July 2008

Witness for the Prosecution . Maybe

Samuel Sommers casts a social psychologist's skeptical eye on the criminal justice system

By Helene Ragovin

About 15 minutes into the inaugural Lerman-Neubauer lecture by Samuel R. Sommers, an assistant professor of psychology, a young man strode down aisle of the auditorium and removed a folder from a low table at the front of the room.

"Memory is not videotape. Quite often, it is a reconstructed process," says Samuel Sommers, assistant professor of psychology and recipient of the 2007 Lerman-Neubauer Prize for outstanding teaching and advising. Photo: Zara Tzanev

It's safe to say those in the audience didn't give him another thought until 30 minutes later, when Sommers, who specializes in the areas of social perception and judgment, particularly as they relate to the legal system and racial diversity, turned to the subject of eyewitness memory.

Could those in attendance, Sommers asked, pick out the same young man they had seen half an hour before? He showed a photo lineup of eight "suspects" on a projection screen and asked for a show of hands-which one took the folder? The result was hardly unanimous. Audience members were split in their identifications, variously choosing six of the eight in the lineup. But none was correct-the folder-snatcher wasn't even among those pictured.

It was a vivid example of how eyewitness identification can prove unreliable, one of the weaknesses in the criminal justice system that Sommers discussed in the April 24 lecture.

Another is the ability to accurately judge others' honesty. Sommers recruited six volunteers from the audience and asked each to tell a story about an embarrassing incident from childhood. Some were secretly instructed to tell a true story; others to fabricate one. The audience was then asked to decide which stories were true, which false.

There was little consensus among the audience about the veracity of each story-in each case, at least half of the audience was swayed by a made-up story, or remained unconvinced by a true tale. Most listeners, for example, were hoodwinked by a student who recounted an incident involving a boy named Joey, who pulled up a girl's skirt during a school play-a story that turned out to be from an episode of Friends.

The Tufts students in the audience were no less astute than most others. Research in a controlled experiment showed that only 52 percent of college students were able to distinguish deceptive responses from truthful ones. But don't feel bad for them. They performed only marginally worse than people who are supposed to be pros: police officers, judges, psychiatrists and those trained to give polygraph tests. Only Secret Service agents had a significantly better track record, as a group, in uncovering deception.

© The New Yorker Collection 2000 Michael Maslin from cartoonbank.com. All Rights Reserved.

"This is a fascinating area of practical importance," Sommers said. The psychology of law enforcement and the justice system is a rich field for investigation, he noted. It ranges from exploring what motivates criminal behavior and analyzing the performance of children as witnesses to issues of memory and eyewitness testimony and understanding group interactions-such as a jury's decision-making process.

Take the psychological dimension of eyewitness identification. In the courtroom, "we know from interviews with jurors, there are few things as persuasive to a jury as eyewitness identification," Sommers said. But research by the Innocence Project, an organization devoted to exonerating wrongfully convicted people through DNA testing, shows that in more than 150 cases in which a convicted person was later exonerated, 75 percent included incorrect eyewitness identification.

"How do we explain that?" Sommers asked. The psychological factors at work here involve issues of memory, perception and social influence, among others, he noted.

When making an eyewitness identification, there are two types of variables at work, according to Sommers. The first are "estimator variables," which occur at the time of the event, such as the lighting; the distance between the perpetrator and the witness; how long the witness saw the perpetrator and the race of each party involved.

Then there are "system variables" that come into play when a witness is making a criminal identification, such as how a lineup is conducted or whether a lineup uses live subjects or photographs. (To illustrate, Sommers played a clip from a Seinfeld episode in which Kramer, in his own inimitable way, wreaks havoc with a police lineup, nodding, winking and twitching in the direction of the supposed guilty suspect.) All these variables can influence the accuracy of an eyewitness identification, Sommers said.

In addition, the U.S. Supreme Court has ruled that the confidence of an eyewitness can be used as a good indicator of the accuracy of a witness's account. But research by psychologists has shown that an eyewitness's confidence is "malleable"-things said to a witness at the time of the initial identification can increase or decrease the confidence with which an eyewitness makes an identification.

"Memory is not videotape," Sommers said. "Quite often, it is a reconstructed process."

The importance of this type of research is significant, Sommers says. "In many respects, research on eyewitness memory is one of the success stories of the field of psychology and law," he notes.

"Several years ago, a working group of psychologists and law enforcement officials put together a list of recommendations, which several jurisdictions have since adopted. One suggestion was to make sure that the official who administers a lineup is 'blind,' meaning that she or he does not know who the chief suspect is, or, for that matter, whether a suspect is in the lineup," Sommers said. This prevents the problems with suggestibility and feedback effects.

Another suggestion was to instruct witnesses that the culprit may or may not be in the lineup, which is intended to prevent some of the problems with relative judgments.

"Another option is through the testimony of psychologists as expert witnesses during trial," Sommers added. "Many psychologists have now testified in cases, not just regarding the potential problems of eyewitness evidence, but also more specifically regarding the precise circumstances under which such problems are most likely to emerge."

Helene Ragovin can be reached at helene.ragovin@tufts.edu.

Article Tools

emailE-mail printPrint