Emotional Intelligence and Deception Detection: Why Most People Can't 'Read' Others, But a Few Can

Emotional Intelligence and Deception Detection: Why Most People Can't 'Read' Others, But a Few Can

In this chapter we will consider why most people seem impervious to the many nonverbal cues that they could use to understand the thoughts, feelings and intentions of others. Evidence wi11 be offered that there are such cues that can be used to detect deception, as well as some initial findings from a small group of lie detection 'wizards' which suggest that at least some people are able to use these cues in understanding others.

Although other research areas (e.g., social cognition, personality as sessment) could be surveyed, this review focuses on the relevance of nonverbal cues to detecting deception as a particular example of what is currently called emotional intelligence (Goleman, 1995; Salovey & Mayer, 1990), but which has also been referred to as social intelligence (Thorndike, 1920), empathy (Lipps, 1926; Ickes, 1993; Mehrabian & Epstein, 1972), social insight (Chapin, 1942), behavioral intelligence (O'Sullivan & Guilford, 1975), applied intelligence (Sternberg, 1986) or Intra- and Interpersonal intelligence (Gardner, 1993).

It is well known that nonverbal clues are involved in 'reading' peo ple as trade books such as 'Reading people' (Dimitrius & Mazzarella, 1999) presume. Books with titles like 'Never be lied to again' (Lieberman, 1998) and 'Conquering Deception' (Nance, 2000) sug­gest that understanding nonverbal clues involved in deception is as easy and natural as learning to speak your native language. This chap-ter will demonstrate that this is not true.

The first part of the chapter will review the evidence, both theoreti­cal and empirical, for the relevance and importance of nonverbal be­havior in emotional intelligence and in detecting deception. The middle part of the chapter will review a wide variety of explanations for why most people don't use these cues. It may seem counter-intuitive, in a book about applications of nonverbal communication, to discuss reasons why most people are unable to make such applications. My hope is that pointing out the difficulties involved in using nonverbal cues may suggest techniques for alleviating them. The last part of the chapter will describe an ongoing study with a small group of expert lie detectors who avoid the many difficulties experienced by most people and are able to use nonverbal behavior (and other information) accu­rately to determine whether a person is truthful or not.

ONE HUNDRED AND THIRTY YEARS OF RESEARCH
ON NONVERBAL BEHAVIOR AND EMOTIONAL/SOCIAL
INTELLIGENCE: A BRIEF SURVEY

In 1872, Darwin's description of the nonverbal behaviors involved in the emotional communications of humans and animals (1998) was not only one of the earliest, but has been the most influential. Other early psychologists also described nonverbal behavior. In his theory of emotion, for example, James (1884) emphasized the centrality of the body, so it is no surprise that he was an astute observer of what we now refer to as nonverbal communication. 'Can one fancy the state of rage and picture no ebullition of it in the chest, no flushing of the face, no di­latation (sic) of the nostrils, no clenching of the teeth, no impulse to vig­orous action, but in their stead limp muscles, calm breathing, and a placid face?' (p. 194). Similarly, Wundt (1897) was well aware of ' expressive movements (that) correspond exactly to the psychical ele­ments of emotions and their fundamental attributes: intensity quality, and ideational content' (p. 173).

In the first part of the twentieth century, many eminent psycholo­gists attempted to understand and measure the ability to understand others through the use of nonverbal behavior. Among these was E. L. Thorndike (1920) who wrote that people 'had varying amount of dif­ferent intelligences' (p. 228). He distinguished three of them: Abstract (verbal and symbolic) intelligence, mechanical intelligence and social intelligence, or 'the ability to understand others, to manage (other peo­ple) wisely.' His article contained photographs of a woman posing various facial expressions of emotion, though he noted the limitations of such stimuli and underscored the importance of real-life and inter-active stimuli in composing tests of social intelligence.

Many leading psychologists of the era attempted to measure the ability to understand the expressive behavior of others. Boring and Titchener (1923) devised a schematic profile that was disappointingly unreliable. Guilford (1929) studied individual differences in the abil-

ty to recognize facial expressions of emotions while Frois Wittmann (1930) produced a series of photographs of himself posing a variety of facial expressions of emotion. But the most ambitious attempt to mea­sure emotional/social intelligence was made by Moss and his col-leagues (Moss, 1931; Moss, Hunt, Omwake, & Woodward, 1955) who developed the George Washington University Social Intelligence Tests, some of which are still in use today. A series of factor analyses of the Moss tests (Thorndike & Stein, 1937) demonstrated that the tests showed convergent validity (they were highly inter correlated with one another), but they did not demonstrate discriminant validity (they were also highly correlated with measures of verbal ability).

From the 1930s to the 1960s little work on individual differences in expressive or nonverbal behavior was done. One exception was a 1924 dissertation by Wedeck published in 1947. He devised a series of tests based on the assumption that nonverbal behavior is essential to what he called 'psychological ability ' He did not pursue this research and so his work never received the attention it merited.

Interest in expressive (nonverbal) behavior never totally died out, however. A few researchers remained interested in the face, the voice and gesture. Efron (1941) studied the transmission of hand gestures across generations in two groups of New York immigrants. His de­scription of emblems, nonverbal gestures that substitute for words, is still used. Engen, Levy, and Schlosberg (1957) studied the Marjorie Lightfoot series of photographs, which was used in many early studies of emotion recognition.

Attempts to measure social/emotional intelligence also continued. Examples of measures from this early period that are still in use are Chapin 's Social Insight Test (1942), which describes interpersonal problems and asks the examinee to choose the most accurate or wisest comment about it.

Mehrabian & Epstein (1972) developed a self-report instrument of empathic accuracy that asks examinees how astute they are in re­sponding to nonverbal behaviors. Although termed accuracy, this is a self-report measure and score variations may or may not correlate with scores on tests of empathic ability or aptitude. More contempo­rary measures of self-reported social skill have been provided by Riggio (1989) and Bar-On (Bar-On, Brown, Kirkcaldy, & Thorne, 2000).

Lipps' (1926) analysis of empathy focused on the internal feelings that one person has in response to another person rather than on their perceiving and responding to another person's nonverbal (external) behavior. Although it is likely that nonverbal behavior is at least a me­diator of this process, Lipps' analysis emphasizes the internal pro­cesses involved rather than issues of accuracy. His terminology survives in Ickes' (1993) work on empathic accuracy. Ickes and Simp sons (1997) discussion of the processes involved in empathic accuracy is relevant to understanding some of the problems described later in

this chapter concerning why most people are inaccurate in detecting deception, a kind of empathic accuracy.

With the increasing availability of photographs and videos, the 1960s and 1970s saw a marked increase in attempts to measure so cial/emotional intelligence. Guilford (1956) initiated a series of studies based on his Structure of Intellect model. O'Sullivan and Guilford (1975) devised 23 different measures of 'cognitive behavioral intelli­gence' based on Guilford's model of intelligence and the belief that ear­lier social intelligence tests failed to achieve discriminant validity because they were mostly verbally stated.

They devised tests that were totally nonverbal, except for orally pre sented instructions. Although the tests defined the hypothesized fac­tors and some of them continue to be used, there has been limited work relating them to real-life criteria.

More successful in this regard is the Profile of Nonverbal Sensitivity (PONS; Rosenthal, Hall, DiMatteo, Rogers, & Archer, 1979), 90 video taped items showing a young woman posing social situations such as giving directions, comforting a lost child, praying, and getting angry. Research on the PONS was characterized by the creativity with which it demonstrated construct validity. Discriminant validity was evidenced in the lack of correlation with measures of IQ. Convergent validity was supported by predicted correlations with performance as a foreign service officer, being the mother of a pre-verbal child and other, equally intriguing criterion measures. Convergent validity with other measures of nonverbal sensitivity, however, has been limited.

Archer & Akert (1977) designed the Interpersonal Perception Test (IPT) which has a number of salutary features. Chief among them is the undeniable accuracy of its scoring key; the items of the IPT show real life situations for which the truth is known. One item shows two adults playing with a child. The question is: Which adult is the parent? An-other item shows two people talking: Which is the supervisor? In both cases, the correct answer is indisputable. Two of the 15 items of the IPT also show two people lying and telling the truth about events in their lives. The task is to pick the scenario in which each person is ly ing. The IPT is still available for use (Costanzo & Archer, 1993); how ever, its strength is also its weakness. It is likely that many different kinds of social/emotional intelligence are measured by this instru­ment, but each is represented only by a few items. The total score probably reflects a variety of social skills.

Buck (1976) developed the Communication of Affect ReceivingAbil ity Test in which men and women were videotaped watching neutral or emotionally arousing slides. Given the difficulty and ethical restraints of arousing emotion in the laboratory, most of the facial expressions shown were subtle or ambiguous, so the meaning of scores on this measure is unclear.

Ekman and Friesen (1975) developed a BriefAffect Recognition Test (BART), in which photographs of prototypic facial expressions were

shown for 1/15th or 1/30th of a second. Almost everyone can accu­rately identify facial expressions of basic or prototypic emotions (such as happiness, fear, or anger) if they are presented for a second or lon­ger. With briefer exposure times, a nearly normal distribution of accu­racy scores results. The technology available when BART was produced meant that in order to simulate micro-momentary facial ex­pressions, the photographs had to be presented tachistoscopic ly, which presented several difficulties (O'Sullivan, 1982).

More recently, Ekman (2003) has developed two CD's that are com­bination training tools and micro-facial-expression recognition accu­racy measures. The Measuring Emotional Expressions Tool (Ekman, 2003) contains 56 colored photographs in which a neutral face is the background on which a prototypic facial expression of the same per-son is then flashed at speeds of 1/15th or 1/30th of a second. The Sub­tle Emotional Expression Tool (Ekman, 2003) contains black and white photographs of one young woman posing subtle variations of many different facial expressions. The tool allows the user to self-test at various speeds.

Ekman has also developed several measures of the ability to detect deception, a more narrow aspect of emotional intelligence, but one which may identify people with emotional intelligence in other areas as well. His first test (Ekman, Friesen, O'Sullivan & Scherer, 1980) shows nurses lying or telling the truth about whether they are watching a pleasant nature film or a gruesome surgical one. Two other deception detection measures (Frank & Ekman, 1997) show young men being in­terviewed about whether or not they stole $50 or are telling the truth about a controversial opinion. These measures were used to identify the expert lie detectors described at the end of this chapter and will be referred to, respectively, as the emotion, crime and opinion tapes since these are the topics about which the participants lied or told the truth. Behavioral measurements of all three tapes indicate that there are sig­nificant nonverbal clues that could be used in accurately assessing truthfulness or deception.

O'Sullivan and Ekman (O'Sullivan, 1983) developed the Affect Blend Test (ABT), in which people of different ages posed facial ex­pressions in which two or more different emotions were combined. The photographs were produced based on criteria developed from the Facial Action Coding System (Ekman & Friesen, 1978). TheABT is of interest because it relates to the work on deception that Ekman and his colleagues have done over the last thirty years, in which leakage of emotion is important. That is, while attempting to portray one emotional state (i.e., pleasant unconcern) the 'real' emotion (fear of apprehension, or guilt, or delight at duping the interviewer) will 'leak' in fragments of the facial expression that the liar cannot fully control. The ABT simulates this phenomenon by showing com­plex mixes of emotions in which the constituent parts are readily identifiable.

