Will new Texas' eyewitness ID law reduce false convictions?

Photo via Slate

In a short piece at the Austin Chronicle on Texas' new eyewitness ID law and the draft model statute, Jordan Smith wonders "how quickly the policy might be adopted" at Austin PD, leading me to point out that, by statute, local departments must adopt eyewitness ID policies by September 2012. That's why the Sam Houston State Law Enforcement Management Institute of Texas has to produce a "model policy" by the end of the year to give them time to develop their own policies, using the components in the SHSU model as guidance. (See SHSU's draft policy; they're presently soliciting public comments between now and Nov. 30 before the policy is finalized in December.)

Smith mentions as one flashpoint the odd assertion in the model policy Grits disputed earlier, that "live lineups are preferable to photo IDs because 'witnesses typically view perpetrators of crimes in three dimensions.'" I agree, I hope that's omitted from the final version. But Jordan failed to point out that, with that notable exception, the rest of the draft model policy, though perhaps not written as clearly or organized as well as one might prefer, mostly contains the key "best practice" elements which national standards and memory researchers have encouraged departments to include in such policies. It could stand to be tweaked but they didn't do a terrible job.

However, Smith expresses skepticism about the new law's ultimate impact that I do not share, though I know where it comes from (indeed, it comes in part from my boss at the Innocence Project of Texas, Jeff Blackburn, who's been spreading this meme): I just consider it a result of courtroom lawyers' myopia. Smith's article concluded: "Without any enforcement mechanism, it is unclear how quickly the policy might be adopted – and whether the new procedures will reduce the instance of mistaken identifications. Indeed, the new law contains no provision to keep out of evidence identifications made outside the parameters of the model policy."

What she means is, although a defense attorney can question police in front of a jury about why they didn't follow their policy, the eyewitness testimony still gets in. Regrettably, and for reasons I'll never fully understand, back in 2009 when the bill language was being negotiated, the Texas Criminal Defense Lawyers' Association vigorously opposed inclusion of a "jury instruction" in lieu of the exclusionary rule (which the Governor's office had said would draw a veto) as a remedy. Exhibiting all the subtlety and foresight of a kamikaze pilot, after working as hard as they could to neuter the bill, TCDLA then opposed it for being too weak. (In 2011, a different lobbyist for the group supported the watered-down version of the bill that eventually passed.)

Chart from this report (pdf) from the national Innocence Project

When a trial lawyer says there is no remedy, they mean that, at trial, if eyewitness evidence gathered via flawed procedures is introduced and the defense objects, they cannot get it excluded or even be certain of a jury instruction, though under a recent precedent they possibly can get an expert witness to explain problems with the lineup and the sources of eyewitness error. So jurors may hear experts question the validity of a problematic lineup or photo array under Texas law, but courts are not required to exclude testimony gathered under questionable circumstances, nor even to instruct the jury that failing to follow best practices increases the likelihood of error.

That's a legitimate critique, and from the perspective or a courtroom advocate, I might agree with it. But here's what it misses: 97% of criminal convictions are plea bargained and never reach a jury trial. Many more are dismissed after eyewitness evidence is discredited on the front end. Plus many eyewitnesses are wholly credible: E.g., witnesses who previously knew the suspect don't suffer the same memory problems as those trying to identify strangers. So in the vast majority of cases involving eyewitness evidence, that lack of a courtroom remedy for the 3% of cases that go to trial doesn't really matter. The number of cases in which such a remedy might come into play would be symbolically important but statistically de minimis.

Meanwhile, 88% of Texas departments have no written policy at all on how to handle eyewitness identifications, and only a few of those with written policies follow anything close to best practices. Forcing departments to think through these issues, examine SHSU's model policy and others, decide on a method, then train on it over time will be a big improvement compared to the current practice, where methods differ not just from department to department but literally from detective to detective.

So from a public policy perspective, this statute will encourage consistency, uniformity (at least within departments and to a large extent across agencies) and training on a subject presently governed by departmental lore supplemented by examples from TV detective shows. From past experience, you can expect the majority of Texas 1,000+ departments that perform lineups (with the likely exception of some of the largest ones, which will write their own from scratch) to adopt the LEMIT policy wholesale, or at least in large part. So it's important for LEMIT to end up with the best model policy possible.

Once local policies are in place next September, my employers at the Innocence Project of Texas will be gathering as many as possible under the Public Information Act and grading local policies to identify those following best practices and those whose policies are deficient. So there will be at least some external oversight to ensure local policies aren't a complete joke, even if it's not from a government agency. I'm looking forward to that project.

Certainly Grits would prefer that the law had more "teeth" - if police failure to comply with proper procedures garnered a jury instruction, for example, or if departments didn't have the option to adopt a weaker policy. But this was the compromise achieved at the Lege, and compared to the status quo the law at least provides a starting point for improving police practices, certainly compared to the absolute void that preceded the bill. Departments with no policies (88% of them) where procedures vary from officer to officer will now have to decide how they want to perform lineups, write it down, and train everybody on the same method. Even if it's not precisely the method Grits, the Innocence Project, the National Institute of Justice, or LEMIT might prefer, once it exists in writing, it's possible to argue to change it. And of course it's always possible to go back to the Lege to add a remedy in the future if departments react with extreme noncompliance. While this bill was an important first step, I don't suspect it will be the last word on eyewitness identification procedures at the Texas Legislature.

See related Grits posts:

• Onus on state, local departments to improve eyewitness testimony
• Court of Criminal Appeals: Trial court abused its discretion by disallowing eyewitness ID expert
• Model policy under development on eyewitness ID procedures
• Slow but steady progress toward improving eyewitness identification
• SCOTUS to reevaluate eyewitness IDs
• How much do eyewitnesses really see?
• Eyewitnesses and the 'feeling of knowing'
• Eyewitnesses in staged test only 8% accurate
• More on the fallibility of eyewitness testimony
• Eyewitnesses miss big changes in their environment, like the person in front of them 
• Study: 88% of police and sheriffs have no written policy on eyewitness ID procedures, even fewer follow best practices
• Why Carl Reynolds would make a lousy witness