In a court room, few pieces of evidence carry as much power eyewitness testimonies. Yet decades of research in social science and psychology– not to mention a growing list of exonerated inmates– indicate that witness recollections can be strongly biased by flawed law enforcement procedure. We explore why institutions have lagged behind social science, and new developments in state and federal courts.


  • Brandon Garrett Roy L. and Rosamund Woodruff Morgan Professor of Law at the University of Virginia. Author of "Convicting the Innocent: Where Criminal Prosecutions Go Wrong" (Harvard University Press, 2011).
  • Steven Penrod Distinguished professor of psychology, John Jay College of Criminal Justice
  • Dana Schrad Executive Director, Virginia Association of Chiefs of Police

Eyewitness Testimony Policies in D.C., Maryland, and Virginia

Rebecca Brown of The Innocence Project outlined some of the policy differences between D.C., Maryland, and Virginia regarding eyewitness testimony. The Innocence Project argues that one of the most important policies law enforcement officials should implement is that of “blind administration,” which means that the officer conducting the ID procedure is not aware of who the suspect in the case is. They also argue that it is crucial that the witness in the case is told that the officer doesn’t know who the suspect is.

1. D.C. does not have any policies, written or otherwise, on eyewitness identification that comport with best practices. Those best practices are blind administration, confidence statements, instructing the witness, making sure that the non-suspect photos or lineup members match the description provided by the eyewitness and not the suspect. So, DC has nothing along those lines.

  1. Maryland passed a law a few years ago that required written policies of all law enforcement agencies in the state and required that those written policies comply with Department of Justice guidelines that were devised back in 1999 which are quite strong, but fall short in terms of a complete endorsement of blind administration. They sort of indicate that further study might be accomplished, but what we did learn was that of course over the intervening years there’s been uniform support by the scientific community for blind administration of lineups. So, there are some uneven policies in Maryland as far as we understand because all that was required was that there be written policies that minimally comport with the Department of Justice guidelines. So that may be something that will be revisited in the coming months or years.

  2. In Virginia, they also have a similar law on the books. Policies are required, sequential presentation seems to be used uniformly, but blind administration is not used everywhere. And that’s a huge problem, because it’s worse than the traditional method of non-blind simultaneous lineups because a sequential, non-blind lineup is quite suggestive. And so, this past session the Virginia legislature compelled the Department, or the Division of Criminal Justice Services in Virginia to craft a new model policy to kind of address these issues, and they are in the midst of doing that.

Related Video

The Innocence Project’s video explores some of the procedural problems involved in the process of eyewitness identification:


  • 13:06:43

    MR. KOJO NNAMDIFrom WAMU 88.5 at American University in Washington, welcome to "The Kojo Nnamdi Show," connecting your neighborhood with the world. It's something in human nature. There's almost nothing as convincing to a jury as an eyewitness pointing the finger and identifying a suspect. And yet, our memories are far from infallible. Eyewitnesses in criminal cases are wrong for a third of the time.

  • 13:07:18

    MR. KOJO NNAMDIAnd of the first 250 convictions overturned by DNA evidence, 75 percent of those were based on eyewitness testimony. Studies over the past few decades have gone a long way to explaining why and how memory can be influenced and what can be done in police departments and courtrooms to reduce misidentification. But the law has been slow to catch up to the research. A landmark case in New Jersey recently took up the issue and set new rules for courtrooms and police departments.

  • 13:07:49

    MR. KOJO NNAMDIAnd the Supreme Court takes up one aspect of eyewitness identification in November. Joining us to discuss this contentious issue is Brandon Garrett. He's a professor of law at the University of Virginia. His new book is "Convicting the Innocent: Where Criminal Prosecutions Go Wrong." Brandon Garrett joins us from studios at the University of Virginia. Thank you very much for joining us.

  • 13:08:13

    MR. BRANDON GARRETTThank you for having me.

  • 13:08:14

    NNAMDIYou, too, can join this conversation by calling 800-433-8850, 800-433-8850 or you can go to our website, Have you ever been an eyewitness in a criminal case? How reliable do you think memory is? How about in a stressful situation or on a dark night? 800-433-8850. Brandon Garrett, first, we'd like to follow-up on a case that we discussed on the air last week. Troy Davis who was death row in Georgia was executed last night after the Georgia Board of Pardons rejected a plea and after an 11th hour temporary stay by the Supreme Court. What happened? And what exactly was the Supreme Court's role last night?

  • 13:08:59

    GARRETTWell, last night, the Supreme Court apparently was considering the case and then after a couple of hours decided not to issue a stay. You know, what I was most disturbed at from the perspective of someone who cares about getting eyewitness memory right, getting eyewitness procedures right, was that the Georgia Board of Pardons had a psychologist, one of the leading experts on eyewitness memory there at the hearings on Monday and decided not to let her, Jennifer Dysart, speak to them. And I think this is -- Troy Davis' case is a case about a lot of things. But one of the important features of the case was that there are a number of eyewitnesses. And his team says they got it wrong.

  • 13:09:40

    NNAMDIAs you mentioned, the Troy Davis case rested on eyewitness testimony. Tell us about that. We kept hearing last night that seven of nine eyewitnesses in that case had either recanted or changed their testimony.

  • 13:09:55

    GARRETTSure. And there are also jailhouse informants in the case, really disturbing case. What bothered me the most, looking back at the original trial materials where you have these eyewitnesses describing how they came to identify Troy Davis is that they weren't shown a lineup or a photo array right after the crime. The police prepared the photo array right away. They had a photo of Troy Davis. They displayed it all over the neighborhood in wanted posters. It was played on the media all around Savannah, Georgia at the time.