By the 1980s, as interest in cognitive psychology increased, interest in nonverbal behavior per se waned. TWo major theoreticians of hu­man intelligence, Sternberg (1988) and Gardner (1993) described un­derstanding self and others as an aspect of human cognitive ability in which there were marked individual differences. Sternberg empha­sized what he called 'practical intelligence,' the ability to understand what is necessary in a situation, to be able to try different solutions to solve interpersonal problems and then to have the practical wisdom to leave a situation if one's best efforts have failed. Although Sternberg and his students published interesting work related to academic pro-motion and business success, the measures of practical intelligence were not easily accessible, and the measures of nonverbal behavior (Sternberg, 1986) that were provided did not seem superior to similar ones already available, such as Costanzo and Archer's IPT. Gardner's theory of multiple intelligences includes both an interpersonal and an intrapersonal ability Although no psychometrically validated scales for these abilities have been published, his ideas have been well-re­ceived in educational circles.

The fulcrum for research in emotional/social intelligence, however, was the publication of Goleman's best-selling book (1995) on the topic. Based largely on Salovey and Mayer's (1990) work on emotional intelligence, Goleman's contribution was to link the idea of under-standing the emotions of others to the concept of understanding and mastering one's own emotions for pro-social purposes. His popular­ized presentation of these concepts spurred an explosion of research on emotional intelligence over the next several years.

Although Goleman's book was long on exhortation and description, it was short on measurement methods. So, relatively hoary instru­ments like the O'Sullivan-Guilford tests, the Chapin Social Insight Test, and the IPT were resurrected. More recently, several new mea­sures have been constructed. Mayer, Salovey and Caruso (2002) devel­oped the MSCEIT, a multifactorial measure of emotional intelligence in which the criterion is determined largely by consensual validation. Cronbach (1955) distinguished differential accuracy (what most ear­lier measures of emotional intelligence attempted to assess) and ste­reotypic or consensual accuracy, which is the basis for the MSCEIT. Another unusual aspect of the MSCEIT is that sensitivity to artistic ex­pressions of emotion is measured alongwith sensitivity to human non-verbal expressions of emotions. Nowicki (Nowicki & Duke, 2001) has produced several measures of emotional intelligence that emphasizes differential accuracy of affect recognition in face and voice. Separate measures are available for adults and children. A number of studies have been conducted using these measures. The DANVA (Diagnostic Analysis of Nonverbal Accuracy) is not highly correlated with general intelligence (discriminant validity) and shows some evidence of con­vergent validity through prediction of school achievement (not school aptitude) and at least one significant correlation with the PONS.

Although conducted over a 130 year period, these research efforts have at least one commonality. They assume that understanding other people's thoughts, feelings and intentions depends on under-standing their expressive, nonverbal behavior. This book presents a number of approaches to understanding nonverbal communication. Let us now consider what is known about nonverbal cues to under-standing one aspect of another person-whether she is lying or telling the truth.

NONVERBAL CUES TO DECEPTION

Nonverbal behavior is an integral and distinguishing characteristic of the self-presentation and coherence of each person (DePaulo, 1992). So, each individual varies in the ways in which she behaves when lying, or puzzled, or anxious. Researchers have identified some of the non-verbal behaviors that distinguish lying from truthful communication, but those behaviors don't occur in every case, and they must always be interpreted in the context of the unique individual displaying it.

Although other reviews have been done (Zuckerman & Driver, 1985), the current gold standard in reviews of nonverbal cues to de­ception is the meta-analysis performed by DePaulo, Lindsay, Malone, Muhlenbruck, Charlton, and Cooper (2003). They identified 158 cues to deception that were studied using 1,300 different estimates from 120 independent samples. Like all meta-analyses, theirs evaluated, by summing over similar cues, whether there were recurring and signifi­cant differences in the nonverbal behavior shown by truth-telling and deceiving people.

The meta-analyses done by DePaulo and her colleagues are notable for their clarity, inclusiveness and organization. The effect sizes are presented in an accessible and engaging fashion, with 83 of the 158 cues discussed in the light of DePaulo's self-presentational theory. A drawback of this approach, however, is that some cues (those based on smaller numbers of estimates, or not relevant to the self-presenta­tional theory) were included only in an appendix. Also, the cues are not separated into verbal and nonverbal categories. For the purpose of this chapter, therefore, I have taken the liberty of re-ordering the 158 cues. Rather than discussing the relevance of the cues to a particular theory of deception, they are listed based on whether they are nonverbal cues (in whole or part) or verbal cues.

A complete list of the nonverbal and mixed cues is given in Tables 10.1 and 10.2. The mixed cues are those that involved both verbal and nonverbal elements, such as 'verbal and vocal immediacy.' The decep­tion research literature, as synthesized by DePaulo and her colleagues, suggests that somewhat more than half of the cues that have been stud­ied as cues to deception are nonverbal or mixed (n = 87); the rest are verbal (n = 71 ). In terms of significant effect sizes, about 25% of each of the two domains of cues (nonverbal and verbal) is significant, sug-

Nonverbal and Mixed Cues to Deception With Significant Effect Sizes'

a.     Adapted from DePaulo et al., (2003). d's were computed by subtracting the mean for truthful persons from the mean for liars and dividing that value by the average of the standard deviations obtained by the two groups (truthful and deceptive). Positive d's mean that the liars obtained significantly higher scores; negative d's that the truthful persons obtained higher values. Values for d greater than 0.50 (in bold) represent large effect sizes.

b.     Id's are the numbers assigned by DePaulo et al.

* Measured using subjective methods such as ratings or overall impressions.

* * Vocal tension and pitch were measured by both subjective and objective methods.

gesting that liars and truth tellers can be distinguished on the bases of both nonverbal and verbal behaviors.

Verbal cues included measures such as word and phrase repeti­tions, plausibility of the statements made, and contextual embedding. The effect size used in the DePaulo meta-analysis was ' d, defined as

Nonverbal Cues to Deception With Non-Significant Effect Sizes'

(continued)

a. Adapted from DePaulo et al., (2003).

b. Id's assigned by DePaulo et at.

* Variables measured using subjective methods.

* * Measured using both subjective and objective methods.

the mean for the deceptive condition (i.e., the lies) minus the mean for the truthful condition (i.e., the truths), divided by the mean of the stan­dard deviations for the truths and the lies . Positive ds therefore indi­cate that the behavior occurred more often during the lies than the truths, whereas negative ds indicate that the behavior occurred less of-ten during lies than truths' (DePaulo et al., 2003, p. 89). Table 10.1 lists the nonverbal cues that had significant effect sizes, ordered ac-cording to the size of the effect. The table also gives the identification codes (from 1 to 158) and the value of the effect sizes reported by the DePaulo group. Table 10.2 gives the same information for the nonver­bal cues that did not have significant ds. A d of .50 or greater is consid­ered large. By that criterion, although 21 significant effect sizes were found for the nonverbal cues to deception, only six of them qualify as large effects. These are:

1.  Liars show more changes in their foot and leg movements. (d = 1.05)

2.   Liars show more changes in the size of their pupils. (d = 0.90)

3.   Liars seem indifferent or unconcerned. (d = 0.59)

4.   Truthful people show more genuine (Duchenne) smiles than li­ars. (d = -0.70)

5.  Truthful people seem more cooperative. (d = -0.66)

6.   Truthful people have more verbal and vocal immediacy. (d = .55)

The organization of the nonverbal cues presented here was dictated by the focus of this chapter. The estimates of some of the ds are based on more independent samples than others and may, therefore, be more reliable. In general, but not always, cues with lower numbered identification codes (83 or less) are based on more independent sam­ples than cues with codes greater than 84.

THE ISSUE OF DISCREPANCY

Although many deception theorists (e.g., Bugental, Kaswan & Love, 1970) have suggested that discrepancy, either within a channel or among verbal and nonverbal channels is a salient cue to deception, only one such clue (discrepant, ambivalent, d = 0.34) is represented among the 158 cues surveyed. This variable, surveyed over several studies, is described as 'Speakers' communications seem internally inconsistent or discrepant; information from different sources (e.g., face vs. voice) seems contradictory; speaker seems to be ambivalent ' (DePaulo et al., 2003, p. 113). Obviously, given the difficulties of measuring nonverbal behaviors, the added difficulty of reliably deter-mining discrepancies among them seems to have limited researcher's enthusiasm for doing so. The fact that the one variable that did assess

discrepancy was significantly discriminating of liars and truth tellers is noteworthy.

Individual Differences and Deviations From Baseline

Several writers have urged human lie detectors to ground their obser­vations of others in the baseline behavior of the person being ob­served. Zuckerman, DeFrank, Hall, Larrance and Rosenthal (1979) described what they called the demeanor bias, a tendency for people to be seen as always honest or always deceptive regardless of their ac­tual veracity. Ekman (2001) cautioned against the idiosyncrasy er­ror-always interpreting a particular nonverbal behavior, such as changes in foot and leg movements, without determining whether this is a recurring aspect of a person's ordinary behavior. On average, liars are significantly more likely to show more changes in their foot and leg movements than truthful people (see Table 10.1), but not ev­eryone does this when lying, and some truthful people may character­istically jiggle their feet. As with discrepancies, although theorists recommend attending to such behavior changes, few of the 158 cues surveyed in the DePaulo meta-analysis measured such changes. Seven of the 158 deception cues reflect change from baseline and of these, two (changes in foot movement and pupillary change) have the highest ds reported: 1.05 and 0.90 respectively.

THE GOLDILOCKS PHENOMENON:
TOO BIG, TOO SMALL, AND JUST RIGHT

There Are Smiles and There Are Smiles

Another finding, more noticeable when the data are grouped as they are in Tables 10.1 and 10.2, is the importance of the level of measure ment used in describing nonverbal behavior. On the one hand, for dis crete behavioral entities, it seems better to assess variables in a highly specific, well-defined manner. For example, genuine or Duchenne smiles had one of the higher effect sizes (d = -0.70), suggesting that truth tellers wi11 show more of these kinds of smiles than liars. On the other hand, feigned smiles were more frequent among liars, although the d associated with this difference was insignificant. Obviously, if the two kinds of smiles are summed, they will cancel each other out, and show no overall effect, since one kind of smile is more frequent in truth tellers and the other is more frequent in liars. In fact, that is what hap­pened. Undifferentiated smiles had an insignificant effect size. Lance's study (chap. 7, this volume) is another example of the im­portance of differentiating among different kinds of smiles. She video-taped the nonverbal behavior of applicants who were harassed during their job interview. Applicants who responded to the harassment with

non-Duchenne (feigned) smiles were rated by as less competent and less likely to be hired than women who showed felt enjoyment (Du­chenne) smiles.

