MS. CHRISTINA BELLANTONIFrom WAMU 88.5 at American University in Washington welcome to "The Kojo Nnamdi Show," connecting your neighborhood with the world. I'm Christina Bellantoni of the PBS News Hour sitting in for Kojo. Later in the hour, early American tourists once threatened the very sites they came to visit, cutting off pieces of the Star Spangled Banner and chipping off chunks of Plymouth Rock. We'll get a history of the souvenir industry, but first, preventing wrongful convictions.

MS. CHRISTINA BELLANTONISince 1989, 16 people have been freed from Virginia prisons thanks to DNA testing, wrongfully convicted of such serious crimes as rape and murder. And as investigators and lawyers work backwards to figure out how innocent people serve time for crimes they didn't commit, a common culprit has emerged, eyewitness identification or misidentification, which played a role in 13 of those 16 cases. Two years ago Virginia published a new model policy for eyewitness testimony designed to eliminate the subtle biases that lead to such mistakes. But a new report finds only a handful of local people departments are following it.

MS. CHRISTINA BELLANTONIJoining me to discuss is the author of that new report Brandon Garrett. He's a professor of law at the University of Virginia and author of the book "Convicting the Innocent: Where Criminal Prosecutions Go Wrong." Brandon is joining us from the phone in Charlottesville, Va. Hi, Brandon. Thank you for being here.

MR. BRANDON GARRETTHello. Thanks for having me.

BELLANTONISo let's just get right to it and talk about what happens when a witness picks out the wrong suspect, might send the wrong person behind bars. Can you explain how law enforcement procedures can end up playing a role in misidentifications like this?

GARRETTSure. I think we all know that it's really hard to remember the faces of strangers. We probably -- you know, we all -- you starts to say hi to someone and realize that's actually not my friend. That's some random person. But I think most of us figure that, okay, in a criminal case, even if you don't see the culprit for very long, even if they run away quickly, even if it's dark, police have a way to screen out the situation where the eyewitness really didn't get a good look at the face and didn't remember much. And that's by using a lineup. And we've all seen TV portrayals of lineups, people standing in a row, looking really uncomfortable.

GARRETTThat's not how lineups work today. They use photos instead. And what a lot of people don't appreciate is that the way the lineup is done can dramatically affect the accuracy of the lineup. And they can in fact cause an eyewitness to be really confident that they picked the right person when they have picked the face of an innocent person.

BELLANTONISo introducing human error perhaps can lead to this. Do you have a couple specific examples of some of these wrongful convictions that we've seen?

GARRETTSure. You know, one of the best known DNA exonerations in the country is that of a man named Ronald Cotton. He was exonerated in North Carolina. And one of the reasons it's so well known is that he wrote a book with Jennifer Thompson-Cannino and she is the victim of the rape who wrongly identified him. And she talks about how, you know, the police weren't trying to frame Ronald Cotton in any way, but they used really bad lineup procedures. They repeated him from the first procedure to the next. And that obviously signaled that he was the only person that they cared about.

GARRETTEven still she was really unsure, when she viewed the second lineup, of her identification. And immediately afterwards she was hesitating and finally went with Ronald Cotton. And the cop told her, you know, good job. That's the one we showed you before. And she describes how she just felt this wave of reassurance. She just felt so much better. She was struggling so hard to do the right thing and to bring this rapist to justice. And that feedback from the officer made her absolutely confident.

GARRETTAt the time of the trial, no hesitation at all. Ronald Cotton was the man. She later saw the person who DNA later showed was the actual perpetrator. And she describes how she has no memory of that face now. Ronald Cotton's face is in her mind, even though she knows he's innocent. Tons of social science studies have been done of lineups. Lineups are experiments and so they're easy to do experiments to replicate. And they find predictably that if you do things like repeat only one person, if you give someone feedback before or after saying good job, even if you just have someone running the lineup who knows which one is the suspect, all that has a powerful affect on the accuracy of the eyewitness identification.

BELLANTONIAnd which of those are you seeing specifically in Virginia, some of these practices that can lead to that kind of misidentification?