  • 13:10:28

    GARRETTAnd they waited and they waited to show these eyewitnesses the photo array. If there's one thing we know about eyewitness memory is that it fades. So, why did they wait? Well, they took four of the key witnesses back to the crime scene, this Burger King parking lot where the shooting took place. And they did this unusual, really a regular reenactment, where they had them work out with each other, where did everyone stand.

  • 13:10:51

    GARRETTYou know, a disturbing effort maybe to get them to get their stories right. And certainly their stories became more consistent after the police staged this reenactment. And it was only after that that many of these eyewitnesses ever saw the photo array. And so maybe no surprise that by the time of trial, after seeing his image in the wanted postings, after the reenactment that they can say, oh yeah, it was Troy Davis.

  • 13:11:13

    NNAMDIThe rules in Georgia in terms of the procedures have changed since this case was prosecuted 20 years ago. Do you think the rules in place today would likely have changed the outcome?

  • 13:11:26

    GARRETTWell, certainly Georgia, like many other states, has learned some important lessons from these DNA exonerations. There are a number of really disturbing cases in Georgia, where eyewitnesses got it wrong and not because of, you know, more suggestive conduct like that reenactment, that staging in the Troy Davis case but just regular lineups, where maybe even unconscious cues from police they have led an eyewitness astray.

  • 13:11:50

    GARRETTAnd so, Georgia has redone its policies and training. I think that certainly would have made a big difference in a case like Troy Davis' and hopefully those procedures will help to prevent future mistakes from happening. A lot of states haven't made those changes.

  • 13:12:03

    NNAMDII was about to say, could a case like Troy Davis happen today in other states of jurisdictions?

  • 13:12:09

    GARRETTSure, you know, more and more jurisdictions have been rewriting the rules. But for a long time, there are just no rules at all, nothing in writing about how to do a lineup. Lineups are memory tests. But if the test is done wrong, you can change someone's memory. You can fail to preserve someone's memory or you can have a guilty person go free, which happens a lot when eyewitnesses keep picking fillers because of bad lineups.

  • 13:12:31

    GARRETTSo, you know, getting lineups right is really important for the police. But traditionally, there are really no rules. And you still have a lot of departments around the country that don't have the best practices in place. You know, New Jersey has done it. North Carolina and Ohio have legislation. Other states like Georgia and Virginia now are sort of writing the rules within law enforcement. But there's a lot of work to be done.

  • 13:12:53

    NNAMDIWe're talking about eyewitness IDs with Brandon Garrett. He's a professor of law at the University of Virginia. His new book is called, "Convicting the Innocent: Where Criminal Prosecutions Go Wrong." Joining us now by phone from Dobbs Ferry, NY is Steven Penrod. He is a distinguished professor of psychology at the John Jay College of Criminal Justice. Steven Penrod, thank you for joining us.

  • 13:13:14

    MR. STEVEN PENRODMy pleasure.

  • 13:13:15

    NNAMDIAlso joining us by phone from Henrico, Va. is Dana Schrad, executive director of the Virginia Association of Chiefs of Police. Dana Schrad, thank you for joining us.

  • 13:13:26

    MS. DANA SCHRADIt's my pleasure.

  • 13:13:26

    NNAMDISteve Penrod, what do we know now that we did not know 30years about the reliability of eyewitness identification?

  • 13:13:36

    PENRODA great many more things than I could tell you in a matter of a few minutes. I suppose among the most important are that it's now clear that lots and lots of witnesses make mistakes when attempting identification. Excellent data out of the United Kingdom, a new study just published or released this week looking at over 800 identifications here in the United States. One of the things that's clear from these studies of actual witnesses is that 1 out of 3 of the positive identifications that are made by witnesses are clearly wrong, 1 out of 3.

  • 13:14:16

    PENRODWe know this because the witnesses are picking fillers or foils, the innocent individuals that have been put into photo arrays or live lineups. So, at the outset, we know that lots of witnesses are prone to make errors. They appear to be prone to make probably good faith guesses, but nonetheless guesses that are wrong. These numbers also raise the question of just how often the police actually have the right person in hand.

  • 13:14:49

    PENRODA lot of eyewitnesses will reject lineups. It's unclear from these studies whether that means that the witness has sufficiently poor memory that they're not willing to make an identification or the police have got it wrong. They've got an innocent person in the array and the witnesses correctly stating they don't see anybody that they recognize. Now, we, those of us in the research community, run experimental studies where we can gauge whether a person gets it right or wrong because we set it up knowing who the target individual or perpetrator.

  • 13:15:28

    PENRODAnd in the experimental studies, we likewise see a great many errors by witnesses and a great many guesses, many of them incorrect. And occasionally, some of them correct.

  • 13:15:39

    NNAMDIDana Schrad, what are some of the typical procedural factors that can lead to misidentification?

  • 13:15:47

    SCHRADWell, I think in many cases it depends on the timing of when the police lineup, whether it's a photo lineup or a physical lineup is held and the proximity and time between the event and when you actually are able to bring an eyewitness in and can conduct the lineup. And it takes time. And depending on the type of case we're talking about, to develop a suspect lineup and then bring that individual in. If the individual's under a great deal of stress, they may feel that they cannot, you know, participate in a lineup at that time.