Hand and Feet Movements

Another example of the importance of specificity in measuring nonver­bal behavior is the very high effect size (d = 1.05) for changes in foot and leg movements ('Changes in the number of foot or leg movements over time (absolute value),' DePaulo et al., 2003, p.117) and specific hand and arm movements (d = -.0.36) ('Hand movements that do not include arm movements and finger movements that do not include hand movements.' DePaulo et al., 2003, p. 116). Notice, however, that these significant effect sizes go in opposite directions! Liars are signifi­cantly more likely to show changes in foot and leg movements and sig­nificantly less likely to show the specific hand and arm movements described. Further complicating this issue is Vrij's report (this vol­ume) that some liars show particular hand and arm movements.

Illustrators vs. Emblems

This principle of specifying exactly what behavior is of interest is also demonstrated by the findings with two different hand gestures-illus­trators and emblems. Illustrators, hand gestures accompanying speech, were significantly more frequent among truth tellers (d = -0.14), but emblems, gestures that can substitute for speech, are not significantly different in liars and truth tellers (d = 0.01). It is likely that a generic 'hand gestures' category would not have distinguished liars and truthful people whereas illustrators do.

Fidgeting

On the other hand, overall ratings of the amount of fidgeting ('Object fidgeting and/or self-fidgeting and/or facial fidgeting (undifferenti­ated)' DePaulo et al., 2003, p. 115) are significantly higher for liars than truthful people (d = 0.16). The component parts of this variable (object fidgeting, self-fidgeting and facial fidgeting) are not significant in their own right, but are when included in a summary measure. The d associated with overall fidgeting is small, albeit significant. This sig­nificance may be due merely to the large number of estimates (n = 14) on which it was based.

So, why is the summary measure significant for fidgeting while the components are not? My speculations areas follows: Differences between smiles and between illustrators and emblems are theoretically based (Ekman & Friesen, 1969) and were predicted beforehand to relate differ­ently to truthful and deceptive behavior. A change from one's baseline be-

havior has also been postulated as a marker of discomfort, or emotional or cognitive shift so the significant finding that changes in foot and leg movements tend to characterize liars and not truth tellers was predicted. (The finding with specific hand and finger movements is not predicted by the baseline thesis, however. See Vrij, chap. 4, this volume.)

Fidgeting is generally interpreted as a sign of nervousness or anxi­ety, although there are no data relating fidgeting to self-reported anxi­ety or physiologically-related arousal. There is no theory, however, about what kind of fidget goes with what kind of lie, what kind of emo­tion or what kind of cognitive activity. So whether the fidget involves an object, the self or the body is irrelevant and probably relates only to id­iosyncrasies of the individual rather than the lying per se. at is im­portant is to get a reliable overall estimate of fidgeting and this is best done, it seems, by summing across all fidgeting occurrences. Other sources of information for which a general measurement seems more productive are attributional ratings.

Trait Ratings of Liars and Truth Tellers

Some of the more significant ds were obtained with overall subjective ratings of attributes of the person being judged. As seen in Table 10.1, ratings of cooperative, indifferent or unconcerned, verbal and vocal immediacy, seems planned or not spontaneous, discrepant, or ambiv­alent and nervous, tense (overall) significantly discriminated liars and truth tellers. Note, however, that all of these descriptions relate to a particular gestalt-of people who are open, concerned, immediate, spontaneous, consistent, and generally relaxed on the one hand (i.e., truth tellers) and uncooperative, indifferent, distant, non spontane­ous, ambivalent, and tense or nervous on the other (i.e., liars). Ratings not included in this particular gestalt, such as attractive, competent or friendly did not significantly differentiate liars and truth tellers. More recently, O'Sullivan (2003) reported that there were no differences be­tween liars and truth tellers in how observers rated their likeability in­telligence or attractiveness. Also, one of the more perplexing findings in the field of deception research is that most people cannot accurately identify when others are lying or telling the truth. In other words, rat­ings of 'honest' or 'trustworthy' do not significantly differentiate liars and truth tellers. Reasons for this particular rating error are dis­cussed below. At this point, what should be noted is that subjective overall ratings of personal characteristics with relevance to honesty (but not ratings of honesty themselves) significantly distinguish liars and non-liars. Such ratings, by necessity, are general summations of many aspects of a person's behavior. The inability of raters/observers to make the summary conclusion that an individual who they perceive as uncooperative, non-immediate, non-spontaneous, tense, and am-bivalent may well be telling a lie is one of the more intriguing puzzles in the field of deception detection.

But why are attributional ratings included in a chapter on nonverbal behavior? Just as emotional intelligence researchers assume that knowledge of another's inner life must be mediated by observable exte­rior cues, i.e., nonverbal expressions, most interpersonal perception theorists would accept the idea that judgments about other people's at-tributes must come from observable aspects of them and their behav­ior. Self-ratings, on the other hand, are probably more dependent on internal processes than one's external nonverbal behaviors of which most of us are unaware.

Although the particular nonverbal behaviors that are being empha­sized in this chapter reflect dynamic cognitive, motivational and emo­tional aspects of people, other nonverbal behaviors are also available for use in detecting deception. As Sherlock Holmes (Doyle, 1892) rou­tinely demonstrated, general appearance provides a great deal of infor­mation about people. Visual cues include information about age, which implies information about values, health, interests, experiences and a host of other conclusions, many likely to be accurate. Visual clues also permit conclusions about attractiveness, vanity, ethnicity, concern for fashion, place of residence, health, social status, tidiness, occupation and paternity. Clothing, rings, pins, and other jewelry con­vey information about recreational activities, marital and sexual orien­tation, organizational memberships, educational level and religiosity. Vocal cues, in addition to emotional, cognitive and motivational pro­cesses, may give information about the country or region of origin, age, energy level, health, education and sexual orientation. Even olfactory cues can be used in judging others. Odor was a diagnostic indicator for physicians entering homes to treat the ill in the early part of this cen­tury. Pheromones affect sexual attraction, menstrual cycling and in­fants suckling. Although we are not consciously aware of these nonverbal indicators of biological processes, our bodies are and that information may guide our judgments more than we know. Thus, it seems appropriate to include those summary judgments that must be based on nonverbal cues with other, more objectively measured nonverbal cues to deception

WHAT KIND OF LIE IS IT?

In early work, Ekman and his colleagues (Ekman, et al., 1980), found no difference between liars and truth tellers in the number of shrugs that each group made. In more recent work (Frank & Ekman, 1997), in which men lied about strongly held opinions, shrugs that were incon­sistent with what was being said were more likely to occur when the men were lying. This suggests that the kind of lie told may affect the kind of nonverbal behavior involved in different kinds of deception. A related finding is discussed in the last part of this chapter describing work with expert lie detectors, in which the kind of lie told affected the lie detectors' accuracy.

DePaulo and her colleagues (2003) were also interested in the im­pact of the kind of lie told on nonverbal behavior. They reported that when the lie was about a transgression, liars had more changes in foot or leg movements, spoke at a faster rate, did more eye blinking and were rated as more nervous and tense.

In Lie Detecting, Ignorance Is Not Bliss

The DePaulo meta-analysis provides evidence that many culturally-held beliefs about what people look like when they lie are often incor­rect (Zuckerman & Driver, 1985; O'Sullivan, 2000; Vrij, this volume). In the United States, eye gaze aversion is widely assumed to be a cue to deception. But, overall, eye gaze aversion is not a significant difference between liars and truth tellers. This is an example of misinformation. Missing information is exemplified by nonverbal behaviors that are re-liable cues to deception that most people ignore. Examples include pitch (frequency) of voice, the quality of verbal and vocal immediacy, pupillary change and pupil dilation, and specific hand and finger movements. Because many of these behaviors are fleeting (micro-mo­mentary), people either ignore them or do not notice them at all.

at this review of nonverbal cues to deception suggests is that re-search in this area must specify exactly what kind of nonverbal behav­ior is of interest. The utility of greater specificity may be seen in Vrij's work on a particular kind of small hand movement in a particular situ­ation (chap. 4, this volume) and LaFrance's distinction between Du­chenne and non-Duchenne smiles in hiring (chap. 7, this volume). The theory behind predicting the occurrence of a particular kind of nonver­bal behavior should also be made clear, as Ekman, Friesen, and O'Sullivan (1988) did in predicting different kinds of smiles when peo­ple are lying or telling the truth. Ekman (2001) and others (Burgoon, Buller, White, Afifi, & Buslig, 1999) have emphasized the importance of the context within which the nonverbal behavior occurs. Ekman em­phasizes the baseline demeanor of the individual; Burgoon, the inter-personal relationship within which the deception or truthfulness occurs. O'Sullivan (2003) suggested that the cognitive heuristics involved in understanding other people should also be specified.

WHY MOST PEOPLE CAN'T TELL WHEN OTHERS ARE LYING

Although it is clear that there are significant mean differences be­tween honest and deceptive individuals in terms of the nonverbal be­haviors they show, most people are unable or unwilling to use them as a basis of detecting deception (Ekman & O'Sullivan, 1991; Malone & DePaulo, 2001). Many reasons for this difficulty in social cognition have been suggested or can be inferred from the literature. For ease of discussion, they have been organized into six categories: 1) strategic errors, 2) cognitive biases, 3) knowledge deficiencies, 4) characteris-

10. EMOTIONAL INTELLIGENCE AND DECEPTION DETECTION 23 I

tics of the liar or truth teller, 5) motivations of the lie catcher, and 6) evolutionary biases.

Strategic Errors

The first strategic error, and the one most relevant to the focus of this book, is most people's inattention to the nonverbal information avail able to them when attempting to detect deception. The fact that people disregard nonverbal information that could be useful in understand­ing others was demonstrated in an early study of how observers judge honest and deceptive behavior. Ekman and his colleagues (1980) found that although observers could not accurately identify liars and truth-tellers, they used different strategies in describing them. Verbal and nonverbal channels (facial expressions, hand gestures, body pos tures and vocal qualities) were used equally often when the behavior being judged was honest. When the behavior was deceptive, however, judges paid more attention to the speech (verbal) channel. Ekman and his colleagues offered three sources of evidence to support their con tention that one of the major differences between good and poor lie de­tectors is the greater utilization of nonverbal clues by good lie detectors. These sources are:

1) Self-reports of accurate lie detectors, who indicate using either nonverbal clues alone, or a combination of verbal and nonverbal clues. This was found in a large group of medical school respon dents (Ekman, private communication, June, 2003) as well as in groups of college students and fraud investigators (O'Sullivan, 2000). Less accurate lie detectors reported using mostly verbal clues;

2) A positive correlation between lie detection accuracy and a mea­sure of the ability to recognize micro momentary facial expres­sions of emotion (Frank & Ekman, 1997) and;

3) Greater accuracy in lie detection by people with left-brain lesions which force them to use nonverbal clues to make judgments of others (Etcoff, Ekman, Magee, & Frank, 2000).

The second strategic error is related to the first. When judging someone who is lying, people tend to pay attention to the content of speech rather than to vocal quality (DePaulo, Rosenthal, Rosenkrantz, & Green, 1982). In a follow-up to the study described above (Ekman et al., 1980), O'Sullivan, Ekman, Friesen, and Scherer (1985) demon­strated that observers paid more attention to the content of speech (what was said) than its vocal or nonverbal quality when they rated de ceptive behavior, but not when they rated honest behavior. In the mate­rials used, vocal pitch was significantly higher in the deceptive condition, so ignoring the nonverbal aspects of speech was not a wise

strategy. This finding illustrates two paradoxes. The first is that al-though the judges were not accurate in labeling people as honest or de­ceptive, they used different impression formation strategies in describing those who were honest and those who were deceptive. Judges used both verbal and nonverbal channels equally in describing honest behavior, but depended principally on the content of speech in describing their impressions of deceptive behavior. The second para­dox is that most people can control the content of their speech more easily than they can control their facial expression, voice quality or body language. Yet, in the face of inconsistent information, observers switch from a complex processing mode in which they weigh many in-formation sources, both verbal and nonverbal, to one in which they weigh most heavily the information channel that is most easily controlled.

In addition to the strategic errors of under-utilizing nonverbal clues and over-emphasizing the content of speech, a third strategic error is to believe that a single nonverbal clue always or nearly always indicates the presence of deception. Ekman and Friesen (1969) termed these de ception clues. They and other lie detection researchers (DePaulo et al., 2003) now know that a dedicated deception clue, like Pinocchio's nose, is just a fairy tale. There is no behavior, nonverbal or otherwise, that always occurs when someone lies. This does not mean that some individuals won't betray that they are lying with typical and consistent nonverbal behaviors. Champion poker players report being able to spot the 'tells' of their opponents. (Tells are behaviors that players show when they are trying to bluff and pretend they hold better cards than they actually do. Poker players might also be pleased or disap­pointed with the cards they have been dealt and have to hide those reactions as well.)

Champion poker players do not commit the fourth strategic error, the 'idiosyncrasy error.' Most people fail to allow sufficiently for the many unique and bizarre behaviors most people display as their base line behavior (Ekman, 2001). Even professional interviewers like Tom Brokaw will say 'I don't look at a person's face for signs that he is lying.

at I'm after are convoluted answers or sophisticated evasions.' (as cited in Ekman, 2001, pp. 90-91). There is ample evidence that many people will not show this behavior when they lie, so depending on an idiosyncrasy of some to detect the deceptiveness of all is a strategic er ror. Another example of the idiosyncrasy error is what Ekman (2001) termed the Othello error. In Shakespeare's play, Othello believes that his wife, Desdemona has been unfaithful. He interprets her fear as proof that she has lied to him. The truth, however, is that she fears that he will not believe her. He does not and kills her. Desdemona's fear was well-founded. People differ in terms of their reaction to not being be­lieved. Fear of not being believed can lead some innocent people to look guilty, leading as Othello tragically discovered, to the wrong conclusion.

Cognitive Biases

Although psychology has been in the throes of the 'cognitive revolu­tion' for more than 40 years, knowledge from that field has not been sufficiently incorporated into our understanding of the processes in lie detection. Another difficulty is that the names given to certain cognitive or motivational errors in deception research are not the same as those widely used by social-cognitive psychologists and so their existence, al-though well-documented, is not usually included in general theorizing about social cognition.

Truthfulness Bias (and Deception Bias) a.k.a. Availability Heuristic. Zuckerman and his colleagues (1979) were among the first to suggest that most people have a truthfulness bias-they tend to judge others as truthful most of the time. In the ordinary course of events, most people do not expect to be lied to; they presume that most people are honest. Consider the many interactions we have during the course of the day. 'What time is it?' 'Did you see my keys?' 'Listen to what I just read in the newspaper.' If people tell two lies a day (DePaulo, Kirkendol, Kashy, Wyer, & Epstein, 1996), this is a trivial percentage of the thou-sands of interactions that occur each day. With such a high base rate of honesty, it makes sense for people, in the ordinary conduct of their life, to presume that most people are telling the truth. When a significant proportion of individuals are in fact lying, as they are in most decep­tion studies (usually about half of the targets to be judged are decep­tive) unless observers switch their base rate assumptions, which is difficult for most people to do (Pronin, Lin, & Ross, 2002), they will be inaccurate in detecting lies, i.e., they will judge others as truthful when they are not.

Ekman (2001) described a related deception bias among some law enforcement personnel who frequently rated others as lying. Since the underlying error involved in the truthfulness bias is that of availability it makes sense that one group that does not show this bias is police of­ficers. Most police officers interview people who lie to them, either as perpetrators of a crime or as witnesses to it, so the base rate of lying that they are exposed to is much higher than that of most people. (Some businesses and some cultures might also have a higher base rate of lying.) It is adaptive given the availability of information to ad-just one's cognitive strategy to that base rate and so we see many police officers showing a deception bias rather than a truthfulness bias.

Anchoring, a.k.a. Representativeness. Zuckerman, Koestner, Colella, and Alton (1984) demonstrated that the baseline or anchor behavior that observers use in judging individuals will significantly affect their judgments of truthfulness. O'Sullivan, Ekman, and Friesen (1988) demonstrated that if the first sample of a person's behavior is honest, and the second sample is deceptive, accuracy in detecting the change

from the first sample to the second increased significantly. If, however, the first sample of behavior is deceptive, because of the truthfulness bias, observers seem to assume that the sample shows the person be­ing honest. Therefore, when they are later shown a sample of honest behavior they call it deceptive and their accuracy decreases. Their 'an­choring' or assuming that the sample they have been given is represen­tative of that person's ordinary (i.e., honest) behavior misleads them into believing that the change they have observed is from honest to de­ceptive, when, in fact, the change is in the opposite direction.

The Boy-Who-Cried-Wolf Effect, a.k.a. Fundamental Attribution. An-other cognitive heuristic that distorts the accurate utilization of non-verbal behavior in detecting deception is a variant of the fundamental attribution error (FAE; Ross & Nisbett, 1991). O'Sullivan (2003) has argued: 1) that most people make automatic and almost instanta­neous trait judgments about people s credibility (fundamental attribu­tion error); 2) that these judgments, like all fundamental attribution errors, will be difficult to change (Pronin et al., 2002) and, if corrected, will only rebound with greater strength (Yzerbyt, Corneille, Dumont, & Hahn, 2001); and 3) that the confusion between the trait of credibility or trustworthiness and the state judgment of honest or truthful is a particularly pernicious variant of the FAE. I called this version of the FAE the boy-who-cried-wolf effect after Aesop's fable (1793) about the shepherd boy who lied about a wolf stalking his sheep so often that the townspeople no longer believed him. Their attribution of him as a liar (enduring trait characterization) undermined their ability to recognize when he told the truth (temporary, state or situational reality). Once someone has decided that another person is generally credible or gen­erally untrustworthy (i.e., they have attributed enduring characteris­tics to them related to honesty), it will be extremely difficult for them to see the credible person as deceptive, or like Aesop's shepherd boy, the liar as telling the truth. My data suggest that this error is problematic even for accurate lie detectors. They are able to judge generally trust-worthy people as deceptive, but even they are unwilling or unable to be­lieve that someone they think is generally untrustworthy will tell the truth in a specific instance. The implications of this finding will be ad-dressed below in the section on accusatory reluctance.

Inaccurate Clue Paradigms a.k.a. Representativeness. Another reason for poor accuracy in detecting deception is the inaccurate paradigms most observers have about what lying behavior looks like. The mis­match between subjects' beliefs about deception clues and the actual behavior that occurs in lying is well documented (DePaulo et al., 1982). For example, most Americans believe that liars won't look you in the eye. Since most Americans know about this display rule (Ekman & Friesen, 1969), in laboratory studies of deception (Riggio & Fried-man, 1983), when people lie, they sometimes do more eye gaze than

when they are telling the truth. There is little research on the reasons that people give for thinking that others are lying or telling the truth (O'Sullivan, 2000; Vrij, this volume) and it is not clear whether these misunderstandings about the appearance of truthful and lying behav­ior are cognitive heuristics, in which case they will be difficult to change (Yzerbyt et al., 2003), or merely a lack of correct information (discussed below) which can be significantly affected by training and education (Devine, 1989).

Knowledge Deficiencies

Many people do not accurately assess truthful and deceptive behavior because they lack the requisite information. Reasons for this incompe­tent information be include: 1) lack of feedback about accuracy, 2) in-accurate or incomplete information about the appearance of lying and truthful behavior, 3) limited experience with the different kinds of lies that people can tell, and 4) inadequate social or emotional intelligence.

Ignorance Is Bliss-No Feedback About Accuracy. Ekman (2001) sug­gested that low accuracy in detecting deception was due principally to the fact that most of us do not get good feedback about our lie detection hit rate. We will say 'I can always tell when my child lies to me.' 'My friend is a terrible liar. I can always tell when she is trying to put one over on me.' But, of course, we only know the lies we have caught. If we have been lied to successfully, we are blissfully ignorant of this fact. As with lie detecting ability, lie-telling ability is probably a continuum, which if not normally distributed, is at least symmetrical, with fewer people at the ends of the distribution than in the middle. Very few peo­ple are extremely bad or extremely good liars. The really terrible, al­ways-detected liar is probably the model, the representation, which most of us have in our minds about what a liar looks like. This infor­mation serves not only as a short-cut heuristic that will bias our deci­sions (an inaccurate cue paradigm as discussed earlier), but it is only incorrect information, a knowledge deficiency.

Perceptual Inadequacies. Lack of feedback is one reason for knowl­edge or informational errors. Another is misperception or no percep­tion at all. Although there are many cues to deception, as indicated earlier in this chapter, most people seem to be unaware of them. Either they do not know that such cues may indicate cognitive or emotional changes that may be related to deception (Ekman & Friesen, 1974) or the nonverbal cues are so subtle or so brief that they do not see or hear them. Alternately, they may perceive them but not interpret them accu­rately, for reasons suggested below.

Different Lies May Involve Different Cues.  In one of the earliest stud­ies of honest and deceptive behavior, Hartshorne and May (1928)

found that people's tendency to lie did not generalize across different situations. Honesty depended on the kind of lie required, the situation in which the lying or cheating occurred. Detecting deception accurately may also be situation-bound, and most deception detection studies use a rather narrow range of lie types. DePaulo, Stone, and Lassiter (1985), for example, reported findings that illustrate this phenome­non. The attractiveness of the truth teller or liar affected the accuracy with which they were detected. Ekman and Frank (1997) have argued for a general lie detection ability across different kinds of lies. They re-ported a significant correlation (r = .37, p < .05) between accuracy scores on two different kinds of lie detection tasks and identified sev­eral groups of professional lie catchers who are highly superior in two kinds of lies (Ekman, O'Sullivan, & Frank, 1999). But both lie tasks in­volved people talking. There may be other kinds of lie detection abili­ties, e.g., spotting loan defaulters, shop-lifters or husband-stealers, which may depend on other kinds of lie detection abilities.

Although Frank and Ekman (1997) reported a positive correlation in lie detection accuracy across two different kinds of lies, the size of the correlation was moderate, suggesting that different kinds of lies may be involved in different kinds of verbal and nonverbal behaviors. Shrugs, for example, which suggest some negation of what is being said, may be more common when discussing opinions or beliefs than when discussing feelings. Emotional leakage, in the face and in the voice, may be more common in:

lies about feelings

serious lies, not polite, 'white' lies or false compliments

unauthorized lies, not those involved in bargaining, gaming or acting

lies told to people who are loved, admired, respected or feared

lies told to protect the self as opposed to lies told to protect others

lies told to be helpful or supportive

lies about transgressions

This list of lies is suggested by the emotional requirements of the relationship between the liar and the target of the lie. Anderson, DePaulo, and Ansfield (2002) presented data showing that the gender pairings (male to male, male to female, etc.) as well as the sample (col­lege students vs. community adults) will make a difference in the fre­quency with which people wi11 report telling each of several different kinds of lies.

Ekman (2001) suggested that how the lie is told is one way of classi­fying lies. He distinguished concealment, falsification and telling the truth falsely. Different strategies in lying should result in different be-

havioral cues. As noted above, people attend to different cues in judg­ing others. If the cue they are aware of and accurate in detecting is relevant to one kind of lie but not another, they will be differentially accurate in detecting them.

Differences In Social or Emotional Intelligence. The thesis of this chap-ter is that detecting deception is one manifestation of emotional or so­cial intelligence. Riggio, Tucker, and Throckmorton (1987) did the earliest work on this hypothesis by examining the relationship be­tween emotional/social intelligence as measured by Riggio's Social Skills Inventory (1989) and lie detection accuracy. Costanzo and Ar­cher (1993) by including lie detection items in their Social Interpreta­tion Task also agreed with this premise. It is likely that emotional intelligence is distributed as general intelligence is, i.e., most people are average and only a very few are very emotionally intelligent or very emotionally unintelligent. By extrapolation, this suggests that few peo­ple will be very good at detecting deception. This conclusion is sup-ported by many years of lie detection accuracy research (Ekman & O'Sullivan, 1991; Malone & DePaulo, 2001) as well as the study of ex-pert lie detectors described below.

Characteristics of the Liar or Truth Teller

Even if the lie catcher is socially and emotionally intelligent, has accu rate information about the dynamic clues to deception, has received feedback about his accuracy in detecting lies, and is not hampered by cognitive heuristics and strategic errors, there are still pitfalls to be avoided in searching for the truth. Some individuals, by virtue of their culture, appearance or personality will be misinterpreted.

Cultural Differences. Despite obvious differences, there has been surprisingly little work on cross-cultural differences in beliefs about ly ing, sanctions against lying and the appearance of lying in different cul tures (Bond, Omar, Mahmoud, & Bonser, 1990). We know that one class of nonverbal behaviors, emblems (hand and facial gestures that are substitutes for words) are markedly different in every culture sampled (Ekman & Friesen, 1969). Although certain facial expressions of emo tions will be recognized in all cultures (Ekman et al., 1987) there are also subtle variations among these expressions even after people have lived in the United States for many generations (Tsai, personal commu nication, May, 2003). Since nonverbal decoding is one of the bases for recognizing whether someone is lying or telling the truth, these subtle differences between generations and among cultural groups within a country need to be acknowledged and the behavior of the person being observed evaluated with that standard in mind. It is likely, based on the information presented earlier concerning cognitive heuristics and

knowledge deficiencies, which members of out-groups (Devine, 1989) are more likely to be judged as lying than members of in-groups. The re-search on these questions, however, remains to be done.

Idiosyncrasy Errors. The idiosyncrasy error has already been dis­cussed as a cognitive heuristic (representativeness). In that discussion it was described as an aspect of the lie catcher's world view that might contribute to his or her inaccuracy. It is included here because some individuals^' honesty may consistently be misjudged based on unique aspects of their appearance or behavior. People who are more socially skilled will be judged as more honest, whether they are or not (Riggio, Salinas, &'Ricker, 1988). Frank and Ekman (in press) have described a 'credibility generalization' related to dynamic aspects of the facial ex­pressions of truth tellers and liars. Zuckerman and his colleagues (1979) reported a similar finding except that they measured a more static overall assessment of 'demeanor.'

Other researchers have also reported the effect of static appearance cues on judgments about honesty. Zebrowitz, Voinescu, and Collins (1996) demonstrated that 'baby-faced' subjects, with high foreheads and widely-spaced eyes were more likely to be judged as honest when compared with non baby-faced peers. Bond, Omar, Pitre, and Lashley (1992) found that if targets were weird-looking, or 'fishy-eyed,' ob­servers described them as liars when other labels, such as mentally in-competent, were not available. Observers tend to misinterpret deviations from the norm as signs of deception, so people with ordi­nary non-verbal behavior and 'baby faces' who are not fishy-eyed or weird-looking are more likely to be judged as truthful.

Personality. Riggio and his colleagues (1988) reported that extro­verted individuals were more likely to be described as truthful, even when they were lying. Riggio and Friedman (1983) also reported indi­vidual differences in personality that corresponded with differences in nonverbal behaviors when people were lying and telling the truth. This suggests that understanding the basic personality of the person whose honesty is being evaluated may be related to accuracy in detecting de­ception.

Motivations of the Lie Catcher

The motivations of lie catchers may also affect their accuracy. Four dif­ferent motives can be distinguished: 1) cognitive laziness, 2) socializa­tion practices, 3) accusatory reluctance and 4) collusion and other self-deceptions.

Cognitive Laziness. The literature on social cognition is replete with examples of the importance of motivation in the processes involved in one person's understanding of another. Fiske (1992) has written about

cognitive pragmatism or 'satisficing,' in which observers settle for 'good enough' in making decisions about others. This laissez-faire at­titude may be even more prevalent when the judgment concerns hon­esty or deception. A number of researchers have reported that most people think many kinds of lies are trivial. Feldman, Forrest, and Happ (2002) found that even in a brief, casual conversation, college students reported many instances of exaggeration, misstatements, and omis­sions. In a study of lies in romantic relationships, O'Sullivan's (1999) subjects reported that they were extremely likely to tell a variety of lies to their romantic partner and that they thought most of these lies were not very serious. If lying is not considered serious, spending cognitive capital to determine whether lying has occurred may not be regarded as a wise investment for most people who seem to husband their cogni­tive resources (Baumeister, 1993).

Socialization Practices. In order to live together, human beings must subjugate personal desires for the common good (Freud, 1938). Chil­dren are taught early to dissemble (Saarni & Weber, 1999) to feign in­terest in class, delight at an unwanted present, or forgiveness of a mischievous sibling. The lessons are two-fold. Children are overtly in­structed in how to manage their behavior, but, covertly, they are also being instructed to accept the deceptions that others offer to them, in the guise of unfelt thanks, unmeant apologies, insincere compliments. Lewis, Stanger, and Sullivan (1989) demonstrated quite sophisticated and successful lie behavior in children as young as three years of age. It is likely that skills learned so early are not highly amenable to con­scious control or awareness in adulthood. Although most societies discourage deception, they encourage politeness and other misleading impression-management strategies. Learning to cooperate in this so­cial choreography, by pretending to believe white lies or overlooking the social mistakes of others, may undermine the skills needed to de­tect deception.

Accusatory Reluctance. Social life seems to provide not only positive reinforcement to those who engage in the semi-lies of courtesy, but also negative reinforcements both to those who lie (or are caught lying) and those who detect lies and let others know that they have done so. Ear­lier, I described a study (Ekman et al., 1980) in which subjects who were inaccurate in labeling truthful and deceptive behavior with the la­bels 'lying' and 'truthful' nonetheless used different sources of verbal and nonverbal information in describing their impressions of other people. The observers seemed reluctant to accuse people of lying, even though, at some level, they were aware of a mismatch in their behavior, since they ignored the nonverbal behavior that occurred during decep­tion, concentrating instead on the content of speech (O'Sullivan et al., 1985). More recently, DePaulo (1998) described a related phenome­non. Observers, who are at chance in labeling people as truthful or ly-

ing, achieve significantly better accuracy if they are asked instead to rate how comfortable the people looked.

DePaulo and Rosenthal (1979) questioned whether being too accu rate in lie detection might be a liability in modern day social life. There is little cultural consensus about how to behave when lying is ob­served. And the costs incurred, in embarrassment, if incorrect in one's accusation, anger from the one accused, and damage from the fraying of the social fabric is quite high. One example of the high price of de tecting lies is the ubiquitous fate of whistleblowers, those individuals who reveal the dishonest practices of their business colleagues or gov­ernment officials (Johnson, 2002). Most whistleblowers lose their jobs, thereby incurring substantial economic and social losses. For those who are highly sensitive to lies and deception, knowing too much about other people may make it difficult to have ordinary social interactions with them.

Earlier, the-boy-who-cried-wolf-effect, a variant of the fundamental attribution error was described. Although this heuristic was evident in the judgments about both the liars and the truth tellers, it seemed to be particularly intransigent with respect to judgments in which people were described as generally untrustworthy. Although accurate lie catchers would frequently rate generally trustworthy individuals as having lied in a particular situation, if they had labeled someone as generally untrustworthy, they rarely described that person as truthful, even when he was. The intransigence of the label 'liar' (once affixed, never removed) seems to make even astute observers of others loathe to label them as dishonest.

A related aspect of accusatory reluctance is the truthfulness bias de scribed earlier. The difference between them is this: The truthfulness bias is based on a particular world view, and functions like a represen tativeness heuristic. Accusatory reluctance is a response variable. At some level, subjects observing lying behavior are aware of discomfort and the disjunction between verbal and nonverbal behavior. They are unwilling, however, to label such observed discomfort or such discrepancies as lying.

Collusion and Other Self-Deceptions.though no empirical evidence could be found, common sense and everyday observation suggests that people often cooperate in being deluded. Freud (1938) argued that some truths are too painful to know, so we forget them, transform them, project them, sublimate them or distort them into a form we can bear. Examples of this lie detection difficulty range from Chamber­lain's believing Hitler's incredible protestations (Ekman, 1988) to spouses' overlooking blatant infidelity (Baumeister, 1993). Although self deception and collusion are distortions of reality they help people to maintain their sense of themselves and to organize what might oth­erwise be ambiguous or threatening information.

EVOLUTIONARILY BASED BIASES

The social cost for identifying liars, whether correctly or incorrectly has been of concern to evolutionary psychologists for some time. Bond and his colleagues (Bond, Kahler, & Paolicelli, 1985; Bond & Robin-son, 1988) and other evolutionists have argued that humans have de veloped acute lie detection abilities. This assertion is inconsistent with the resounding evidence that most people are only at chance in recog­nizing the truthful and lying behavior of others. On the other hand, Ekman (1996) argued that the cost for detecting lies in our evolution ary history was probably so severe that there was no widespread selec­tion for this ability. Socialization practices and reinforcement schedules related to courtesy training and accusatory reluctance are more consistent with Ekman's speculations than with Bond's.

