On the afternoon of June 29, 1998, Trina Gomez and Maximilian Patlan were closing a branch of Fidelity Financial Services in Fullerton, California, when two men knocked on the door. They asked if they could make a payment, so Patlan let them into the bank. Once inside, one of the men punched Patlan in the face and ordered him to lie on the ground. The other man pulled out a shotgun and forced Gomez to collect all of the available cash. Minutes later, the two thieves walked out of Fidelity with more than $5,400 in cash and personal checks. The armed robbery and assault happened so quickly that Gomez and Patlan later had difficulty identifying the perpetrators.
A few months later, police arrested Bernard Teamer as the getaway driver in the robbery after a witness identified him. Police, however, had no leads on the two men who actually robbed the bank. So, they began to investigate all of Teamer’s acquaintances and neighbors who fit the description that Gomez and Patlan had provided. The investigators eventually whittled the group down to eleven suspects.
Police used photos of these eleven people, along with so-called “filler” photos of 43 other men who were not believed to be involved in the crime, to assemble nine photo lineups for Gomez and Patlan to examine. The witnesses made nearly a dozen tentative identifications, but ultimately decided that Guy Miles, a 34-year-old black man, was one of the robbers.
The rest of the case against Miles, however, was flimsy. Six people provided alibis for him, swearing that when the robbery took place he was at his home in Las Vegas — 250 miles from Fullerton. There was no physical evidence, such as fingerprints or DNA, linking him to the crime. And although Miles and Teamer had been members of the same street gang, they both said they had only met once — about ten years before the Fidelity robbery.
The photo lineups that police put together also contained several major flaws. Only a handful of the filler photos matched the robbers’ general description, making it easier for Gomez and Patlan to select a suspect who looked like one of the men who robbed the bank. And the lineup test was administered by the investigating officer, who knew the identities of the eleven suspects, and thus could have influenced Gomez and Patlan to pick out at least one of them.
“It was almost as though [police] didn’t care who [Gomez and Patlan] picked, just as long as they picked someone,” said Alissa Bjerkhoel, an attorney with the California Innocence Project, a legal group that works to exonerate people who have been wrongfully convicted and is now working to free Miles.
Throughout the trial, Miles fervently maintained that he was innocent. But Gomez and Patlan’s testimony ultimately convinced jurors of his guilt, and he was convicted of two counts of second-degree robbery and one count of being a felon in possession of a firearm. Because of his prior criminal record and California’s harsh sentencing laws, Miles received a 75-year-to-life sentence.
“It all seemed surreal, like I was beside myself … watching a bad scene in a movie,” he recently wrote to me in a letter from San Quentin State Prison. “It was the most helpless feeling I ever experienced in my life.”
In 2009, a decade after Miles was sentenced, the California Innocence Project, which had agreed to investigate his case, unveiled strong evidence that Gomez and Patlan had identified the wrong man. In fact, two other people confessed that they had robbed Fidelity, and Teamer eventually swore under oath that Miles was not one of his accomplices. Nonetheless, Gomez and Patlan remain steadfast in their belief that Miles is guilty, and so he remains behind bars for a crime that, in all likelihood, he did not commit.
Contrary to popular belief, people are often sent to prison based solely on eyewitness testimony, and forensic evidence is seldom used to solve most serious crimes. Unfortunately, numerous studies have shown that eyewitness memory can be unreliable. One in-depth study found that witnesses identify innocent people as being guilty in nearly two out of every ten cases. And of the 310 people who have been freed nationwide using DNA evidence, roughly 230 of the false convictions were based on incorrect eyewitness identifications, according to data collected by the Innocence Project.
While some witness mistakes can be attributed to how the brain stores and retrieves information, many false IDs are the product of poor police practices. “We have to treat eyewitness memory like a crime scene, which can be trampled through by the wrong practices,” noted Maitreya Badami, a lead attorney with the Northern California Innocence Project.