  • 13:16:19

    SCHRADIt's so much psychology, you know, involved with photo lineups and relying on that as a key piece of evidence in the development of an investigation that it's so critical that the timing of a lineup be optimal that you have good materials, whether it's a photo lineup and you've got good materials that sort of neutralize the process so that you have a good array of photos for the witness to look at and do the elimination of the person that he or she does not believe is the suspect. And then also just dealing with, in some cases, being able to facilitate optimal conditions such as a double blind presentation.

  • 13:17:09

    NNAMDIAnd exactly what is a double blind presentation?

  • 13:17:12

    SCHRADMaking sure that you have police personnel there that are administering the lineup. That have not been involved in the case because people pick up obviously on nuances of inflections in your voice or any kind of words you might say or how you might structure or how you even, you know, relate to the witness when that person is trying to through and eliminate, you know, the suspect from the fillers, as they're called.

  • 13:17:43

    SCHRADAnd so, having an officer in there who would not know who the suspect is helps to sort of neutralize that process. The problem we went into in law enforcement is the vast majority of our agencies in Virginia are not the large agencies, and this is true across the country. It's very small agencies where it can be operationally a challenge to have a second officer there to be that double blind officer to administer the lineup.

  • 13:18:11

    SCHRADSecondly, in small agencies in rural areas where you may not have lineups very frequently, keeping people actively trained up to speed on proper procedures to conduct a lineup in the best professional way means you have that repetition of activity and we don't have that very often in areas we don't have serious crimes that require an eyewitness identification.

  • 13:18:38

    NNAMDIBrandon Garrett, in your book, you examined the first 250 cases in which DNA evidence overturned convictions. What did you find?

  • 13:18:50

    GARRETTWell, I found that those cases involved a lot of eyewitness mistakes. About a third of the cases where there were eyewitnesses involved multiple eyewitnesses. And people might think, you know, in Troy Davis' case where we have a whole group of eyewitnesses. It's one thing if one eyewitness is uncertain. But if you have a whole group of them all identifying someone, that seems much more powerful, right?

  • 13:19:13

    GARRETTBut we know that multiple eyewitnesses can get it wrong, maybe because they experience the same flawed lineups as each other. Or if they communicated or even saw the photo arrays together, which is really inappropriate. But what I saw was that those eyewitnesses, when they got to the courtroom, they were absolutely sure that they were identifying a guilty person. They said, you know, it's either him or it's his twin brother. I'm 100 percent sure. I'm 120 percent sure.

  • 13:19:38

    GARRETTThat image is burned into my mind like a photograph. They described seeing the image in their mind like a photograph. And we now know they were wrong, which makes that testimony that seemed so powerful to the jury all the more troubling in retrospect. And what was also interesting, just one quick note, is that those eyewitnesses remembered having not been so sure when they first looked at the lineups.

  • 13:19:59

    NNAMDIWhile we know of cases that were overturned with DNA evidence, Brandon, do we know how many cases were confirmed by DNA evidence?

  • 13:20:10

    GARRETTNo. Although we know that DNA testing is sometimes done and plenty of the time it confirms that the person convicted was the right one. We also know that DNA clears many thousands of people a year, long before a conviction which is good. You know, during criminal investigations, a suspect is identified and the DNA testing clears them.

  • 13:20:30

    GARRETTMaybe it shows that the eyewitness was wrong. But the case never goes to a trial. And that's a wonderful thing except very few criminal cases have usable DNA. DNA can't provide a solution in most criminal cases. So that's why it's so important to get the lineups right.

  • 13:20:45

    NNAMDIIn the...

  • 13:20:45

    GARRETTOne little note on lineups, by the way. Small departments can do a double blind lineups, easily. All they need is six manila folders. You put the photos in folders, you shuffle them, you hand them to the eyewitnesses. The eyewitness can open up the folders without the police officer being able to see what photo they're looking at. So as long as a small department can afford six manila folders, they can do double blind.

  • 13:21:05

    NNAMDIWhat do you say to that Dana Schrad?

  • 13:21:07

    SCHRADYes, that is one of the options. I think another option that we've seen used in areas is task forces, sort of, put together where you would have officers trained in neighboring agencies and if you want to bring another officer in to facilitate the lineup procedure, for you, that's not involved with the case at all and it can serve as your neutral double blind officers, so to speak, then that helps as well.

  • 13:21:31

    SCHRADAnd, yes, that is absolutely an option. I mean, on one of the other areas now is, it's electronic. You put them on a laptop and you set the person in front of a laptop and the officer may not even be, you know, physically in the room with them. But all of these things are -- there's so many factors that go into setting up, sort of, the best possible scenario for assisting people through an identification procedure and recognizing that you have all kinds of people.

  • 13:21:54

    SCHRADPeople who are young, old, children, people who are under a great deal of stress. You have a great deal of variety of crimes that you're talking about using eyewitnesses for. And I think that police are challenged to determine, in some cases, well what would be the best scenario in this particular case to make this witness comfortable and to make this witness feel not threatened by the process and comfortable enough to make a clear minded identification of a possible suspect.

  • 13:22:20

    NNAMDIGot to take a short break. When we come back, we will continue this conversation on eyewitness IDs in the wake of the execution last night of Troy Davis in the state of Georgia. If you have a comment or a question, you can call us at 800-433-8850. Why do you think eyewitness testimony is so powerful in a courtroom even though we know how fallible memory can be, 800-433-8850? I'm Kojo Nnamdi.

  • 13:24:47

    NNAMDIWe're discussing eyewitness IDs with Dana Schrad, executive director of the Virginia Association of Chiefs of Police. She joins us by phone from Henrico, Va. Steven Penrod is a distinguished professor of psychology at the John Jay College of Criminal Justice. He joins us by phone from Dobbs Ferry, N.Y. And Brandon Garrett is a professor of law at the University of Virginia. His new book is called "Convicting the Innocent: Where Criminal Prosecutions Go Wrong." He joins us from studios at the University of Virginia.

  • 13:25:15

    NNAMDIWe're taking your phone calls at 800-433-8850. Have you ever been an eyewitness in a criminal case? You can also join us at our website or send a tweet to kojoshow. Steven Penrod, Juries tend to over believe eyewitnesses, why are eyewitnesses so convincing to a jury?

  • 13:25:36

    PENRODWell, I think, one of the reasons is that most of us, in our everyday experience, don't have many opportunities or situations where we clearly misidentify another individual. Most of us probably have had the experience of walking down the street, seeing someone a block or two away and thinking it's someone that we know and later realized that it's not. Even in that situation, we get it right once we get close enough to the person to see who they are.

  • 13:26:08

    PENRODSo I think that jurors and looking at eyewitnesses who present themselves as having made a correct identification are not -- don't have a background of experience that allows them to be as skeptical as they probably ought to be about identifications of strangers under most crime-like situations.

  • 13:26:30

    NNAMDIHere is...

  • 13:26:30

    PENRODThe other...

  • 13:26:31

    NNAMDIGo ahead, please.

  • 13:26:31

    PENROD...I was going to say, the other feature that I think is problematic for jurors is that witnesses, as Brandon pointed out, frequently come in and express high levels of confidence about their identifications. Now, there's a growing body of research that indicates that witnesses, for a variety of reasons, can grow more confident as time passes after they've made an identification.

  • 13:26:53

    PENRODSo that by the time they get to the courtroom, either because they've told their story over and over again or they've made repeated identifications of the same individual, their presentation is really very confident. And I think it's difficult for jurors to step away from that confidence and think about the circumstances under which the original identification was made and what sort of factors may have contributed to errors when that identification was made.

  • 13:27:20

    NNAMDIWhen you say what sort of factors, what are considered suggestive factors that can influence someone's identification of a suspect?

  • 13:27:28

    PENRODWell, I would first point to problems that undermine correct memory or formation of a solid memory that would allow a person to make a correct identification. Factors such as stress, that I think you've mentioned, presence of a weapon, relatively brief opportunity to view, the lighting conditions, the actual presence of multiple perpetrators, cross race identification are among the many factors that have been subject to investigation, that we know undermine the quality of memory.

  • 13:28:04

    PENRODNow, couple that with, as Dana pointed out, the passage of time, the loss of information from memory is most rapid in the first minutes and hours following exposure to a face. So time is very critical, the passage of time. In terms of police procedures, there are several things that can be problematic but they don't have to be.

  • 13:28:28

    PENRODWitnesses should certainly be strongly admonished that if they're being presented an identification parade -- a photo array or a life lineup, it should be admonished to remember that the perpetrator may not be there. This proves to be very important in minimizing guessing by witnesses. I think it's also important that, even in blind situations where the officer is blind, that the witness know that the officer is blind. So that there's no tendency on the part of the witness to look to the officer for guidance about who the suspect might be.

  • 13:29:06


  • 13:29:06


  • 13:29:07

    NNAMDIGo ahead, please.

  • 13:29:07

    PENROD...opposition of a raise is very important they need to be fair and ideally, I think, one of the big problems in the United States is that we use relatively small arrays, tends to be six person arrays and other countries of the world, there's often nine or 10 people. And that's important because those spoils draw away bad guesses by witnesses who are inclined to make guesses and afford significant additional protection to an innocent suspect.

  • 13:29:36

    NNAMDIWell, before I go to the phone, Dana Schrad, police procedure is clearly the first piece of this. What actually happens when an eyewitness comes forward? Most of us only know this from television crime shows. Can you explain some of the basic procedures police use with eyewitnesses?

  • 13:29:54

    SCHRADWell, it's -- I guess, it really does vary from case to case, from agency to agency. It depends on whether or not it's a case in which, as it's been mentioned, there's multiple suspects, there's multiple witnesses. So I think you really have to treat each case uniquely. However there are still standards of conduct in terms of conducting eyewitness investigation -- eyewitness interviews and then presenting them with photo or physical lineups to help them narrow down and hopefully perfect their memory in terms of identifying a potential suspect.

  • 13:30:28

    SCHRADBut I think, in many cases, the -- if you have other independent evidence that will -- that steps -- that helps you support your case, particularly involving a particular -- a suspect, that those are key pieces of information and key pieces of evidence that are integral to the investigation. And when you have that larger array of evidence to help you build your case, then that has an impact on just exactly how important eyewitness identification is.

  • 13:31:01

    SCHRADIf that's the case, then you already have other evidence. You have a gun with prints on it that you can scientifically test. If you have, you know, other DNA evidence at the scene. All those kind of things, then mean that the actual eyewitness identification doesn't become as pivotal in the entire investigation as does that scientific evidence. So then that has an impact on how much the officer might rely on the eyewitness identification.

  • 13:31:28

    SCHRADAnd it also, I think, another reason it varies is, you get into a small town and let's face it, not every crime involving eyewitness identification involves a stranger. And it may very -- in many cases, it's someone that you -- that the witness does know and has seen this person before in the community and so has no trouble identifying that individual. So I think it's very difficult to say that there's, sort of, a one size fits all approach although the key thing is, to neutralize the process, as much as possible to reduce the stress on the witness and to have as timely an identification as you can to...