Clore suggested (personal communication, August, 2002) that what has developed is the ability to detect chronic cheaters who can then be sanctioned. Ekman and Clore's hypotheses are in concert with the many reasons outlined here for why most people do not, can not, or will not label liars. at they offer are evolutionarily based reasons, rather than socially or cognitively based ones. Of course, both ap­proaches could be correct. The social cognitive biases discussed here could have developed as ways of implementing the biologically-based motivation to avoid the high costs of detecting deception.

THE LIE DETECTION WIZARDS

Given the many reasons why most people are inaccurate in detecting deception one might reasonably ask: Can anyone accurately detect de­ception from verbal and nonverbal clues? Ekman, Frank and I (Ekman & O'Sullivan, 1991; Ekman et al., 1999) have described sev eral different groups, who, as groups, were significantly above chance in their lie detection accuracy. We noticed that within each group there were some individuals who were extraordinarily accurate, scoring 70% or higher on various tests of deception accuracy. As would be ex pected from the binomial distribution, across all the groups we stud­ied, not just the highly accurate ones, about 10% of people score significantly above chance. All the groups we studied were rather small in size. When we (see O'Sullivan & Ekman, in press, for more details) tested the lie deception accuracy of 1200 therapists, we found that the same level of accuracy characterized the scores of this large group. Rel atively few of the therapists obtained very high scores, but, given the large size of the sample, there were enough of them to encourage us to do an idiographic analysis of highly accurate lie catchers.

As we continued to lecture to law enforcement and other profes­sional groups we asked the participants in our workshops to raise their hands if they obtained scores of 90% or higher on a videotaped lie

detection measure that showed ten men lying or telling the truth about a strongly held opinion. (For more information about this measure, see Frank & Ekman, 1997.) We then asked these very high scorers if they were willing to participate in a research project on expert lie detec­tion. Those participants who gave us their contact information were then sent two additional lie detection accuracy tests, one of ten men ly­ing or telling the truth about whether they had stolen $50 (the crime video; Frank & Ekman, 1997) and the other of ten women lying or tell­ing the truth about their feelings as they watched either a pleasant na­ture film or a gruesome surgical film (Ekman et al., 1980). To qualify as 'ultimate' experts, participants had to obtain scores of 80% or better on both tests. (See O'Sullivan & Ekman, in press, for more de-tails.) As of August 2003, 14 such experts have been identified. Over the last 15 years, we estimate that we have tested over 12,000 people, most of them adults working in professions for which lie detection is relevant.

An additional 15 experts, having received a score of 90% on the ini­tial screening test (opinion video), received a score of 80% or greater on only one of the two other tests. We noticed that the nine therapists in this group were highly accurate on the test showing lies about feeling, but not on the test in which the lies involved a crime. The opposite pat-tern was found with six law enforcement personnel. They were highly accurate (scoring 80% or better) when the lie or truth concerned a crime, but not when it concerned feelings. A chi square analysis of these data was highly significant (x2 = 11.429, p < .00072). Because this error pattern was linked to professional experience, we included these 'penultimate' experts in our sample, as well. As of August 2003, we have identified 15 such penultimate experts, giving us a total 'wiz­ard of deception detection' group of 29.

THE PROTOCOL

After the experts are identified, they meet with one or both of the exper­imenters (O Sullivan & Ekman) to review their responses to each of the three videos. They are instructed to say aloud anything that occurs to them as they watch the video. They are encouraged, in other words, to 'think aloud'(Ericsson & Simon, 1998). This procedure was usually the first activity we did with the experts because we were trying not to influence their recall or their reporting of the process they used in de­tecting deception. This initial review procedure lasts from one to two hours and is tape-recorded and transcribed.

A semi-standardized interview about personal and life history infor­mation was then recorded. Early on, we found that this interview was adequate for gaining factual information and, occasionally, informa­tion about career choices, mentors, and the like, but, especially with the police personnel, it did not yield much information. Consequently,

when possible, we try to spend time with the experts and their friends and family members as well.

In addition to the think-aloud procedures and the gathering of data about their personal and professional lives, a series of psychological tests will be administered to the experts. These wi11 include the NEO personality inventory, a short measure of verbal comprehension as an indirect measure of IQ, and, perhaps a measure of attributional or cognitive complexity.

THE CONTROL GROUP

One of the dilemmas of this project has been the problem of defining an adequate control group. The members of the control group would have to be non-expert lie detectors, by our criterion, but also be inter­ested to participate in the research project. They would have to be sim­ilar to the experts in terms of social class, educational level, geographic location and age. There was no pre-existing group that met these re­quirements, so we decided to use each expert's spouse or close family member as his or her control. In most cases, the family member is in­terested in the project, and in their spouses' involvement. In some cases, before we had decided to use spouses as controls, husband or wives of the potential expert took the test, too, out of personal interest. In one case, the wife was the target and did not qualify. Her husband took the test, and did qualify!

At this point we have just started to interview the lie detection ex­perts, so the following observations are merely that-observations based on interview data. The comments will be organized in line with the reasons for detection inaccuracy outlined above.

STRATEGIES Nonverbal Cues

The deception detection wizards seem more aware of nonverbal be­havior and attend to it more closely, base judgments on it more fre­quently and have more sophisticated and unusual descriptions of nonverbal behavior than other people we have interviewed. Every one of them has spontaneously described nonverbal behavior and discrep­ancies between verbal and nonverbal behavior. This does not mean that they use nonverbal behavior in every instance. They do not. But nonverbal cues are an important part of their deception detection ar­mamentarium.

The wizards also observe and use types of nonverbal behaviors that have not been studied in the research lab. One law enforcement officer, for example, who had not yet been identified as an expert, was attend­ing a conference at which several already-identified experts were teaching. She called the groups' attention to a subtle, micro momen-

tary lip stretch, occurring at an important point in the interview, that none of the other experts had seen, until she called it to their attention.

Another wizard, comparing two truthful men, noted that they made similar circular head movements, as though saying, with their head, 'Well, that about wraps it up.' It was a loose, relaxed head movement, in synchrony with the pace and content of their story. This same expert also differentiated the quality of the eye gaze of the truth telling and the deceptive men. She did not depend merely on an eye gaze vs. no-eye gaze distinction, but rather evaluated the quality of the eye gaze, whether the individual was intently watching the interviewer to see whether he was being believed, or merely looking, in the ordinary way of conversational partners.

Verbal Cues

Many of our wizards are lawyers, or professional interrogators, so they,. pay a great deal of attention to the nuanced use of language. They ob­serve slips of the tongue, rather than making sense of them, or excus­ing them as the average observer tends to do. They also use language to assess the education and intelligence of the people they are observing, thereby forming an assessment of the person as an individual. This in­dividual assessment then seems to be used as a kind of baseline for as­sessing the nonverbal behavior that occurs.

No Pinocchio Noses

Other than the over-all category of 'nonverbal behavior' and discrep­ancies within that behavior, there is no single clue that every wizard uses. On the other hand, each of the wizards seems to have a finite number of behaviors to which they attend with great precision and in-tensity. For one, it might be voice quality; for another (noted above) the combination of head and eye movements. If these preferred behaviors are not shown by the person they are observing, however, they switch to other clues or depend more on their overall assessment of the indi­vidual's personality. It is also interesting, that like the rest of us, there are many reliable clues to deception and truthfulness in the videos we reviewed together of which the wizards were blissfully unaware. They seem to have developed a set of tools that works very well for them, but there are other tools that other people use with equally good results. So, even the most accurate lie catcher has something still to learn. Inci­dentally, although there were a few 100% scores among the 29 wizards, no one got 100% on all three tests.

Cognitive Biases

Obviously, by virtue of their having received very high scores, the ex-pert lie detectors do not have either a truthfulness bias or a deception

bias, since that would prevent their obtaining high scores. In fact, an analysis of the errors made by the 29 wizards shows an equal number of incorrect lying and truthful answers. The wizards also did not show evidence of either the anchoring or the boy-who-cried-wolf effect. Many times, even during a one minute interview, they would change their as­sessment or evaluation of the person they were watching. One expert watched the entire interview, took in all the available information, and only then made her judgment even though she had seen the video be-fore. Also, the wizard's base rate of information (representativeness) seems to be more accurate than that of most observers. As will be noted below, many of them have a wide range of life experiences which seems to have enriched their knowledge about people in general.

Knowledge and Motivation

In his realistic accuracy model of judging personality, Fonder (1999) suggests that there are three aspects to accurate appraisal of others: ability motivation and knowledge. The ability to understand whether someone is truthful or deceptive seems also to reflect these elements. We have already given some examples of the kinds of acute sensitivity to both verbal and nonverbal behaviors that characterize the deception detection wizards. In addition, these highly emotionally intelligent wiz­ards are characterized by both exceptional motivation and unusual knowledge.

In our early studies of groups of experts (Ekman & O'Sullivan, 1991; Ekman et al., 1999), we speculated that superior lie catchers were distinguished by their motivation to do well at the task. We sug­gested that the very fact of their taking a workshop on the topic of de­tecting deception when others in their professions did not was evidence of this motivation. This observation is even more apt for the lie detection wizards. They not only attended workshops, or contacted Ekman in response to media coverage of his work, but they agreed to participate in a research project and most of them have continued through every step of the research process. The wizards, like the ex­perts identified earlier, seek out information to validate their impres­sions and improve their performance. One wizard called me at 10 pm. He had been watching the TV show, American Justice, and was excited because he was sure the person featured on the show was telling the truth, and he wanted to let someone know his opinion before the 'truth' was revealed on the program. This man retired from his law en­forcement many years ago. Nonetheless, he was highly motivated to test himself, to be measured against an objective reality, to be proven wrong. And if he had been wrong, he wanted to discuss it with someone and learn from it.

As a group, the wizards are concerned about excellent performance, whether it is lie detection, playing hockey, singing in the church choir

or riding their Harley. In completing the think-aloud protocol, a few wizards changed some of their responses, giving an incorrect answer. They were quite distressed about this, mentioning it several times, and still referring to it days later, attempting to understand why they had changed their response, why they made the wrong assessment.

Ericsson (1996), in his review of expertise in professions as widely varying as grand master chess champions and concert violists, argued that the characteristic that distinguishes the highly expert from the merely good was an extended period of intensive practice, usually last­ing ten years, at the start of their careers. Recently, Brassington (2003) reported similar findings with prima ballerinas and members of the corps de ballet. Prima ballerinas reported more practice, more visualiz­ing about their performance, more planning. Given that only a third of the wizards have been interviewed, at this point it seems that an intense focus and investment in their performance, and concentrated attempts to improve it, are characteristic of most, if not all, of the wizards. They are a highly motivated group.

An aspect of their performance that also fits under the rubric of knowledge is what I have termed the Miss Marple effect. Miss Marple is a character invented by Agatha Christie (1985) who solves crimes through her analysis of the personalities of suspects. She compares suspects to people she has known. She has a broad and accurate 'rolodex' of relevant 'types' against which to compare the individuals she is trying to assess. Many of the wizards have a similar capacity. One described a truthful man as looking 'like a choir boy who has not been assaulted.' Another wizard described a different young man this way. 'He has trouble with authority, but he is an honest man. He has been well-taken care of.' Although some of the wizards will mention stereo-types of various sorts, they do not let such views interfere with their at-tempts to understand a particular young woman or a particular black man when they are in their professional mode.

