Eyes might not have it
Witness testimony sometimes wrong, DNA shows


November 5, 2001

Incorrect eyewitness testimony convicted, or helped to convict, most former prisoners who have been cleared by DNA in recent years.

Among them are three Virginians - Wilson Snyder, Ed Honaker and Troy Webb, all freed after being imprisoned for rape.

They were identified the traditional way, from photo spreads. The practice is now coming under increasing scrutiny.

Psychologists and law enforcement experts have come up with a way they believe reduces mistaken identifications in photo presentations and police lineups without reducing correct identifications.

Earlier this year, the attorney general for New Jersey called for the new method to be used statewide.

But, for the most part, police in Virginia have been slow to adopt the new procedure, even though its proponents argue that it would help to protect the public as well as the innocent.

In New Jersey, police feared they would get fewer "hits" using the system.

Barry Scheck, who, along with New York lawyer Peter Neufeld, founded the Innocence Project, agrees there will be fewer hits, but said that is because many incorrect identifications will be prevented without cutting down on correct identifications.

"Each time an innocent person is arrested, indicted, convicted or, God forbid, executed, a guilty person is on the street out there committing more crimes. It is in everyone's interest - it is just good law enforcement - to try to get this right in the first place," said Scheck.

The new technique is called sequential presentation.  It involves showing a witness a single photo- graph or a single suspect at a time rather than a traditional simultaneous presentation, which often involves showing the witness six photographs, a "six-pack," at a time.

Scheck, speaking to a group of commonwealth's attorneys earlier this year, said that, in a simultaneous presentation, "you're implicitly saying to the witness, 'Tell me which of these individuals most resembles the person who committed the crime,' and the witness is led to believe one of the six is guilty.

"If you don't have the person who committed the crime in your lineup or in your photo array . . . you may get a wrong hit.

"Now, if you use this sequential presentation technique, what the studies continually show is, you do not in any way reduce the number of correct identifications, but you dramatically drop the number of mistaken identifications.

"That is a win-win for law enforcement," Scheck said. 

Police don't necessarily see it that way.

Dana Schrad, executive director of the Virginia Association of Chiefs of Police, said most departments across the state are sticking with the traditional, simultaneous approach for now. 

Police, she said, "don't jump into the deep end on changes like this until they feel like it's accepted by the prosecutors and the court systems and that they've covered all those bases. If there's been any hesitance, it's usually because they like to err on the side of caution."

First Sgt. Roger Broadbent, of the Virginia State Police, was on a 34-member panel of experts from across the nation and Canada that researched the issue for the U.S. Department of Justice's National Institute of Justice. Their report, "Eyewitness Evidence A Guide for Law Enforcement," was issued in 1999.

While it does not recommend one method over the other, it notes that scientific research indicates that identification procedures such as lineups and photo arrays produce more reliable evidence when the individual lineup members or photographs are shown to the witness one at a time rather than simultaneously.

"Old habits are hard to change sometimes," said Broadbrent. He said that he and others in law enforcement on the panel are convinced sequential presentation is a good system that will cut down on false identifications.

"Is it the cure-all for everything? No. But I think it's something that each department needs to look at," he said.

One officer, who did not wish his department to be identified, said not enough scientific evidence has been assembled to prove the sequential method is better. He said Virginia courts readily accept the traditional method, which also has built-in safeguards.

But Pulaski Police Chief Gary Roche said his department uses both types of presentations for photo lineups. "We have not had a problem in court with either method," he said.

Lt. David L. Martin of the Richmond Police Department said it generally prefers the simultaneous presentation, in large part because the department has a sophisticated computer system that can generate a good lineup of photos.

But there are times that a witness or a victim will be handed a stack of photos. Martin said he was not aware of studies that show the sequential method produces fewer false hits.

"I haven't seen that many false hits," Martin said. "We probably have the same number of hits [both ways] and the same number of people" who are unable to spot the suspect in the photos.

Likewise, Fairfax County Commonwealth's Attorney Robert F. Horan is not aware of studies that show one method is superior to another. "I've seen them pick them both ways, and I've seen them reject them both ways," he said.

Rod Lindsay, an expert on eyewitness identification and a psychology professor at Queen's University in Ontario, Canada, is among those who have conducted tests that show how a properly performed sequential presentation cuts down on mistakes.

Lindsay used volunteers who viewed a crime enactment and then were asked to identify the perpetrator using traditional photo spreads and lineups in which the perpetrator was not present.

Some 20 percent to 40 percent of the eyewitnesses identified one of the innocent people they were asked to view.

When a sequential presentation was conducted, fewer than 10 percent of the witnesses made a mistake.

In the sequential tests, witnesses were not told how many people would be presented; the person presenting the lineup did not know which person was the suspect; and the witness had to make a clear "yes" or "no" before moving on to another photo or lineup member.

Testing also showed that witnesses who were able to correctly identify the "criminal" were able to do so equally well in both the sequential and the simultaneous presentations.

According to Lindsay, sequential lineups limit the witnesses' ability to compare suspects. Therefore, they are less likely to simply choose one who looks most like the suspect.

Dr. Ronald P. Fisher, a professor in the Department of Psychology at Florida International University, has studied eyewitness identification and agrees that the sequential presentation elicits fewer false identifications.

Fisher said he believes that eyewitness testimony carries much weight with jurors because most people tend to overestimate how often witnesses can identify another person who is not known to them from just a brief encounter.

In a criminal case, the witness, especially if a bystander, is attempting to identify someone he does not know, and that is much more difficult than recognizing a familiar face. Jurors may fail to distinguish between the witness recognizing a stranger and their own experience of recognizing a familiar face.

A second problem is that witnesses often appear to be very confident in their identification, but witness confidence is not a good predictor of accuracy.

"It is very compelling for jurors to want to believe witnesses who exude much confidence. But it is a mistake to give that much weight to confidence," he said.

Scheck said that, as of May, 92 inmates had been cleared of crimes by DNA testing. A study of 71 of their cases revealed that 84 percent were convicted on mistaken eyewitness identification.

"That's long been known to be the single greatest cause of the conviction of the innocent," he said.

Contact Frank Green at (804) 649-6340 or fgreen@timesdispatch.com

Eyewitness ID
Truth in Justice