Jurisprudence

The Dark Dangers of Tunnel Vision

Why police and prosecutors suspect the wrong person—and then dig in.

Man in handcuffs.

Man in handcuffs.

Photodisc/Thinkstock.

The best magazine piece I’ve read this year is by Pamela Colloff, in Texas Monthly, about the murder conviction of Michael Morton and the 25 years he spent in a Texas prison as prosecutors insisted he’d killed his wife, despite the mounting proof that he hadn’t. You really have to read this story for yourself—it’s movie material; it’s that riveting. I want to pull on one thread of Colloff’s narrative, because I’ve seen it in the weave of many other wrongful conviction cases. I’m talking about tunnel vision: the tendency of investigators to seize on an early piece of evidence that appears to implicate the defendant, and to hold on to their belief in his guilt even as other evidence points to his innocence. It’s a problem that by definition emerges in hindsight. What’s scary is how tenaciously police and prosecutors cling to their initial assumptions—and how much this reflects basic human tendencies.

On an August afternoon in 1986, Christine Morton was found lying on her bed, bludgeoned to death. A neighbor had seen her 3-year-old son Eric walking around the family’s front yard by himself; she searched the house and called the police.

When Sheriff Jim Boutwell arrived, he found this note, signed “I L Y,” for I love you, and “M,” for Michael, who was Christine’s husband.

Chris, I know you didn’t mean to, but you made me feel really unwanted last night. After a good meal, we came home, you binged on the rest of the cookies. Then, with your nightgown around your waist and while I was rubbing your hands and arms, you farted and fell asleep. I’m not mad or expecting a big production. I just wanted you to know how I feel without us getting into another fight about sex. Just think how you might have felt if you were left hanging on your birthday.

Colloff explains that based on the note, which established Michael’s anger with his wife in the hours before she was killed, “Boutwell treated Michael not like a grieving husband but like a suspect.” Boutwell read Morton his Miranda rights and began questioning him. Michael answered without emotion, which “did not help to dispel the sheriff’s suspicion that the murder had been a domestic affair.” Colloff continues:

Odd details about the crime scene only reinforced his hunch. There were no indications of a break-in, a fact that Boutwell would repeat to the media in the weeks to come. (Though it was true that there were no signs of forced entry, the sliding-glass door in the dining area was unlocked.) Robbery did not appear to have been the motive for the crime; Christine’s purse was missing, but her engagement ring and wedding band were lying in plain sight on the nightstand. Other valuables, like a camera with a telephoto lens, had also gone untouched.

As Boutwell’s suspicion of Michael deepened, he chose to ignore the physical evidence that pointed to an outside intruder: Fingerprints on a dining room door frame— and elsewhere—that matched no one who lived in the house. A fresh footprint in the fenced-in backyard. And a bandanna, found by Christine’s brother the day after her death, when he searched a construction site behind the Mortons’ home. (Why weren’t the cops the ones searching? Because they’d already zeroed in on Morton.) The bandanna was stained with blood. When the brother handed it over to the police, they failed to test it or to further search the area. They also disregarded a neighbor’s sighting of a man in a green van on the street on the morning of the murder—and, Morton learned much, much later, his young son’s statement to Christine’s mother that he’d seen a “monster” hit Mommy and “break the bed,” when his father wasn’t there.

Two and a half decades later, after the advent of DNA testing and an agonizing, protracted battle by Michael Morton and his lawyers to submit the bandanna to a crime lab for analysis, it became clear that the blood on it came from Christine and an unknown man. Still, the district attorney’s office that had prosecuted Morton refused to admit he’d been wrongly convicted. And then Morton’s lawyers (who include my friend Nina Morrison of the Innocence Project) succeeded in finding a match for the mystery DNA in the FBI’s database. And then it turned out that the man with the matching profile—Mark Alan Norwood—had lived around the corner from a woman who looked like Christine and who’d been bludgeoned to death in her bed in an unsolved murder. Finally, Michael Morton was released from prison and Norwood went on trial for the two murders for which he’d escaped punishment for so many years. (Ken Anderson, one of the prosecutors in the case, now a Texas judge, is about to go before a state court of inquiry for his alleged misconduct. It’s a rare show of accountability, as Joe Nocera points out in the New York Times.)

