IPOT study to grade Texas eyewitness ID policies at local police, Sheriffs

The Texas Tribune this morning published a story about a research project your correspondent is spearheading on behalf of my employers at the Innocence Project of Texas to analyze new eyewitness identification policies which Texas law enforcement agencies were supposed to adopt by September 1 of this year. The story by Brandi Grissom opened:

The Innocence Project of Texas is preparing to grade about 1,200 law enforcement departments statewide on their compliance with a law that requires police agencies to adopt eyewitness identification policies.

“Unless somebody is really grading their papers, nobody knows whether the law is really being implemented,” said Scott Henson, a policy consultant for the Innocence Project.

Last year, Texas legislators approved a measure that required police agencies to adopt policies meant to prevent faulty eyewitness identification in criminal cases. Under the law, departments were required to adopt a written policy by Sept. 1. Last week, the Innocence Project sent the departments letters requesting copies of their lineup policies.

Faulty eyewitness identifications are the leading cause of wrongful convictions, according to the New York-based Innocence Project. In 297 DNA exonerations across the nation, the Innocence Project reported, mistaken identifications contributed to 75 percent of the wrongful convictions.

In Texas, faulty eyewitness identification contributed to wrongful convictions in 40 of the 52 DNA exonerations, according to the National Registry of Exonerations.

“There’s almost nothing more powerful in a courtroom than eyewitness testimony,” Henson said.

The story quoted Assistant Chief Bryan Carlisle of the Shenandoah Police Department, "who has been traveling the state conducting training for the Texas Police Chiefs Association," declaring that “We really thought as a profession we had been doing right,” but “Now, science has caught up and said, ‘Hey, there really is a better way to do this.’” He hoped, as do I, that "what they’ll find is that most agencies are in compliance with the law."

As background, regular readers may recall that:

Legislators instructed the Law Enforcement Management Institute of Texas at Sam Houston State University to develop a model policy on eyewitness identification that departments could use as a template.

That policy sets out guidelines for conducting lineups in a way that does not suggest to witnesses whom they should select. Those guidelines are the criteria against which the Innocence Project of Texas plans to judge the policies that departments have adopted, Henson said.

Among other things, the criteria include ensuring that the person presenting the photos does not know who the suspect is, asking witnesses how confident they are that the person they identified is the same one they saw at the crime scene, giving witnesses instructions that include letting them know the perpetrator may not be among the choices presented, showing potential suspects sequentially instead of simultaneously, and choosing subjects for the lineup who have similar characteristics to one another and to the suspect described.

Texas' new law allows agencies to diverge substantially from the "model policy" when crafting their own local standards. So it's possible for departments to comply with the law but still avoid adopting best practices. The IPOT analysis will grade departmental policies based on compliance with the LEMIT model policy, weighting its components based on the most critical elements identified in the research on eyewitness identification practices. Anecdotally, many departments have welcomed LEMIT's legislatively mandated advice, while others have been more resistant. Until the policies come in and are analyzed, though, there's no way to know which how many agencies adopted key elements from the model policy and how many failed to do so.

IPOT's open records requests apparently began to hit yesterday. As of this morning, we'd already received policies from 77 departments at a dedicated email address set up for the project.

Notably, though the best practices promoted in the model policy will significantly reduce eyewitness errors, but they will not eliminate them. That's because eyewitnesses, especially when they did not know the perpetrator before the crime event, tend to make relative judgments, and one lineup member will always look more like the perpetrator than the others, even when the actual perpetrator is not in the lineup. The most comprehensive field study on the topic found that, even using best practices including sequential presentation, 12.2% of eyewitnesses chose a filler instead of the suspect. (And of course, it's impossible to say how many suspects chosen were really the wrong person.) So requiring new policies won't be a panacea, but it's an important first step toward reducing eyewitness errors and, by extension, the rate of false convictions based on them.