Colin McCallin Jan. 22, 2019

Convicting the Innocent

According to The Innocence Project, a nationwide organization working to exonerate, win compensation for, and rehabilitate people wrongfully convicted of crimes, around seventy percent of the hundreds of convictions the organization has overturned so far using DNA evidence have resulted from eyewitness misidentification. Traditional police methods promote such error, in part because of the authority police wield: distracted by an impulse not to hinder figures in power, a witness tends to confirm the suspicions of investigators rather than hold true to his or her original recollection. Also, the police themselves sometimes misidentify suspects, and in those cases their authority seeps upward, moving prosecutors to keep alive investigations they might otherwise terminate. There is room for improvement here. At what stages of current procedure is error most likely to occur, and why, and what tweaks, if any, could be made right away?

What to Do about It

First, we need to clarify something: law enforcement agencies are full of hardworking, decent people performing their jobs as well as they can with what they know at a given time. Are there bad cops? Yes. Does that say something about cops in general? No. Show us a group of human animals, and we’ll show you a subgroup of human animals behaving reprehensibly. It’s even true of defense attorneys! Crazy, we know! Eyewitness misidentification therefore should not be ascribed to some special proclivity for evil among the officials who investigate crimes. The problem most likely lies instead in the way we form memories and respond to the opinions of others. Data from The Innocence Project indicate that people are most likely to misremember details of an incident when they are steered, on purpose or otherwise, toward a certain result. If a police officer asks a witness to pick a suspect out of a lineup of five people, for example, but somehow gives away which of them the officer thinks is guilty, the witness’s memory will be tainted by that information. And such giveaways are in fact a source of error in real cases. For this reason, one change to police practice The Innocence Project recommends is the institution of “double-blind” lineups, in which neither the person conducting the lineup nor the person viewing it know who the suspect is. Many police departments across the country have already made that change in advance of any legislation requiring so. Other ways to make eyewitness identification more reliable include recording lineups on video, taking statements from witnesses as to their confidence in their assessments at the time those assessments are made, rather than once further information has come out, and ensuring that participants in a lineup all match the witness’s description of the perpetrator, to avoid making any one participant stand out artificially.

The Bottom Line

These are all cheap, simple adjustments that together could make misidentifications much less likely. Eyewitness testimony in criminal law is a crucial class of evidence without which a great many cases would remain in doubt forever, and eyewitness identifications are part of that evidence. As we strive to recreate events for the sake of justice, we cannot do without the resource of human memory, of course, but we must acknowledge its quirks and limitations. Otherwise, we end up with the kind of tragedies The Innocence Project has been set up to rectify.

For more information: