I really think you should take time to view the PBS documentary The Confessions. It tells the shocking story of the grave injustice done to four sailors in Norfolk, Virginia in 1997. (http://www.pbs.org/wgbh/pages/frontline/the-confessions/)
A young navy wife, Michelle Bosko, had been found raped and murdered in her apartment. Based on the “hunch” of one of her friends, the police started investigating a Navy enlisted man, Daniel Williams, for the crime. Williams at first denied any involvement, but after an interrogation that started late at night and ended 9 hours later, he signed a confession to the rape and murder of Michelle Bosko. The police had found their man and further investigation stopped, at least until months later when a DNA test showed that Williams’ DNA did not match the blood found at the murder scene.
Instead of releasing Williams, the police decided that Williams must have had an accomplice. So a second sailor was arrested. He too at first denied any involvement, but after another long interrogation, he also signed a confession, one which implicated Williams. But then the DNA report for the second defendant also proved negative. The police response was to look for a third and finally a fourth perp. And, since each new confession contradicted the earlier ones, they too had to be reworked to appear consistent.
All four defendants eventually were either convicted or pleaded guilty. Williams and two others were sentenced to life without the possibility of parole. The fourth was sentenced to 20 years for rape.
Why did the police continue to believe Williams was involved even when that conclusion became more and more implausible? Social science can help us with that question. Michael Lewis’new book The Art of Unknowing tells how two Israeli psychologists, Amos Tversky and Daniel Kahneman, upended conventional wisdom by showing that the maxim “To err is human, to forgive divine” is at least half right. Their basic message is that human beings tend to decide many important issues by relying on intuition and human intuition is often wrong.
Take the initial judgment that Williams was the murderer-rapist. It turns out that when we consider the truth of a hypothesis like “Williams did the murder” our minds tend to automatically focus on information that supports its truth, thereby making it appear more probable than it actually is. This “confirmation bias” would have led the police to convince themselves of the truth of Michelle Bosko’s friend’s “hunch” and led them to elicit a confession from Williams to provide support for that conclusion.
And once they had a firm belief they had their “bad guy,” the police were, like most experts, quite good at creating scenarios that reconciled that belief with later evidence, like the negative DNA tests, that undermined it. There must have been an accomplice– or, if necessary, three.
But it wasn’t only the police who erred; there was also the prosecutor who handled the case; he seems to have had no problem with the DNA evidence contradicting the prosecution theory. Then there was the judge who refused Daniel’s lawyer’s motion to suppress the confession, and the appellate court that denied his appeal. Our legal system seems to have a systemic commitment to resolving cases that sometimes outweighs its fear of convicting innocent people.
I don’t doubt that all the officials involved wanted to convict a guilty person, but the system seems to have a greater need to convict someone. The overworked police want to close the case and move on. The prosecutor looks to his or her “conviction rate.” The judge tries to clear a crowded docket. And the appellate courts are too busy to carefully review every allegation of error, and therefore assume that serious errors have been corrected below.
There are also practical reasons for turning a blind eye to questionable convictions. Every admission of error encourages more charges of misconduct, leading to lawsuits and a growing loss of public confidence in the criminal justice system.
I recognize that what happened to Williams and his co-defendants represents an extreme example of a system run amok, but sometimes dramatic cases highlight problems too ignored in less extreme circumstances. If you have any doubt about whether police convinced of a defendant’s guilt are willing to use coercion to get a confession, just take a look at the recent Netflix documentary Making of a Murderer.
Unfortunately involuntary confessions are not the only fallible tool used to get convictions. Eyewitness identifications are also often mistaken. And let’s not forget that the plea bargains involved in the vast majority of all convictions can easily be improperly induced by over-charging or refusal of bail.
Our criminal process sees itself as a well-oiled machine producing correct decisions, but the truth is that it relies heavily on judgments, like which witness to believe, that are based on intuitions that Tversky and Kahneman warn us are often wrong.
It is time to rethink our approach to crime and punishment in our world of uncertainty. There are reforms that can reduce errors in confessions, eye witness identifications, and the plea bargain process, and we should adopt them. But I think we must also face a larger problem. “Guilty” really only means labelled so by a system quite capable of error. So while we should attempt to improve the process, we should also soften the consequences of a mistaken verdict.
A good place to start would be the death penalty. As the the PBS documentary makes clear, the threat of the death penalty induces false confessions. More importantly, it just doesn’t make sense in a world of uncertainty to impose a penalty that can never be corrected if later proven to be mistaken.