GARRETTWell, the most important protection is to have the person running the lineup be blind, be a police officer who just doesn't know which one is the suspect. That way there's no danger of giving suggestion or feedback. There's no possibility of it because the officer doesn't know the correct answer and so can't cue. And what the studies find is that it's really hard not to give even unintentional cues. If you're showing photos one at a time you can be saying, oh is it this one, oh is it this one? You might linger just with your voice or with your gesture over the picture that you care about.

GARRETTIt's really easy to just have someone different do the lineup, or put the photos in folders and shuffle them so that you can't see which photos the eyewitness is looking at. It's really easy to make a lineup blind. And that was the crucial reform that was introduced in these model policies in Virginia. But unfortunately, most departments haven't adopted that reform.

BELLANTONIAnd you detail a lot of that in a new report that came out just this week, and we're going to get a lot more of that in this conversation. But first I want to bring in Dana Schrad. She's the executive director of the Virginia Association of Chiefs of Police. And you raised the issue at your three-day conference that started this Sunday -- you're joining us on the phone -- at this meeting with state police chiefs now. Thanks so much, Dana Schrad, for joining us. What is your response to what this report is finding?

MS. DANA SCHRADWell, I apologize that I can't stay on the program very long. I've got a conflict, but one of the things that we have done recently is, we just completely our annual conference with the chiefs. And we reemphasized to them the need for them to adopt modern policies. I'm not making excuses for them because this is a priority for us to get them compliance with the code of Virginia that requires them to have a policy.

MS. DANA SCHRADBut let me just give you some caveats about what might be part of the problem. First of all, the code requires all the public agencies to have policies. And we have a number of private policy departments in Virginia that are not -- don't have to be compliant with this code. However, we are encouraging these practices with them as well. So that's another area that we have to reach out to and say, even though the code doesn't mandate you, it's a best practice so as a professional law enforcement agency you should adopt.

MS. DANA SCHRADSecondly, the code requires them to have a policy. Many of them have policies they just want abated to the more recent research. And so they are in fact compliant with the code. They just didn't have more modern policies. And something else that did happen while we were working Professor Garrett on the collection of policies, some of the agencies, was at the same time there was a foyer request to all the agencies from a UVA law student who was doing a separate study. And it did create some confusion. I got a lot of phone calls about it. But that's water under the bridge.

MS. DANA SCHRADWe're going to move forward with asking all the agencies, first of all, to look at the model policy that DCJS, the Department of Criminal Justice Service has done -- it's an excellent policy -- and have them adopt it. I got a call today from one of our chiefs who needed more explanation of the blindfolded policy. And so I'm getting some information to him as well. I've also told the chiefs that we're going to survey all of them this fall, either September or October, depending on when we think that we're at a point where we've got a lot of them assisted with developing their policies, so that I have a better compliance rate that I can report to the general assembly in 2014.

BELLANTONIThank you, Dana Schrad. That's very -- a lot of detail there,. And just to point out, we also have on the phone Brandon Garrett who wrote this study, taking a look at some of the things that are being implemented in Virginia. But also has been cooperating in working with law enforcement to help updated policies and has done quite a bit of that since writing the book "Convicting the Innocent: Where Criminal Prosecutions Go Wrong." So just one more question for you, Ms. Schrad, is looking at sort of the timeline. You say you're going to survey them in September or October, but how long does it take to implement such changes?

SCHRADWell, it all depends on the researches of your agency. For one thing, we have some really small agencies with some new chiefs. And we are constantly trying to provide professional training for them in both leadership and management in the adoption of policies, particularly those that are mandated through state code. So for some of the very small ones that don't understand,, for example, what options they might have that would make them be able to customize the policy to their specific needs, we're offering them that kind of technical assistance.

SCHRADSo the larger agencies have the resources to be able to do it without much change from the DCJS model policy. They've got plenty of officers. They don't have any problem bringing in an officer who's unfamiliar with the investigation. When you get into smaller towns where everyone knows everybody, where everyone in the agency knows everybody and also you have some small departments that don't have the resources to do criminal investigations, in some of those cases they actually might turn their investigation over either to the state police or the sheriff's office .

SCHRADBut they still need to have a policy in place in those cases that they have those directives for their officers so they know what to do if they're not themselves going to do the lineup for their own agency.