  • 13:32:08

    NNAMDIWell, from...

  • 13:32:09 the process.

  • 13:32:10

    NNAMDI...from the police to the court. Here is Lisa in Washington, D.C. Lisa, you're on the air, thank you for waiting. Go ahead, please.

  • 13:32:17

    LISAYes, thank you. My question is, how can eyewitness testimony put an innocent man to death? For example, if I and a group of five other people say that you, Kojo, killed a victim in a dark alley at night when you were no near the scene of the crime, could we put you on trial, our testimony, and have you condemned to death?

  • 13:32:36

    NNAMDIBrandon Garrett.

  • 13:32:39

    GARRETTWell, we just had an execution, right? And we had a -- there's a case in Maryland too, a well known case, the Kirk Bloodsworth case. The first person who is freed by DNA, off of death row, in the United States. And that was a case with five eyewitnesses. Some of the witnesses were a little shaky, even at trial. In fact, Bloodsworth tried to bring in a psychologist like Steve Penrod to testify a trial and explain some of the problems of eyewitness memory to the jury.

  • 13:33:11

    GARRETTThe Maryland supreme court said, no, no, no. That would not help the jury. That decision is still a good law in Maryland, but Maryland has done something different since then. Maybe in part in reaction to the Bloodsworth debacle, they now limit the death penalty so that a case based solely on eyewitness testimony cannot be eligible for the death penalty.

  • 13:33:29

    NNAMDISteve Penrod, Steve...

  • 13:33:30

    PENRODYou know, 35 years ago, in the...

  • 13:33:32

    NNAMDI...I was about to say, Steve, researchers have been doing studies for decades on the unreliability of eyewitnesses. Why do you think the law has taken so long to catch up to the science?

  • 13:33:43

    PENRODWell, I think, law, unlike medicine, relies on people outside the law to do research. You know, medicine moves along fairly briskly in adopting new practices, more effective techniques. And I think that in part, it's because people in medicine are trained in science and many are trained to be doing the research, in a sense, in house. That's not been so true with policing.

  • 13:34:08

    PENRODSo 99 percent of the research on eyewitness reliability has been conducted by psychologists outside the system. And moving the products of that research into the consciousness of law enforcement and the courts is not that easy. There aren't courses on this in law school, there are increasingly courses for police, in their training, that draw up on the signs. But that's been fairly recent in terms of adoption of scientific perspective on identification.

  • 13:34:49

    PENRODIn addition in general, the law is a rather more abund -- enterprise. It relies on president, it is slow to adopt new practices, new procedures, new ways of thinking. And one can mount a defense for that but it does mean that's in areas where there might be fast moving research, it's not necessarily going to be adopted with any degree of alacrity.

  • 13:35:14

    NNAMDIHere is...

  • 13:35:14

    SCHRADAnd that's...

  • 13:35:15

    NNAMDIOh, please, go.

  • 13:35:15

    SCHRAD...that's a really important point when you think about just the history of our legal system in this country and you go back, not that long ago, eyewitness testimony was not only the most reliable, it might be the only kind of evidence you would have. So the introduction of scientific evidence that we now have through DNA and printing and ballistic evidence and all the other kinds of things we now have available to us to verify evidence in a case, is relatively new in our legal system of -- which is 400 years old.

  • 13:35:42

    NNAMDIHere is Arman (sp?) in Chevy Chase, Md. Arman, you're on the air. Go ahead, please.

  • 13:35:47

    ARMANGood afternoon, can you hear me okay?

  • 13:35:49

    NNAMDIYes, we can.

  • 13:35:51

    ARMANThen I wanted to make a point which draws on some of the scientific literature around face perception. And it struck me that it might be relevant of this issue of witness identification. And a kind of well recognized phenomenon is the race effect in which people from a given race have greater difficulty discriminating individuals from the race that they've not been exposed to in their sort of upbringing.

  • 13:36:14

    ARMANSo a white individual is trying to discriminate a Caucasian, you know, the kind of popular folk line of they all look the same. And I was just wondering, to what extent, that phenomenon causes problems for the way witness identification is conducted with specific respect to the racial profile of witnesses as opposed to the racial profile of the population of people they're trying to distinguish?

  • 13:36:39

    NNAMDIAny research on that Steven Penrod?

  • 13:36:41

    PENRODYes. We now have, as I recall, one of the first studies appeared in the late 1960s. So we have a long trail of research studies, scores of such studies and they do indeed confirm the notion that cross-race identifications are more difficult then same race. And it's a problem both with respect increased errors being made when the target individual or perpetrator is not there. And fewer correct identifications being made when they target individual is present.

  • 13:37:19

    PENRODAll of which, those contribute to overall, greater effectiveness for witnesses when attempting to make an identification of someone of their own ethnic background or racial group. Now, that's not true of everybody, some people will get it right when it's cross-race. It's just that it's significantly more difficult then when it's the same race.

  • 13:37:44

    NNAMDIArman, thank you very much for your call. We got this e-mail, Brandon Garrett, from Joseph in Washington, D.C. "I'm very troubled by flawed witness testimony. But the gentle take-away so far on today's show seems to be that the value and accuracy of such testimony is very limited and is usually wrong. I was an eyewitness in the case of a violent attack where my role was to confirm that the suspects were at the scene.

  • 13:38:08

    NNAMDII watched them exit the house where the attack occurred. And then to associate them with evidence they dropped, some of which was stolen from the house and some of which had blood evidence on it. I could not place them inside the house and no one but they and the victim witnessed the attack, but my testimony provided critical corroboration that was factually correct and lead to a valid conviction.