For decades, experts have been pushing for reforms that would make IDs more accurate. The National Institute of Justice, a branch of the US Department of Justice, believes that so-called “double-blind sequential lineups” produce the most reliable identifications. Double blind means that the officer administering the lineups doesn’t know the suspect’s identity so he or she cannot influence witnesses. Sequential lineups involve police showing photos to a witness one at a time rather than six at a time, a procedure that allows a witness to compare each photo to his or her memory rather than to the other five photos.
According to the US Department of Justice, double-blind sequential lineups are twice as reliable as standard lineups. “It’s a much more careful and forensically appropriate way to collect eyewitness evidence,” explained Badami.
In recent years, several Bay Area police departments — including those in Richmond, Pleasant Hill, and San Francisco — have adopted the policy of using double-blind sequential photo lineups. All police agencies in Santa Clara County also use double-blind sequential lineups, and police agencies throughout Alameda County are in the process of adopting them. Police agencies throughout Contra Costa County are also considering making the reform.
This trend follows in the footsteps of the landmark legal principle formulated by the famous 18th-century English judge and jurist Sir William Blackstone, who once said, “It is better that ten guilty persons escape than that one innocent suffer.” That idea — that people must not be imprisoned for crimes they did not commit — has guided Western jurisprudence for more than 250 years, and forms the basis of legal protections enshrined in our criminal justice system.
And yet most California police agencies still oppose double-blind sequential lineups. They contend that these procedures make witnesses more hesitant to identify suspects, and thus would result in fewer guilty criminals going to prison. Lobbying by powerful law enforcement groups in Sacramento also has beaten back legislation that would make double-blind sequential lineups mandatory statewide.
As a result, some innocent people will continue to suffer so that the guilty do not escape punishment.
Traditionally, people have likened human memory to a tape recorder: The brain records an event, and when someone wants to recall it, all they need to do is hit the play button. The human mind, however, is much more complex than that. Through decades of research, scientists have discovered that we store memories of past events, also called episodic memories, in multiple parts of our brain. The hippocampal complex and amygdala process emotional responses, the prefrontal cortex helps assign meaning to past events in our lives, and visual memory is created by electrical signals jumping between multiple parts of the brain. When someone wants to recall a memory, neurons in all of these segments of the brain are triggered and sent down a specific neural path.
This nuanced process is also incredibly susceptible to error. For example, psychologists have used minor verbal cues in crime-scene simulation studies to convince witnesses that clean-shaven men had mustaches, straight hair was curly, and that a stop sign was a yield sign. In other experiments, scientists have planted into people’s minds false memories — sometimes as traumatic as an animal attack or a hospital stay, according to research compiled by UC Irvine social scientists. As Stanford University neuroscientist Anthony Wagner stated at a 2012 symposium on memory and the law: “The normal functioning of the brain allows for false memories to be created.”
Witnesses, as a result, often accuse innocent people of committing crimes. Moreover, research shows that once witnesses misidentify a suspect as being the perpetrator of a crime, they often refuse to change their minds — even if conclusive evidence eventually surfaces that proves them wrong. “The witness is not lying; they honestly believe they are identifying the right person,” noted longtime Alameda County defense attorney Charles Denton. “The problem is their own eyes can deceive them.”
Furthermore, jurors often give great weight to eyewitness identifications — even when other evidence contradicts it. “Once a person has made an incorrect identification, it is very difficult to show a jury that they are probably wrong,” said Badami.
Defense attorneys sometimes put experts on the stand to testify about the different factors that can cause witness mistakes. Jurors, however, typically disregard such testimony. “When you’re looking at a person who appears to be telling the truth, it’s very hard to take the generic instructions that they should ‘take into account that it was dark,’ or that ‘they should take into account that it was foggy,’ because you still have the witness saying, ‘I’m sure that it was him,'” said Badami.
Yet despite these problems, police and prosecutors continue to rely heavily on witness testimony. Richmond Police Lieutenant Andre Hill told me that witness and victim identifications form the “basis” for solving the “majority of shooting and violent crimes in the city.”
