suspects from abuse in police interrogations. The earliest of these decisions prohibited violence and torture. The first concern was to prevent confessions that are “unreliable”—that is, false.
By 1966, false confessions seemed like a rare problem. Fifty years later, we have seen hundreds of exonerations of innocent defendants who confessed to terrible crimes after they received Miranda warnings.
It’s a good time to take stock.
Ernesto Miranda's Police File Photo, 1963
Do innocent people really confess without torture?
Why would an innocent person ever confess to a murder or some other terrible violent crime?
Torture would explain it. That was the issue in Brown v. Mississippi in 1936, the first case in which the Supreme Court excluded a confession from a state court prosecution. Three suspects had been tortured for days. Asked how severely one defendant was whipped, the deputy in charge testified: “Not too much for a Negro; not as much as I would have done if it were left to me.”
Between 1936 and 1966 the use of torture to extract confessions declined greatly, a major accomplishment by American courts and criminal justice reformers. When Miranda was written, a shift was underway to more “modern” methods of interrogation: isolation, deception, manipulation and exhaustion rather than beating. Without torture or threats of death or violence, it seems implausible that an innocent suspect would confess to a serious crime. That is precisely why confessions are such powerful evidence of guilt. But we know it happens, time and again.
Henry Shields, Arthur Ellington, and Ed Brown, the defendants in Brown v Mississippi. Photo Credit: The Crisis Magazine, 1935.
The National Registry of Exonerations has collected data on 1,810 exonerations in the United States since 1989 (as of June 7, 2016). They include 227 cases of innocent men and women who confessed, 13 percent of the total, all after receiving Miranda warnings (at least according to the police). Nearly three quarters of those false confessions were in homicide cases.
But these exonerations deeply understate the extent of the problem.
First, most suspects who falsely confess—probably the great majority—are never convicted at all. In a classic 2004 study, Steven Drizin and Richard Leo identified 125 proven false confessions in the United States from 1971 through 2002. Only about a third were cases of exoneration after conviction. In most, charges were dismissed before trial or never filed at all because of indisputable proof of innocence.
Second, few convictions based on false confessions are cleared by exoneration. That’s true for all wrongful convictions, but especially for those based on confessions. It’s very hard to convince people that a defendant who confessed is innocent. We see this in the cases: Exonerations of defendants who confessed are more likely to depend on the most unassailable evidence, DNA, to overcome the weight of a confession. Forty-two percent of exonerated defendants who had confessed were cleared by DNA tests, compared to only 21 percent of exonerees who had not confessed.
In some cases, even exculpatory DNA evidence doesn’t help. In October 1992, after a grueling four-day interrogation, 19-year-old Juan Rivera falsely confessed to the rape-murder of an 11-year-old girl in Lake County, Illinois. In fact, he confessed twice. His first confession was so riddled with factual errors that the detectives made him do it again to “clear up” the inconsistencies, even though Rivera was plainly in a state of mental collapse.
Rivera was convicted of murder in 1993, and again in 1996 after his first conviction was reversed for a host of legal errors. In 2005, DNA tests proved that a different man was the source of semen recovered from the body of the victim. Rivera’s conviction was vacated but the prosecution took him to trial again, and in 2009, despite the DNA evidence, Rivera was convicted a third time. Finally, in 2011, the Illinois Appellate Court ruled that Rivera’s conviction was “unjustified and cannot stand” and dismissed the charges.
Juan Rivera barely overcame his false confession even with conclusive DNA evidence of innocence. Without it, he’d be in prison today—together with other innocent defendants who confessed but did not have DNA tests to rescue them.
False confessions by co-defendants
In many cases, innocent suspects who confess implicate others who are also innocent. Some do it because that’s the story their interrogators want to hear. John Kogut, for example, not only falsely confessed to his own involvement in murder, he also said he did it with two friends Dennis Halsted and John Restivo, both of whom (like Kogut) spent 20 years in prison before they were exonerated in 2005.
And some innocent suspects who confess blame others to deflect responsibility and reduce their punishment. Richard Ochoa, for example, was facing the death penalty for the murder of Nancy DePriest in Austin, Texas in 1988. He confessed, named his roommate Richard Danziger as the actual killer and agreed to plead guilty and testify against Danziger. Both were convicted and sentenced to life in prison. Both were exonerated by DNA in 2002.
The Registry includes 195 exonerations with confessions by co-defendants who implicated the exonerees, 11 percent of all exonerations. The net result is that in 19 percent of all exonerations in the United States—and in 34 percent of homicide exonerations—the innocent defendant confessed or was implicated by a false confession of a co-defendant, or both.
