PITTSBURGH -- The following report by Team 4 investigator Jim Parsons first aired May 11, 2009, on WTAE Channel 4 Action News at 5 p.m.

---

How well do you trust your eyes and memory? Team 4 investigative reporter Jim Parsons put that question to the test for 21 local college students who witnessed a mock crime. Parsons didn't tell them ahead of time that the crime they were witnessing wasn't real.

Team 4 wanted to test the ability of witnesses to identify the suspect of a crime using a photo lineup. But first, we needed a crime that we could capture on videotape.

So, we set it up ourselves with the help of administration and faculty at Duquesne University. They knew what was coming, but the students did not.

We want to hear from you. Please share your feedback in the "Add your comment" box at the end of this story.

It's business as usual in a classroom at Duquesne University, but something very unusual is about to occur, and 21 senior students are not expecting it. In the next few seconds, a laptop on the podium will be stolen.

As the lecturer addresses the class, a stranger enters the room, strolls calmly toward the podium, and then...

Lecturer: "Hey, hey, hey, that's mine."

Now, the question is, how well will these witnesses remember the face of the thief who stole that laptop? Eyewitness identification is not as simple as you might think.

Kirk Bloodsworth, death row inmate exonerated and freed by DNA testing: "It's so frustrating. You're an innocent man. Nobody wants to believe you."

Bloodsworth was convicted of raping, strangling and beating to death a 9-year-old girl near Baltimore. The most compelling evidence against him was the testimony of five eyewitnesses who said they saw him with the girl. All five had the wrong man and Bloodsworth went to prison.

Bloodsworth: "Eight years, 11 months and 19 days. Every day, people hollering in the fence, 'We're going to do to you what you did to that little girl,' over and over and over, and nobody should suffer that. Nobody."

John Rago , Duquesne University law professor: "If you look at the single highest indicia of wrongful conviction error, eyewitness testimony is at the top of the scale."

Rago is an expert on wrongful convictions, and he says he is certain about two things with regard to eyewitness identifications.

Rago: "We're making mistakes with eyewitness identifications, and that there are procedures to reduce error rates."

Those procedures for reducing errors in photo lineups were spelled out in a Justice Department report 10 years ago, and some police departments across the nation adopted all of them, but many still have not -- including Pittsburgh and Allegheny County police.

We put those procedures to the test with half of the students who witnessed our mock crime at Duquesne. The other half of the witnesses was shown photo lineups the traditional way.

Parsons: "OK, I want you to take a look at the six photos that I've laid out and I want you to tell me, can you identify any of these people as the person who came into the classroom and stole the laptop?"

Student: "Yes."

The man he just fingered for the laptop theft was the same person picked by eight out of 10 witnesses in this group. The only problem is he didn't do it. Another man did, but his photo wasn't included in the photo array.

Why did they pick the wrong guy?

Student: "You're assuming that he's in the lineup at all. Process of elimination. One of them has to be the guy. So you eliminate the people you know it's not and pick whoever is left."

Parsons: "What made you make that assumption?"

Student: "I don't know. I think it's an assumption most people would make."

Rago: "That's human nature. We want to find an answer, and the studies suggest this transference takes place. One way to eliminate that is something called a double-blind sequential test."

And that's just the method we used with the other group of witnesses. That photo lineup was administered by Ben Wecht, of Duquesne's Wecht Insititute of Forensic Science and Law. Unlike my group, Wecht informed his witnesses that the real culprit might not be in the lineup.

Wecht: "The suspect may or may not be in this sequence of photos, so don't feel compelled to make an identification. Rest assured that regardless of whether you make an identification, the police will continue to investigate this case. OK?"

Instead of seeing all the photos at once, these witnesses viewed them one at a time. Experts say witnesses who view sequential lineups are more likely to compare the photo to their memory rather than to the other pictures.

Witness: "Uh, this looks like him."

This photo lineup did include the real culprit, and six out of 11 witnesses chose him. Compare that to the traditional lineup group, where eight of 10 witnesses picked an innocent man.

Team 4 showed our eyewitness identification test to Harvard-educated law professor Bruce Antkowiak, a former criminal defense attorney and federal prosecutor.

Antkowiak: "This illustrates the failings of eyewitness identification -- the potential failings of it, the fact that it is not as simple as we sometimes hope it is."

Antkowiak says it's time for police departments to accept that there is a better way to run photo lineups.

Antkowiak: "I think police officers, if they understood that, would want to make sure that their identification tactics were reliable and would give to them the best possible identification testimony that they could get."

A change in lineup techniques is Bloodsworth's fondest hope.

Bloodsworth: "The biggest thing in this, Jim, is I do not want this to happen to anybody else, whether it's in Maryland, Pittsburgh or any other place."

Pittsburgh police and Allegheny County police say they have no written policy on administering photo lineups. Detectives are free to decide from one case to the next whether to show suspect photos all at once or one at a time.

We asked both departments why they do it that way and why they haven't implemented all of the Justice Department guidelines. Pittsburgh police did not answer that question. County police Superintendent Charles Moffatt told us that he believes it is time to revise the county policy on conducting photo lineups and that he intends to look into it.

Sen. Jim Ferlo has authored a bill that would require police departments across the state to implement some of the Justice Department guidelines, and another bill being considered by the Legislature would direct Pennsylvania courts to instruct juries about the potential fallibility of eyewitness testimony. Currently, that is not done in Pennsylvania.