A man with a low IQ confesses to a gruesome crime. Confession in hand, the police send his blood to a lab to confirm that his blood type matches the semen found at the scene. It does not. The forensic examiner testifies later that one blood type can change to another with disintegration. This is untrue. The newspaper reports the story, including the time the man says the murder took place. Two witnesses tell the police they saw the woman alive after that. The police send them home, saying they “must have seen a ghost.” After 16 years in prison, the falsely convicted man is exonerated by DNA evidence.
How could this happen? “False confessions can corrupt other evidence, both from laypeople and forensic experts,” says John Jay College of Criminal Justice psychologist Saul Kassin, summarizing a new study conducted with Daniel Bogart of the University of California Irvine and Nova Southeastern University’s Jacqueline Kerner. The findings, which will appear in an upcoming issue of Psychological Science, a journal published by the Association for Psychological Science, have far-reaching implications for judges and juries, prosecutors and defense attorneys.
Confessions, when true, are an important tool in convicting criminals. But false confessions frequently play a major role in convicting innocent people. Experiments show that juries and potential witnesses are influenced by confessions even if they know they were coerced. Also in the lab, experienced polygraph examiners, fingerprint experts, and other experts, when informed of a confession, see what they expect to see—that is, evidence of guilt.
To back up these findings with real-life data, the psychologists thoroughly reviewed the trial records of 241 people exonerated by the Innocence Project since 1992. Of these, 59—or 25 percent—involved false confessions, either by the defendant or an alleged accomplice. One-hundred eighty—or 75 percent—involved eyewitness mistakes. The analysis revealed that multiple errors turned up far more often in false confession cases than in eyewitness cases: 69 percent versus fewer than half. And two thirds of the time, the confession came first, followed by other errors, namely invalid forensic science and government informants.
Kassin believes the findings “greatly underestimate the problem” because of what never shows up in court: evidence of innocence. Told the suspect confessed, “alibi witnesses back out, thinking they’re mistaken,” police stop searching for the real culprit. “We show that confessions bring in other incriminating evidence that is false. What we don’t see is a tendency to suppress exculpatory evidence.”
The study throws doubt on a critical legal concept designed to safeguard the innocent: corroboration. Appeals courts uphold a conviction even if a false confession is discovered, as long as other evidence—say, forensics or other witness testimony—independently shows guilt. “What these findings suggest is that there may well be the appearance of corroboration,” says Kassin, “but it is false evidence that was corrupted by the confession—not independent at all.”
Already, many states require that interrogations be taped, so that confessions are not coerced or taken when the suspect is in psychological distress. With this study, “Juries and judges have more reason to critically evaluate the conditions under which that other evidence was taken, too.”