Press Release

NSF PR 97-1 - January 3, 1997

This material is available primarily for archival purposes. Telephone numbers or other contact information may be out of date; please see current contact information at media contacts.

False Identification: New Research Seeks to Inoculate Eyewitnesses Against Errors

More than 75,000 people become crime suspects each year in the United States based on being identified from lineups and photo spreads. Some identifications will be false and lead to mistaken arrests and imprisonments.

A recent study* reveals the extent of this problem. In 24 of a sample 28 cases involving Americans who were released from prison in the last few years based on DNA evidence that proved their innocence, eyewitnesses had made false identifications from photo spreads and lineups. No one knows how many others may be unjustly jailed every year due to false identifications.

New research sponsored by the National Science Foundation (NSF) is now helping to understand the circumstances that lead to false identifications and to find ways to reduce their occurrences. The results should eventually help police and jurors who must decide the fate of a suspect based on the confidence of an eyewitness account.

Particularly troublesome, says Gary Wells, a psychologist at Iowa State University who has conducted NSF-funded research in this area, is evidence that an objective question such as "How certain are you that the person you identified is the person you saw commit the crime?" elicits a similar response regardless of whether the eyewitness' memory is accurate or not.

This discovery surprised Wells and his team of researchers, who hoped a direct question would counteract certain influences on an eyewitness' memory. They knew that confidence can be manipulated easily if eyewitnesses receive information about a suspect after making an identification. Their earlier research indicated that casual remarks ("Yes, you've identified the same suspect we picked up for questioning") uttered by police station personnel administering suspect lineups or photo spreads can bolster the confidence of an eyewitness and distort the witness' memory.

The latest research funded by NSF reveals that, once an eyewitness' memory has been distorted in this way, a straightforward cross-examination often fails to produce an accurate recollection. In particular, eyewitnesses began to change their answers to questions about how much attention they had paid to the culprit, how good a view they had of the culprit's face, and other factors surrounding the event. As a result, Wells believes stronger steps are needed to "inoculate" eyewitnesses' memories, especially over the weeks and months that may stretch between the crime and a courtroom trial.

"We know the nature of the problem," says Wells. "Now we need to look at ways to prevent manipulation. Part of the solution is to require blind testing, where the person administering line-ups or picture spreads does not know who the suspects are" and therefore cannot bias the eyewitness, he says.

Wells also suspects that more accurate testimonies may result from posing a series of deliberate questions immediately after identification to probe an eyewitness' confidence and memory. Questions such as "How clear a view did you have of the suspect?" "How long did you look at him?" "How easy was it for you to make the identification from the photo spread?" may help reduce later distortions of their answers, he says. This is his next line of study.

To test the accuracy of eyewitness testimonies, Wells and his team staged thefts and possible acts of fraud before unsuspecting potential eyewitnesses in offices, stores and waiting rooms; the eyewitnesses then were asked to make identifications under various conditions which isolated factors that could affect the confidence of their memories.

National Science Foundation
Office of Legislative and Public Affairs
4201 Wilson Boulevard
Arlington, Virginia 22230, USA
Tel: 703-292-8070
FIRS: 800-877-8339 | TDD: 703-292-5090




This article comes from Science Blog. Copyright © 2004
http://www.scienceblog.com/community

Archives C