The Oakland Police Department also relies heavily on eyewitness IDs to solve violent crimes. The department’s crime lab has traditionally been overwhelmed with work, so even when there is reliable physical evidence at a crime scene (which there usually is not), it’s often not tested.
“We have many violent crimes, [even] attempted-murder cases, which are charged solely on the basis of a positive photo identification, and they are among the most unreliable forms of evidence that we see in the criminal justice system,” said longtime Alameda County Assistant Public Defender Ray Keller, who retired earlier this month. Keller added that because eyewitness testimony carries so much weight in the legal system, when a positive identification is made, police often consider the case to be closed. It effectively “cuts short a thorough investigation,” he said.
This past year, the Northern California Innocence Project helped exonerate two Oakland men who were imprisoned due to faulty eyewitness identifications. Johnny Williams was freed in March after being locked up for thirteen years for child sex abuse. His conviction was based almost entirely on a false ID made by a nine-year-old girl. DNA evidence later proved that the girl had made a mistake. Ronald Ross was also recently freed after being imprisoned for seven years for attempted murder. In that case, one of the key witnesses admitted that an OPD officer had pressured him to identify Ross; the witness had owed the officer a favor so he did what was asked of him.
In the case that ultimately put Guy Miles in prison, police used an identification process that is commonly employed throughout California. Typically, it involves six-person photo lineups, in which officers show witnesses a single document that includes a photo of a suspect along with five filler photos. According to the Berkeley Police Department’s lineup instructions, which are standard in most California cities, the fillers are supposed to share similar physical characteristics with the suspect, and the officer administering the ID test is instructed to have minimal verbal or physical contact with the witness. The BPD’s lineup instructions state that these guidelines are meant to “reduce the likelihood of prejudicial suggestion.”
However, there are still countless examples of police influencing witnesses. UC Irvine professor Elizabeth Loftus, who is a leading expert on eyewitness ID procedures, told me that she testified in a case in which a witness was unable make an identification, and “the officer administering the [test] said, ‘What do you mean? I see your eyes drifting down to number six.’ And before long the witness was identifying number six.”
According to attorneys with the Innocence Project, there have also been cases in which officers have tapped on the suspect’s photo, scowled after one of the fillers was picked, and suggestively marked the photo spread. In one rape case, the suspect’s photo was marked with a bold “R,” while the filler photos had no markings.
Such practices can influence a witness’ identification and potentially foster a false sense of certainty in them. One of the best-known examples of mistaken identity occurred in 1984, when Jennifer Thompson-Cannino accused Ronald Cotton of raping her in her bedroom in Burlington, North Carolina. Thompson-Cannino was initially unsure that Cotton was the man who raped her, but over the course of the investigation, she became “absolutely, positively, without-a-doubt certain” he was the man who raped her, she stated in a 2009 book she co-wrote, Picking Cotton. Thomspon’s confidence likely stemmed from feedback that police gave her during the investigation, and her testimony persuaded the jury to convict Cotton. In 1985, the judge sentenced him to life plus fifty years. In 1987, Cotton secured a new trial after an appellate court overturned his conviction. However, during the appeal Cotton was accused of raping a second woman, Elizabeth Watson. A jury convicted him of both rapes and he was sentenced to two life terms plus 54 years.
Eventually DNA evidence was used to exonerate Cotton. The actual perpetrator had gone on to rape six other women. “It’s a human system,” Thompson told reporters after Cotton was freed. “We are fallible. We make mistakes.” Thompson and Cotton now speak publicly together about reforming eyewitness identification procedures, and Cotton was Thompson’s co-author on Picking Cotton.
East Bay police agencies routinely use another lineup procedure that many experts contend leads to false IDs as well. As of 2010, the most commonly used lineup in Alameda County was “a field showup,” a highly suggestive procedure in which police show a witness a single suspect — who is oftentimes handcuffed in the back of a police car — and ask them, “Is this the person you saw?” Although field showups allow a witness to ID a suspect when the crime is still fresh in his or her mind, the process is inherently problematic. It implies that “‘we just got someone that we think did it, do you think he did it, too?'” explained Denton. “That can have a very powerful effect on some people’s identifications.”
