Thanks to the efforts of organizations like Innocence Project New Orleans, Louisiana has the second highest rate of exonerations per capita in the country. Such a statistic gives the state something to be proud of. But what if this high exoneration rate reflects a more worrying possibility — that Louisiana is at the low end of the curve for getting verdicts right in the first place?

Serial rapist went free while wrongfully convicted man spent decades behind bars — until his exoneration last week

We’re not talking about an abstract or long-ago scenario. Take the case of Archie Williams, whose conviction was cleared this March after he served more than 36 years in prison for a 1982 Baton Rouge rape that he had nothing to do with. Fingerprints from the scene were run through a national database nearly four decades later and matched to another man, who had gone on to rape five other women in the meantime. This sad outcome is not unusual. When an innocent person goes to prison, the actual criminal is free to commit more crimes. In about half of the DNA exonerations cases where the real criminal was ultimately identified, they had gone onto commit more crimes.

Letters: Clarifying law on death penalty cases

Williams had been convicted largely on the victim’s identification of him. Clearly, as it turns out, she made a terrible mistake, as did the jury. In this sense, Williams’ case is absolutely typical. Close to three-quarters of the many hundreds of DNA exonerations across the country have involved mistaken identifications. And crimes are typically associated with factors known to exacerbate these errors (for example, the presence of weapons, identifications involving someone of another race, multiple viewings of an innocent suspect during the investigation, and other problematic police practices).

One important way we can improve our criminal justice system is to better understand this inherently flawed form of evidence. Cognitive scientists like me have conducted tens of thousands of experiments and have established a large body of reliable findings about eyewitness memory. We now know just how bad people are at identifying strangers, even under the best of circumstances (and the circumstances of a typical crime are far from the best of circumstances). Beyond this, we tend to be swayed by others who make confident identifications. Indeed, that moment when an eyewitness points to the defendant and confidently says, “That’s him,” is very compelling to jurors, who tend to be convinced by it. But such confident identifications are easily manufactured as the investigation continues and as memory related to the crime itself fades into the past. The issue isn’t whether the witness is lying or guessing; it’s that memory and confidence are easily tainted in highly predictable ways.

Jurors have the daunting job of evaluating the quality of evidence before them. If the evidence concerns identification, jurors enter the courtroom ill-equipped to do this, sometimes harboring demonstrably wrong intuitions. Nearly 40 years after Archie Williams’ wrongful conviction, we have a well-developed science about exactly this, and it’s obvious we should equip jurors with findings that will help them avoid these same mistakes. It’s the jurors who will make important decisions about guilt or innocence, and it’s in all of our interest that these decisions be truly educated.

I and other identification experts are routinely called to testify in courtrooms around the country about specific factors that impact eyewitness accuracy, but we are currently barred from courtrooms in Louisiana. This is the only well-established science I know of that has been sidelined in this way, and this is precisely the science that addresses the vast majority of wrongful convictions. It’s a truly puzzling state of affairs.

Luckily, House Bill 226, currently under consideration, would go a long way to correct this problem. It will give discretion back to judges so that they can decide on a case-by-case basis whether eyewitness expert testimony can help jurors to reach a fair verdict. Trusting judges to exercise that role, and allowing them to trust jurors to appropriately consider the current state of scientific knowledge, would put Louisiana on a footing with 48 other states in the U.S. that allow expert testimony. House Bill 226 represents an enormous step forward for criminal justice in Louisiana, and I urge lawmakers to support it.

Nancy Franklin is associate professor of cognitive science, Department of Psychology at Stony Brook University in New York.