How reliable is eyewitness testimony?

The reliability of eyewitness testimony continues to be an issue for law enforcement and the legal system. According to the Innocence Project, eyewitness misidentification "was a factor in more than 70 percent of post-conviction DNA exoneration cases in the U.S., making it the leading cause of these wrongful convictions."

Judges in several states now instruct juries about the potential inaccuracies in eyewitness testimony or identification. In New Jersey, the jury instructions explain factors that can affect such testimony, like distance, lighting, stress and the difficulty of cross-racial identification.

To discuss the ongoing science of eyewitness testimony and how courts are responding, MPR News host Kerri Miller spoke with Brandon Garrett, a law professor at the University of Virginia, and Sandra Guerra Thompson, director of the Criminal Justice Institute at the University of Houston Law Center.

"People have always known that it's hard to remember the faces of strangers, and that there are problems with eyewitness memory," Garrett said, citing legal decisions from 50 or even 100 years ago. But a widespread push to reduce such errors actually had an opposite effect.

"We entered a period in the 70s where people though we can make eyewitness identifications more scientific: We can do these line-ups, and we can weed out the people who don't have much of a memory of what they saw," Garret said. "It turned out those very line-ups that we started relying on to make the whole thing more scientific introduced even more error. Police departments started using line-ups that were suggestive, that allowed police to sort of cue which was the suspect they had in mind, so things probably got a lot worse for 20, 30 years."

When DNA became a crucial part of the legal system, and hundreds of people were exonerated using DNA evidence, "people realized something terrible was going on" with eyewitness testimony, Garrett said.

One of the factors contributing to the problem was the difficulty of cross-racial identification.

Numerous studies have shown, Thompson said, "that people are less able to identify a stranger's face when they are of a different race."

But even with continued research on the subject, most states still don't include such findings in their jury instructions.

"Plenty of judges still treat [eyewitness testimony] the way they always did," Garrett said. "They say: 'Jurors, use your own common sense to decide whether this eyewitness's memory is good or not.'"

For the full discussion on the reliability of eyewitness testimony, use the audio player above.