The new frontier in blunders in Texas' criminal justice system: witnesses identifying the wrong person.A new national registry detailing nearly 900 exonerations shows that for cases of robbery and sexual assault, the vast majority of wrongful convictions came from mistakes in eyewitness identification."In many cases, that's all [the evidence] you have," said James Doyle, visiting fellow at the National Institute of Justice at the U.S. Justice Department. "It's a real problem."In Texas, which is third in the nation in the number of exonerations, 84 percent of DNA exonerations came in cases in which witnesses had made false identifications.Now, flaws in eyewitness identification procedures are taking center stage as the deadline approaches for Texas law enforcement agencies to comply with reforms to state law.As of Sept. 1, the agencies must adopt methods to ensure the reliability of their procedures for eyewitness identifications, as part of House Bill 215. The Legislature approved the measure last year in memory of Tim Cole, the Texas Tech University student from Fort Worth who died in prison after he was wrongfully convicted of rape based on a false ID.State Rep. Rodney Ellis, D-Houston, author of the bill, says he plans to introduce tougher legislation if police agencies make only flimsy efforts to reform."We'll have to revisit the issue to ensure that all Texans can have faith in the reliability of the evidence in our courtrooms and a justice system they can trust," he said.Problems with identification practices were also exposed this month in a university study that criticized Corpus Christi police for conducting a "showup," in which eyewitnesses were allowed to identify a suspect in a robbery-murder as he sat handcuffed in the back seat of a squad car."It's a textbook example of why we need reform," Ellis said.Police officials in Fort Worth and Arlington say they have been working for years to adopt "best practices" that set standards for police photo and live lineups.Fort Worth Police Chief Jeff Halstead did not respond to written questions from the Star-Telegram, but a Police Department official said it "had a written policy in place prior to the new legislation based on case law and research from the scientific community."She added that the department is updating the written policy in response to the new legislation.In Arlington, homicide detective Sgt. Kraig Fryer said he has received special training since 2009 on eyewitness identification.Arlington is also reviewing its policy to make sure it is compliant with state law, he said. The hope is that the local standard will be more rigorous than model protocol, he said."Every police agency should feel the same way," said Fryer, who was responsible for conducting the research and making policy revisions."We want to make sure we arrest the right person," he said. "We never want to put an innocent person in jail and have the perpetrator on the street. And we want to do everything we can to make it as transparent as possible."Danger of showupsRecent furor over mistaken eyewitness identification stems from the case of Carlos DeLuna, who was executed for the 1983 convenience store robbery-murder of Wanda Lopez.DeLuna was identified in a "showup" -- in which a single suspect is shown to a victim or witness soon after the crime.The procedure is the fastest and easiest of all identification methods. All the witness has to do is give a "yes or no" statement of recognition, said Dan Simon, professor of law and psychology at the University of Southern California Law School.It can be "advantageous," he said, if the showup can be arranged quickly before the witness's memory fades and the appearance of the perpetrator changes.At the same time, Simon and other criminologists have all kinds of hang-ups about showups.Showups are a major source of misidentification, said Simon, who wrote a book on that and other ID practices, In Doubt: The Psychology of the Criminal Justice Process, due out in June.Showups and lineups produce similar rates of correct identifications, he said. On the other hand, showups are more likely to yield false identifications, laboratory studies found.And showups have strong suggestive effects because of the lack of "fillers" or choices for the witness."The thing that makes them very dangerous is that peoples' memories aren't like video tapes," said Doyle, the attorney at the National Institute of Justice. "People's memories change ... and it's quite possible that the person they see at a showup will replace in their mind the person they saw at the time of the crime."Other data show that showups can be a dangerous source of evidence. For example, archive data from Sacramento County shows that the rate of identification of suspects in showups is considerably higher than in other types of lineups (76 percent versus 48 percent, respectively), and is highest in photographic showups (83 percent), Simon said.Fryer, the Arlington homicide detective, said Arlington continues to perform showups but makes sure that witnesses and victims are instructed on procedures.In addition, he said, witnesses are asked to say how confident they are in their ID."We document right then what happened," Fryer said.The showups and lineups are also videotaped, he said.National registryIt's easy to understand why Ellis and other lawmakers are concerned about improper eyewitness identification.The website of the National Registry of Exonerations profiles 873 of the more than 2,000 people who were falsely convicted of serious crimes from 1989 to 2012. (Exoneration is a legal process in which the convicted person, after new evidence is found, is relieved of all legal consequences by a prosecutor, governor or a court. The registry's website keeps a live count of exonerations, which since the report was compiled has climbed to 885.)Of those 873 profiled exonerations, eyewitness misidentifications led to mistakes in about 80 percent of rape and robbery cases. (False convictions for robbery are also almost always caused by eyewitness misidentifications, but there are few exonerations because DNA evidence is hardly ever useful in robbery cases.)The report underscores the need for good identification protocols, Doyle said."One of the things that the registry does is it tries to do a very careful look at exonerations that are not DNA exonerations but exonerations from other methods," he said.The registry, a joint project of the University of Michigan Law School and the Center on Wrongful Convictions at Northwestern University School of Law, also points out that there are many more false convictions and exonerations that are not yet known."It used to be that almost all the exonerations we knew about were murder and rape cases. We're finally beginning to see beyond that," said Michigan law professor Samuel Gross, editor of the registry and an author of an accompanying report. "This is a sea change."Yamil Berard, 817-390-7705Twitter @yberard
Of the 873 exonerations from January 1989 through February 2012:
8 percent pleaded guilty (71/873), and the rest were convicted at trial -- 87 percent by juries and 8 percent by judges.
37 percent were cleared at least in part with the help of DNA evidence (325/873).
63 percent were cleared without DNA evidence (548/873).
More than 75 percent had been in prison for at least five years, and half for at least 10 years.
The defendants collectively spent more than 10,000 years in prison for crimes for which they should not have been convicted, an average of more than 11 years each.
Of these exonerees:
93 percent were men (816/873), and 7 percent women (57/873).
Race of the defendants was known in 92 percent of the cases (802/873): 50 percent were black; 38 percent were non-Hispanic white; 11 percent were Hispanic; and 2 percent were American Indian.