The US Supreme Court has declared field showups as being dangerously suggestive, the US Department of Justice has recommended regulating when they can be used, and, in 2011, the California Legislature considered changing state law so that police officers could only conduct showups if they did not have probable cause to arrest a suspect. Still, showups are a cornerstone of most criminal investigations in Alameda County, and Assistant District Attorney Paul Hora told me that District Attorney Nancy O’Malley has no plans to ask East Bay police agencies to stop using them.
Our memories often fail us. We forget where we put our keys, we distort the details of a conversation, and, sometimes, entire events in our life seem to disappear completely. We sometimes forget things because we don’t have the right physical or verbal cues to spark our memories. The same is true with criminal investigations: When a witness is unable to recall the details of a crime, sometimes a police investigator just needs to ask the right questions or show the right photos and the witness’ memory will sharpen. Conversely, if the investigator asks the wrong questions or shows the wrong photos, it can introduce errors in the witness’ memory.
According to many social scientists, displaying photos one at a time rather than six at a time provides the most effective visual cues for witnesses. The theory is that when people view photos in six-packs it triggers a mental process called “relative judgment,” in which the witness compares each of the photos against each other and not against his or her memory of the crime. “Our minds cannot help but look at it as though it were a multiple-choice test,” said Badami. “We are looking for an answer. We are looking for who looks the most like the perpetrator.”
This mental process is innocuous when the true perpetrator is in the lineup. But if he or she isn’t in the lineup, the witness sometimes picks whoever looks the most like the perpetrator.
When witnesses view photos one at a time, scientists believe it sparks something called “absolute judgment,” in which the witness weighs each photo against his or her own memory of the crime. An extensive 2006 field study conducted in Greensboro, North Carolina by the American Judicature Society found that witnesses identified innocent people as being guilty 18.1 percent of the time when shown photo arrays in standard six-packs. In sequential photo lineups, the false ID rate was 12.2 percent.
In recent years, Georgia, New Jersey, North Carolina, and Wisconsin have started using double-blind sequential lineups. Some cities such, as Dallas, Seattle, and Denver, have also adopted the procedure. However, as of 2010, only 8 percent of California’s police departments used double-blind sequential lineups, most of which were in the Bay Area.
In 2011, San Francisco Democratic Assemblyman Tom Ammiano introduced AB 308, which would have made double-blind sequential lineups mandatory throughout the state. The bill was based on recommendations from the National Institute of Justice and the California Commission on the Fair Administration of Justice and was supported by a consortium of defense attorneys, civil liberties groups, and social scientists. But the legislation died in the state senate.
Law enforcement groups played a key role in sinking AB 308, arguing that some witnesses have trouble identifying any suspects when viewing double-blind sequential lineups, thereby allowing perpetrators to avoid justice.
In recent years, a handful of social scientists have also begun to question the benefits of double-blind sequential lineups. Relatively new research by Steven Clark, a professor of psychology at UC Riverside, found that, under some conditions, double-blind sequential lineups can lead to far fewer correct IDs. “So the question is: How do we evaluate the tradeoff?” Clark said. “How many correct IDs are we willing to lose in order to protect the innocent?”
Most criminal justice experts, however, continue to support double-blind sequential lineups, even though it might reduce the total number of positive identifications. “We have often decided as a society [that] it is a far worse error to convict an innocent person than to let a guilty one go free,” noted UC Irvine professor Loftus.
Every year, the California Innocence Project (CIP) receives between 1,500 and 2,000 letters from inmates who claim they’ve been wrongfully imprisoned. Due to the group’s shoestring budget and largely volunteer staff, the CIP can only investigate a very small percentage of these cases. “In terms of innocence, we’re more cynical than prosecutors a lot of days. We have to be, we can only take so many cases,” explained Justin Brooks, a professor at the California Western School of Law and the project director of the CIP.
