In the 1985 wrongful rape conviction of Joe Jones, there were problems with how the Topeka Police Department handled the eyewitness identification, said Brandon Garrett, a University of Virginia law professor who examined Jones’ case as part of a 2009 large-scale study sponsored by the Innocence Project.

“This case looks exactly like the others I’ve looked at,” said Garrett, whose study found that half of 250 wrongful conviction cases examined included a witness misidentification.

But the biggest issue was how the victim in the case — who spent the most time with the attacker — identified Jones after the assault.

At the time, Topeka Police used what’s known as a “six-pack” photo array, during which a witness or victim is shown one sheet, with six suspects. In the Jones case, the victim picked a photo of someone other than Jones.

“That’s a huge red flag,” Garrett said.

But the two other eyewitnesses spotted Jones on the street the night after the attack and called police. They were confident Jones was the rapist.

Then at a preliminary hearing, the victim in the case saw Jones in the courthouse. She immediately picked Jones as her attacker, even though she hadn’t selected him in the photo lineup.

From then on, the victim insisted Jones was the rapist, and testified to that at trial.

“The trial ID may be very powerful,” said Garrett, but cautioned such an identification is prone to mistakes.

Jones was handcuffed, and the victim expected to see the attacker in court.

And there’s the two other witnesses who could have influenced the victim’s choice, Garrett said.

With mounting evidence that eyewitness misidentifications have led to scores of wrongful convictions over the years, police are changing practices.

“There’s a lot more awareness about this,” Garrett said.

“Eyewitness identification has for a long time been questionable,” said Topeka Police Chief Ronald Miller, whose department recently instituted new policies aimed at reducing potential misidentifications.

Topeka Detective Larry Falley, explained how they now approach witness identifications.

First and foremost, Falley said, is a change from the one-sheet six pack photo array to a one-photo “sequential” lineup, in which the witness is shown photos one at a time. This has several advantages, Falley said. In sequential lineups witnesses are less likely to guess than when there are multiple photos shown at once. And the photos themselves are much larger — and also include a side view — giving witnesses the opportunity to better examine facial features.

The department also use an “independent administrator” for photo lineups. The person showing the witness the lineup is not involved in the case, and has no idea who the suspect is. This “blind” approach removes any chance the person showing the photos can influence the process, Falley said.