Featured Stories

Other Pamplin Media Group sites


Best fix for photo lineups? Police, ID experts disagree

Techniques revised to more correctly pick out criminals


TRIBUNE PHOTO: JAIME VALDEZ - Portland police detective Joe Santos shows a folder with one photo that will be one of six folders presented to a witness. In the past, police showed witnesses six packs--one sheet of paper with six photos. But that led to misidentifications, according to experts.Forgive Sgt. Joe Santos if he’s a bit confused. The Portland Police Bureau detective has been asking witnesses to identify suspects for 16 years.

He wants to get it right: arrest the perpetrator, get the witness to objectively identify him or her, and conduct an interrogation without fear a defense attorney will later say he coerced a confession.

But it’s getting harder.

In recent years, a growing body of research has begun to show that an almost incomprehensible number of suspects confess to crimes they didn’t commit. Furthermore, research on the reliability of eyewitnesses’ memory is casting doubt on another of the tools in prosecutors’ arsenal. Even the slightest, most nuanced police instructions to eyewitnesses can affect what they do or say, according to experts — and the experts don’t always agree.

New job, new procedures

Last October, Multnomah County District Attorney Rod Underhill assigned long-time deputy Russ Ratto to a new position: post-conviction deputy. Ratto’s job is to look at past convictions to make sure they were properly obtained, and develop protocols to ensure that in the future the DA does not pursue convictions that later could be overturned.

The purpose, Underhill says, is to figure out what work police and prosecutors can do up front so they don’t need to defend convictions later. The timing, with the newly formed Oregon Innocence Project starting to review cases of Oregon prisoners who might have been wrongfully convicted, is not coincidental, Underhill says. In fact, Innocence Project attorneys have been working with Multnomah County prosecutors in their review of police procedures.

But Ratto has found, like Santos, that turning the latest research into on-the-ground protocols that every cop can follow isn’t easy. Witness identification, for example. For years, Portland detectives showed witnesses what they call a “six-pack” lineup of photographs. One of the pictures showed their suspect.

That changed two years ago when Portland police, with the help of Reed College psychology professor Dan Reisberg, author of “The Science of Perception and Memory,” adopted a new procedure based on the latest memory and identification science. Now police show each of six photographs one at a time.

The idea is to avoid what Reisberg calls “comparison shopping.” Witnesses shouldn’t be comparing the photos at all, they should be determining if one or none of the six is the person they saw at the crime scene. But human nature being what it is, witnesses fall into all sorts of traps that police should try to avoid setting, Reisberg says.

But even sequential photo lineups present problems, at least according to lab studies. By the time witnesses get to the fourth picture without selecting one as the offender, they start feeling pressure to choose somebody. So their subconscious makes them more likely to choose one of the photos later in the sequence.

The science says comparison shopping is minimized if witnesses are instructed to say yes or no to each photograph before they are allowed to see the next one, according to Reisberg.

Finally, Reisberg says research indicates police should not show a lineup of photos more than once because by the second or third run through, witnesses are comparison shopping. More laps result in more picks, but many of them are the wrong picks, Reisberg says.

High stakes, lower identifications

The stakes in the ID process are high — sometimes the difference between a suspect spending his or her life in prison or not. But the research doesn’t always work so well in the real world, says detective Santos. For instance, consider what often happens, according to Santos, after a witness has been told not to make an identification until he or she has viewed all six photos.

“You’ll say, ‘Which one is it?’ and they’ll say, ‘I don’t know, but I saw him,’” Santos says.

Defense attorneys argue that if witnesses can’t pick out a photo after looking at it one time, the perpetrator probably isn’t in the lineup. Reisberg says state rules enacted two years ago (see sidebar) “completely botch the sequential lineup. It’s doomed from the start.”

Detective Santos says state rules aren’t practical, especially when witnesses say they can’t remember which of the six photos looked like the perpetrator.

State guidelines be damned, Santos says. Portland detectives are starting to leave out the “wait until you’ve seen them all” instruction — which Reisberg says is a good idea. But when a witness says he or she can’t recall which photo looked right, Portland police are showing the six photos again, and they’re allowing the witness to stop and point when they reach the photo that they think matches their memory. Bad protocol, Reisberg says. Regardless, police are working with the Multnomah County District Attorney to draft new witness ID protocols to use in the future.

All this science-informed police procedure has certainly had an impact. Santos estimates that since Portland police abandoned six packs and started sequential lineups, witnesses identifications have dropped at least 50 percent and probably more. The first 30 times police used sequential photo lineups, witnesses failed to choose any picture. Previously, Santos says, about one in three witnesses would at least choose somebody.

