False Confessions

Scaring Suspects to Death

by Edwin Dobb

Amnesty International NOW magazine, Winter 2002

When police interrogators manipulate suspects into confessing to crimes they didn't commit, innocent people end up on death row.

Early this year, a man with slight mental retardation was freed after serving a year and a half in an Illinois jail. The 25-year-old's statement that he had killed his mother, while convincing, was nonetheless fabricated. In July, prosecutors in Philadelphia dropped murder charges against four young men, one of whom, defense attorneys insisted, had been coerced into accepting responsibility for a crime he didn't commit. A month later, a DNA test proved that a Detroit resident Eddie Joe Lloyd, could not have raped and killed a teenage girl in 1984. Lloyd was living in a mental hospital when he approached police and offered to confess in order to help them find the actual killer. The 17 years he served before his release were the result of police trickery and, it would appear, his own fateful gullibility.

Lloyd was the 110th person since 1989 to have his innocence established by post-conviction DNA testing, according to the Innocence Project of New York's Benjamin N. Cardozo School of Law, which spearheaded the effort to exonerate Lloyd. Remarkably, false confessions have played a central role in about one-fifth of reversals. In other words, in the past few years more than two dozen demonstrably innocent people admitted guilt for the most heinous of acts.

What might once have been dismissed as a regrettable aberration is looking to some legal experts like an even more regrettable trend and the most compelling example of what can go wrong in the interrogation room. "We've moved from force to fraud," says Richard Leo, a University of California, Irvine criminologist. He is referring to a steady shift-from outright physical abuse to sophisticated psychological techniques-that was accelerated by the 1966 Supreme Court decision known as Miranda.

That ground-breaking and controversial ruling guaranteed that people arrested had the right to remain silent, to have an attorney present, including a court-appointed one if needed, and to be given the warning that anything said during questioning could be used in court. Miranda's protections were meant, as Chief Justice Earl Warren famously put it, "to dispel the compulsion inherent in custodial surroundings."

But coercion continues, though in forms so subtle-and removed from public scrutiny- that they are unrecognizable as such. Indeed, within the modern interrogation setting, police routinely compel suspects to incriminate themselves. Certainly, investigative zeal-especially when combined with sloppiness and pressure from prosecutors to make arrests- contributes to the phenomenon of false confession. But so do broadly accepted policies and procedures that amount to an institutionalized violation of the Fifth Amendment, which is the constitutional underpinning for Miranda. And therein lies the real threat to the criminal justice system and to the human rights of suspects.

As might be expected, certain people-the uneducated, the poor, the emotionally insecure, the mentally impaired, the young-are especially susceptible to manipulation in the interrogation room. In cases involving the death penalty, such vulnerability is especially dangerous. A sobering example, captured on police videotape and obtained by lawyers, is the recent harrowing experience of three teenage boys in Escondido, Calif. Four years ago, after exhausting rounds of police bullying, threatening, and Iying that lasted up to 10 hours at a stretch, Joshua Treadway, 15, and Michael Crowe, 14, admitted to murdering Crowe's sister. Crowe repeatedly said he was fabricating most of what he said, but the detectives encouraged him to continue Iying.

Crowe: The only reason I'm trying to lie here is because you presented me with two paths, one I'm definitely afraid o£ I'd rather die than go to jail.

Shortly after making this statement, Michael was charged with murder. Later, he made a sworn statement saying police had told him that

his parents blamed him for his sister's death and didn't want to see him again. The boy was apparently too disoriented, too exhausted to believe otherwise.

Crowe: Like I said, the only way I even know I did this is that she's dead and that the evidence says I did it. You could find that someone else did it, and I pray to God someone else did. I think it's too late for that. I think I did it.

But no forensic evidence linked Michael and his two friends to the murder. And incredibly, a more likely suspect had been identified before the teenagers were questioned. Then in January , DNA evidence all but proved that the real perpetrator was a mentally ill transient with an extensive criminal history well known to the Escondido police. Almost 12 months after the murder, the teenagers were released, but three innocent families, including that of the murder victim, had been terribly damaged.

Miranda was designed to prevent the kind of interrogation that produced Crowe's confession. And indeed, 36 years ago, when the decision was handed down, those on the front lines of law enforcement had vigorously argued that the ruling would handcuff them and result in far fewer confessions and thus convictions. Today, those who once complained have come to embrace Miranda. Why the change of heart?

The answer lies behind the long-locked doors of the interrogation room. What the Warren Court did not foresee was how its decision would lead to new, questionable interrogation methods. Most people know the "good cop-bad cop" routine popular with police drama writers (and less so with police themselves).

Less familiar are a number of more refined approaches that are amply evident in law enforcement training manuals. None is more revealing nor more influential than the latest edition of a book Earl Warren drew on for his conclusions-Criminal Interrogation and Confessions by Fred Inbau and John Reid. And thanks to social psychologist Richard Of she of the University of California, Berkeley, and criminologist Leo, studies of actual police practice-including the eponymous Reid technique-also now exist. By analyzing hundreds of interrogations, videotape recordings, and transcripts, Ofshe and Leo have drawn aside the curtain that customarily obscures the interrogation room, exposing the often mesmerizing, sometimes disturbing theater that unfolds therein.

