Sunday, January 20, 2008

Eyewitness ID: Reliable Or Risky?

Kyle Martin

Published: January 19, 2008

The wrong place at the wrong time can bring a world of trouble.

For Joseph Caggiano, it was Dec. 17 at a Sunoco gas station, where he briefly stopped to buy a pack of cigarettes.

Within 24 hours he was the primary suspect in a string of armed robberies.

The evidence was damning.

A handprint lifted from a paper on the Sunoco's door matched his and even more important, the clerk picked him as the perpetrator out of six photos.

Caggiano turned himself in when he learned there was a warrant out for his arrest, but he clammed up and asked for a lawyer when a detective started asking questions.

In a strange turn of events, the detective came across the actual suspect in person as he continued his investigation. An arrest was made and by the next day, 19-year-old Caggiano was a free man.

But not everyone is so lucky.

There are no firm figures on how many people end up in Caggiano's situation, but consider that eyewitness misidentification accounts for 75 percent of the 210 wrongful convictions overturned by DNA evidence.

Statistics compiled by the New York-based Innocence Project also show that most victims served an average of 12 years in prison and 15 faced the death penalty before their innocence was proven.

James Doyle, author of True Witness, calls it a "tragedy without villains."

"Cops are going by the book, generally the witnesses are sincere," he said, adding that both parties "want to get the right guy" but the methodology is flawed.

Researchers urge reforms in the methods detectives use to secure eyewitness testimony. Years of experiments and scientific scrutiny of current practices show there are simple methods to cut down on false IDs.

Principal among their targets are photo packs like the one used to identify Caggiano.

But the changes are not a magic bullet either and carry their own risks. And the validity of eyewitness identification varies on a case by case basis, says Assistant State Attorney Bill Gladson.

The prosecutor has successfully argued before the Florida Supreme Court that it's up to the jury to decide whether witness testimony is credible, not an expert.

"That's what cross examination is for," Gladson said, "to flush it all out."

A case for reform?

What investigators should consider is that witness or victim identification is unlike any other evidence because it's solely in one person's mind, says Robert Shomer, an eyewitness expert who has testified in more than 400 trials.

"Under the best circumstances, (eyewitness identification) is the least reliable," he said.

In instances like the Caggiano case, the suspect pulled up his shirt to reveal what appeared to be a handgun tucked into the waistband, investigators say. Life-threatening situations further corrupt the memory, according to Shomer, in part because the weapon serves as a magnet for the eyes and draws attention away from the suspect's face.

Local defense attorney Jimmy Brown represented Caggiano and said the clerk likely picked out his client because he was a regular at the Sunoco.

"People are prone to pick out someone who looks familiar and that's what we had," he said.

Doyle's solution is for investigators to treat a witness' memory with "CSI seriousness." By his standards, memory would fall into the same category as blood and semen in that they are "very difficult to recover and easy to contaminate."

Offering True or False

As far back as the late 1800s, scholars were questioning the accuracy of eyewitness identification. The landmark publications "On the Witness Stand" in 1907 and "Convicting the Innocent" in 1932 took a hard look at witness' memories and their role in convictions.

Even legendary FBI director J. Edgar Hoover weighed in on the subject to sell the concept of collecting fingerprints as a safeguard against misidentification.

But it's over the past three decades that scientific studies really kicked off; many got a boost by the advent of DNA exonerations beginning in the early 1990s.

Researchers now suggest that photos be shown one at a time, versus spreading them all out on a table for a witness to choose from.

That compels a witness to draw on the memory for each picture instead of comparing pictures and choosing what looks best, according to Doyle.

Using that method gives "six true or false (options) instead of one multiple choice," he said.

Experts also encourage the adoption of the "double blind" method, meaning someone not tied to the investigation show the pictures. They worry that an investigator's non-verbal cues, however inadvertent, can be leading for a witness.

A sigh of relief or a pat on the back can boost the confidence of the victim from that day on through to the trial when the suspect is identified in court.

They also recommend that the investigator forewarn the witness that the actual suspect "might or might not" appear in the photos.

Shaky Results In Field Tests

At first, these methods were hailed as the cutting-edge cure for misidentification. But when the Chicago Police Department tested them out in 2004, the results revealed they were not the panacea promised.

That study found that witnesses were 9 percent more likely to choose an innocent person and only 45 percent chose the real suspect when using the sequential, double-blind method. That's compared to a 60 percent rate of choosing the right suspect when using the traditional lineup method.

Advocates of the reforms claim that Chicago's experiment was tainted because it was supervised by police officers.

Field tests in other jurisdictions have also found it's difficult to pull unbiased staff off the job to show the witness pictures in the double-blind method.

Gladson, the Hernando County prosecutor, argues that eyewitness identification cannot be completely discredited just because some deem it flimsy evidence.

He successfully convinced the Florida Supreme Court considering the appeal of a convicted murderer that the trial judge had the right to exclude an eyewitness expert.

"Lab research doesn't apply on a case by case basis," Gladson said.

County's Practice With Photo Packs

Field tests not withstanding, New Jersey, North Carolina and several smaller jurisdictions are embracing the double-blind sequential method.

At the Hernando County Sheriff's Office, the proposed reforms are followed to a certain extent.

According to the lead detective in Caggiano's case, Philip Lakin, forensics technicians pull together the photo packs and place six photos on a single, glossy sheet.

The technicians do their best to find similar-looking suspects and avoid drawing photos with different backgrounds, according to Lakin. For example, one mug shot should not be included among five passport photos.

After telling the witness the suspects' photo might not be included, Lakin said he steps out of sight when the witness is looking over the photos to avoid interfering with their decision.

The witness or victim writes the number of the suspect he has chosen and includes it in a brief, sworn affidavit. Then the whole photo pack is placed into evidence, according to Lakin.

The detective makes it clear that "photo packs do not solve cases. It's one of many tools."

Reporter Kyle Martin can be reached at 352-544-5271 or

No comments: