illustration by PJ Loughran
One morning in April 1981, a Dallas woman woke to a man on top of her wielding a butcher knife. She tried to escape, but he cut her hand, both sides of her neck, and her buttocks. He then raped her and left. On two occasions, police showed the woman photo lineups. The first time she couldn’t identify the rapist; the only light in her room the day of the crime, save for dawn breaking outside her window, came from a digital alarm clock. The second time she saw a lineup, the woman fixated on Larry Fuller, a 32-year-old, two-tour Vietnam vet. For his service, Fuller received the Air Medal and was honorably discharged. After the war, he was convicted and sentenced to three years for robbery. Once released, he began to turn his life around. He pursued a degree at Dallas Baptist University and helped his girlfriend raise her two children.

In August 1981, a jury convicted Fuller of aggravated rape, based on the victim’s identification of him and a blood type that matched the semen from the rape kit. Fuller, who is black, was sentenced to 50 years in prison. In the mid-1990s, he wrote to the New York-based Innocence Project, which took on his case. In October of last year, DNA testing was conducted on the victim’s rape kit. Fuller’s DNA wasn’t found. After spending 25 years in prison, he was a free man.

In January, so, too, was Andrew Gossett, convicted of an aggravated sexual assault he didn’t commit. Later that month, James Douglas Waller was exonerated for a rape he never committed. The three became the 10th, 11th, and 12th men freed by DNA evidence in the past five years in Dallas County, which leads the United States in exonerations at the county level. Illinois and Texas tie for the most exonerations at the state level, with 26 each.

The 12th man won’t be Dallas’ last, either. District Attorney Craig Watkins, a criminal defense attorney before his election last November, will not erect the barriers to DNA testing his predecessor Bill Hill did. Watkins has permitted the Innocence Project to review 354 Dallas County convictions, dating back to 1970. As D Magazine went to press in early March, a 13th exoneration looked imminent. Watkins won’t comment on the reviews other than to say, “We need to change the focus from convictions to justice.” He supports a bill from State Sen. Rodney Ellis (D-Houston) that would establish an independent commission to review all exoneration cases, looking for similarities among them. But you don’t need a commission to find similarities. Ten of the 12 Dallas County convictions resulted from eyewitness identifications—sometimes only from eyewitness identifications.

And that’s the problem. Despite their persuasiveness in court, eyewitness accounts are notoriously unreliable. Studies repeatedly show that memory is flawed when stressed. The new DNA reversals only accentuate the implications of error.

Is there a way to fix the system so that innocent people stop going to prison based on false witness? Research from law enforcement officials and psychologists says yes. And the analyses point to the same place, the root of most of the conviction errors: the police lineup. Change three things about the lineup process, the experts say, and we can reduce the rate at which the innocent are tried and convicted by as much as 50 percent. Led by New Jersey, nine states, as well as many counties and cities across the nation, including Boston and Minneapolis, have switched procedures, variously incorporating tenets first written in a Department of Justice manual in 1999. Wisconsin and North Carolina have issued voluntary guidelines promoting the changes. The Dallas Police Department could be next.

The three lineup changes, collectively   known as the sequential, double-blind method, have a common thread: the role of memory. Among the more famous studies of how we form memories and recall them is the mock prisoner of war experiment. Yale professor Charles Morgan conducted the study in 2004, with assistance from a U.S. Army survival school. Five hundred active duty military personnel were placed in mock POW camps and interrogated, sometimes physically, for 40 minutes at a time. They were held in “captivity” for 12 hours. The interrogators made no effort to hide their facial features. But the next day, asked to choose their interrogator from a photo lineup, 68 percent of the harassed subjects picked the wrong man.

“Trauma, stress, fear, and high arousal do not help with the formation of memories,” says Gary Wells, a professor of psychology at Iowa State University and perhaps the world’s foremost leader on eyewitness identification. Wells helped develop the Department of Justice’s national guidelines for eyewitness evidence and just received the Presidential Citation from the American Psychology Association. “What happens is that a stressful, arousing event leads people to never forget that it happened,” he says. “But the details that they end up forming in their minds are the result of ‘filling in the blanks,’ which they confuse with having a good detailed memory.” These “filled-in blanks” are thought to be formed at the time of the event but in fact are not. This is why many eyewitnesses are wrong but are so sure they are right.

The eyewitness is susceptible to outside influences, too, especially the questions and comments of the person administering a lineup based on photos. In Dallas County, that’s the detective working the case. And that’s a problem, Wells says. Through no fault of his own, the detective has a vested interest in arresting a suspect: he wants to get bad guys off the street. Since he knows who the suspect is, he could, knowingly or otherwise, persuade the eyewitness. It might be as blatant as including the suspect in multiple lineups (which happened in Larry Fuller’s case). It could be as subtle as a detective staring at one photo above all others.

The first and most essential reform calls for instituting the “sequential” part of the sequential, double-blind photo presentations. In a traditional, nonsequential lineup, an eyewitness is asked to look at six photos on a piece of paper all at once—three on the top, three on the bottom. The detective says the suspect may or may not be in the lineup, and the eyewitness is under no pressure to choose anyone. He then asks the eyewitness to identify the person she saw. The problem occurs when that person isn’t there. In a 1993 lab study, Wells found that when the “perpetrator” was excluded from a lineup, 68 percent of participants identified someone else anyway, despite assurances that they could identify no one at all. Why? Because the participants in the study had matched their memories of the perpetrators against the six people in the lineup, comparing, contrasting, and ultimately skewing the image they had originally seen.

Sequential lineups alleviate this problem by showing one photo at a time. The eyewitness has only to match a photo against his or her memory, instead of the five other photos on the page. Numerous studies have shown that this results in fewer misidentifications.

The second reform, the “double-blind” part of this procedure, further insulates a witness from the arresting officer or detective. Not only is the officer with a vested interest kept away from the lineup ID, as in the aforementioned example, he is replaced during that time by an officer with no knowledge of the investigation. Bias is twice-removed.

The third reform deals with an eyewitness’s confidence. Sometimes, psychologists say, a witness can be shaky about positive identification. That is, until it’s reaffirmed by the administrator of the lineup. Then he or she forgets any previous doubts and very likely says on the stand that the ID of the accused was never in doubt. Under this change, the witness must write a statement describing her degree of confidence immediately after her selection. The result is a paper trail of certainty either the defense or prosecution could use.

The beauty of the sequential, double-blind method is that it addresses all three deficiencies in police lineups. This is why in 2001, after the state of New Jersey implemented the sequential method, other counties and cities followed—all in response to DNA exonerations. “Our judiciary basically sent a strong message to our attorney general,” says Deputy Attorney General Lori Linskey. “[Which was] ‘Find a way to make this better, or we’re going to find a way to limit eyewitness evidence.’” At first, smaller police departments complained that they didn’t have the staff for double-blind administration. Larger departments and even the attorney general’s office helped out, and today the reformed lineups are a “non-issue” with departments statewide, Linskey says. Prosecutors may like them most of all. “They take away an argument for the defense,” she says.

This is not to say sequential double-blind is without faults. A sequential presentation yields fewer identifications than a normal police lineup, and one study has claimed sequentials are no better than standard lineups. (Although that study failed to disclose its methodology to the public.) Even Wells acknowledges the sequential method “is not a silver bullet.”

Last October, the American Judicature Society proposed a pilot study in which six cities would test the sequential and standard methods head-to-head. Dallas Police Chief David Kunkle is in favor of testing it. Assistant Chief Ron Waldrop has worked on the issue since October. “We need to find out what the best practice is,” he says. The Innocence Project is looking to secure funding as early as this month for the study. It would continue until 300 lineups have been catalogued—150 using the sequential method, 150 using the standard one. If the DPD likes what it sees, it may change the way it conducts its lineups.

Ultimately, it would be a far-reaching preventive measure. The exonerated men were lucky. Their cases included DNA evidence. Lineup reforms might nip in the bud the convictions of innocent defendants in the other 90 percent of cases, the ones that don’t have any DNA to test.