MADELEINE BRAND, host:
From NPR News, it's Day to Day. You know, it's been know for decades that eyewitness testimony in legal cases is unreliable. Innocent people have been put away for years based on it. There is new research that shows just how shoddy eyewitness testimony can be. Slate's legal expert Dahlia Lithwick is here now with more. And Dahlia, there have been countless cases of these eyewitness testimony identifications gone awry. Tell us about the case of Jennifer Thompson.
DAHLIA LITHWICK: Sure. She was raped when she was 22. And she went out of her way to study her attacker's face because she wanted to be able to help the police catch him someday. She later identified Ronald Cotton in a photo ID lineup, and then again in a physical lineup. The detective who was in charge of the lineup said, hey, you pick the same guy you picked out in the photo ID, for which Ronald Cotton served 10 years in jail. He was later exonerated by DNA evidence. It turned out the actual rapist was someone called Bobby Poole, who looked a lot like Ronald Cotton. This turned into yet another case of a DNA exoneration that showed that eyewitness testimony, eyewitness identification is just not reliable and needs to be rethought.
BRAND: So, how does that happen? How is it that someone who, like Jennifer Thompson, was convinced that she had the right guy, how could she be so mistaken?
Ms. LITHWICK: Well, science is starting to prove - and again, the whole field of eyewitness identification, the science of it, is quite new. Much of the really interesting research has happened since this 1977 Supreme Court case. And the research shows that our eyes simply deceive us, that our brains do a lot of work manipulating our memories and reframing our memories. One of the things that's interesting about that particular case is that long after Jennifer Thompson knew that Ronald Cotton wasn't her actual rapist, that Bobby Poole had done it, she still could close her eyes and see Ronald Cotton, in her head, committing the rape.
BRAND: So, this has been said for many, many years, decades in fact, by legal experts and by people who observed this, and yet nothing has changed. Why?
Ms. LITHWICK: Well, the Supreme Court hasn't heard a case on this in over 30 years. And Gary Wells, who's a professor who's been studying this for decades, says that what happened is in 1977, when the court last looked at this, they simply relied on old science about the underpinnings of the reliability of eyewitness identification. And because they haven't revisited it, police departments haven't brought their practices up to a modern standard. He says that eyewitness ID evidence should be treated more like trace evidence, like blood evidence or fingerprint evidence that can be corrupted and that can be manipulated over time. And so his point is, we really need to rethink the science underpinning the 30-year-old science that we've relied on for a long time.
BRAND: In that study that you mentioned by Gary Wells, he notes that of the more than 230 people that were wrongly convicted in the U.S. and exonerated by the DNA evidence, almost three quarters of them - or more than three quarters of them - involved cases of mistaken eyewitness identification.
Ms. LITHWICK: That's exactly right. Seventy-seven percent are the result of mistakes in eyewitness ID. And what he says is that's the single largest factor now that's contributing to wrongful incarceration, wrongful conviction, it's something we really need to reassess going forward. Because every time we put an innocent guy in jail, of course, we're letting a guilty guy go free.
BRAND: Dahlia Lithwick is senior editor for Slate.com and the legal analyst there. Thank you, Dahlia.
Ms. LITHWICK: It's always a pleasure.
BRAND: Stay with us, Day to Day returns in a moment.
NPR transcripts are created on a rush deadline by a contractor for NPR, and accuracy and availability may vary. This text may not be in its final form and may be updated or revised in the future. Please be aware that the authoritative record of NPR's programming is the audio.