In terms of their own personality, most of the wizards seem to be in­troverts. This is inconsistent with research showing that extroverts are more socially skilled than introverts (Riggio et al., 1987). Lieberman and Rosenthal (2001), however, suggest that introverts do well on so­cial tasks if that is their focus. And the wizards are exceptionally fo­cused individuals. Another aspect of the 'introversion' description was amplified by a wizard who said 'I am quiet, but I am not shy.' As we explore aspects of personality related to emotional intelligence, distinctions like this should be made.

Another surprising finding from this research is the deep, wide and unusual experiences with other people that the wizards have had, usually in connection with their work, but sometimes as a con-sequence of a family situation. The basis for friendship and mar­riage is almost always similarity, so most of us spend our lives with people much like ourselves. The wizards, by choice or necessity, seem to have a much wider base of social experience, based on inter-

action with many different kinds of people. One sheriff, for example, spent his career, first with psychiatric patients in the county jail, then with prostitutes and drug addicts of both genders in one of the most dangerous areas of Los Angeles. But he spends his free time helping his daughter find Daisy Duck collectibles at garage sales. An arbitrator, over the course of his career, has dealt, in significant ways, with mentally retarded blue collar workers, wild-cat coal mine operators in the Appalachians, and CEO's of international companies in the Mid-West. Some of the wizards' breadth of experi­ence comes from challenging childhoods either because of abusive parents or an unusual social situation-being part of an isolated im­migrant group or having a working mother at a time when that was uncommon.

Because of their high level of motivation the wizards seek informa tion to correct inaccurate clue paradigms through feedback, reading and participation in relevant workshops and experiences, such as par ticipating in this research. They have knowledge of many kinds of peo ple in many kinds of situations, so are better prepared to identify lies of many sorts.

On the basis of their ability to interpret nonverbal behavior accu­rately, the wizards are highly emotionally intelligent. Many of them, if not all, are very talented role players as well. When a situation calls for it, these reserved, sensitive people can be as outrageous as nec­essary. In terms of managing their own emotions, an aspect of emo­tional intelligence that some theories suggest, the picture is more mixed. All of the wizards seem highly professional in their work lives, but several of them have had difficult love relationships. Some of the wizards seem able to turn off their acute sensitivity in non work situations; others do not, or can not and seem to be scanning the world all the time. Elfenbein and Ambady (2002) have demon­strated the differential effect on supervisors' ratings of sensitivity to different kinds of nonverbal cues. They argued that noticing subtle evidence of negative emotion that co-workers or supervisors were attempting to hide impaired workers' perceived job effectiveness. Based on this perspective, in order to achieve professional success, the lie detection wizards would need to manage their own behav­ior-to not indicate knowledge of others' emotional states when such knowledge was unwelcome.

In terms of accusatory reluctance, an analysis of the few errors that the experts made in the three lie detection tests showed no truthful ness bias. There were as many errors where they incorrectly labeled the person as deceptive as honest. In terms of collusion with the liar and other forms of self-deception, the wizards do not usually do this when they are focused on the task of detecting lies. In their personal lives, however, they can collude with their loved ones as well as the rest of us, and deceive themselves about relationships, goals, and life outcomes.

CONCLUSION

In sum, the expert lie detectors are extraordinarily emotionally intelli­gent people. They observe the emotions of others accurately. They are aware of their own emotional reactions to others and can use this in-formation in understanding others, especially with respect to detect­ing deception. For Olympic athletes, talent is a necessary, but not a sufficient condition for athletic excellence. The development of even the greatest gifts takes practice, feedback and motivation. It is likely that the same is true for the development of the ability to understand others. The lie detection wizards do not use 'tricks' that can be taught in a seminar. They do not 'do' lie detection; they listen and watch peo­ple in order to understand them and then, having understood them, they are able to accurately determine their truthfulness. And they seek to improve their already considerable ability to do this. Through the kinds of careers that most of them have chosen, by their willingness to participate in a research project investigating their rare and precious gifts, they are choosing to use their talents for the welfare of all. God bless them, every one.

REFERENCES

Aesop. (1793). The fables of Aesop: With a life of the author and embellished

with one hundred and twelve plates. London: John Stockdale. Anderson, E. D., DePaulo, B. M., & Ansfield, M. E. (2002). The development of

deception detection skill: A longitudinal study of same-sex friends. The So‑

ciety of Personality and Social Psychology, 28(4), 536-545.

Archer, D., & Akert, R. M. (1977). Words and everything else: Verbal and non verbal clues in social interpretation. Journal of Personality and Social Psy­chology, 35, 443-449.

Bar-On, R., Brown, J. M., Kirkcaldy, B. D., & Thome, E. P. (2000) Emotional expression and implications for occupational stress; An application of the Emotional Quotient Inventory (EQ-I). Personality and Individual D(er­ences, 28(6), 1107-1118.

Baumeister, R. F. (1993). Lying to yourself: The enigma of self-deception. In M. Lewis & C. Saarni (Eds.), Deception in everyday life (pp. 166-183). New York: The Guilford Press.

Bond, C. F., Kahler, K. N., & Paolicelli, L. M. (1985). The miscommunication of deception: An adaptive perspective. Journal of Experimental Social Psy­chology, 21(4), 331-345.

Bond, C. F., Omar, A., Mahmoud, A., & Bonser, R. N. (1990). Lie detection

across cultures. Journal of Nonverbal Behavior, 14(3), 189-204.

Bond, C. F, Omar, A., Pitre, U., & Lashley, B. R. (1992). Fishy-looking liars: De‑

ception judgment from expectancy violation. Journal of Personality and

Social Psychology, 63(6), 969-977.

Bond, C. F., & Robinson, M. (1988). The evolution of deception. Journal of Nonverbal Behavior, 12(4, R. 2), 295-307.

Boring, E. G., & Titchener, E. B. (1923). A model for the demonstration of fa­cial expressions. American Journal of Psychology, 34, 471-485.

Brassington, G. (2003, May). Mental skills distinguish elite soloist ballet dancers form corps dancers. Paper presented at the annual meeting of the Western Psychological Association Convention, Vancouver, B.C.

Buck, R. (1976). A test of nonverbal receiving ability: Preliminary studies. Hu-man Communication Research, 2(2), 162-171.

Bugental, D., Kaswan, J., & Love, L. (1970). Perceptions of contradictory meanings conveyed by verbal and nonverbal channels. Journal of Personal­ity and Social Psychology, 16, 647-655.

Burgoon, J. K., Buller, D. B., White, C. H., Mill, W, & Buslig, A. L. S. (1999). The role of conversational involvement in deceptive interpersonal interac­tions. Personality and Social Psychology Bulletin, 25(6), 669-685.

Chapin, F S. (1942). Preliminary standardization of a social insight scale. American Sociological Review, 7, 214-225.

Christie, A. (1985). Miss Marple: The complete short stories. New York: Dodd, Mead.

Costanzo, M., & Archer, D. (1993). The Interpersonal Perception Task-I5 (IPT-15). Berkeley, CA: University of California Center for Media and Inde­pendent Learning.

Cronbach, L. J. (1955). Processes affecting scores on 'understanding of oth­ers' and 'assumed similarity.' Psychological Bulletin, 52, 177-193.

Darwin, C. (1998). The expression of the emotions in man and animals (3rd ed.). With an introduction, afterward and commentary by P. Ekman. New York: Oxford University Press.

DePaulo, B. M. (1992). Nonverbal behavior and self-presentation. Psychologi­cal Bulletin, 111(2), 203-243.

DePaulo, B. M. (1998, May). eceiving and detecting deceit: Insights and over-sights from the first several hundred studies. Invited address. American Psychological Society, Washington, D.C.

DePaulo, B. M., Kirkendol, S. E., Kashy, D. A., Wyer, M. M., & Epstein, J. A. (1996). Lying in everyday life. Journal of Personality and Social Psychol­ogy, 70(5), 979-995.

DePaulo, B. M., Lindsay, J. J., Malone, B. E., Muhlenbruck, L., Charlton, K., &

Cooper, H. (2003). Cues to deception. Psychological Bulletin, 129(1), 74-118. DePaulo, B. M., & Rosenthal, R. (1979). Telling lies. Journal of Personality

and Social Psychology, 37(10), 1713-1722.

DePaulo, B. M., Rosenthal, R., Rosenkrantz, J., & Green, C. R. (1982). Actual and perceived cues to deception: A closer look at speech. Basic and Applied Social Psychology, 3(4), 291-312.

DePaulo, B. M., Stone, J. I., & Lassiter, D. G. (1985). Telling ingratiating lies: ef­fects of target sex and target attractiveness on verbal and nonverbal deceptive success. Journal of Personality and Social Psychology, 48(5), 1191-1203.

Devine, P. G. (1989). Stereotypes and prejudice: their automatic and controlled components. Journal of Personality and Social Psychology, 56, 5-18.

Dimitrius, J., & Mazzarella, M. C. (1999). Reading people. New York: Ballantine Books.

Doyle, C. A. (1892). The adventures of Sherlock Holmes. London: G. Newnes. Efron, D. (1941). Gesture and environment. Oxford, England: King's Crown Press.

Ekman, P (1988). Self-deception and detection of misinformation. In J. S. Lockard & D. L. Paulhus (Eds.), Seif-deception: An adaptive mechanism? (pp. 229-257). Englewood, NJ: Prentice-Hall.

Ekman, P (1996). Why don't we catch liars? Social Research, 63(3), 801-817. Ekman, P (2001). Telling lies: Clues to deceit in the marketplace, politics, and marriage (3rd ed.). New York: W. W. Norton.

Ekman, P (2003). The Micro Momentary Expression Recognition Tool. Avail-able from: http://www.emotionsrevealed.com

Ekman, P (2003). The Subtle Expression Recognition Tool. Available from: http:/.emotonsreveed.com

Ekman, P. & Friesen, W. V. (1969). Nonverbal leakage and clues to deception. Psy‑

chiatry Journal for the Study of Interpersonal Processes, 32(1), 88-106.
Ekman, P, & Friesen, W. V. (1969). The repertoire of nonverbal behavior: Cate‑

gories, origins, usage, and coding. Semiotica, 1, 49-98.

Ekman, P, & Friesen, W. V. (1974). Detecting deception from body or face. Journal of Personality and Social Psychology, 29, 288-298.

Ekman, P & Friesen, W. V. (1975). Unmasking the face. Englewood: Prentice Hall.

Ekman, P. & Friesen, W. V. (1978). Facial action coding system. Palo Alto, CA: Consulting Psychologists Press.

Ekman, P, Friesen, W. V. & O'Sullivan, M. (1988). Smiles when lying. Journal of Personality and Social Psychology, 54, 414-420.

