People are notoriously bad at spotting lies: unless privy to information that directly contradicts a spurious story, past research suggests the average listener pinpoints a fib less than half the time. A study in June in the Proceedings of the National Academy of Sciences USA now indicates that groups do a bit better at detecting dishonesty—but only when group members confer with one other before coming to a conclusion.

University of Chicago business researcher Nicholas Epley and doctoral student Nadav Klein conducted four experiments to compare lie perception by individuals or groups. In each scenario, hundreds of people were assigned into groups of three to watch a series of 10 video clips containing some speakers who were telling the truth and others who were trying to deceive them.

Participants then weighed in on whether they believed the speaker in each video lied—some people offered individual judgments right away, and others discussed the case with group members first. In each scenario, the discussing groups had a slight edge, detecting lies up to 62 percent of the time compared with the individuals, whose lie-detection average topped out at 54 percent in one of the experiments but was typically lower.

Researchers believe there is more at work than the so-called wisdom of crowds effect because conglomerating judgments from a few or many individuals did not increase lie detection in the absence of the discussions. Instead they suspect there are still to be determined synergistic elements that bolster groups' baloney detection. The researchers plan to follow up further by studying the conditions and group characteristics that produce the lie-detection boost.

“It's not that every group is better than every individual. But what is it that gives groups this nudge?” Epley wonders. He cautions that the effect sizes are small and variable but says the results hint at the importance of group discussions in settings where people are asked to uncover lies—from deliberating in a jury to ferreting out insurance fraud.

Looking for Lies

Falsehoods are not easy to spot. According to an analysis of more than 200 studies, people with no training are able to distinguish a lie from truth just over 50 percent of the time (about the same as guessing). Those with training fare somewhat better, with an accuracy of about 65 percent. Experts have developed a range of detection methods to help with the task, but no method is foolproof because there is no one behavior that indicates deception. Here is the truth about five methods that have shown some promise for detecting lies:

Click or tap to enlarge