Oh what a tangled web we weave, when trying to determine who deceives. Virtually everyone, even those experienced at dealing with deceivers, detect others’ lies no better than would be expected by chance.
Those sobering conclusions come from the first large-scale analysis of individual differences in deception detection. It takes two to tangle in deceptive encounters, note Charles Bond Jr. of TexasChristianUniversity in Fort Worth and Bella DePaulo of the University of California, Santa Barbara. The two psychologists say their analysis of the findings to date suggest some people are relatively easy to read, while others shroud their intentions in mystery.
A person’s perceived credibility, as reported by volunteers on questionnaires, rather than honesty, plays a major role in whether that person gets branded as a liar, Bond and DePaulo report in the July Psychological Bulletin. Certain people appear either honest or dishonest from the get-go, whether or not they’re telling the truth, the psychologists assert. Earlier research has found that baby-faced people seem credible whereas people who look nervous or avert their gaze typically get labeled untrustworthy.
The new analysis shows that participants more often believe liars perceived as high in credibility than truth-tellers regarded as low in credibility.
“When all the evidence is statistically analyzed, deception judgments depend more on the liar than the judge,” Bond says.
The new investigation challenges a view, championed by psychologists Maureen O’Sullivan of the University of San Francisco and Paul Ekman of the University of California, San Francisco, that a small number of individuals with considerable experience in unraveling certain kinds of lies do so with great accuracy. O’Sullivan and Ekman have found that a minority of psychotherapists quickly discerns lies about what a person says he or she is feeling, whereas insightful police officers readily discern a suspect’s crime-related deceits.
“There are significant differences among individuals in lie detection accuracy if you pick your subjects appropriately,” O’Sullivan says.
Bond and DePaulo disagree. They devised a new statistical method for estimating the range in the percentage of lies and truths that groups of volunteers would accurately identify if a lie-detection test was infinitely long. The technique corrects for measurement errors that occur on standard lie-detection tests, especially those requiring only a few true-or-false judgments.
The researchers applied this statistical tool to data from 142 earlier laboratory studies of lie detection. In these investigations, 19,801 judges assessed the veracity of 2,945 people conveying either true or false information. Many studies involved only college students as either judges or potential liars, but a substantial minority consisted of people with real-world lie-detection experience who were making deception judgments relevant to their professions.
Overall, participants accurately detected lies an average of 54 percent of the time, when an overall average of 50 percent would be expected by chance. This figure aligns with what researchers already knew.
But Bond and DePaulo focused on an individual’s performance, not a group average. They found that the highest detection rate achieved by an individual in these studies, which peaked at about 75 percent, did not exceed the maximum rate that guessing would have yielded, the researchers say. Individual differences in lie-detection accuracy were small, with scores clustering near the overall average of 54 percent correct.
Experienced judges displayed no lie-detection advantage over inexperienced ones. Neither did judges show greater accuracy in evaluating highly motivated liars, such as crime suspects, compared with less-motivated liars, such as college students pretending to have stolen money.
The researchers also found that the tendency to label someone as a liar also depended on whether a judge regarded other people as generally truthful or not.
Bond and DePaulo call for experiments that examine the complexity of real-world lie detection. Outside the laboratory, people infer deception from many lines of information, not just a person’s immediate behavior and speech, they say. In these situations, lies get identified over days, weeks or longer, rather than at the time a lie is told.
O’Sullivan also sees a need for research that addresses such issues. But she maintains that some people, due to their professional experiences, can quickly detect certain types of lies. In a new study submitted for publication, she and her colleagues find that experienced police officers rapidly identify high-stakes lies told by actual crime suspects far more often than they identify low-stakes lies told by students.