Hart: Can’t count on eyewitnesses to ID criminals

“Why do journalists care so much about criminals?” That was the simple question posed in an email I recently found in my inbox. The author didn’t tell me the grounds for her premise, but it arrived right after a column I had written about providing lawyers for poor people accused of crimes. I thought about the email this week after reading Chronicle reporter Allan Turner’s story about Carlos DeLuna, who was executed by the state of Texas for a 1983 robbery-murder in Corpus Christi that he almost certainly did not commit.

Journalists are better at asking questions than answering them, so I’ll pose a few of my own: How do you know that someone is a criminal?

From eyewitnesses?

A new study of DeLuna’s case by the Columbia University Human Rights Law Review concluded that flawed police procedures led an eyewitness to identify the wrong man for a convenience store clerk’s murder.

Very likely that mistake cost an innocent man his life.

In DeLuna’s case, an eye¬≠witness initially gave a description of the clerk’s assailant that did not match DeLuna’s appearance.

Then police brought DeLuna to the scene, handcuffed in the back of a squad car, and asked the witness if he could identify the young Hispanic man as the assailant. After being assured that DeLuna had been apprehended nearby, the witness agreed he was the killer. Later, the witness told investigators he was only 70 percent sure of his identification, and that if police had not told him DeLuna had been found in the area, his level of certainty probably would have been only 50 percent.

Unreliable testimony

Compounding the error, DeLuna was assigned an inexperienced lawyer who didn’t uncover witnesses who could have saved his client’s life. Columbia Law investigators discovered that a police informant named Carlos Hernandez, who would later die in jail, confessed to relatives on the day of the crime that he had killed a woman named Wanda (the store clerk’s name).

He later confided to an acquaintance that he had killed a woman but “Carlos DeLuna took the fall.”

The case underscores academic and scientific studies demonstrating the unreliability of eyewitness testimony.

During the last session of the Texas Legislature, a bill by Sen. Rodney Ellis, D-Houston, was passed into law requiring the Law Enforcement Management Institute of Texas at Sam Houston State University to adopt a model policy for police lineups, to reduce the possibility of injecting error into the process.

The bill was recommended by the Timothy Cole Innocence Commission, named for a Texas Tech University student who was convicted of a rape he did not commit and died in prison of an asthma attack before he was exonerated by DNA evidence. The rape victim in the case was shown Cole’s picture repeatedly until she selected it, and a police officer congratulated her for her selection.

Earlier this year, LEMIT finalized its model policy. Among the recommendations: Eyewitnesses should be shown suspects’ photos one at a time, instead of all together. That, the experts say, reduces subconscious pressure to select a person who looks most like the criminal. Each suspect stands alone.

HPD resisting

That’s not how the Houston Police Department currently conducts photo lineups. According to a 2009 policy, photo spreads are shown to witnesses.

When LEMIT took testimony regarding its recommendations, Houston police Detective Mark Holloway argued there wasn’t enough research to abandon photo spreads in favor of showing the photos sequentially (one-at-a-time) to witnesses.

But a prestigious think tank, the American Judicature Society, says otherwise. In a recent survey of research, the AJS concluded that “decades of laboratory research” show that the “sequential procedure reduces mistaken identifications with little or no reduction in accurate identifications.”

Under Ellis’ legislation, police departments can adopt their own policies, and are not obligated to choose the LEMIT “model” policy in total. Friday, a spokesman for the HPD said the LEMIT policy was under review.

Only two years ago, Houstonian Michael Anthony Green was released from prison after serving 27 years for a rape he did not commit, proven by DNA testing. His conviction was based on the testimony of an eyewitness, who picked his picture out of a photo spread, and later out of a live lineup.

Fewer apologies

And back in 2007, former appellate Justice Michol O’Connor urged the Houston City Council to force the HPD to adopt sequential photo lineups.

“City leaders,” wrote O’Connor, “should be weary of making apologies to innocent people who were convicted of crimes they did not commit.”

We humans have a terrible track record in picking out the real “criminals” out there. If HPD doesn’t choose the LEMIT model policy, maybe the City Council will heed Justice O’Connor’s advice and show leadership on the issue.

patti.hart@chron.com

Posted in: News Articles • Tagged with: