ROBERT SIEGEL, HOST:
An eyewitness identification of a suspect at one time seemed to be solid evidence to take to court. But scientific studies of the vagaries of human memory and DNA exonerations have shown how often they can be wrong. The National Academy of Sciences reviewed 30 years of research into the subject and recently issued new recommendations for police and court procedures. Professor Brandon Garrett of the University of Virginia Law School took part in that review and joins us. Welcome to the program.
BRANDON GARRETT: Thank you. Thank you so much for having me.
SIEGEL: Is it fair to say that one way of summing up what the National Academy of Sciences wants from, say, a lineup or an identification from a photo array is it should be conducted as if it were a scientific experiment?
GARRETT: Yes. I think just like in an experiment, the National Academy recommends that the person administering the lineup not know the answer or not be able to convey the answer to the person taking the test. And so having the lineup be blind is one of the most important recommendations.
SIEGEL: So you wouldn't want to have the police officer running the lineup know whether it's number two or number five that's the correct answer?
GARRETT: Yeah. What the researchers have found is that people are really powerful nonverbal communicators, even through gestures or just the tone of their voice. The other aspect of it is that the eyewitness is going to be looking to the police officer for reassurance. It's stressful to be making this momentous potential identification of a suspect. They want to know that they're doing OK, and they'll be looking to the police officer for confirmation.
SIEGEL: Another recommendation deals with recording the identification of the suspect.
GARRETT: Yeah, so if you have a clear documentation of what happened, then you know - how quickly did the eyewitness make a choice? And what did they say when they were asked how confident were they? Did they act with a lot of hesitation? Traditionally, it was either written down as an identification or not, and you'd sign your name next to the photograph you picked.
SIEGEL: How many cases of exonerated defendants have you found in which there was an eyewitness identification that evidently persuaded somebody in the courtroom that it was accurate?
GARRETT: I studied the first 250 DNA exonerations in this country. And over three quarters of them involved eyewitnesses who made mistakes. And in fact, many of those cases involved multiple eyewitnesses who all made the same mistake. Many of those identifications were also cross-racial, which is also what the social scientists would predict. It's harder for people to identify faces cross-race.
SIEGEL: Now, these recommendations are addressing problems that aren't novel, or the research isn't novel. How common are best practices if we were to look from one jurisdiction to another around the country?
GARRETT: You know, there are some states that require statewide that good lineup procedures be used. But for the most part, you have police departments doing lineups the way they did 20 or 30 years ago. We have seen many more states each year have legislation introduced on this topic. Judges have become much more open to bringing in experts to explain eyewitness memory. So slowly but surely, we are seeing a real change. But it's awfully slow when you consider that this research has been quite active since the 1970s.
SIEGEL: Professor Garrett, I want you to describe a distinction that you've made in writing about the law. You've written about two different kinds of evidence or testimony that involves an eyewitness identification. One of them is the evidence that happened at the lineup or the photo array ID, which is introduced in evidence. The other one is when the witness is asked, and do you see the man who attacked you sitting in this courtroom? Yes, I do. It's that man sitting over there - the Perry Mason moment. The two moments are treated differently, I gather.
GARRETT: Well, in court - there's normally not a lineup in the courtroom, so you see the guy sitting at the defense counsel's table next to his or her lawyer. It's not a real test of the eyewitness's memory, but it's incredibly dramatic. You know, in these DNA exonerees' trials, it was often saved for the very end of the prosecutor's direct examination. You know, can you - do you see that person in the courtroom? And are you sure? So it's incredibly dramatic, but everyone knows that it's drama. What's much more important is what did this witness say when they were looking at an experiment where there were choices?
And what I found in these innocent people's cases is that often at trial they were completely confident. That's the one. That's him or his twin brother. I'm a hundred percent sure. I'm a hundred twenty percent sure even. But when they were confronted with those lineups initially at the police station, they really couldn't make a choice very easily.
And that just goes to show how confidence can really increase over time. It's easy to do it in the courtroom and look really convincing, but what the jury really needs to hear about is how sure was this person when they first looked at the images or the people in the lineup?
SIEGEL: Professor Garrett, thank you very much for talking with us today.
GARRETT: Oh, thank you so much for having me.
SIEGEL: That's Professor Brandon Garrett of the University of Virginia Law School who took part in a National Academy of Sciences review of research into memory and eyewitness identification. Transcript provided by NPR, Copyright NPR.