  • 13:38:31

    NNAMDIPlease remember to confirm the fact that witness testimony is often legitimate and keep the honest criticism in your conversation in perspective." How would you respond to that, Brandon Garrett?

  • 13:38:43

    GARRETTWell, of course, we want eyewitnesses to come forward, to work with the police. Police, you know, depend on all of us in the public to report crimes and to work with them. And that's why it's so important that police do right by crime victims and eyewitnesses by using the most careful lineup procedures and identification procedures they can. So that whatever memory is there, can be carefully preserved. That's why...

  • 13:39:09

    SCHRADBut it's also...

  • 13:39:10

    GARRETT...this Henderson...

  • 13:39:12

    SCHRAD...go ahead, I'm sorry.

  • 13:39:12

    GARRETT... (word?) all I was going to say is that's why this Henderson opinion is so important, from New Jersey. That provides a new model really, for how we can not only do the lineups right but handle them properly in the courtroom.

  • 13:39:22

    NNAMDIGoing to get to that in a second but Dana Schrad?

  • 13:39:24

    SCHRADIt's also one of the reasons why law enforcement works so hard to conduct crime prevention programs in their communities with young people, with the elderly, to help them learn to be more observant about different things around them. It's one of the reasons why you have the height markers in the doorways of stores. So that if someone leaves the store, you can immediately identify, possibly, what their height is.

  • 13:39:45

    SCHRADAnd so there are tools out there in our crime prevention program, sort of, help all of our citizens to be better watch persons and to be more vigilant at the moment that they might observe a crime. But obviously, you can't take away the fact that the stress that occurs, the adrenaline rush that occurs. Sometimes it helps people, it hones their and improves their identification skills. For other people who go into sort of almost a shock at the moment it might impede their identification skills. And that's just a human trait we have to deal with.

  • 13:40:14

    NNAMDIDana Schrad, back up for a second. You mentioned height markers in stores. I never noticed them.

  • 13:40:19

    SCHRADWell, if you go into a convenience store there you go. You need to be more observant. Look in many of the doorways and so forth and you might find something akin to a tape measure in -- that is part of the doorframe or whatever. And it certainly helps the stores in the event that someone were to come through they can identify, well, at least this person is maybe 5'6" or 6'. And it certainly helps overall in the identification of the individual. So there are different tools we use as rudimentary as they are. But they all are critical pieces in helping people become better witnesses.

  • 13:40:52

    PENRODCould I also address your questioner's point? It certainly is the case that a great many witnesses do get it right. The difficulty confronting prosecutors and juries and judges is differentiating those who get it right from those who get it wrong. We've actually known for quite some time that mistaken identifications are the primary source of erroneous convictions, good studies dating back to the 1930s establishing that.

  • 13:41:18

    PENRODAnd in fact there was a parliamentary commission in Great Britain in mid 1970s headed by Lord Devlin which recommended that no prosecutions proceed solely on the basis of an eyewitness identification. They were concerned that the problems of erroneous convictions arising from such instances were just too great. That's a fairly sound recommendation and, in fact, it probably would not cost that many cases in the big wide world of eyewitness identifications were we to adopt such a practice.

  • 13:41:53

    PENRODGood research showing that many eyewitness identifications are, in fact, accompanied by corroborating evidence that would under such a rule permit prosecutions to proceed.

  • 13:42:03

    NNAMDIGot to take another short break. When we come back we will talk about that case in New Jersey that addressed a number of issues in eyewitness identification. If you're calling and the lines are busy go to our website and join the conversation there. We're talking about eyewitness IDs. I'm Kojo Nnamdi.

  • 13:44:07

    NNAMDIWelcome back to our conversation on eyewitness IDs. We're talking with Dana Schrad, Executive Director of the Virginia Association of Chiefs of Police. Brandon Garrett is a Professor of Law at the University of Virginia. His book is called "Convicting the Innocent: Where Criminal Prosecutions Go Wrong." And Steven Penrod is a distinguished professor of psychology at the John Jay College of Criminal Justice in New York.

  • 13:44:29

    NNAMDISteve, a recent case in New Jersey Supreme Court addressed a number of the issues in eyewitness identification. Can you tell us a little bit about that case?

  • 13:44:39

    PENRODYes. The crux of the case concerned what courts are supposed to do when something suggestive occurs during the attempted identification by a witness or an identification by a witness. And the New Jersey Court essentially backed away from an old and well-established practice of looking for other so called indicia of reliability, other factors that might indicate that the witness made a correct identification despite the fact that the procedures were somewhat suggestive.

  • 13:45:15

    PENRODAnd in particular they repudiated relying upon such things as representations by the witness that they were paying close attention to the perpetrator, that they had a good memory, that they were not distracted during the crime. What the court did was to recognize that those factors coming from the witness also are factors that end up being shaped by or tainted by suggestive procedures. So that a witness who, for example, received some sort of feedback from police about their identification indicating that it might be correct now become more confident, now believe that they had a better opportunity to view the perpetrator and so on.

  • 13:46:05

    PENRODThe New Jersey Court instead turned to the large body of scientific research and enumerated a number of factors. It may have been 25 or so factors which they encouraged the courts to look to to gauge whether or not an identification made under suggestive circumstances is likely to be reliable or unreliable. And I think they did a very sound job. They missed the mark on a couple of instances but they were right 90 percent of the time about their characterizations of the research findings and how they bear upon eyewitness accuracy.

