The San Diego Police Department is changing the way it conducts lineups to identify suspects after results from a national study indicate that the method of reviewing suspect photos the department uses has an 18 percent error rate.
Related Story: SDPD Changing Lineup Practices
CAVANAUGH: This is KPBS Midday Edition. I'm Maureen Cavanaugh. Much of the national controversy surrounding last week's execution of Troy Davis in Georgia was centered around the fact that seven of the nine eyewitnesss in his case recanted their testimony. The reliability of eyewitness testimony has been questioned for quite some time. And the U.S. Supreme Court is expected to hear a case in November which challenges the methods used by some police departments in conducting lineups. The San Diego police department is now planning review its follow-up procedure after participating in a national study in which photo-lineup method is best. James Collins is captain of investigations for the San Diego police department. Captain Collins, welcome to the show.
COLLINS: Thank you.
CAVANAUGH: And California innocence project director Justin brooks is with us too. Hi Justin.
BROOKS: Good afternoon, Maureen.
CAVANAUGH: Captain Collins, what is the current procedure that the San Diego police department follows when conducting photo-lineups?
COLLINS: Our procedure for the past many years has been a simultaneous lineup or what we call a six pack. The individual is shown six photographs at one time and asked if they recognize any one of the people that is in that lineup.
CAVANAUGH: Now, let me take a step back. I think a lot of people believe lineups are conducted the way they see them on TV and in movies. You stand behind glass and people come in and they stand before you. Is that done anymore?
COLLINS: Yes. That's a live lineup. But considering the photo-lineups, this is how we do our photo-lineups.
CAVANAUGH: What's the sort of ratio between the two of them.
COLLINS: I'm not sure. But a lot more photo-lineups than live lineups.
CAVANAUGH: So in this study, it was found that showing witnesses, instead of that six-pack that you just told us about, sequential photographs, one at a time. It reduced inaccurate IDs from taken% to 12% am were you surprised by that?
COLLINS: We had hears the information that was out there. We were waiting for the study to be completed. Of we were willing participants of this, and obviously if there's something there that's going to be better for justice and reduce the false implications, we're more than happy to change our procedures.
CAVANAUGH: I guess what I'm asking you, does it take any intuitive sense to you that showing people photographs one at a time would make it easier to correctly identify someone?
COLLINS: Well, yes, because you're asking them to look at that one picture and say do you know this person or not? And they look at that one picture. They may not even know how many pictures they're going to be shown. So they don't feel the need to say, yeah, he looks most like it.
CAVANAUGH: Now, what is the San Diego police department doing in response to the study findings?
COLLINS: Developing a committee right now to develop a new procedure for the sequential lineups. We're already encouraging our detectives to go ahead and switch to that method. If they feel comfortable with it. Some have received some training in sequential lineups, and have been doing it for a while.
CAVANAUGH: Do you anticipate this is going to be a difficult transition?
COLLINS: No. I don't think it's going to be difficult. It's just coming up with our policy and procedures to make sure everybody is doing the same thing. And I know chief Lansdowne is did going to take this to the county chief and sheriff's association and present it to them.
CAVANAUGH: Justin, what are the problems that the innocence project has found in the cases that it's reviewed with the way that some police departments conduct lineups?
BROOKS: Maureen, bad identifications are the number one cause of wrongful convictions. And the reason is, an identification is very powerful evidence. When a witness in a courtroom points at a suspect and says that's the guy, juries convict. Sometimes with just that evidence alone. And there's a lot of problems with identifications. It starts with things like cross-racial identifications are typically bad. Identifications made in fearful situations have problems. When time has elapsed, there's problems. So there's always some problems with identifications. When you couple that with Brad probability coles, you get misidentifications. And it's supposed to be a scientific process, so it should follow scientific protocols. And a bigger problem than the sequential identifications is actually the double blind process.
CAVANAUGH: Which I'm going to talk about in a minute. But I think most people who have been victims of crime really want to see the right person punished. They don't want to go and misidentify someone. So how does this seem to happen so often?
BROOKS: There's so many causes of I misidentification. They want to get the right person, and sometimes that leads to a misidentification. Studies have shown that there's not a correlation between confidence and accuracy. And sometimes that person who says I'm 100% sure this is the guy, that's the identification you've got to question. Where's another person who says the hair looks the same, the height is about the same, that's the person who's really doing some level of analysis, not being led to that suspect. And the six-pack photo arrays have been shown that what happens is the witness does a comparative analysis in their head and picks the person who looks most like, not the person, and actually the five other photos become a process of distracting and distorting the identification as opposed to as we've spoken about already, showing one photo at a time.
CAVANAUGH: When I said, why does this happen so often, I want to be absolutely clear, even with the six-pack photo method, a misidentification only happened in this study 18% of the time. While that is significant, it is not an overwhelming majority of the identifications being false.
BROOKS: 18% of the identifications, that's 18% of people who go to prison for the rest of their lives as a result of it. What you want is either no identification or the right identification.
BROOKS: And the problem is now sometimes you get the right identification and wrong identifications.
CAVANAUGH: What has -- how difficult is it, Justin, for -- when people recant their eyewitness testimony as happened in the case of Troy Davis who was executed last week in Georgia? How difficult on that evidence alone -- people say no, I was wrong. I identified the wrong purpose, to get a new trial?
