Law and Disorder: The Psychology of False Confessions

At 9:45 PM on November 10, 1984, 16-year-old Theresa Fusco finished up her shift at the roller skating rink in the Long Island village of Lynbrook. She never made it home that night. She was reported missing, but nearly a month passed before her body was found, naked, in a wooded area not far from the rink. She had been strangled and covered up with leaves and debris. There was semen in her vagina.

Fusco was not the first young woman to disappear in the vicinity, and police were under tremendous public pressure to make an arrest. They eventually charged three men with the rape and murder, but the case was built largely on the confession of just one, a 21-year-old landscaper named John Kogut. Kogut steadfastly maintained his innocence, and even provided an alibi: He had been at a party for his girlfriend the night of the crime, with several witnesses. Nevertheless, after 18 hours of interrogation, the police produced a confession, written out by an officer and signed by Kogut.

The confession stated that Kogut and the two other men had given Fusco a ride in a van owned by one of the others. She entered the van voluntarily, but was then raped by the two others. Kogut confessed to murdering the teenager with a coil of rope. Based on this confession, police searched the van and found several strands of hair, which were introduced as forensic evidence. The alibi witnesses, now aware of Kogut’s signed confession, started questioning the certainty of their own memories.

Kogut was convicted in May of 1986 and sentenced to 31½ years to life in prison. His two accomplices got similar sentences in a separate trial. The sentences were vacated in 2003, based on new DNA evidence, and Kogut was retried. In 2005, a judge found him not guilty, and set him free. Kogut had spent nearly two decades in prison.

Confessions are powerful and damning evidence, which is a good thing if the defendant is guilty. But what if the defendant is innocent, as Kogut was? It’s not clear just how the police extracted Kogut’s false confession, but the statement clearly had a powerful corrupting effect on his trial. In part, that’s simply because people tend to believe confessions. It doesn’t make any sense that someone would take the blame for murder and rape if they didn’t commit the crimes. But there is also a second, more insidious, explanation for the power of false confessions: Confessions can taint others’ perceptions of potentially exonerating evidence. Were Kogut’s alibi witnesses swayed by his confession? How about the forensic experts? The jurors? The judge?

Fully 25 percent of exonerations based on DNA evidence reveal a false confession, and experts are beginning to believe this is just the tip of the iceberg. Psychological scientist Saul Kassin of New York’s John Jay College of Criminal Justice has been especially vocal on this issue, and in a recent study he and two colleagues decided to systematically reexamine the evidence in exoneration cases that revealed a false confession. They compared them to exoneration cases without a false confession, to see if evidence other than the confession itself was tainted—and contributed to wrongful imprisonment. Were false confession cases more likely to have other evidence errors? What kind and how often? And, importantly, which came first—the other evidentiary mistakes or the false confession?

The scientists examined the case files of the Innocence Project, an organization dedicated to helping prisoners prove their innocence with DNA evidence. Since 1992, the Innocence Project’s efforts have led to the exoneration of 273 prisoners, including 17 who served time on death row. Kassin and his colleagues examined 241 of these cases, including police reports, witness statements, trial testimony, and other court records—basically any evidence that might have led to conviction. Two independent coders examined these records, enumerating the contributing causes of conviction, erroneous eyewitness testimony, bad forensic evidence, and reliance on informants. They also noted the order in which the confession and other evidence were gathered.

The results were troubling. As reported on-line in the journal Psychological Science, multiple errors were discovered in three out of four cases involving a false confession, compared to fewer than half of cases without a false confession. These additional errors included, in order of frequency, invalid or improper forensic science, eyewitness mistakes, and incriminating snitch testimony. Two-thirds of the false confession cases also had forensic errors, and a third had at least two of these additional errors. What’s more, false confessions were much more likely to come before (rather than follow) forensic missteps or informant errors. The timing strongly suggests that the false confession actually corrupted the other evidence.

Laboratory evidence has already highlighted this problem. Confessions have been shown to bias not only other witnesses but also trained polygraph examiners and fingerprint experts. This study is the first to document the same kind of bias and misjudgment among forensic experts in actual criminal cases. The findings reiterate the conclusions of the National Academy of Sciences, which in 2009 was highly critical of many forensic disciplines, including ballistics, handwriting analysis, fingerprint analysis and—most relevant here—hair and fiber analysis. Indeed, the expert hair analysis introduced at Kogut’s original trial was so inept that it was thrown out at his retrial.

How does a false confession distort perceptions of other evidence? It’s not entirely clear, but one possibility is the well known confirmation bias: Belief in guilt leads people to see only evidence of that guilt. It’s also possible that knowledge of a confession motivates people to help the police and prosecutors bring the guilty to justice.

The scientists believe that this analysis underestimates the problem of evidence corruption. It also demonstrates that the criminal justice system’s safeguards are inadequate. Many states require that confessions be corroborated by other evidence in order to be admissible, but this doesn’t help if the corroborating evidence itself has been biased by the false confession. And courts allow convictions to stand if a coerced confession is considered “harmless”—that is, if the remaining evidence is enough to convict. But if the remaining evidence is tainted, as these findings say, that calls into question any possibility that a confession is truly harmless.

Wray Herbert’s book, On Second Thought, is available in paperback. Excerpts from his two blogs—“We’re Only Human” and “Full Frontal Psychology”—appear regularly in The Huffington Post and Scientific American Mind.