In the criminal justice system, eyewitness identifications can point police to a suspect and lead to a conviction. But human memory can be fallible, and many factors can affect our memories. On the latest episode of the podcast Short Talks From The Hill, Todd Price speaks with the chair of the University of Arkansas Psychological Science Department, Jim Lampinen. They discuss his work on memory and facial recognition in witness identification.
Price: Most of us, thankfully, have probably had little direct contact with the criminal justice system, and our ideas of law enforcement are shaped by movies and television shows. Personally, I love old episodes of Law and Order, but what is it like to actually be an eyewitness to a crime? Let's imagine I just witnessed a crime. I saw someone walk into our building, grab a laptop, and run out. The police are called. What happens next?
Lampinen: Typically— I watch those shows, too— It depends a little bit on the nature of the crime and also on the jurisdiction. But typically a witness gets interviewed multiple times. So you might make a call to 911 and you'll be asked some questions during that phone call. There will be an initial response from the police. Usually a uniformed officer will come and take a statement. So that's an interview.
At that point, it's sometimes the case that you'll be asked to make an identification. So there's a procedure called a show up where after the police take an initial statement from you, they find someone in the area who kind of looks like who you described, and they may hold that person for a short period of time, drive you up to the person and say, is that the guy?
After the initial interview, you might be taken to the police station and talk to a detective. So that's another interview. So we're up to three interviews now. You may testify at a hearing. So now you're asked the question a fourth time. You may again make an identification then. But that's called an in-dock identification. It's kind of like you see on the Law and Order show where they say, do you see the person who stole the laptop in the courtroom, and there's only one person that could possibly be. You may, in a serious crime, have to do what's called a deposition, where the defense attorneys can ask you questions ahead of time to kind of see what your story is, and then you may have to testify again at trial. So altogether, multiple times you're being asked similar questions.
Price: How often is witness identification alone the main evidence used to convict someone?
Lampinen: We don't have great estimates of that. But there's by now a fairly old study that estimated there are about 80,000 cases in the US every year where eyewitness identification is the main evidence against the defendant.
Price: In general, is memory just more fallible than we think, typically? And does law enforcement give it too much credit?
Lampinen: What we mistake about memory is that we kind of think memory is like a video recorder, where we can play back scenes in our head. Memory is instead more like a jigsaw puzzle, where we're kind of putting pieces of evidence together to come up with the story that makes the most sense to us. That is a process that's usually useful because if we can't remember the information, at least we have a reasonable guess of what happened. But in high stakes situations like a crime, it can also introduce systematic errors.
Price: In the book "The Psychology of Eyewitness Identification," you and your co-authors talk about several variables. Do they hedge? Are they certain? How long does it take them? Estimator values are factors outside the control of either the witness or law enforcement. Was it dark when the crime occurred? How far away was the suspect? How long did the witness see the suspect?
The third group of variables, which I want to focus on, are system variables. These are things that law enforcement can affect, like how a lineup is set up, who conducts the lineup, what instructions are given. One system variable is the testimony of others. You and the authors talk about this. Explain how other accounts from fellow witnesses or social media can affect or distort the memory of a witness.
Lampinen: Well, there's a well-known phenomenon called the misinformation effect in psychology. And the misinformation effect is basically a finding that your memory is ultimately a combination of what you actually saw, but also information you gathered from other sources. So imagine you are the victim of the laptop theft we were talking about before. You're actually able to see the lineup. You read a newspaper account that includes the statement of another witness, and the other witness mentions a blue tattoo on the left side of the person's neck. You didn't actually see the left side of the person's neck because you saw the profile from the other direction. But when you remember it, you may well remember that blue tattoo because your memory has combined what you saw with what other people saw.
So that's why police are encouraged when they interview witnesses to separate them, so that one witness isn't influencing the other witness's story. Sometimes when that happens, witnesses give different accounts because maybe they saw it from a different perspective, or maybe one witness's memory is wrong or another witness's memory is right. But what you don't want to happen is a situation where all the witnesses are telling the same story, not because they saw the same thing, but because they've kind of talked to each other about it. The problem with that is it may amplify an incorrect memory of one particular witness, and it may seem especially convincing to the jury because three people all saw that blue tattoo on the left side of the person's neck.
