Wizards of deception detection are rare individuals who achieve scores of 80% or higher on at least two of three videotaped lie detection tests. Most people’s accuracy on these tests is about 50%, as would be expected by chance alone. Of more than 15,000 people tested, only 47 have been so classified. Although these individuals are termed “wizards,” their accuracy is not due to magic but to a particular kind of social-emotional cognition coupled with a strong motivation to discern the truthfulness of others.
Although the exact distribution of the ability to detect deception is not known, increasing evidence suggests that it is distributed mesokurtically (normally), like many psychological and physical variables. Among a hundred randomly selected people, most will be average in height. Only a very few will be exceptionally short or exceptionally tall. So, too, with lie detection. Most people are average in lie detection ability, but a very few (i.e., truth wizards) will be highly accurate.
Much of the research on lie detection has focused on identifying behaviors that differentiate between honest and deceptive behaviors. Implied but not stated in such research is the belief (or hope) that such behaviors can be used to detect automatically whether someone is lying or not. Certainly, there is evidence that some behaviors are more or less likely to occur in deception than in honesty. To date, however, no single behavior has been identified that always or usually occurs when someone is lying. Although some people have “tells”—behaviors they exhibit when they are lying, such tells vary from person to person, and not everyone has them. Another complication is that verbal and nonverbal behaviors related to deception do not occur in isolation. They are part of an expressive system that communicates a variety of information, such as emotions, thoughts, feelings, habits, social class, health, age, and many other aspects of individuality. The behaviors of liars and truth tellers must be evaluated in terms of their appropriateness for the individual, the situation, the statement being made, the relationship with the person discussing the veracity of the information, the stakes in the situation, and the rewards or punishments involved. Consistency among behaviors and the authenticity of any given behavior must also be evaluated. Thus, the task of detecting deception shares many characteristics with other judgments under uncertainty, including those involved in social cognition and social-emotional intelligence.
Most truth wizards are exceptionally sensitive to verbal and nonverbal clues to emotion and cognition. They notice facial expressions, including micro expressions, which most people do not. They are sensitive to nuances of language. They are aware of vocal clues—pitch, resonance, and respiration. They do not use just one of these cue domains but several of them. Average lie detectors attend to a more limited array of behaviors. Expert lie detectors are also more sensitive to baselines—whether the baseline is the person’s usual behavior or the person’s personality, social class, gender, or ethnicity. Wizards use these baselines to evaluate the behavioral clues they have perceived. On the other hand, no wizard uses all the available deception clues, and no single wizard is 100% accurate. Wizard accuracy, like that of most people, is affected by emotional disruption (e.g., someone looks like an ex-girlfriend) or lack of familiarity with a particular kind of lie.
A defining characteristic of almost all truth wizards is the motivation to know whether someone is lying or not, coupled with extensive experience in observing many kinds of people and obtaining feedback about their behavior. Most people are not highly motivated to know the truth. In fact, most people have a truthfulness bias, a tendency to call a higher percentage of people truthful than the base rate would suggest. But some people, including most wizards, because of their profession or because of events in their personal life, report a drive to know the truth. They do not show the cognitive laziness that most people exhibit when making social judgments.
Wizards range in age from 25 to 75, although most are middle aged. They include extroverts and introverts, liberals and conservatives, believers and atheists, heterosexuals and homosexuals, men and women, and people of many ethnic groups. Some are intellectuals with advanced degrees; others are high school graduates.
Truth wizards were identified after testing thousands of college students as well as professional groups with an interest in accurate lie detection. Among such unselected groups only one in a thousand qualified as a wizard. The discovery of several highly accurate groups (e.g., Secret Service agents, federal judges) suggested that focused testing of such professions would produce a much higher percentage of truth wizards. In such preselected targeted groups, the yield of wizards ranged from 5% to 20%.
The ability to predict the professional groups within which wizards are more likely to occur is one demonstration of the construct validity of the identification method used. Like all measurement methods, however, this method has limitations. Few expert lie detectors are equally good at detecting every kind of lie, even with the small sample of lie types used in the wizard research. In addition, there are many individuals who are good at lie detection in real life whose talent will not be assessed accurately by watching a videotape of someone else’s interview. So a videotape-test method will be subject to false negatives. But some people’s ability (including that of truth wizards) can be measured accurately in this way. The construct validity of the procedure used is bolstered by the professional achievements of many of the wizards (some of them have been featured in books and TV shows for their lie detection abilities and “people sense”) as well as the increasing efficiency in identifying the groups in which they are located.
Intense examination of the processes used by truth wizards in evaluating truthfulness has uncovered behavioral and attributional clues that have not yet been studied in other research on lie detection. The methods of person perception used in real life by truth wizards can be used to test the theories of interpersonal sensitivity and social cognition developed in the laboratory and to develop better methods for training lie detection professionals.
- Ekman, P. (2001). Telling lies (3rd ed.). New York: Norton.
- Ekman, P., O’Sullivan, M., & Frank, M. (1999). A few can catch a liar. Psychological Science, 10(3), 263-265.
- O’Sullivan, M. (2003). The fundamental attribution error in detecting deceit: The boy-who-cried-wolf effect. Personality and Social Psychology Bulletin, 29(10), 1316-1327.
- O’Sullivan, M. (2005). Emotional intelligence and detecting deception. Why most people can’t “read” others, but a few can. In R. Riggio & R. Feldman (Eds.), Applications of nonverbal communication (pp. 215-253). Mahwah, NJ: Lawrence Erlbaum.
- O’Sullivan, M. (2007). Unicorns or Tiger Woods? Are expert lie detectors myths or rarities? A response to: On lie detection “wizards” by Bond and Uysal. Law and Human Behavior, 31, 117-123.
- Detection of Deception: Cognitive Load
- Detection of Deception: Nonverbal Cues
- Detection of Deception in Adults
- Detection of Deception in High-Stakes Liars
- Statement Validity Assessment (SVA)