Who falsely confesses?
All sorts of people falsely confess, but two groups are particularly vulnerable: young suspects and those with mental disabilities.
In 1983, for example, Earl Washington, a 22-year-old black man with an IQ of about 69, was arrested in Culpeper, Virginia, for burglary and malicious wounding. Over two days of questioning, Washington “confessed” to five separate crimes, four of which were not pursued because his confessions did not match the actual crimes and the victims could not identify Washington as the criminal.
Washington’s fifth confession, however, was to a murder, that of Rebecca Lynn Williams. His initial version—before police officers cleaned it up—was riddled with errors. He did not know the race of the victim (white), the address where she was killed, or that she was raped. Nonetheless, Washington was convicted and sentenced to death in January 1984. He was exonerated by DNA 16 years later, in 2000.
Overall, of exonerees with reported mental illness or intellectual disability, 72 percent had confessed.
Young suspects fared almost as badly. Forty percent of exonerees who were under 18 at the time of the crime falsely confessed, including 53 percent of 14- and 15-year olds, and 86 percent of the few who were 13 years old or younger. By comparison, only 7 percent of adult exonerees without reported mental disabilities falsely confessed.
Why do all these innocent defendants confess?
Innocent suspects confess because they are terrified and confused and exhausted; because they are deceived or tricked; because they don’t understand what they are doing; because they feel hopeless and helpless and isolated. But what leads to this desperate predicament? Miranda sets the stage.
In part, Miranda was a step in the Supreme Court’s campaign to eliminate violence in interrogations. But Miranda also ratified the “modern practice of in-custody interrogation [which] is psychologically, rather than physically, oriented.” Miranda described how this is done:
The officers who conduct “modern” interrogations may lie about the evidence and tell the suspect that his fingerprints were found at the scene; that a codefendant already confessed and put the blame on him; that he was seen by an eyewitness. They routinely say that they already have him dead to rights and that this is his only chance to tell his side of the story and help his cause; that the victim must have provoked him; that what he did is understandable. They may describe dire consequences if he does not come clean, perhaps the death penalty, and imply leniency if he does. This can go on for days, in isolation, with police officers constantly repeating that they know the suspect is guilty, that the evidence is overwhelming, that this is his only chance to help himself.
The Supreme Court recognized that this process “exacts a heavy toll on individual liberty, and trades on the weakness of individuals,” but it did not forbid any of these practices. As a result, Miranda is regularly cited as authority for the legality of all of these coercive techniques.
Instead of regulating the process of non-violent interrogation, the court required police to give warnings before they start, and then only continue if the suspect waives his right to silence. But most do waive their rights at the outset of the ordeal; it’s hard to tell an officer who has you under arrest that you won’t talk to him. After that, the issue almost never comes up again.
By the time they confess, Miranda is a distant memory, if not entirely forgotten. The process works. Many suspects confess after Miranda warnings and most are guilty; that’s why these techniques are used and trusted. But some are innocent.
Can we do better? Here again, Miranda is a good starting point.
The court noted that it’s difficult to regulate interrogations because we don’t know what goes on: “Interrogation still takes place in privacy. Privacy results in secrecy, and this, in turn, results in a gap in our knowledge as to what, in fact, goes on in the interrogation rooms.”
That’s changing. Fifty years ago almost no interrogations were electronically recorded. The FBI, for example, prohibited recording. Now the FBI requires it, as do 23 states and many local police forces, at least in homicide cases. It should be universal. Recording greatly helps us evaluate any claim that a confession was false, and it has taught us how to improve the conduct of interrogations. It’s a good start.
Samuel Gross is the editor and co-founder of the National Registry of Exonerations and a professor at the University of Michigan Law School. Maurice Possley is a Pulitzer Prize-winning journalist and senior researcher at the registry.
The National Registry of Exonerations is a project of the Newkirk Center for Science & Society at University of California Irvine, the University of Michigan Law School and Michigan State University College of Law. It was founded in 2012 in conjunction with the Center on Wrongful Convictions at Northwestern University School of Law. The Registry provides detailed information about every known exoneration in the United States since 1989—cases in which a person was wrongly convicted of a crime and later cleared of all the charges based on new evidence of innocence. The Registry also maintains a more limited database of known exonerations prior to 1989.
We welcome new information from any source about exonerations already on our list and about cases not in the Registry that might be exonerations.