BELLANTONIBrandon Garrett, you're a professor of law at the University of Virginia and author of "Convicting the Innocent: Where Criminal Prosecutions Go Wrong." So back in 2011 you worked with Virginia to revive this model policy we've been talking about. What changes did you recommend and how did you see police departments move forward? What kind of adoption rate have you seen?

GARRETTWell, Virginia now has really a top of the line model policy. I think it's the best in the country. It's really comprehensive. A lot of people might not have heard of the Department of Criminal Justice Services, but it's the criminal justice policy agency in Virginia. And they reached out to law professors like me, to social scientists, to police chiefs around the state and drafted something that was really clear and laid out, like Dana Schrad was saying, options for all different type of departments.

GARRETTIf it's a small department they can use the folder system if they can't bring in another detective to do the lineup. An idea was to also just make the instructions that you give to an eyewitness really clear, lay them out in detail and also explain why some of the stuff is important. The reason why that model policy came about is that we've had so many of these DNA exonerations in Virginia that involved eyewitness mistakes. And there have been even more of them.

GARRETTDating back to 2005, there was a law, like Dana mentioned, that says departments have to have something in writing on lineups. And actually I think a lot of people might be surprised that something as important as lineups, traditionally departments didn't have anything in writing. Around the country that was the case. Departments have a lot of rules about use of force, what happens if a gun goes off, searches and seizures because they can get sued over that stuff. If a lineup goes wrong, it might be years before anyone realizes that someone innocent was convicted. And it's pretty unusual for DNA to be available to test. And I think it was just kind of a neglected subject.

GARRETTIn 2010 it came to our attention, because of good work by Dana and the crime commission, that a lot of department really weren't adopting the best practices. And so the idea was, okay, you know, Delegate Alexander proposed legislation. Why not just require them to have best practices? That's what North Carolina has done. That's what West Virginia did last year. And the response was, no, no, no. We can trust departments to do it on their own. Why wouldn't they? It doesn't really cost anything to do a lineup right.

GARRETTAnd so the approach back in 2010 was provide an even better model policy that's really detailed, roll out training on it, provide every resource so that department know how to do a lineup right to make sure that wrongful convictions don't happen, to make sure that criminal cases don't fall apart over eyewitnesses or are encouraged to guess by bad policies. And we just sort of trusted that that would work. I wanted to follow up this year to find out whether that voluntary approach was working. And I was really disappointed to find out that it hasn't been.

BELLANTONISo Dana Schrad, go ahead. We know you have to get back to your meeting but, Dana Schrad, we'd love to have you weigh in. And also just note if you're worried about the general assembly getting involved.

SCHRADWell, we're -- and let me -- I just want to back into Professor Garrett's comments there to tell you a couple of other things we're doing. First of all, our budgets have been really stripped in terms of being able to provide a lot of training opportunities. And so we are looking forward to utilizing some of the folks who have come forward in offering some free training for us, if we can actually get our folks to be able to go because their travel funds have absolutely been stripped.

SCHRADSo we're looking at other alternatives for how we can offer this training, whether it's through webinars or some kind of online training, if that's a possibility. And I'm working with the Department of Criminal Justice Services on that. The other thing that we are doing is many of our agencies are involved in our state accreditation program, which is also a voluntary program but it's what we term a program that helps to develop good management practices. And part of that accreditation process is to review all policies.

SCHRADAnd our chiefs who are on the accreditation commission are making sure that the updated policy is part of the mandate to be able to receive accreditation. So that's another step forward we're trying to take to close this gap on compliance.

BELLANTONIVery interesting. And about the general assembly and whether they get involved?

SCHRADWell, certainly we're always very much concerned I think. One of the things that we want to avoid is the general assembly putting detailed instructions into state code as far as policy goes. It's not a good management practice for the general assembly to be operationalizing law enforcement instructions through state code.

SCHRADSo we much prefer that they -- it's fine that they have a policy mandate. Leave the particulars of the policy to the agencies because otherwise when we have a change in best practice then we will find our people out of code compliance. And we can't move the code changes as quickly as we can move policy changes. So it's just a better practice to have the general policy mandate in the code and let the particulars of the policy be changed as best practices change.

BELLANTONIDana Schrad, the executive director of the Virginia Association of Chiefs of Police. Thank you for taking a moment away from your meeting to join us. We appreciate it.