“Even when the right guy is in there, they don’t see it,” DA Underhill says.

On the other hand, the new techniques may be doing their job: eliminating wrong IDs.

“When they pick someone now they’re more certain,” Santos says.

That’s precisely what should be happening, according to Reisberg. “You lose some good IDs at the same time you avoid a much larger number of bad IDs,” he says.

Worth it? Reisberg thinks so, since research shows the number of missed identifications is small compared to the wrong identifications avoided. He quotes 18th-century English jurist William Blackstone: “It is better that 10 guilty persons escape, than that one innocent suffer.”

“If you use the Blackstone rule the sequential lineup looks really good,” Reisberg says.


Lawson case sparked Oregon change

Oregon is ahead of most states in trying to eliminate bad witness identifications, according to Reed College psychology professor and memory expert Dan Reisberg. But, Reisberg says, it took a high-profile, wrong ID murder case to get there.

Last year, Samuel Lawson was released after spending nine years in prison on an aggravated murder charge. He had been convicted in the death of Noris Hilde in rural Douglas County. Hilde was shot and killed in the summer of 2003. His wife, Sheri Hilde, identified Lawson as the perpetrator, but later court actions showed that the process that led Hilde to make that identification was greatly flawed, and without it there was not enough evidence to convict Lawson.

Reisberg says that sheriff’s deputies influenced Sheri Hilde to identify Lawson. Though Lawson was convicted, the Oregon Supreme Court in 2012 overturned the conviction, released Lawson and set in motion a process that established statewide guidelines for police to follow when conducting witness identifications — the same guidelines many Portland detectives find ill-conceived.

Today in Oregon, largely because of the Lawson case, the burden of proving a police identification procedure has been properly conducted falls on prosecutors. In most states, Reisberg says, it is up to defense attorneys to prove police procedure was flawed.


Sure today, but how about tomorrow?

So you’ve picked your perp out of a police photo lineup. But trial could be months away. How confident you feel and remain about your identification could be crucial to the case. According to Reed College memory expert Dan Reisberg, there is very little connection between how you felt when making your choice and the likelihood that your choice is correct.

“People will commonly say things like, ‘That event is burned into my brain.’ And those common-sense statements are wrong,” Reisberg says. “Stress is bad for memory, not good. And memory for highly stressful events is worse, not better.”

Incorrect witness identifications have a role in nearly seven in 10 wrongful convictions, Reisberg says. Sometimes, a bad ID has nothing to do with how a police present a lineup. For instance, studies show that cross-race IDs are particularly unreliable. Whites tend to pick the wrong photo when looking at a lineup of black faces and blacks often misidentify whites.

But what police do or don’t do affects a witnesses’ confidence level, Reisberg says. An average witness will report a confidence level of about 40 percent after making an identification. But if an officer finishes up a photo lineup by saying, “You’ve identified the suspect,” witnesses report an 80 percent confidence level.

Reisberg recently testified in an Idaho case in which four people were convicted 12 years ago of carjacking and assault. The conviction, he says, rested largely on an eyewitness identification by the victim. But the ID was made almost two years after the crime and in the period between the crime and the identification, the case received national attention, including a television re-enactment on “America’s Most Wanted.” The victim, Linda LeBrane, even participated in the re-enactment.

In addition, Reisberg says, police failed to document precisely how they conducted the original identification, and LeBrane confessed to having smoked marijuana at the time of the assault. Reisberg says, she was “completely traumatized” by the event — which could affect her recall.

According to Reisberg, the victim admitted to police that by the time of her identification, she was having trouble keeping straight which faces were in her memory because of the crime and because of the TV re-enactment. And yet, Reisberg says, the original judge in the case allowed the identification into the record.

Sarah K. Pearce, one of the four people convicted in the crime, has been released without her conviction being overturned. Reisberg says the case is a classic example of why police need to conduct eyewitness identifications that are above reproach.

Portland detectives don’t give hints, says Sgt. Joe Santos, a Portland police detective. Each of the six photographs is handed to the witness in a separate envelope and the six envelopes have been shuffled so even the detective doesn’t know which photo is being viewed at any moment.

“We absolutely are trained to not say, ‘Hey, good job,’ or, ‘You got the right person,’” Santos says.

But making a wrong ID does more than let down police. “Most of the wrongful convictions are a double error,” Reisberg says. “You’ve got somebody in jail, and the real culprit is still out there.”

JW_DISQUS_ADD_A_COMMENT