New research questions the assumption that police should interview witnesses to a crime separately

See the column on the Wall Street Journal site

If you witness a crime, what’s the best way to recall what happened? Minutes to months later, police might ask you the color of the perp’s eyes, the design of his tattoo or how long it took him to pull out a gun after he entered the room. Would it be better to recount your story on your own or alongside the person you were with at the time?

Anyone who has ever watched a police procedural can answer that question. Witnesses are always interviewed alone, in a dismal, windowless holding cell—that is, if the interview takes place in Hollywood.

But the isolation of witnesses is not just for dramatic effect. Psychologists have long warned police that one witness can contaminate another’s testimony. Social pressures can make someone change his tune, or errors might be introduced into the testimony.

Contagion and the power of suggestion might also create the feeling that events that never even happened actually occurred. That was the case during the 1980s epidemic of “repressed” memory syndrome, which focused on false claims of childhood sexual abuse. In 2013, researchers at the Massachusetts Institute of Technology were even able to plant false memories in lab mice, leaving neurochemical traces indistinguishable from the neural footprints left by real experiences.

Still, even if memory is highly malleable, two new studies show that there are big benefits in bringing witnesses together to collaborate on testimony. The findings could shake up decades of practice in legal circles.

One study, published in the journal Memory this past May, shows that witnesses who are interviewed together do, in fact, influence each other. But they also correct and amplify each other’s accounts of the same event, increasing accuracy in the process. The opportunity to edit each other’s memories allows pairs of witnesses to make fewer errors than witnesses who are questioned on their own.

Led by the Dutch legal psychology professors Annelies Vredeveldt and Peter van Koppen and colleagues at VU University Amsterdam, the researchers asked people who had recently seen a play to describe a violent, emotional scene. Of 53 adults who saw the same play on three separate nights, 36 came to the theater as couples and were interviewed together afterward. They ranged from spouses to one pair that had just met for the first time. The 17 others were interviewed as individuals. The researchers then compared the two groups. Who would produce a more accurate chronicle of a rape-and-murder scene acted onstage the week before?

The people who came together corrected each other’s errors, as we know. But a second, more intriguing finding surfaced, too: The nature of a couple’s communication skills influenced how much detail they remembered. “People who repeated, rephrased or elaborated on what their partner just said remembered more,” Prof. Vredeveldt told me.

It wasn’t how long they had been married or how well the couple knew each other that mattered. It was their skill in creating a joint narrative. If the husband described the crime victim as wearing “some type of dress,” the wife might add “one that opens at the front.” When the man agreed and added that the dress was white, the wife concurred, then specified that it was “a dirty white.” Because the researchers didn’t examine the features of a pair’s relationship, they plan to find out what happens if the couples are strangers or are asked to remember an event in groups.

A second study, published in Legal and Criminological Psychology in June with some of the same authors, used a larger sample and more controlled conditions—and had the same findings: People interviewed together made fewer errors.

Prof. Vredeveldt still believes that people should be interviewed individually. “But instead of sending them home after that, you might generate more leads and fewer errors if you put witnesses together.” Because when couples are bouncing ideas off each other, “the whole is greater than the sum of its parts.”