And make no mistake, it is theater. Those who developed the Reid Technique characterize their brand of interrogation as "the undoing of deception," on the assumption that all suspects employ subterfuge. But what the manual promotes is a more encompassing and, at times, insidious form of deceit: It teaches police to create a make-believe world of ever-increasing disorientation and discomfort from which the suspect's only hope of escape is to admit guilt. As befits a stage, the interrogator mounts a performance, and the greater the "histrionic skill," as another manual puts it, the more likely a confession. The initial step is always the same: misrepresenting the nature of questioning. As Inbau and Reid baldly advises, "Avoid creating the impression that you are an investigator seeking a confession or conviction." In other words, do precisely what the Warren Court had thought it was curtailing-convince suspects that they are not facing an adversary but instead someone who is looking out for their interests.

Small wonder that once law enforcement personnel understood the disjunction between the ruling's intent and its effect, they made Miranda their own. Rather than hamper police or force an overhaul of the interrogation process, the controversial decision legitimized it. Now, as long as suspects could be persuaded to waive their rights-and most do-interrogators could proceed largely as before. Nor is it surprising that interrogators started to deliberately manipulate the way the warnings are given. In a study of an urban police station in California, Leo found that most detectives employed psychological tactics such as small talk and friendly gestures to build rapport and condition the suspect to respond favorably. They downplayed Miranda rights, camouflaging the warnings within casual conversation, dismissing them as a mere formality, or jokingly referring to television shows. Police also used the particularly compelling device of persuading suspects that only by waiving their rights-usually to tell their own side of the story-could they help themselves.

Once the interrogator obtains a waiver, the drama officially begins. It can be divided into three basic elements: The first-call it the helping hand ploy-extends the ruse that what is transpiring is simply a problem-solving exercise carried out in a spirit of cooperation and for the benefit of the suspect. Detectives show sympathy and frequently make promises. Although courts have held that explicit pledges of leniency are coercive, the routinely admit confessions made in response to indirect suggestions aimed at raising a suspects' expectations or making them fearful.

Dante Parker was one of three men accused of murdering nine people at a Buddhist temple in Phoenix in 1991. Investigators misled Parker by implying he could gain favor by confessing to a less serious version of the crime. They then threatened to send deputies to arrest his brothers, against whom they had no evidence of criminal behavior, and to humiliate them in front of their children. Parker relented and confessed. But two months later, standard police work led to the apprehension of the actual killers and Parker was freed.

Additionally, investigators commonly fabricate evidence, a technique designed to destroy a suspect's confidence. The courts have been slow to establish boundaries for such deception, in effect giving police carte blanche. They can lie about statements from a witness or accomplice or about the presence of blood or other evidence, and to pretend that tests given by Computer Voice Stress Analyzers or other so-called lie-detecting machines are acceptable in court.

And last, whether in response to the helping hand ploy or evidence fabrication, or both, and usually after prolonged, withering questioning, a time may arrive when the suspect actually welcomes the opportunity to construct a story that minimizes the crime or his involvement in it. Such mitigating narratives may assume several forms but the cleverest and, according to police training manuals, most effective turn on ethical misrepresentation. "We offer the suspect psychological justification for the commission of the crime," current Reid and Associates president Joseph Buckley says in a training videotape. "We offer him a moral excuse that will minimize or justify in his own mind [the alleged act]." Buckley goes on to explain that if an interrogator is to have any hope of eliciting a confession, he must squelch denials and deflect objections. And one of the best ways to accomplish that, he argues, is by proposing a story that presents the suspect in a favorable light, suggesting, say, that an obviously deliberate assault was in fact inadvertent or that his rape victim was "asking for it," or that low wages and miserable working conditions made embezzlement inevitable. The culmination of the mitigating narrative arrives when suspects believe they have to choose between a "good reason or a bad reason" for the crime. "We make it easy," Buckley explains in the tape, "for [the suspect] to accept the good reason and make that first admission of guilt." The reason, of course, doesn't count. Only the confession does.

Consider the case of Edgar Garrett of Goshen, Ind., who was suspected in 1995 of his daughter's murder, despite the absence of direct evidence linking him to the crime. Garrett's interrogators lied egregiously, saying that several witnesses saw him with his daughter shortly before she disappeared. They also said that a polygraph test had proved his guilt. Then one of the detectives put it to the accused, who sometimes drank heavily, that he could have experienced a blackout. He reminded the suspect that he had once struck his daughter while in an alcoholic haze. Garrett's confidence in his memory began to falter and, though increasingly upset and confused, he ceased insisting that he had not seen the girl shortly before she disappeared.

Garrett: I just don't remember if I went out, if I did talk to Michelle....l can't remember fighting with Michelle on Sunday.

Detective: You did. Not only did you fight but you thumped her. You didn't mean to hurt her.

