CHICAGO -- Jeffrey Pierce was working as a maintenance man at an apartment complex in Oklahoma City on a day in 1985 when a woman was stabbed and raped in one of the units. Looking for the rapist afterward, a police officer saw Mr. Pierce and asked the victim if he might be the man who attacked her. She said he wasn't.
Ten months later, her recollection changed. Seeing his police lineup photo, she identified Mr. Pierce as the rapist. Never mind that his co-workers said he was having lunch with them at the time. He was indicted, convicted and sent to prison for 15 years. Twice he was denied parole because he refused to admit what he had done.
But last week, Mr. Pierce walked out of prison, thanks to a DNA test proving his innocence. He became one more argument for the widespread use of DNA analysis -- and one more example of the fallibility of witness identification.
In recent years, scores of inmates convicted of serious crimes, including many on death row, have been exonerated by evidence that was discovered or analyzed only years later. Almost all of them were incriminated by a convincing witness whose memory later turned out to be totally wrong.
Such testimony, of course, is one of the most powerful forms of evidence deployed in criminal trials. Few things are more persuasive to a jury than a victim or other witness who takes the stand, looks straight at the defense table and declares, "That's him. That's the man who did it. I saw him with my own eyes."
But few types of evidence are less trustworthy, either. In another rape case, the victim and two other people placed Herman Atkins at the scene of the 1986 crime -- testimony persuasive enough to earn him a 47-year prison sentence. But last year, a DNA test exonerated Mr. Atkins.
It's not that witnesses deliberately lie. Experts have established that memory is susceptible to all sorts of capricious influences. Police may inadvertently or intentionally nudge witnesses to choose one person in a lineup or set of photos. Witnesses may pick the person who looks most like the culprit, even if he's not in the group.
Iowa State scholars Gary Wells and Amy Bradfield did an experiment showing 352 participants a surveillance video of a man who, they said, had shot a security guard. When presented with photos of five different men, all 352 fingered one or the other of them -- even though the villain was not, in fact, among the five. Some prosecutors are leery of cases that rest on the testimony of a single witness simply because they know such testimony is often mistaken.
False memories can come from almost anywhere. In the Atkins case, the victim identified him after seeing his photo on a wanted poster at the police station.
In another case, reports the Chronicle of Higher Education, an Australian psychologist was implicated by a rape victim. But he had a perfect alibi: He was being interviewed live on TV at the time of the crime. So how did he get accused? It turned out the victim had seen him on TV as she was being attacked, and her mind somehow converted him into the rapist.
Juries, however, are likely to be swayed by someone who identifies a criminal under oath. So obviously, some method is needed to offset that impulse. Last week, New York's highest state court ruled that trial judges may admit expert testimony about the shortcomings of witness identification, which courts there have generally barred in the past.
The theory was that juries don't need to hear from experts on subjects that they understand from personal experience -- like, duh, remembering a face you've seen. But the New York Court of Appeals concluded, "Despite the fact that jurors may be familiar from their own experience with factors relevant to the reliability of eyewitness observation and identification, it cannot be said that psychological studies regarding the accuracy of an identification are within the ken of the typical juror."
Some states already allow such testimony, and it's generally permitted in federal court. But in most places, the criminal justice system has been slow to adapt to evidence suggesting that witness identifications should not always be believed.
One of the simplest correctives is for police, prosecutors and juries not to put too much trust in witnesses. They should keep in mind what the rape victim said as she testified against Jeffrey Pierce at his trial: "I will never forget his face."
Steve Chapman is a columnist for the Chicago Tribune.