BROOKS: Extraordinarily difficult. That's what I spent the last 12 years of my life mostly doing. Most of my cases are cases of bad identifications that I'm trying to get reopened. Our United States Supreme Court has said innocence alone is not a constitutional claim. So the mere fact that you're innocent isn't a reason to reverse your conviction. You need to show there are problems in the processes, bad lawyering, some kind of misconduct in order to get a case reopened. And the problem with identification is, sometimes the police are absolutely doing their job, and everybody's doing their job, and you still end up having a bad ID come out as a result of it.
CAVANAUGH: Captain Collins, one of the most interesting parts of this study, I thought, was the idea that they also recommended that the person who shows, the detective who shows these photographs in sequential order should not know who the suspect is. They call it a double blind. Why would that affect any kind of identification as far as you can figure out?
COLLINS: Well, you could have the allegation that the officer that is presenting the lineup is pointing, making some indication which person should be selected as the suspect. Obviously officers object to that because they say you're questioning my integrity. And if we can eliminate a question that could come up in court later on, why not do it? Let's make sure that everything is done as best we can to make sure that we get the right person in jail and don't put the wrong person in jail.
CAVANAUGH: Now, the idea that a detective who knew who the suspect was could indicate that to someone looking at those photographs is actually a little bit more subtle than, you know, here's the guy. You know?
CAVANAUGH: Sort of pointing at the person. How can someone indicate that without even trying to?
BROOKS: That's not generally the problem. I can understand the defensiveness of police officers believing that to be accused of misconduct. But most communication is nonverbal. If people don't believe that, they should try to speak a second language over the phone. Any time you try to do that, you cannot communicate anywhere near as well as in person. So it could be the police officer without knowing it maybe holds the photo a little longer, maybe is looking at it a little longer, we're constantly communicating in ways we don't know wee communicating. The processes have to follow a scientific model, and good science is blind science. We trust scientists as well, but when they're doing experiments, we know the best way to get good results is when the scientists don't know exactly what's going on.
CAVANAUGH: And that diagonal blind aspect of the sequential photo-lineups is also being considered now as part of a new policy for the San Diego police department; is that right?
COLLINS: Yes, we're going to take a look at everything. See what is gonna fit best with our department. Obviously the sequential rather than simultaneously photos, we'll start doing that. We're going to take a look at probably doing the double behind most of them too. There may be times when you have one detective out on a case in the middle of the night, and get a suspect, so it might not be possible. We're going to explore the possibility of putting it all on a laptop computer that records everything, and the officer that is doing the double blind, just presents the computer, hits the start button, and the person just follows the cues.
CAVANAUGH: You've been in the police department for quite some time, and I'm wondering, you've been hearing all of the problems that people have identified with the eyewitness identifications and so forth. How much credibility does the police department give eyewitness testimony today in light of that?
COLLINS: We still give it a lot of credibility. But we always look for more evidence rather than just the eyewitness identification. We know we're going to have trouble getting the case presented to the District Attorney's office. We know there can be a lot of issues with the case. We always try to find additional information. We also try to corroborate the person's alibi.
CAVANAUGH: I see. And Justin, I'm wondering how you would like to see these procedures changed. In other words, for someone who comes in a victim or someone who has seen a crime happen to point out or to identify a particular suspect that has been brought in as the possible perpetrator.
BROOKS: I'm very pleased that San Diego is looking at their procedures. Talking about it here for a long time. We talked before we went on the air. I was down in Bolivia last month trying to get their police departments to switch over to this system. And I would like to see San Diego ahead of Bolivia. How we want to change them, there are simple changes that also don't cost money, which I think this is one area of the criminal justice system that maybe we can be successful, 'cause it doesn't cost money to show people one photo at a time. And it doesn't cost money to have an officer who's not the investigating officer on the case to do the identification procedures. They're very simple processes we need to change. And they're in everyone's best interests. We all want to make sure that the right person is identified and that wrong people are not identified.
CAVANAUGH: And are there any other tweaks that you would like to see perhaps the down the road introduced into the idea of eyewitness identification?
BROOKS: I think the problems of using photos -- we had a client that was convicted based on pay photo that was ten years old. He had a solid alibi, was working in a factory at the time this crime was committed, and he still got convicted. I think we need the public to understand the fundamental problems with identifications, that even when the procedures are followed -- for example, people have a great deal of difficulty identifying people not of their own race. That's not racism, it's reality. And it's based on your life exposure. And so -- and also people have trouble making identifications when there's weapons involved because they so focus on the weapons. As long as we recognize that these things, there's always going to be problems. It's not DNA. It's not this hard science. It's people trying to reconstruct memories and that's a difficult process.
CAVANAUGH: It must be hard on the police's end when someone comes in who's been the victim of a serious crime, and they are absolutely certain that they are pointing to the right person to take that with a grape of salt and step back and say maybe yes, maybe no.
COLLINS: It is difficult. Especially when the personal is very positive. And again, we try not to rely just on that eyewitness identification. We want other corroborating evidence whenever possible.
BROOKS: I was going to say there's a great documentary about that called what Jennifer saw, about Jennifer Thompson, and she put a man away for more than ten years based on identification, and then DNA came out, and exonerated him. They go around the country talking about that. She was 100% sure. And she was wrong.
CAVANAUGH: Do you know how long it's going to be before the policy changes?
COLLINS: Probably more than a couple months to have our policy and procedure. But I'm already encouraging detectives to use the sequential rather than simultaneously on their own.
CAVANAUGH: I want to thank you boat. I've been speaking with San Diego police captain James Collins and California innocence project director Justin brooks. Thank you both.
BROOKS: Thank you very much.
COLLINS: Thank you Maureen.