In 2002, after being sentenced to 75 years to life in prison, Guy Miles wrote a letter to the Innocence Project asking for help. Among the thousands of queries that the group received that year, Miles’ plea stood out. “It’s relatively rare that our office comes across a true case of wrongful conviction,” said Bjerkhoel. “It’s hard to get a better innocence case than Miles’.”
After a seven-year investigation, the CIP uncovered convincing evidence that Miles had been wrongly convicted. The group found Jason Steward and Harold Bailey, who then both confessed that they had robbed Fidelity Financial Services alongside Bernard Teamer, who had previously been convicted as the getaway driver in the crime. All three men said Miles had nothing to do with the heist.
Initially, Steward didn’t want to discuss the robbery with CIP attorneys. He had recently been released from prison, and he thought that he could be recommitted if he admitted to the crime. But he eventually agreed to testify on Miles’ behalf. No one promised him anything in return. In fact, his testimony only could have hurt him. “Regardless of the consequences, he came forward because he genuinely felt the need to correct the situation. He knew the truth, he knew who was involved, and he knew Miles was not one of the robbers,” the CIP attorneys wrote in their petition to overrule Miles’ conviction. In his declaration, Bailey wrote, “[I]t was mind boggling knowing that [Miles] had so much time for something he did not do.”
In 2010, the CIP filed a Petition for Writ of Habeas Corpus in Orange County Superior Court alleging that the new evidence undermined Miles’ conviction. An evidentiary hearing commenced in May 2011, at which Steward gave a detailed description of the robbery: He knew which employee opened the door; the fake name that the robbers had used; and the amount of money that they took from Fidelity. He also drew a precise map of the Fidelity branch.
Teamer also testified that Miles wasn’t involved in the robbery. He had initially kept quiet about the crime to avoid being labeled a “snitch,” but he had turned his life around after being released from prison, and he decided that it was best to come clean. He told CIP attorneys that he hoped his testimony would rectify the damage he had caused and that one day Miles would forgive him “for his actions and his silence,” according to court documents.
Bailey did not testify at the evidentiary hearing, but he wrote and signed a sworn declaration admitting to the crime and absolving Miles of any involvement in it. Bailey stated that he thought long and hard about coming forward, and ultimately agreed to because he “never wanted what happened to Mr. Miles to happen to anyone else.”
California State University Los Angeles professor Mitchell Eisen, an expert on memory and lineup procedures, also testified that the witnesses’ identifications — which were the only pieces of evidence linking Miles to the crime — were deeply flawed. For starters, one of the witnesses, Maximilian Patlan, initially told investigators that he wouldn’t be able to recognize the man he described to police as the “stocky” robber. But then he later identified Miles in the courtroom as that robber. The other witness, Trina Gomez, had identified Miles in the photo lineup, but was unable to identify him in the courtroom during the original trial. “I’m sure that that’s him in the photo, but I’m not sure if that’s him over there,” she said, while standing at the counsel table and scrutinizing Miles from various vantage points. “He looks different.”
However, after Gomez spoke with a prosecutor during a courtroom recess, she returned to the witness stand and pointed her finger at Miles saying, “That’s him.”
“Police and prosecutors could have reliably ruled me out as a suspect through proper handling [of the case],” Miles wrote in a letter to me from prison. “But their actions demonstrated a bias to win.”
Aside from the flawed identification process, numerous psychological factors also could have influenced Gomez and Patlan’s memories. The use of a gun in a crime can cause a perception problem called “weapons focus,” in which the witness concentrates on the weapon rather than the perpetrator. Two guns were used in the Fidelity robbery.
Studies have also shown that witnesses have a harder time distinguishing people across racial lines. According to data collected by the Innocence Project, about 40 percent of overturned convictions have involved cross-race identifications. In the Miles case, both Gomez and Patlan made cross-race identifications: He is black and they are not.
Yet despite the compelling evidence of innocence, Judge Frank Fasel denied Miles’ habeas petition. Judge Fasel wrote that the evidence presented did not “undermine the prosecution.” He also cast doubt on Steward, Bailey, and Teamer’s declarations because of their previous criminal histories and gang ties.