Ekman, P, Friesen, W. V., O'Sullivan, M., & Scherer, K. R. (1980). Relative im­portance of face, body, and speech in judgments of personality and affect. Journal of Personality and Social Psychology, 38(2), 270-277.

Ekman, P, Friesen, W. V. O'Sullivan, M., Chan, A., DiacoyanniTarlatzis, I., Heider, K., Krause, R., LeCompte, W. A., Pitcairn, T., Ricci-Bitti, P E., Scherer, K. R., Tomita, M., & Tzavaras, A. (1987). Universals and cultural differences in the judgments of facial expressions of emotion. Journal of Personality and Social Psychology, 53, 712-717.

Ekman, P, & O'Sullivan, M. (1991). Who can catch a liar? American Psycholo­gist, 46(9), 913-920‑

Ekman, P, O'Sullivan, M., & Frank, M. G. (1999). A few can catch a liar. Psy­chological Science, 10(3), 263-266.

Elfenbein, H. A., & Ambady, N. (2002). Predicting workplace outcomes from the ability to eavesdrop on feelings. Journal of Applied Psychology, 87(5), 963-971.

Engen, T., Levy, N., & Schlosberg, H. (1957). A new series of facial expressions. American Psychologist, 12, 264-266.

Ericcson, K. A. (1996). The acquisition of expert performance: An introduc­tion to some of the issues. In K. A. Ericsson (Ed.), The road to excellence: The acquisition of expert performance in the arts and sciences, sports, and game (pp. 1-50). Hillsdale, NJ: Lawrence Erlbaum Associates.

Ericcson, K. A., & Simon, H. A. (1998). How to study thinking in everyday life: Contrasting think-aloud protocols with descriptions and explanations of thinking. Mind, Culture and Activity, 5(3), 178-186.

Etcoff, N. L., Ekman, P. Magee, J. J., & Frank, M. G. (2000). Lie detection and language comprehension. Nature, 405(6783), 139.

Feldman, R. S., Forrest, J. A. & Happ, B. R. (2002). Self-presentation and ver­bal deception: Do self-presenters lie more? Basic &Applied Social Psychol­ogy, 24(2), 163-170.

Fiske, S. T. (1992). Thinking is for doing: Portraits of social cognition from Da­guerreotype to laser photo. Journal of Personality and Social Psychology, 63(6), 877-889.

Freud, S. (1938). The basic writings of Sigmund Freud (trans. & ed. by A. A. Brill). New York: The Modern Library.

Frank, M. G., & Ekman, P (1997). The ability to detect deceit generalizes across different types of high-stake lies. Journal of Personality and Social Psychology, 72(6), 1429-1439.

Frank, M. G., & Ekman, P. (in press). Appearing truthful generalizes across dif‑

ferent deception situations. Journal of Personality and Social Psychology. Frois-Wittmann, J. F. (1930). The judgment of facial expression. Journal of Ex‑

perimental Psychology, 13, 113-151.

Fonder, D. (1999). Personality judgment: A realistic approach to person per­ception. San Diego, CA: Academic Press.

Gardner, H. (1993). Multiple intelligences: The theory in practice. New York: Basic Books.

Guilford, J. P. (1929). An experiment in learning to read facial expressions. Journal of Abnormal Social Psychology, 24, 191-202.

Guilford, J. P. (1956). The structure of intellect. Psychological Bulletin, 53, 267-293.

Goleman, D. (1995). Emotional intelligence. New York: Bantam Books. Hartshorne, H., & May, M. A. (1928). Studies in deceit. Book l. General meth‑

ods and results. Book H. Statistical methods and results. New York:

MacMillan.

Ickes, W. (1993). Empathic accuracy. Journal of Personality, 61, 587-610. Ickes, W., & Simpson, J. A. (1997). Managing empathic accuracy in close rela‑

tionships. In W. Ickes (Ed.), Empathic Accuracy (pp. 218-250). New York:

Guilford Press.

James, W. (1884). at is an Emotion? Mind, 9(34), 188-205.

Johnson, R. A. (2002). Whistleblowing: When it works and when it doesn't. Boulder, CO: Lynne Rienner Publisher.

Lewis, M., Stanger, C., & Sullivan, M. W. (1989). Deception in 3-year-olds. De­velopmental Psychology, 25(3), 439-443.

Lieberman, D. (1998). Never be lied to again. New York: St. Martin's Griffin. Lieberman, M. D., & Rosenthal, R. (2001). Why introverts can't always tell who

likes them: Multi-tasking and nonverbal decoding. Journal of Personality

and Social Psychology, 80(2), 294-310.

Lipps, T. (1926). Psychological Studies. (Trans. by H. C. Sanborn). Oxford, UK: Williams & Wilkins.

Malone, B. E., & DePaulo, B. M. (2001). Measuring sensitivity to deception. In J. A. Hall & F. J. Bernieri (Eds.), Interpersonal sensitivity: theory and mea­surement (pp. 103-124). Mahwah, NJ: Lawrence Erlbaum Associates.

Mayer, J. D., Salovey, P, & Caruso, D. (2002). Mayer-Salovey-Caruso emo­tional intelligence test user's manual. Toronto, Canada: Multi-Health Systems

Mehrabian, A., & Epstein, N. (1972). A measure of emotional empathy. Jour­nal of Personality, 40, 525-543.

Moss, F. A. (1931). Preliminary report of a study of social intelligence and exec­utive ability. Pub. Personnel Studies, 9, 2-9.

Moss, E. S., Hunt, T., Omwake, K. T., & Woodward, L. G. (1955). Social Intelli gence Test. Washington, DC: George Washington University.

Nance, J. (2000). Conquering deception. Kansas City, KS: Irvin-Benham. Nowicki, S. J. (2001). Nonverbal receptivity: The Diagnostic Analysis of Non-verbal Accuracy (DANVA). In J. A. H1& F. J. Bernieri (Eds.), Interpersonal

sensitivity Theory and measurement (pp. 183-198). Mahwah, NJ: Law­rence Erlbaum Associates.

O'Sullivan, M. (1982). Measuring the ability to recognize facial expressions of emotion. In P. Ekman (Ed.), Emotion in the human face (2nd ed., pp. 281-317). Cambridge, UK: Cambridge University Press.

O'Sullivan, M. (1983). The Affect Blend Test: A new facial recognition test. Symposium on Social Intelligence; American Psychological Association Convention, Anaheim, CA.

O'Sullivan, M. (July 2000). 4Vhy 1 think you're lying. per presented at the In­ternational Congress of Psychology, Stockholm, Sweden.

O'Sullivan,. M. (2003). The fundamental attribution error in detecting decep­tion: The boy-who-cried-wolf effect. Personality and Social Psychology Bulletin, 29(10), 1316-1327.

O'Sullivan, M., & Ekman, P (in press). The wizards of deception detection. In P. A. Granhag & L. Stromwell (Eds.), Detecting deception inforensic con-text. Cambridge, UK: Cambridge University Press.

O'Sullivan, M., Ekman, P., Friesen, W, & Scherer, K. R. (1985). at you say and how you say it: The contribution of speech content and voice quality to judgments of others. Journal of Personality and Social Psychology, 48(1), 54-62.

O'Sullivan, M., Ekman, P., & Friesen, W. V. (1988). The effect of comparisons on detecting deceit. Journal of Nonverbal Behavior, 12(3, Pt. 1), 203-215.

O'Sullivan, M., & Guilford, J. P. (1975). Six factors of behavioral cognition: Un­derstanding other people. Journal of Educational Measurement, 12(4), 255-271.

Pronin, E., Lin, D. Y., & Ross, L. (2002). The bias blind spot: Perceptions of bias in self versus others. Personality and Social Psychology Bulletin, 28(3), 369-381.

Riggio, R. E. (1989). Social skills inventory. Palo Alto, CA: Consulting Psychol­ogists Press.

Riggio, R, E., & Friedman, H. S. (1983). Individual differences and cues to de‑

ception. Journal of Personality and Social Psychology, 45(4), 899-915. Riggio, R. E., Salinas, C., & Thcker, J. (1988). Personality and deception abil‑

ity Personality and Individual Dierences, 9(1), 189-191.

Riggio, R. E., Thcker, J., & Throckmorton, B. (1987). Social skills and decep­tion ability. Personality and Social Psychology Bulletin, 13(4), 568-577.

Rosenthal, R., Hall, J. A., DiMatteo, M. R., Rogers, P, &Archer, D. (1979). Sensi tivity to nonverbal communication: Aprofile approach to the measurement of individual dierences. Baltimore: Johns Hopkins University Press.

Ross, L., & Nisbett, R. E. (1991). The person and the situation: Perspectives of social psychology. New York: McGraw-Hill.

Saarni, C., & Weber, H. (1999). Emotional displays and dissemblance in child-hood: Implications for self-presentation. In P. Philippot (Ed.), The social context of nonverbal behavior (pp. 71-105). New York: Cambridge Univer­sity Press.

Salovey, P, & Mayer, J. D. (1989-1990). Emotional intelligence. Imagination, Cognition, & Personality, 9(3), 185-211.

Sternberg, R. J. (1986). Intelligence applied. New York: Harcourt Brace Jovanovich.

Sternberg, R. J. (1988). The triarchic mind: A new theory of human intelli­gence. New York: Penguin Books.

Thorndike, E. L. (1920). Intelligence and its uses. Harper's Magazine, 140, 227-235.

Thorndike, R. L., & Stein, S. (1937). An evaluation of the attempts to measure social intelligence. Psychological Bulletin, 34, 275-285.

Tsai, J. L., & Chentsova-Dutton, Y. (in press). Variation among European Ameri‑

cans in emotional facial expression. Journal of Cross-Cultural Psychology. Wedeck, J. (1947). The relationship between personality and 'psychological

ability British Journal of Psychology, 37, 133-151.

Wundt, W. (1897). Outlines of psychology (translated by C. W. Judd). Leipzig: Wilhelm Engelman.

Yzerbyt, V. Y., Corneille, O., Dumont, M., & Hahn, K. (2001). The dispositional inference strikes back: Situational focus and dispositional suppression in causal attribution. Journal of Personality and Social Psychology, 81(3), 365-376.

Zebrowitz, L. A., Voinescu, L., & Collins, M. A. (1996). 'Wide-eyed' and 'crooked-faced': Determinants of perceived and real honesty across the life span. Personality and Social Psychology Bulletin, 22(12), 1258-1269.

Zuckerman, M., DeFrank, R. S., Hall, J. A., Larrance, D. T., & Rosenthal, R. (1979). Facial and vocal cues of deception and honesty. Journal of Experi­mental Social Psychology, 15, 378-396.

Zuckerman, M., & Driver, R. E. (1985). Telling lies: Verbal and nonverbal cor­relates of deception. In W. A. Siegman & S. Feldstein (Eds.), Multichannel integration of nonverbal behavior (pp. 129-147). Hillsdale, NJ: Lawrence Erlbaum Associates.

Zuckerman, M., Koestner, R., Colella, M. J., & Alton, A. O. (1984). Anchoring in the detection of deception and leakage. Journal of Personality and So­cial Psychology, 47(2), 301-311.