The clock is ticking. On April 27, the state of Texas is scheduled to execute by lethal injection 53 year-old Melissa Lucio. In 2008, Ms. Lucio was convicted largely on the basis of a confession for the alleged murder of her two-year-old daughter, Mariah. Now her attorneys have filed a clemency petition to the Texas Board of Pardons and Paroles. They point to evidence indicating that the child died from injuries resulting from an accident, not murder.
I did not work on this case and am not privy to the entire case file, so I am not in a position to argue for Ms. Lucio’s actual innocence. But I am in a position to say this: The five-plus hours of police interrogation of this highly vulnerable woman, starting approximately two hours after her child’s death and late into the night, were so psychologically coercive that the risk of eliciting a false confession was substantial.
I know what you’re thinking: It is inconceivable that an innocent person would confess to a crime they did not commit—much less a murder, much less the murder of their own child. It’s inconceivable until you are informed at two levels—the first pertaining to the reality in general of false confessions; the second pertaining specifically to the interrogation of Melissa Lucio.
For three decades now, the Innocence Project has archived 375 DNA exonerations in the United States, dating back to 1989, and including some who had served time on death row. To everyone’s surprise, false confessions (by the defendant or someone else implicating the defendant) have contributed to 29% of these wrongful convictions. Countering intuition, that number climbs to over 60% in homicide cases. And these are only the “lucky” cases that were discovered; they represent an unknown fraction of a larger total.
Research on false confessions points to both personal and situational risk factors. Some individuals are particularly vulnerable (e.g., minors, adults with IQ limitations or a history of mental health problems); certain interrogation tactics are particularly compelling (e.g., the false evidence ploy, minimization tactics that imply leniency upon confession). When these two sets of factors are combined, the effect is devastating. This framework brings us to Ms. Lucio.
On February 15, 2007, two-year-old Mariah, the youngest of Melissa’s twelve children, fell down a flight of stairs while the family was moving into a new apartment. Two days later, they called 911 to report that she was unconscious. She had died. Police arrived and decided on the spot—without seeking medical records or waiting for an autopsy—that the bruised child was murdered and that Ms. Lucio was the culprit. She was taken into custody, presumed guilty and interrogated by four detectives and a Texas Ranger.
Ms. Lucio emphatically denied harming her daughter over one hundred times. But the badgering was relentless. It is hard to imagine the state of grief that would afflict a mother at that time. In addition, Ms. Lucio had a life history of trauma, sexual abuse as a child, and an adult victim of domestic violence. She had mental health issues, substance abuse problems, and cognitive limitations that lead people to be overly compliant and suggestible. On top of that, she was pregnant with twins and sleep deprived, as the interrogation lasted past 3 a.m. (sleep deprivation is yet another risk factor in false confessions).
Although Ms. Lucio had no history of violence, police plowed forward. When asked at trial why he was so convinced of her guilt without any evidence, the ranger cited her passive demeanor. “Somebody with their head down, like their shoulders are slouched forward, and they don’t look at you,” he testified. “They’re hiding—hiding the truth.” Ms. Lucio had just discovered that her child had died, she stood accused of murder, and this investigator cannot imagine an alternative interpretation for passive demeanor? Whether valid or not, tunnel vision had set in.
Over the years, hundreds of social psychology experiments conducted in laboratories all over the world have converged on this conclusion: Eye contact, posture, and other such demeanor cues have little-to-no diagnostic value. Trained or not, police are not mind readers capable of using demeanor to distinguish truth and deception at near-perfect levels of accuracy. But they are often confident in their hunches, and that confidence fills the archives of wrongful convictions.
Once Ms. Lucio was identified as a prime suspect, she faced a perilous process of interrogation, a portion of which was inexplicably not recorded. Right out of the gate, detectives confronted her with an accusation. Using a cluster of tactics designed to maximize the stress of denial, they leaned in and shouted at her. They said they knew exactly what happened and pretended to have evidence of her guilt—evidence now known to be false. They forced her to look at graphic photos of Mariah’s injuries. Then they hinted at a veiled threat if she did not cooperate: “I won’t be surprised if you won’t be able to attend your child’s funeral.”
Using classic techniques, detectives also used “soft” minimization tactics that make it easier to confess. Feigning sympathy, the ranger offered up the “theme” that with so many children to care for, Ms. Lucio must have made a “mistake” born of frustration. Research shows that this offer of moral justification leads people to “hear” leniency—as if an explicit promise was made. This theme culminated in what trained interrogators call the “alternative question,” the ultimate Hobson’s choice: Was this an accident, or are you a cold-blooded killer? she was asked. “I’m innocent” is not an option.
Caught between a cross current of threats and promises, Ms. Lucio eventually and reluctantly uttered the words “I guess I did it” followed by other vague but incriminating admissions. Next she was handed a doll and instructed to hit it, harder, and harder again. The ranger demonstrated how. She succumbed to this command for a reenactment in the same way she succumbed to the demand for an oral admission. As the session closed, she is crying and uttering the words “I wish I was dead.”
My colleague Gisli H. Gudjonsson, a professor emeritus at King’s College in London, started his career as a police officer, then he went on to become a clinical forensic psychologist and leading expert on compliance, suggestibility, and false confessions. At the request of the Innocence Project, Dr. Gudjonsson examined the case materials and was horrified by the tactics used to plow over Ms. Lucio’s denials and break her down into a state of capitulation. In The Independent, he wrote that “Texas is executing an innocent woman.”
You might think that Dr. Gudjonsson and I embrace a psychological approach that is somehow incompatible with a law enforcement perspective. Not so. David Thompson is President of Wicklander-Zulawski & Associates, an interview training firm located in Chicago, reviewed the interrogation materials in this case and concluded that “Ms. Lucio’s admissions are unreliable and have many of the hallmarks of a coerced-compliant false confession.”
For judges and juries, evaluating confession evidence is challenging, to say the least. When a confessor leads investigators to a murder weapon or a body, that proof of firsthand guilty knowledge seals the deal. That’s a good confession. Ms. Lucio provided nothing of the sort.
Shockingly, most proven false confessions appear to be corroborated by independent evidence from witnesses and forensic examinations that were later discredited. As a result of forensic confirmation biases, these errors typically followed—and were tainted by—the confession itself. In this case, two of the detectives who took Ms. Lucio’s confession were literally in the room with the forensic pathologist afterward as she conducted the autopsy.
Recently, several jurors who served on Ms. Lucio’s capital trial jury have expressed doubts about her conviction. They now oppose her impending execution. In the Houston Chronicle, Johnny Galvan, Jr. says the jury was not fully informed about the Ms. Lucio’s interrogation and repeated denials. In his words, “The idea that my decision to take another person’s life was not based on complete and accurate information in a fair trial is horrifying.”
Executing a woman on the basis of a coerced and uncorroborated admission, extracted under extreme duress, a woman whose actual innocence is in question, would constitute a tragic judgment of unforgivable magnitude. This case demands another look.