Meta-Analysis

Meta-Analysis and Behavioral Cues

Researchers have sought cues to deception—behaviors that distinguish liars from truth-tellers. They have examined many aspects of behavior—facial behaviors (such as eye contact), bodily movements (such as torso shifts), aspects of speech content (such as self-references), and vocal variables (such as pitch). Meta-analysts have assessed the strength and direction of the relationship of deception to these various behaviors, which have been examined in more than one study.

As the accumulated evidence shows, there is no infallible cue to deceit. Cues to deception are faint. The least faint cues indicate that liars sound distant and uncertain about what they are saying. They appear nervous and ambivalent. Their remarks lack detail. To each of these generalizations, there are exceptions. Certain truth-tellers, for instance, appear more nervous than certain liars, even though people (on average) appear more nervous when lying than when telling the truth. Meta-analysis yields a precise quantitative index of the difference between liars and truth-tellers in apparent nervousness, a so-called effect size. Effect sizes have been determined for over [Page 665]100 potential deception cues. The more strongly a behavior is diagnostic of deception, the larger is its effect size.

People make judgments of deception every day, inferring, for instance, that a coworker is lying, and that a child is telling the truth. While some investigators have sought cues to deception, others have hoped to identify the bases for deception judgment—cues from which people infer that others are being deceptive. Meta-analysis shows that behavioral variables are more strongly related to deception judgment than to deception itself. Thus, one could better predict from a murder suspect's behavior whether the suspect would be perceived as lying than whether he was, in fact, lying. People are most likely to be perceived as lying if they appear incompetent, ambivalent, or indifferent; if their remarks do not place events in context; and if they fidget. Although there is no perfect sign of apparent deception (no behavior from which every observer in the world would invariably infer deceit), cues to deception judgment are (statistically speaking) stronger than deception cues. Precise effect-sizes for many judgment cues are available.

...