Garrett: What did I thump her with?

Detective: I don't know.

Garrett: I don't know either.

Detective: But you thumped her.

Garrett: Well, I killed my own daughter? Detective: Yeah.

When Garrett renewed his insistence that he didn't remember attacking his daughter, the interrogator once again raised the possibility of alcohol-induced amnesia. Garrett struggled with that grim prospect.

Detective: Okay, then what happened next?

Garrett: I must have left her there.

Detective: Okay.

Garrett: And must have went home.

Detective: All right. What did you do with the stick?

Garrett: It's in the house. I must have took it back to the house.

After 14 hours of such cynical manipulation, Garrett signed a statement that he had killed his daughter and was charged with capital murder. In the months that followed, however, evidence turned up showing that Michelle was slain with a knife, not "thumped" with a stick, and that Garrett could not have been anywhere near the crime scene. He was eventually exonerated.

While it is true that relatively few wrongful convictions-8 to 14 percent, in Leo's estimation-are due to false confession, with the bulk caused by false testimony and mistaken eyewitness accounts and identifications, confession is disproportionately influential. Studies of jury behavior show that how a statement is elicited and whether it has been retracted are far less persuasive than the fact that someone actually uttered the words, "l did it." The confession, Richard Of she says, "is universally treated as damning and compelling evidence of guilt."

It is not only the increased likelihood of conviction that makes police-induced false confession one of the most dangerous of judicial errors. At every step of the criminal process, especially when a violent crime is involved, a person who has officially admitted guilt, if only briefly, is likely to receive harsher treatment. Making bail will be more difficult, which, according to Leo, significantly reduces the chance of acquittal. Prosecutors will bring more serious charges, and in greater number, against those who make admissions, whereas defense attorneys, particularly those who remain unaware of Miranda manipulations and the dynamics of false confession-and there still are many-will tend to advise their clients to admit guilt and accept a plea bargain. Then, following a conviction, which is not only more likely but more likely to be for the most serious charges leveled, the suspect can expect to receive stiffer punishment.

Amnesty International, the Innocence Project, and others are pushing hard for reforms. Currently only Alaska and Minnesota require the videotaping of every interrogation, ~ ~ thereby removing the curtain of secrecy that has long obscured the interrogation drama. But more fundamentally, police, prosecutors, and judges must acknowledge that modern psychological tactics are so powerful that they elicit confessions not only from the guilty but from the innocent as well. Lasting institutional change will require a transformation in the way police are trained and supervised. Ofshe and Leo have proposed teaching detectives how to spot the telltale signs of a false confession (for example, the use of such qualifiers as "probably" and "l must have"). The researchers also advocate barring excessively manipulative deceptions-especially implied threats, which they consider the contemporary equivalent of the rubber hose.

Another significant reform would be training police to assess the "post-admission narrative," which gives interrogators the opportunity to corroborate or impeach a confession by testing the suspect's knowledge of details only the perpetrator would know.

Why would anyone resist this simple and sensible way to assess the viability of a confession? Because it necessitates a massive shift in orientation-away from the exclusive pursuit of "l did it" statements, which one police captain told Leo is "the next best thing to sex." During questioning, the police would have to entertain alternative theories of a crime, thereby remaining alert to evidence of innocence as well as to guilt. In other words, they would be forced to cease playing inquisitor on the interrogation stage and instead perform a role much talked about but rarely seen there: that of truth seeker.

Public support for reform seems to be growing as more and more cases surface that resemble the false convictions of the Escondido, California teenagers. Shortly after the charges against them were dropped, Leo, who served as an expert witness for the defense, appeared with an Escondido police officer on a local radio talk show. In this conservative area where the notion that innocent people might admit to violent crimes had been unthinkable, and where support for law enforcement traditionally runs high, it was the police officer who found himself on the defensive. Try as he might to convince listeners that no one in his department intentionally seeks false confessions (which surely is true), caller after caller expressed doubt. Responses to a detailed expose of the case published in the San Diego Union-Tribune were similar. The events that took place in the interrogation room of the Escondido Police Department had sent tremors throughout the community, creating anger, confusion, and fear-damage that would not be repaired quickly.

"What kind of society have we created," asked one indignant letter writer, "in which those charged with finding the truth are dedicated to Iying, distortions, and other self-aggrandizing tactics in order to obtain convictions?"

That is indeed the crux of the matter. Today especially, when community-oriented policing is being introduced everywhere, it seems not only ethically dubious but self-defeating to engage in practices that at best undermine trust and cooperation between citizens and law enforcement officials and at worst kill innocent people and foster cynicism toward legal institutions. Remedying the policies and procedures that tend to produce false confessions would not only protect the accused from coercion, finally fulfilling the promise of the Miranda decision; it would help to ensure the integrity of the criminal justice system. We should expect no less of a democratic society dedicated to fair play.

Edwin Dobb is a contributing editor of Harper's Magazine and a visiting lecturer at the U.C. Berkeley Graduate School of Journalism.

Police State watch

Index of Website

Home Page