SCHRADThank you.

BELLANTONISo Brandon Garrett , why do you think some of these police departments are taking so long? I mean, is it a matter of technology and training resources?

GARRETTI'm not entirely sure. You know, Dana and I put on a conference at UVA law school in the spring to bring in police chiefs and policymakers and police officers from around the state to talk about lineups, talk about why it was important. And we heard from chiefs in Roanoke and Charlottesville who have had great policies for a long time. And they talked about how successful it's been for them and how, you know, even putting to one side the danger of these horrible wrongful convictions.

GARRETTIt happens all the time that eyewitnesses who are prone to guessing pick out fillers. And that really can discredit them. You know, if you know that they're picking out innocent people, the fillers, the ones that -- the extra pictures that you put in the photo, right, then you know, a jury might not be willing to trust that person if they eventually identify someone who they say is the culprit. And a lot of criminal cases go south because of just guessing by eyewitnesses. These same improvements can also cut down on filler identifications.

GARRETTYou know, we still hear some small departments saying we don't know how to do this. But the model policy really clearly lays out how you use folders. You know, a simple option for small departments. It may just be that a lot of them just haven't read the model policy, haven't looked at it carefully. I wonder though -- what I was seeing, a lot of the departments that had policies that seemed really out of date were policies that really were out of date. And I wonder whether those departments just simply haven't updated their policies in general for a really long time.

GARRETTAnd like Dana was saying, you know, there are departments that are accredited but most of them aren't. And it may be that you have new chiefs in small departments and they just simply haven't looked into lineups. But they haven't looking into a lot of other things either.

BELLANTONISo getting at that, when you talk about small departments, I mean, are we talking about smaller, more rural forces that are downstate or is this happening in Northern Virginia? Where are you seeing specifically in each region?

GARRETTWell, I agreed that I wasn't going to, you know, name particular departments. I'm keeping all the names of the policies -- the departments that drafted them confidential.

BELLANTONIHow about just a general region, though?

GARRETTBut in general, yeah, I didn't see much of a pattern. Plenty of big departments had outdated policies. Plenty of small ones didn't. A lot of the departments that had no policy were sheriff's departments that probably don't do criminal investigations. They should really just have a placeholder that, look, if a lineup situation comes up we refer it to the police. But there are some police departments that didn't have policies at all.

GARRETTYou know, a remarkable number, dozens of departments had these policies that dated back to 1993 that had no detail at all. It just sort of says you can do lineups, you can show pictures or you could just show a single person to the eyewitness. Those are the three options with really no explanation for how in the world they do it at all, ancient policies. And there really wasn't too much rhyme or reason for it. There are plenty of real big jurisdictions that had outdated policies and plenty of small ones that had quite good policies.

BELLANTONIAnd in fact, Virginia law enforcement has been criticized for lack of transparency in the past. And how much does that factor into this?

GARRETTWell, you know, certainly the students -- Dana mentioned that some students at UVA law school took on the project of sending out foyer requests to every police agency in Virginia, which was a big job. But there are plenty of departments that didn't respond to the foyers or that said -- or that refused to -- they responded but refused to provide their lineup policies. I found that surprising. You'd think that just the basic policies of a police department should be public record. So I was surprised that so many departments didn't respond or said that they wouldn’t provide them.

GARRETTBut that said, most departments seemed to respond and provide their policies. And, you know, a lot of the policies were also really, you know, either copied from this ancient 1993 model policy or copied from a 2005 model policy that had some problems and was confusing. And just a tiny number of departments had adopted the state-of-the-art new 2011 policy. It seemed like though for the most part, police departments weren't coming up with their own ways of doing lineups. They want to do the accepted way. It's just that a lot of them aren't up to date with what the accepted way of doing lineups is.

BELLANTONIThank you very much, Brandon Garrett, for explaining this for our listeners. You are a professor of law at the University of Virginia and author of the book "Convicting the Innocent: Where Criminal Prosecutions Go Wrong." We really appreciate you joining us. And thanks again to Dana Schrad with the Virginia Association of Chiefs of Police who was on the phone with us as well. We'll be back after a short break for a conversation about the Smithsonian's Souvenir Nation. Stay tuned.