Public Release: 

With Motivation And Training, Some Professionals Can Catch A Liar

University of California - San Francisco

Sherlock Holmes and Secret Service agents are not the only people who can catch a liar, according to researchers at the University of California San Francisco. UCSF psychologists have found that individuals with a special interest in deception are able to detect a liar based on split second facial expressions, gestural slips, and subtle signs in speech. Their findings will be published in this month's issue of Psychological Science.

"The ability to accurately detect deceit is real," says Paul Ekman, PhD, UCSF professor of psychiatry and principal investigator of the study. "The information is there and we've shown that a few groups of people can utilize it, although most law enforcement officials and most psychologists miss it."

The 627 people who participated in the study included sheriffs, judges, police, intelligence officials, and psychologists. They were categorized into 7 groups based on their profession - either psychologists or law-enforcement agents -- and interest and experience level in deception.

The researchers measured how well each group could detect deceit based upon demeanor. Participants watched videotapes of ten males, ages 18 to 28, who either lied or told the truth about their deeply-held opinions on controversial social issues. The participants then had ten seconds to decide if the subjects had given true or false opinions.

A group of 23 federal law enforcement officers, which included CIA agents, was the most accurate and decided correctly 73 percent of the time, on average. This was significantly better than a group of 84 federal judges (62 percent accuracy, on average) and a group of 36 municipal, state, and federal law enforcement personnel with no special interest in deception (51 percent accuracy, on average).

Clinical psychologists performed better than academic psychologists, most likely because of their greater experience conducting interviews, said Ekman. A group of 107 clinical psychologists highly interested in deception and a group of 209 clinical psychologists moderately interested in deception had average accuracy rates of 68 and 62 percent, respectively. A group of academic psychologists had an average accuracy rate of 58 percent.

Previous studies by the UCSF researchers showed that U.S. Secret Service agents could determine when people were lying about their emotions. The current findings show for the first time that accurate judgments are not confined to selected law enforcement groups, such as the Secret Service, and that it is possible for motivated observers to detect the kind of lies law enforcement and intelligence agents routinely encounter.

Although some of the groups tested better than others, a substantial number of people in every category performed at or below chance. This suggests that even judgments made by professionals with training in deception will not always be accurate, said Ekman.

"Judging deception from facial expressions and body language will probably never be sufficiently accurate to be admissible in the courtroom," says Ekman. "Without the proper training and motivation, most people, even those entrusted by society to assess a person's trustworthiness, do quite poorly."

Perhaps the most surprising result uncovered by the researchers was that the most accurate groups were better at detecting lies than detecting truths. The federal officers group accurately detected lies 80 percent of the time but detected truths only 66 percent of the time. The same pattern held true for both groups of clinical psychologists, and a group of 43 Los Angles County sheriffs.

In addition to Ekman, co-authors of the study include Maureen O'Sullivan, PhD, professor of psychology at the University of San Francisco, and Mark Frank, assistant professor of communication at Rutgers University.

###


Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.