It should be noted that Judge Fasel, who also presided over Miles’ original criminal trial, has a surprisingly high reversal rate. While the average judge in California sees about 5 percent of his or her convictions overturned, Fasel has had roughly 15 percent of his convictions overturned by appellate courts, Bjerkhoel said. In other words, he’s three times more likely to make a reversible courtroom mistake than other judges.
Although neither Gomez nor Patlan testified at the evidentiary hearing, they both continue to stand by their identification of Miles as one of the bank robbers, and their testimony is keeping him behind bars. As Loftus wrote in her 1979 book Eyewitness Testimony, “[T]there is almost nothing more convincing than a live human being who takes the stand, points a finger at the defendant, and says, ‘That’s the one!'”
In a last-ditch effort to free Miles, the CIP asked Governor Jerry Brown to grant him and eleven other inmates clemency. In recent decades, governors have been reluctant to pardon prisoners. Arnold Schwarzenegger, Gray Davis, and Pete Wilson pardoned a combined 28 inmates during their two decades in office. By contrast, Brown pardoned more than four hundred people during his first stint as governor in the 1970s and the early ’80s. “Brown is interested in clemency — he really does want to find the good cases so that he can grant it,” said Bjerkhoel.
To draw attention to his case, Miles’ attorneys walked 712 miles, from San Diego to Sacramento, where they met with the governor’s staff. “We have been banging our heads against the wall for more then a decade, and we’re finally fed up with how the justice system works,” said CIP attorney Brooks, who’s the lead attorney representing Miles. “It’s time to try something different …. If we can get the governor’s attention, maybe we can fix this.”
The attorneys walked for 55 days, and along the way they spoke to people about other potentially innocent men and women still languishing in California prisons. “I can assure you I would not walk 712 miles for Miles if I did not know he was 100 percent innocent of this crime,” said Bjerkhoel.
The group arrived in Sacramento at the end of May, and presented their case for clemency. Top-level officials in the Brown administration were receptive to their appeal. However, California Department of Justice officials are currently working through a backlog of hundreds of clemency petitions, so it could be some time before they review Miles’ case.
Because of the devastating impacts of inaccurate eyewitness identifications, many politicians and academics are trying to educate the public about best-practice lineup procedures. Earlier this year, Assemblyman Ammiano introduced AB 604, which would require courts to educate jurors about the problems inherent in standard lineups. “I would love to see double-blind sequential ID procedures adopted throughout California,” Ammiano wrote in an email to me. “But resistance to change is difficult to overcome. So, we have to approach the changes slowly and carefully.”
Defense Attorney Denton echoed this sentiment. He does not believe that AB 604 is a panacea, but is a step in the right direction. “Social science has come a long way in identifying the problems with eyewitness identification, and law enforcement has lagged behind those advances,” he said. “This bill seeks to make up some of the gap.”
Many civil rights lawyers and social scientists would like to make double-blind sequential lineups mandatory in California. However, inertia is a powerful force and many police agencies don’t want to overhaul their long-standing practices based on lab studies that haven’t been extensively tested in the real world.
Proponents of double-blind sequential lineups, however, often point to Santa Clara County’s experience as proof that the procedure works. In 2002, all police agencies in the county implemented double-blind sequential lineups, and they have reported positive results ever since. “We are very pleased with the protocol, and I think our police chiefs are as well,” said District Attorney Jeffrey Rosen. “This reform has worked well in the big departments and the small departments, in the urban departments and in the rural departments …. Everyone should want to do this because it is the best way to do things.”
Editor’s Note: The original version of this story mistakenly stated that the California Innocence Project helped exonerate Johnny Williams and Ronald Ross. It was actually the Northern California Innocence Project — which, it should be noted, is not a branch of the California Innocence Project but its own independent entity — that helped exonerate the men. In addition, it was a nine-year-old girl, not a six-year-old girl, who made the false ID that helped convict Williams. This version has been corrected.