  • 13:46:43

    NNAMDIAnd they issued new rules, new regulations saying that issues like the police influencing a witness now require that a judge hold a hearing as to whether the evidence should be admitted in court and the like. Dana Schrad, police and district attorney's offices are under an enormous amount of pressure to solve criminal cases. And eyewitnesses have often been key to solving cases. Could stricter guidelines make it more difficult to solve those cases or just make the process more efficient and reliable?

  • 13:47:14

    SCHRADWell, I think it does both. It makes it more difficult because it sets a higher standard for perfecting the evidence before you would go forward. And that means not just perfecting the scientific evidence but perfecting your eyewitness evidence, and making sure that you have reliable eyewitness testimony and something that should be presented in front of a jury and that the judge feels comfortable has been acquired through a proper procedure of neutral and consistent practices.

  • 13:47:43

    SCHRADSo, yeah, it does make it more difficult. And in that respect though it also means it's incumbent on prosecutors and the law enforcement agencies to work together to make sure that they are operating under the same set up procedures. And that they're also developing the case fully so that they have other evidence to bring to bear so that they only -- so that the jury or the judge, if it's a bench trial, doesn't have just eyewitness testimony to rely on, particularly in a case where you may have something as serious as a death penalty imposed.

  • 13:48:13

    NNAMDIBrandon, the decision applies only in New Jersey, but could these guidelines influence other states, Brandon Garrett?

  • 13:48:21

    GARRETTI think it's a national model. And one reason why some jurisdictions have struggled is that they've wondered, well look, let's say we put the best practices into place. Let's say we have double blind lineups, the police are doing the right thing, they're telling the eyewitness look, the suspect may or may not be here. They're recording the confidence of the eyewitness, they're doing everything right. What happens if something goes wrong? What's a court supposed to do?

  • 13:48:44

    GARRETTUntil now it's sort of been an on-off switch. Either the eyewitness comes in and testifies or not. And judges are very, very reluctant to suppress the testimony of an eyewitness who may be the key evidence at trial. New Jersey provides an intermediate option which is to give the jury careful jury instructions so that they're educated about all those factors that Steve just described. That they understand, for example, that the confidence of an eyewitness isn't as important as were the lineups done right, how did they observe what they saw?

  • 13:49:14

    GARRETTAnd so that option of educating the jury is also a really important part of what the New Jersey Court did.

  • 13:49:20

    NNAMDIOnto Chuck in Washington, D.C. Chuck, you're on the air. Go ahead, please.

  • 13:49:25

    CHUCKThank you very much. I love your show. I was a lawyer and I had a client who was subject to an eyewitness. It was a high school riot and he was alleged to have taken the typewriter at the secretary's desk and throwing it on the ground. And the secretary was like 2' away from him, looked right into his eyes and she claimed she knew him because he was a student at the school.

  • 13:49:48

    CHUCKIt turned out that my client had an identical twin brother. I carefully put him in the back of the courtroom as far away from my client as I could put him and the jury acquitted him. My point is, though, that some defendants look a lot like other people and some people are very, very distinctive. And I'm sure that would make a big difference in terms of the reliability of the eyewitness testimony.

  • 13:50:10

    NNAMDISteven Penrod.

  • 13:50:12

    PENRODWell, I agree entirely. And, in fact, when I do presentations about the research I often start out with a series of slides showing side-by-side people who have been misidentified and actual perpetrators, including from several cases that I've been involved in. And it is the case that, in many instances, people bear a strong resemblance to one another, even a remarkable resemblance.

  • 13:50:37

    PENRODI recall when I first started being aware of the research in eyewitness identification in the mid 1970s, Robert Buckhout at Brooklyn College published a piece in Scientific American where he had three photographs, two guys who had been misidentified for crimes committed by the third person. All three of them, strong similarities. So one begins to understand why, for example -- to go back to a point that Brandon was making -- multiple witnesses can make the same mistake. They're confronted by somebody who bears a resemblance to the perpetrator.

  • 13:51:16

    PENRODAnd these problems arise in situations where the witnesses don't have a long time to study the perpetrator. They may be under high stress. A variety of factors that can impair their memory. Put all those ingredients together and that's when you get the problems that give rise to the DNA exonerations.

  • 13:51:37

    NNAMDIYou know, I grew up in a country, Guyana, South America. In the city in which I was living there were twins who were notorious because they were confidence tricksters and that was their scam, the fact that they were identical twins. And so when one perpetrated a fraud on somebody and the person identified them and they went to court, invariably they got off because it was always the other guy who had, in fact, done it. And that was the way they made a living until the law finally caught up with them. Here is Soren (sp?) in Silver Spring, Md. Soren, you're on the air. Go ahead, please.

  • 13:52:10

    SORENHey, Kojo, how you doing?

  • 13:52:12

    NNAMDII'm well.

  • 13:52:12

    SORENI was mugged a couple years ago and I had a gun to the back of my head and a gun in my face. It took the cops a while to show up and gave a report. About a month later, the police showed up with a book of about somewhere between 12 and 20 perpetrators or suspects, sorry. And I couldn't honestly say who did it. I couldn't pick anybody out from the lineup. There was at least three or four guys that I could've said, yeah, that was the guy, but in honesty, I couldn't do it. So I sent them away with no person to, you know, charge.

  • 13:52:51

    SORENOne, why does it take the police so long to respond? And is there a way to shorten that time?

  • 13:52:56

    NNAMDIDana Schrad.

