LESLEY STAHL FAILS A VISUAL MEMORY TEST AND DEMONSTRATES WHY SO MANY INMATES HAVE BEEN WRONGLY CONVICTED BY EYEWITNESS TESTIMONY -- "60 MINUTES" SUNDAY

Eyewitnesses Error Played a Role in 75 Percent of DNA-Exonerated Inmate Cases

When Lesley Stahl fails a visual memory test on this Sunday's 60 MINUTES, she demonstrates how easy it is for eyewitness' memories of a crime to fail them, leading them to identify an innocent person -- and how eyewitness memories can be influenced by the way police conduct identification procedures. Stahl's report helps explain why there have been more than 230 men exonerated by DNA, more than 75 percent of whom were wrongly convicted at least in part because of a mistaken eyewitness. Her report will be broadcast on 60 MINUTES Sunday, March 8 (7:00-8:00 PM, ET/PT) on the CBS Television Network.

Stahl was shown a short, shaky video of a man on a rooftop simulating a crime who quickly looks at the camera twice before running out of frame. She is then shown a six-person lineup and asked to pick out the man she saw. It's a study of eyewitness memory done by Gary Wells, a professor of psychology at Iowa State University -- and like Wells' actual subjects, Stahl failed it. She picked number-five, even though the real "perpetrator" was not one of the men in the lineup. Click here to watch an excerpt. "When the real perpetrator is not in the set - is none of them - witnesses have a very difficult time being able to recognize that," says Wells. In fact, in all of the DNA exoneration cases that involved a mistaken eyewitness identification, the actual perpetrator was not in the initial lineup the witness saw.

Police will often show a witness a photographic lineup before offering a physical lineup. If the witness picks the same person out of both lineups, authorities will often tell them, bolstering the witness' confidence that he or she has picked the right person -- and reinforcing the witness' mistake if he or she has picked an innocent man. Wells analyzed the results of this kind of reinforcement in his study. After getting all of his study subjects to pick an innocent person out of the lineup (remember, the real "perpetrator" was not shown), he divided the subjects into two groups. To one group, he said nothing, then asked them questions about how certain they were of their choice. "Only about three percent are saying they could make out details of the face," says Wells.

But to the other group, he said, "Good, you picked the suspect." "Almost 45 percent of witnesses now report that they were positive or nearly positive," Wells says. This helps to explain why witnesses then go on to say on the witness stand that they were absolutely certain about their identification, even when initially they weren't -- and in a few cases, even when they later see the person who actually committed the crime, but don't recognize him. Wells says eyewitness testimony has two key properties. It's often unreliable, and yet it's highly persuasive to juries. The justice system is quite good at determining whether witnesses are telling the truth or not, "but when someone is genuinely mistaken, the legal system doesn't know how to deal with that," he tells Stahl.

To illustrate the problem of wrongful conviction through mistaken eyewitness identification, Stahl reports the classic story of rape victim Jennifer Thompson, whose identification of Ronald Cotton sent him to prison for 11 years. Despite deliberately focusing on her perpetrator's face as she was being raped, she picked the wrong person, Cotton, out of a photo lineup and then a physical lineup and finally in court - once, in fact, when the real perpetrator was right in front of her in the courtroom.

Thompson, who was devastated after DNA proved that she had made a tragic mistake, has been telling her story to prosecutors, police and defense attorneys in hopes of changing the ways in which the justice system handles eyewitness evidence.

She and Cotton, who forgave her after an emotional meeting and apology, have authored a book about their ordeals, "Picking Cotton," that they hope will advance the reforms in eyewitness evidence they believe are needed. Despite its imperfections, everyone agrees eyewitness testimony is here to stay. As Wells tells Stahl, "We need eyewitnesses....If we cannot convict based on an eyewitness, that's giving a lot of comfort to criminals....We have to find ways to make this evidence better." Reforms along the lines that Wells suggests -- showing lineup photos one at a time rather than simultaneously and having someone administer the lineup who does not know who the suspect is -- have been made in a handful of cities and states, with North Carolina, Thompson and Cotton's home state, in the lead. But in most of the country, police procedures are what they always have been, a situation Thompson and Cotton are trying to change.