KANSAS CITY, Mo. – A photo of Richard Anthony Jones was the 202nd needle in a haystack of mug shots of black men named Richard or Rick.
In 1999, a self-described crackhead pointed at that photo and fingered Jones as the perpetrator of a robbery three months earlier, setting in motion what would become a 17-year nightmare for Jones – and a textbook example of the now well known dangerous unreliability of eyewitness testimony and flawed police photo lineups.
"Since that questionable identification, the police never looked at another suspect," attorneys for Jones said in court documents. "Mr. Jones was the victim of an unnecessarily suggestive police lineup, which is what the other witnesses identified him from."
The victim of the robbery, who along with others identified Jones as the suspect, calls the situation tragic. But for criminal justice experts, it's hardly surprising.
Eyewitness identification is the leading cause of wrongful convictions. Nationwide, of more than 300 wrongful convictions later overturned by DNA evidence, mistaken eyewitness identification played a role in about 71 percent of the cases, according to the Innocence Project.
More police departments around the country began reforming their procedures around eyewitness identifications in the last decade as the research surrounding criminal eyewitnesses has become accepted, proven science.
No less an august organization than the National Academy of Sciences has studied the eyewitness identification issue, and in a 2014 report, it raised concerns about how photo lineups are presented to witnesses.
"Research has consistently shown that the accuracy of these lineups can be skewed or influenced based on how lineups are presented, the type of presentation, how similar the suspect and non-suspects look in the lineup, where the suspect is placed in the presentation, nature of the instructions and any feedback given to the eyewitnesses before or after the identification," according to the report.
Continuing concerns about eyewitness reliability prompted the Kansas Legislature last year to require police departments to adopt protocols for conducting witness lineups. A Missouri lawmaker has pushed reforms but to no avail.
The number of states adopting statewide standards now numbers 19, many of them joining the list since the National Academy of Sciences study, said Michelle Feldman, legislative strategist for the Innocence Project in New York.
"The science is settled," she said.
Dozens of exoneration cases are featured on an Innocence Project website dedicated to Eyewitness Identification Reform.
Among them was a case decided just last month in Indiana. William Barnhouse was exonerated by DNA testing after serving 25 years in prison for rape. The victim in that case had identified Barnhouse as her attacker.
In Missouri, every case involving a DNA exoneration was based on eyewitness identification, according to the Innocence Project.
One exoneration came in 2013 when another Kansas City man was released from prison after DNA testing proved that someone else committed the rape he was convicted of committing.
Robert E. Nelson was charged after the rape victim "tentatively" identified him in a video lineup. She later testified that she was positive after hearing his voice and seeing him in person.
In Kansas, the state's first DNA exoneration case came in 1992 when Joe C. Jones was released from prison after serving seven years in prison for the rape of a woman in Topeka.
The victim and two other witnesses had identified Jones as the attacker.
Even in states that have yet to legislate reform, law enforcement agencies are establishing reform policies, defense attorneys have more ammunition when eyewitness testimony is weak and judges are throwing out bad identifications, Feldman said.
"They're sending the message to law enforcement that they have to follow these practices," she said. "It's in the water."
The enlightenment and subsequent reforms came years too late for Richard Jones, now 41. He lost the opportunity to watch his children grow up.
"When it comes to my kids, it's been a rough ride," Jones told the Kansas City Star.
Jones spent Memorial Day 1999 in Kansas City, celebrating the holiday and his girlfriend's birthday with family and friends.
Across the state line in Kansas City, Kan., three admitted crack cocaine users were driving around short of money and looking to get more drugs.
Near a known crack house, they picked up a man named "Rick" who had them drive him to a Walmart in Roeland Park. There, Rick tried to steal a woman's purse, but she fought back, and he escaped with only her cellphone.
Just over four months later, police arrived at Jones' door and arrested him.
Despite the absence of physical evidence and an alibi backed up by several witnesses, a Johnson County jury found Jones guilty of the crime.
Only the testimony of eyewitnesses who identified Jones as the robber put him in prison.
And it was only through the dogged persistence of Jones and his attorneys that the testimony of those witnesses was found to be woefully flawed.
Two weeks ago, a Johnson County judge vacated Jones' conviction, and on June 12, prosecutors said they were dismissing the case.
His freedom came after his attorneys found another man who bore an uncanny resemblance to Jones who his attorneys alleged was the actual perpetrator. That man was never part of the initial investigation, but he lived near where the robber was seen on the day of the robbery.
And when witnesses who initially identified Jones saw a picture of the other man, they said they could no longer say who committed the crime.
Although the ruling stopped short of finding Jones innocent, District Judge Kevin Moriarty said that based on the new evidence, no reasonable juror would find Jones guilty.
The criminal justice system has recognized for years that eyewitness testimony is notoriously unreliable.
"This court has long recognized that eyewitness identifications' unique confluence of features – their unreliability, susceptibility to suggestion, powerful impact on the jury, and resistance to the ordinary tests of the adversarial process – can undermine the fairness of a trial," U.S. Supreme Court Justice Sonia Sotomayor wrote in a 2012 case.
After the 1999 robbery, investigators quickly traced the getaway car back to three people who told them they had been smoking crack cocaine that day and picked up a man they knew only as Rick or Ricky. It was Rick who committed the robbery, they said.
The victim and witnesses described the robber as a Hispanic or light-skinned black man with a goatee or with his hair pulled back from his face.
Police investigators gathered a series of mug shots of black males with the name of Rick or Richard. The driver of the getaway car was brought in to look at the photos almost three months after the robbery.
When he came to picture No. 202, he told police that it was the Rick who committed the robbery. It was a picture of Richard Jones, whose mug shot was in the system from a previous arrest.
Police then created a lineup using that photo of Jones and photographs of five other black males.
But according to Jones' lawyers, he was the only light-skinned man in the lineup.
"This photo line-up is severely flawed," his attorneys argued court documents. "Richard Jones is the only one who comes close to the description of the suspect."
The judge who ordered Jones' release said in his ruling that although the original photo lineup was no longer available and the photocopy presented as evidence was of poor quality, it appeared that Jones was the only light-skinned man in the lineup.
A detective took that six-picture lineup to a Kansas prison where one of the other men in the car on the day of the robbery was incarcerated. That man identified Jones as Rick.
Nearly five months after the robbery, that photo lineup was shown to a Walmart security officer who had witnessed the robbery. He too picked the picture of Jones.
Jones' attorneys noted in their court filings that the security officer had previously misidentified someone else as having been the driver of the getaway car.
The victim in the robbery, Tamara Scherer, told police in her initial statement after the robbery that she had not focused on the robber's face and did not think she would be able to identify him.
But at the preliminary hearing for Jones almost a year after the crime, she identified him in court as the man who robbed her.
"Completely certain, without a doubt," she testified.
At Jones' trial, Scherer and the security guard once again identified him. But the two men who had initially picked his photo out of the lineup were unable to identify Jones as Rick.
Jones was convicted, and because force was used in the robbery and he had prior felony convictions, he was sentenced to 19 years in prison.
Scherer regrets that she set aside her uncertainty in identifying Jones. The detectives and prosecutors had assured her they had the right man and that there were other eyewitnesses. She was adamant in her court testimony under cross-examination that Jones was the man who had robbed her.
"I don't know where the finger-pointing should go," she said, looking back. "It was a calamity of small things that lined up the wrong way."
The result, she said, was "tragic."
The circumstances of the Jones case were extraordinary and unlike any that Johnson County District Attorney Steve Howe said he'd encountered in nearly three decades in criminal justice.
It was the presence of a lookalike to Jones also named Rick that led to the dropped charges, not the conduct of law enforcement, Howe said.
Prosecutors and investigators have long been aware of the hazards in developing eyewitness testimony, he said.
"We go out of our way to make sure we are not planting our view into their (the witnesses') minds," he said. "Our job is to see justice done."
Jurors heard family members support Jones' alibi – that he was with family that day – but jurors were more inclined to assume they were lying than discredit the eyewitness identifications of Jones, said Tricia Bushnell, executive director of the Midwest Innocence Project.
"We hold up eyewitness identification so highly," she said, " ... (yet) it is the leading cause of wrong convictions."
In an interview with The Star after his release, Jones said that after years of fighting to clear his name, when the other man who looked like him was found, "everybody started scrambling.
"When I saw that picture, it just made sense to me," he said of how people could have mistaken them.
Former prisoners who are found to be wrongfully convicted have later gone to court and sued for damages.
Jones is currently considering such action, Bushnell said, but no decision has been made.
Former Missouri state Sen. Joe Keaveny, a St. Louis Democrat, was alarmed by studies on the unreliability of eyewitnesses and attempted, unsuccessfully, to get similar legislation passed in Missouri.
Law enforcement departments mostly opposed it, he said. Most often, they said they wanted autonomy to control their eyewitness testimony procedures and that they were using best practices.
Lawmakers balked, he said, "because no one wants to look soft on crime."
"But my feeling," Keaveny said, "is we need to find a way to alleviate mistaken eyewitnesses." The more research that is done, he said, "the more unreliable you're going to see it be."
Officers from Roeland Park and the Johnson County Sheriff's Office investigated the Jones case. They asked Kansas City, Kan., police to gather the photographs later shown to witnesses, according to the defense's court documents.
John Cowles, now an attorney in private practice, was the assistant district attorney who prosecuted Jones.
He said he had only a vague memory of the case, according to a written affidavit he provided to Jones' lawyers.
But in reviewing the defense information, he noted that the state's case exclusively relied on eyewitness identification testimony from people who did not previously know Jones.
"In my experience this is important because eyewitness identification becomes more questionable when the witness has not previously known the suspect," he wrote. "Having multiple identifications from several witnesses helps strengthen the state's case, but does not completely remove the concern about stranger identification."
Cowles said it was rare for him to prosecute a case based on eyewitness identification alone.
"Any prosecutor would want at least some corroborating evidence independent from eyewitness testimony," he said.
He said that viewing the evidence obtained by Jones' lawyers, including the photograph of the other man, "has undermined whatever confidence I had at the time that trial of Richard Jones resulted in a just result."
"The guiding principle of all prosecutors should be to achieve a just result, not just to win a case," Cowles wrote.
He was not available for additional comment Friday.
Roeland Park police have adopted a detailed protocol in line with the 2016 change in Kansas law.
For example, the department now shows lineup photos sequentially instead of in an array, as was done in Jones' case. Memory experts say witnesses who see photos one at a time compare each photo to their memory. In an array, the witness compares the photos with each other and is more likely to pick the person who looks most like the suspect, not necessarily the actual suspect.
"We try to get everything required so there are no mistakes," said Roeland Park Police Chief John Morris. "To us, doing a good job means getting it done right the first time; making a mistake is not an option."
A spokesman for the sheriff's office said that every effort is made to make lineups as fair and nonsuggestive as possible. Investigators also are careful to not do anything to prompt a witness or suggest that they are required to make an identification.
"We understand that they have been through a traumatic situation," said Maj. Mike Pfannenstiel. "If you can't definitely make an identification, then please don't."