Price: Well, you mentioned police can try to separate witnesses when they're taking their stories, maybe on the scene. What else can police do to make sure that these witnesses are not being tainted, that their memories aren't being affected by others?
Lampinen: You know, there's limits to what police can do. But police can, of course, instruct witnesses: please don't talk to other witnesses about this, please try to avoid any media about the event. An increasingly large problem is witnesses are starting to do their own detective work. So they will go on social media and try to track down who the person is, or ask around the neighborhood, have you seen anybody with a new laptop, and then look the person up on social media. The problem with that is you might find someone on social media that maybe someone in the neighborhood mentioned. You may look at their face and think, that's kind of what the person looked like who stole that laptop, I bet that's who it is. But you're making that identification without the kind of procedural controls that a well conducted identification would make use of, including properly instructing the witness, selecting fillers in a proper way.
Price: What police tell a witness when they come in for a lineup, or when they're about to look at a photo book, and even signs they might give without realizing it, can influence the witness. So talk to us about that, and how can police avoid that so that they're not inadvertently swaying a witness toward who they think is the suspect?
Lampinen: There are three things that go into that. Current recommendations are that the person who actually administers the lineup should be someone who doesn't know who the suspect is. So instead of the lead detective conducting the lineup, who is aware of who the suspect is and may even have a hunch, like I bet that's who did it, I bet that's who stole that laptop, the lead detective might inadvertently cue the witness. So you could imagine a witness looking through a lineup, they get to picture number three, and maybe the detective smiles a little bit. And that doesn't need to be intentional, that's just human nature. You've got that hypothesis and maybe it's going to be confirmed. So instead, the idea is just have somebody else come in and conduct the lineup who doesn't know who the suspect is. That way, there's no way they could possibly influence the witness.
The second component is witness instructions. We know from research that most witnesses think the police would not be showing me this lineup unless they had a suspect. That belief is actually correct. I mean, they would not be showing you a lineup unless they had a suspect. But the problem is, it makes the witness think, I need to pick someone because I don't want to mess this case up, they know who did it, I need to pick the person who did it, when in reality, the police have a suspect, someone who they think might have did it. And the witness's job is to help confirm or disconfirm that hypothesis. Generally, police are told you should tell the witness the perpetrator may or may not be present before conducting the lineup.
And then there's a third thing, and this is from more recent research. We're actually doing some research on this right now in my lab. Police, sometimes before giving that instruction, that the perpetrator may or may not be present, inadvertently cue the witness that they've caught someone. So you could imagine being called into the police station and being told, hey, remember that laptop you reported stolen, we found the guy who did it, in fact, we found the stolen property in his car, we want to bring you down to the station and see if you can identify him. At that point, it doesn't matter what instructions you're given because you've just been told they have the guy who did it, in fact, there's evidence against the guy who did it. So police are told to avoid making those kind of statements, limit the information you give the witness about what other evidence you may have against the suspect in the lineup.
Price: In general, does the criminal justice system recognize the importance of the work that you and other psychologists are doing, or psychological researchers are doing, to improve their work? Are they incorporating these insights? And if not, why not?
Lampinen: Increasingly so, and at multiple levels. But of course, our system of government in the United States is that there are separate states and separate state governments, and then there are municipalities that have their own kind of rules. So it varies a lot by where you're at. Exactly.
The US Department of Justice, for instance, has published guidelines for federal law enforcement on how to conduct lineups, and it incorporates the procedures that we just talked about. There are a number of states that have developed guidelines for how to conduct lineups in photo spreads. Increasingly, states are developing jury instructions so that when a case goes to trial, the jury is given information about what the science says about eyewitness memory. In Arkansas, for instance, the Arkansas Association of Chiefs of Police have developed a set of model guidelines for how to conduct lineups. These are voluntary, so different agencies can decide to adopt them or not adopt them. But a number of agencies have.
An excerpt from the most recent episode of the podcast Short Talks From The Hill. Todd Price spoke with the chair of the University of Arkansas Psychological Science Department, Jim Lampinen, about his work on memory and facial recognition in witness identification. To learn more about his work, you can visit kuaf.com/shorttalks.
Ozarks at Large transcripts are created on a rush deadline and edited for length and clarity. Copy editors utilize AI tools to review work. KUAF does not publish content created by AI. Please reach out to kuafinfo@uark.edu to report an issue. The audio version is the authoritative record of KUAF programming.