  • 13:52:57

    PENRODWell, I tell you what, I couldn't answer why it happened in your particular case. But I will tell you something that is a critical problem in our country and it's partly due to the economy, it's mostly due to the economy. And that is we are laying off police officers. We are not filling open positions and our agencies are woefully understaffed. We've had a great deal of turnover in our agencies across the country, people who leave the police profession because the salary and the wages are not -- the wages and benefits are not that good. The working conditions were not what they wanted.

  • 13:53:28

    SCHRADAnd so keeping a fully staffed skilled law enforcement force in a police department or a sheriff's office is a challenge across the country. So I couldn't answer why that happened in your case. It does seem like an inordinately long period of time though between filing the report and having you go through a witness' identification.

  • 13:53:48

    NNAMDIThank you very much for your call. We move on to Gabray (sp?) in Springfield, Va. Gabray, you're on the air. Go ahead, please.

  • 13:53:56

    GABRAYThank you, Kojo. Thank you, your guests as well. My question is eyewitness, when they select eyewitness how do they -- what is the process? And second part of the question is do they do a background check on the individuals that are going to testify on the case?

  • 13:54:18

    NNAMDIWhat is the process for selecting eyewitnesses, Dana Schrad, apart from the fact that they are, well, eyewitnesses?

  • 13:54:23

    SCHRADWell, it depends on whether you're talking about a photo lineup or a physical lineup with, you know, live individuals. But it is a delicate balancing act. You want to pick people who have features similar enough to the suspect, if the suspect is going to be in your lineup, that the suspect doesn't stand out like the proverbial sore thumb apart from the other people in the lineup.

  • 13:54:48

    SCHRADOn the other hand, you don't want to have so many similar doppelgangers in your photo lineup that it makes it impossible for the witness to identify the actual suspect. So there is no true science to this. It is a real balancing act of trying to create an array of individuals, whether it's a photo or a physical lineup that are similar enough in features but not so similar as to not allow the witness to pull the actual person that he or she saw from the lineup.

  • 13:55:17

    NNAMDIBrandon Garrett, how do courts and police departments handle eyewitness identification in different jurisdictions? It varies enormously by state and jurisdiction, does it not?

  • 13:55:28

    GARRETTMany police departments still have no written procedures at all on how to do a lineup right. And it's certainly not the kinda thing that should just be sort of done ad hoc based on whatever, you know, the officer feels like doing. More and more police departments are adopting careful written procedures but that's been sort of slow in coming. Some of that has been in response to these high profile DNA exonerations.

  • 13:55:52

    GARRETTNew Jersey though is the first state to join double blind lineups, all the best practices with a framework for the courts to use to figure out how to carefully regulate eyewitness identifications and in the courtroom. So that's something new.

  • 13:56:07

    NNAMDIIt's my understanding that in the District of Columbia where I am there are no written policies on eyewitness identification to comport with best practices. In Maryland they do have a law that requires written policies that comply with department of justice guidelines. Tell us about the 1999 guidelines, Brandon Garrett.

  • 13:56:26

    GARRETTWell, Steve might be able to tell you even more. But those guidelines are somewhat dated and they really didn't reach the issue. Really the most important protection right is to have the lineup be double blind so that the administrator doesn't know who the suspect is to prevent all that powerful kind of cuing that can happen, even if the officer's just unconsciously giving off signals, or if the eyewitness who's looking for reassurance, doesn't want to be there looking at the picture of the attacker. And they sort of perceive cues that weren't intended.

  • 13:56:55

    GARRETTAnd so if the Maryland's recommend policies don't include double blind, than that's really a flawed recommendation and needs to be updated.

  • 13:57:02

    NNAMDISteve Penrod, we have about 30 seconds left.

  • 13:57:06

    PENRODWell, I would say admonish the witness that it's a blind presentation, that the perpetrator may not be there. Get a confidence judgment immediately after an identification is made. The new study that I mentioned earlier comes out I would say fairly strongly supporting experimental research that indicates that sequential presentation showing one face at a time is a superior practice. And ultimately I think we'll probably get to the point where we will be recording all of these procedures so that jurors and prosecutors and judges can see exactly what went on.

  • 13:57:43

    NNAMDISteven Penrod is a distinguished professor of psychology at the John Jay College of Criminal Justice. Steven Penrod, thank you for joining us.

  • 13:57:50


  • 13:57:50

    NNAMDIDana Schrad is the Executive Director of the Virginia Association of Chiefs of Police. Dana Schrad, thank you for joining us.

  • 13:57:57

    SCHRADMy pleasure.

  • 13:57:58

    NNAMDIAnd Brandon Garrett is a professor of law at the University of Virginia. His new book is called "Convicting the Innocent: Where Criminal Prosecutions Go Wrong." Brandon Garrett, thank you for joining us.

  • 13:58:08

    GARRETTThank you.

  • 13:58:09

    NNAMDIThe Kojo Nnamdi Show is produced by Brendan Sweeney, Michael Martinez, Ingalisa Schrobsdorff and Tayla Burney with assistance from Kathy Goldgeier and Elizabeth Weinstein. The managing producer is Diane Vogel. The engineer today is Timmy Olmstead. A.C. Valdez has been on the phones. Thank you all for listening. I'm Kojo Nnamdi.

Related Links

Topics + Tags


comments powered by Disqus
Most Recent Shows

The Politics Hour: April 19, 2019

Friday, Apr 19 2019Delegate Danica Roem joins us to talk traffic, tolls and the 2019 Va. legislative session, and Delegate Dereck Davis tells us why he wants to be the next speaker of Md.'s House of Delegates.