In retrospect, it’s clear that police and prosecutors headed off in the wrong direction and then couldn’t see flaws in the theory they’d chosen, or the clues that led elsewhere. The problem isn’t that they started with Michael Morton. Domestic violence statistics show that women are more likely to be killed by their husbands than by someone they don’t know. The problem, rather, is that they ended with Morton. The same also seems true in the case of Cameron Todd Willingham, who was executed in 2004 for the deaths of his three children, based on the theory that he set the fire that burned down his house and killed them. Read David Grann’s amazing New Yorker exposé about the case, and you come away horrified by the class assumptions and shoddy burn-pattern analysis that prosecutors relied on to convict Willingham. And in his new book A Wilderness of Error, Errol Morris makes a strong case for the wrongful conviction of Jeffrey MacDonald, (yes, the subject of Fatal Vision and The Journalist and the Murderer) who remains in prison for the 1970 murder of his wife and two daughters. Once more, the husband MacDonald was the obvious suspect. And like Morton, he appeared unemotional rather than grief-stricken. Investigators didn’t believe his story about a break-in by a bunch of hippies. Feeding their skepticism was a key piece of initial evidence: A coffee table found in the family’s living room lying on its side, which the investigators said couldn’t have fallen that way given its construction. The table was part of the tableaux of disorder MacDonald claimed he woke up to after he’d lost consciousness when he was wounded; if its toppling over was staged, then that called his entire account into doubt. As Morris tells the story to dramatic effect, a military judge later knocked over the table in a test—and the table fell onto its side, exactly as it had been lying in the crime scene. But for MacDonald, this debunking came too late to halt the snowballing suspicion of him, or to persuade prosecutors to take seriously the confession of a woman who matched the description of one of the intruders MacDonald said he saw.

More common ground that Morton and MacDonald share: Their wives were murdered in placid middle-class neighborhoods, in which fears of a dangerous assailant on the loose was only a source of trouble for local law enforcement. Both Colloff and Morris suggest that arresting the husbands, and sticking with the theory supporting their guilt, helped defuse local tensions and made everyone in the community feel safer.

It’s natural for police and prosecutors to want to ease public fears. And it’s also natural for them to stick with the evidence that supported their preferred explanations. As University of Wisconsin clinical law professor Keith Findley shows in his excellent 2010 article Tunnel Vision, the phenomenon “is the product of a variety of cognitive distortions,” chief among them confirmation bias. In other words, we tend to give weight to evidence that confirms our existing beliefs. “Although such confirmation-biased information is often less probative than disconfirming information might be, people fail to recognize the weakness of the confirming feedback they receive or recall,” Findley writes. He cites studies finding that “police officers who are convinced that a suspect is lying are very resistant to changing their minds” and often “rate disconfirming or exonerating evidence as less reliable or credible than guilt-confirming evidence that supports their initial hypotheses.”

I called Brandon Garrett, the law professor who wrote the book on wrongful convictions and why they happen, and he pointed out that police and prosecutors have no obligation to pursue alternative explanations, or even to follow a particular method of investigation or keep a record explaining the course they’re taking. Which means it’s close to impossible to hold them accountable for their errors. Garrett pointed out that crime labs are different: They write down every step they take. “Even though they have crushing caseloads, they follow that procedure,” he said. And that means we know when there’s a broken link in the chain. Police too, Garrett thinks, should have to record every witness interview they do; perhaps they should also keep notes about how their thinking evolves or doesn’t. Garrett also suggested giving the defense more resources to do their own investigations or requiring police and prosecutors to open their files to the defense. Tunnel vision isn’t going away. But to protect the innocent, we should diminish its dark power.