Play Live Radio
Next Up:
0:00
0:00
Available On Air Stations
Watch Live

Science & Technology

Salk Researchers Uncover Scientific Approach To Eyewitness Identification That Could Reduce Some Bias

Dinh Diep, an author of the Salk Institute eyewitness perception study, views a pair of images and must say which one better resembles her memory of the suspect from a mock crime video, July 13, 2020.
Salk Institute
Dinh Diep, an author of the Salk Institute eyewitness perception study, views a pair of images and must say which one better resembles her memory of the suspect from a mock crime video, July 13, 2020.

Salk scientists say they’ve come up with a scientific method that could make eyewitness identification in a police line-up more effective. Witness identification depends on memory, which can be unreliable and subjective. Scientists say their method helps them objectively analyze witness memories, removing some types of bias.

The research is out July 14, in the peer-reviewed journal Nature Communications. It references data from the nonprofit group, The Innocence Project, which show that since the 1990s around 70% of exonerations with DNA evidence came from cases where witnesses misidentified the perpetrator. Those cases resulted in innocent people serving an average of 14 years in prison.

Salk Researchers Uncover Scientific Approach To Eyewitness Identification That Could Eliminate Bias
Listen to this story by Shalina Chatlani.

Some of the procedures police have used to produce an eyewitness testimony include hypnosis, police line-up, photo arrays, composite sketches and voice recordings.

Advertisement

Tom Albright, a neuroscientist and memory expert at the Salk Institute and a lead researcher on the study, says the problem with a lot of these methods is that an individual’s memory can be influenced by a number of different factors.

Salk Researchers Uncover Scientific Approach To Eyewitness Identification That Could Eliminate Bias

“When people tell you what’s on their mind there’s all sorts of personal biases, social biases, cultural biases that color what they tell you about what’s going on in their brain,” Albright said.

That’s why, he says, a common police practice of lining up a number of people and having a witness choose one may not be that effective. There are too many choices, and that makes it hard for people to know what exactly their recognition memory — or ability to recall places and people — is telling them.

“What if there’s a method to extract those recognition memories covertly so that then you have these memory responses and [you can treat them like] they are objects laid out on the table?” Albright said. “And then you as a scientist can independently evaluate those independent memory responses. That’s what our method does.”

This scientific method is like a standard eye exam. An optometrist will present a patient with two lens options, and ask which one is more clear. He then repeats this process with more lens options, until there’s an obvious winner — or highly ranked choice.

Advertisement

So in this study, Albright showed witnesses a crime in a movie. Instead of a standard police line-up, he showed them faces of potential perpetrators in pairs. Every time a person chose one image out of the pair, it got a point. A scientist could objectively see a winner, based on the image with the highest cumulative score relative to the other images.

“So just asking which [option] is more similar to what you remember is a relatively easy thing for people to do. But if you ask in absolute terms, ‘which one is the highest scoring one?’ that’s a hard thing for people to do,” Albright said.

Now Albright plans to study how to finesse the police line-up process, so that witnesses can make better decisions based on their memory.

KPBS has created a public safety coverage policy to guide decisions on what stories we prioritize, as well as whose narratives we need to include to tell complete stories that best serve our audiences. This policy was shaped through months of training with the Poynter Institute and feedback from the community. You can read the full policy here.