Group and individual accuracy in deception detection

Previous research examining the accuracy of deception detection judgments by individuals has concluded that the ability to detect deception is only slightly better than chance. Research has also found that individuals tend to be over‐confident, truth‐biased, and reliant on nonverbal behavior when making veracity judgments. This study (N = 129) tested if differences in deception detection accuracy, truth‐bias, judgmental confidences, and self‐reported cue reliance exist between individual judges and groups of individuals working in collaboration. No significant differences between groups and individuals emerged for accuracy, truth‐bias, or self‐reported cue reliance. Individuals within groups, however, were significantly more confident in their decisions than individuals working alone.     

Assessing Deception Detection Accuracy with Dichotomous Truth–Lie Judgments and Continuous Scaling: Are People Really More Accurate When Honesty Is Scaled?

Deception detection experiments consistently find that people are only slightly better than chance at distinguishing truths from lies. Interpersonal deception theory research, however, claims that people can accurately detect deception when honesty is assessed with continuous scaling. This article reports an experiment (N = 140) directly testing if the type of measurement meaningfully impacts accuracy results and conclusions. The findings suggest that the difference is largely illusory, and that the differences that are observed are attributable to differential variances in the measures. Substantively, both methods yield similar conclusions including slightly better-than-chance accuracy, persistent truth bias, and a robust veracity effect.     

A probability model of accuracy in deception detection experiments

This essay extends the recent work of Levine, Park, and McCornack (1999) on the veracity effect in deception detection. The probabilistic nature of a receiver's accuracy in detecting deception is explained, and a receiver's detection of deception is analyzed in terms of set theory and conditional probability. Detection accuracy is defined as intersections of sets, and formulas are presented for truth accuracy, lie accuracy, and total accuracy in deception detection experiments. In each case, accuracy is shown to be a function of the relevant conditional probability and the truth-lie base rate. These formulas are applied to the Levine et al. results, and the implications for deception research are discussed.     

To Catch a Liar: Challenges for Research in Lie Detection Training

Can we train people to detect deception? It is the contention of this article that communication scholars should learn how to train law enforcement professionals on how to detect high stake lies, like those faced by police, judges, customs officials, immigration officials, and so forth. It is proposed that in order to know whether we can train or should bother to train people to detect deception, each training study must meet 6 challenges: (1) relevance, (2) high stakes, (3) proper training, (4) proper testing, (5) generalizability across situations, and (6) generalizability over time. Our quantitative review of the literature suggests that training does significantly raise lie detection accuracy rates. Meta-analytic findings indicate a mean effect size of r = .20 across 20 (11 published studies) paired comparisons of lie detection training versus the control group (i.e., those without some type of training). It should be noted that the majority of the studies that attempt to train lie detectors fall short on many of the above challenges. Current research in lie detection training may actually underestimate the ability to train lie detectors due to the stimulus materials employed in most experiments.     

To Catch a Liar: Challenges for Research in Lie Detection Training

Can we train people to detect deception? It is the contention of this article that communication scholars should learn how to train law enforcement professionals on how to detect high stake lies, like those faced by police, judges, customs officials, immigration officials, and so forth. It is proposed that in order to know whether we can train or should bother to train people to detect deception, each training study must meet 6 challenges: (1) relevance, (2) high stakes, (3) proper training, (4) proper testing, (5) generalizability across situations, and (6) generalizability over time. Our quantitative review of the literature suggests that training does significantly raise lie detection accuracy rates. Meta-analytic findings indicate a mean effect size of r = .20 across 20 (11 published studies) paired comparisons of lie detection training versus the control group (i.e., those without some type of training). It should be noted that the majority of the studies that attempt to train lie detectors fall short on many of the above challenges. Current research in lie detection training may actually underestimate the ability to train lie detectors due to the stimulus materials employed in most experiments.     

Ecological Validity and Deception Detection Research Design

The concept of ecological validity is explicated and applied to deception detection research. The ecology of the deception lab typically involves researchers randomly assigning and instructing senders to lie or tell the truth. In lab-based lie detection tasks, receivers are prompted to make explicit truth-lie judgments in real time where truths and lies are equally probable, and where evidence and persuasion-based lie detection methods are precluded. Outside the lab, deceivers self-select, engage deception as a problem-solving activity, and have greater linguistic freedom to deceive in ways other than telling outright falsehoods. Receivers outside the lab face truth-lies base-rates other than 50–50, they are not prompted to consider the possibility of deception by a researcher, decisions about honesty often need not be immediate, and a wider variety of discovery methods are available. The implications of these differences are discussed in contexts of everyday conversation, criminal investigation, high stakes media lies, and security screenings.     

Strong Diagnostic Questioning Effects on Deception Detection Accuracy With U.S. Customs Agents

Recent communication research shows several paths to improving accuracy in deception detection. According to truth-default theory, one promising approach is diagnostic questioning. A sample of elite U.S. Customs agents participated in a deception detection task. Agents viewed senders who were interrogated with one of three different sets of questions that differed in diagnostic utility. The different questioning sets produced a 36-point swing in accuracy from 42% to 78% accuracy. These findings demonstrate that how someone is questioned can make a substantial difference in deception detection accuracy and that improved accuracy is possible with diagnostic questioning.     

Considering organizations as a unique interpersonal context for deception detection: A meta‐analytic review

This article challenges current interpersonal deception literature by summarizing seven studies that examined the ability of members to detect deception in the context of organizational processes. The combined effect (r = .472) indicated that organizational members are able to differentiate honest from dishonest communicators. According to the Binomial Effect Size Display, observers correctly classified honest or dishonest communication in approximately three of four circumstances (73.6%). This finding is important for situations such as employment interviews or other circumstances when organizational members or constituents need to assess the accuracy of information provided by current or potential members. The theoretical implications for interpersonal and organizational deception research are discussed along with the practical ramifications of this study for employment interviewing, manager/employee communication, and retail sales.     

Interpersonal deception: VI. Effects of preinteractional and interactional factors on deceiver and observer perceptions of deception success

Past research on deception success has focused on receiver judgments, ignoring deceiver perceptions, which may influence the ways deceivers choose to behave in interpersonal interactions. The present study investigated how several preinteractional and interactional factors affect both deceiver and observer perceptions of deception success. Preinteractional factors included three that have been found relevant to deceptive communication: social skill, self‐monitoring, and motivation. Interactional factors included receiver suspicion, plus four proposed by the four‐factor theory of deception (anxiety, affect, task difficulty, and behavioral control). Results indicated that deceivers’ perceptions of success were more affected by interactional factors (especially anxiety, interaction difficulty, and conversational normality). Conversely, observers were more affected by preinteractional factors such as deceiver social skill. However, motivation, self‐monitoring, and conversational normality affected both deceivers’ and observers’ assessments: greater motivation and self‐monitoring, coupled with more natural, expected communication, yielded greater believability in deceivers’ own eyes and the eyes of observers.     

The Role of Consistency in Detecting Deception: The Superiority of Correspondence over Coherence

Inconsistency is often considered an indication of deceit. The conceptualization of consistency used in deception research, however, has not made a clear distinction between two concepts long differentiated by philosophers: coherence and correspondence. The existing literature suggests that coherence is not generally useful for deception detection. Correspondence, however, appears to be quite useful. The present research developed a model of how correspondence is utilized to make judgments, and this article reports on four studies designed to elaborate on the model. The results suggest that judges attend strongly to correspondence and that they do so in an additive fashion. As noncorrespondent information accumulates, an increasingly smaller proportion of judges make truthful assessments of guilty suspects. This work provides a basic framework for examining how information is utilized to make deception judgments and forms the correspondence and coherence module of truth-default theory.     

Vocal and visual cue training in behavioral lie detection

The present study tested the effect of training observers to detect deception using different behavioral cues. Participants were trained to detect deception using either:(1) vocal‐only cues, (2) visual‐only cues, or (3) a combination of vocal and visual cues. A fourth group of observers did not receive any training and served as a control. Results indicated that trained observers were more successful than untrained observers in detecting deceptive communication. More specifically, observers trained to use either visual or both visual and vocal cues were more accurate detectors of deception than untrained naive observers. No differences were found between training conditions with respect to determining the veracity of truthful communication.     

How Custom Agents Really Detect Lies

Research has found discrepancies among folk beliefs about lie detection, deception cue utility, and retrospective accounts of detected lies. Elite customs agents (N = 37) were surveyed regarding their overall beliefs about how to detect lies (general strategy), their opinions about best practices (best strategy), and about a successfully detected lie (actual strategy). Responses were coded in three nonexclusive broad categories: (a) behavioral cues, (b) communication content and evidence, and (c) interactive-questioning approaches. Behavioral cues, especially nonverbal cues, were listed most frequently as general beliefs. Opinions about the best approach were evenly split across the three categories. Communication content and evidence was the most frequently reported actual detection method and was mentioned in 97% of the responses. The results are discussed in relation to Truth-Default Theory and their implications for lie detection practice.     

Baseline familiarity in lie detection

Subjects judged the veracity of truthful and deceptive communicators after viewing 0, 1, 2, or 4 case‐relevant baseline exposures (familiarity) of truthful communication. A positive linear relationship was found between detection accuracy and amount of baseline familiarity. More specifically, observers who viewed four samples of baseline information judged deception more accurately than observers who viewed zero samples of baseline information. Results also indicated an increase in the frequency of truth judgments across familiarity conditions. This increase, however, only approached statistical significance. Results are discussed and future directions in lie detection research are suggested.     

The Viability of Using Rapid Judgments as a Method of Deception Detection

Rapid Judgments (RJs) are quick assessments based on indirect verbal and nonverbal cues that are known to be associated with deception. RJs are advantageous because they eliminate the need for expensive detection equipment and only require minimal training for coders with relatively accurate judgments. Results of testing on two different datasets showed that trained coders were reliably making RJs after watching both long and short interaction segments but their judgments were not more accurate than the expert interviewers. The RJs did not discriminate between truth and deception as hypothesized. This raises more questions about the conditions under which making RJs from verbal and nonverbal cues achieves accurate detection of veracity.     

Examining Sender and Judge Variability in Honesty Assessments and Deception Detection Accuracy: Evidence for a Transparent Liar but No Evidence of Deception-General Ability

A round-robin experiment unpacked message veracity, sender believability (demeanor), judge truth-bias, sender transparency, and judge deception detection accuracy. Generally, more variance was observed in senders than in judges. The data were suggestive of the existence of an unusually transparent liar, but the data were not consistent with a deception-general ability. The results highlight the importance of considering variability in addition to central tendency and the importance of individual differences in senders in deception detection.     

Sensemaking and Relational Consequences of Peer Co-worker Deception

This exploratory study examined sensemaking of peer co-worker deception from the perception of the deceived. A total of 58 narrative accounts of deception were collected via face-to-face interviews with 23 employed adults. Analysis revealed four primary narratives of co-worker deception: corrupt system narratives, “cover your ass” (CYA) narratives, personal gain narratives, and personality trait narratives. Perceived motives and consequences were primary considerations in the sensemaking process and employees reported changing their communication patterns to avoid deceptive co-workers or hold them more accountable for their actions. The theoretical and practical implications of these results are discussed and suggestions for future research are posited.     

Exploring sanctioned and unsanctioned lies in interpersonal deception

This paper explores individuals' motivation to lie in experimental lie detection. Recent research has called into question the validity of experiments employing a sanctioned lie (lying to satisfy the experimenter's instructions) to investigate deceptive communication. Participants were randomly selected to tell the truth, tell a sanctioned lie, or tell an unsanctioned lie. Results indicated that interviewees failed to report any perceived differences in impression management, nervousness, or involvement between veracity conditions. Interviewers also failed to notice any differences in interviewees' behavioral displays across veracity conditions. Thus, the present study supports research efforts employing sanctioned lies to examine interpersonal deception. Results are discussed in terms of Lazarus and Folkman's (1984) stress and coping model.     

Lying and deception detection in television families

This study explores the use of lying and lie‐detection among television families in situation comedies. The findings indicate that (a) lying was prevalent on these television shows; (b) lies to spare others were told most frequently; (c) the majority of lies went undetected; (d) the husband/father character told the most lies; (e) children were the most honest characters in situation comedies and (f) peripheral characters were the recipients of many lies. Two conclusions were drawn from these findings. First, this study gives insight into the content of family communication as presented on television. Secondly', the study suggests television families mirror real families with respect to family structure and the ability to detect deception.     

Detecting Deceptive Communication Through Computer-Mediated Technology: Applying Interpersonal Deception Theory to Texting Behavior

Interpersonal Deception Theory (IDT) argues that deception is an interpersonal action where people employ communication tactics to achieve various goals. Deception is a global concept; thus, much investigation has been conducted to determine the roots of such deceptive behavior. Through IDT, deception is categorized as a dynamic process where both the sender and receiver need to be engaged for deception to be successful. In extension of IDT, a study was conducted to determine the degree to which undergraduate students engage in deceptive behavior via text messaging. Through a survey design, undergraduate students' deceptive texting behaviors and deceptive detection traits were assessed. Results indicate that among undergraduate students, text messaging is an interpersonal form of communication that circumvents professionalism and power. Further, undergraduate students were much more likely to engage in deceptive texting with family and friends, but did so with significantly less frequency in the organizational setting.     

The relative impact of violation type and lie severity on judgments of message deceitfulness

This paper is part of a larger program of research assessing variables that underlie quantitative deceptiveness ratings. Several recent theoretical approaches, including Information Manipulation Theory (IMT), propose that deceptive messages are best understood as varying along two or more dimensions. At the same time, researchers have increasingly moved from dichotomous deception judgments to continuous deception ratings. This paper questions the validity of scaling degrees of deceptiveness along a single dimension, and argues that gradations in perceived deceptiveness reflect both the type of information manipulated and the severity of the consequences of the deception. This reasoning was tested with alxl experiment (N = 236) in which both the type of information manipulated and the severity of the consequences were systematically varied. As predicted, the results suggest that false messages (i.e., quality violations) are rated as more deceptive than lies of omission (i.e., quantity violations) when lie severity is low, but this difference diminishes as lie severity increases. In other words, false messages were rated as deceptive regardless of severity, but messages omitting information were rated as deceptive as false messages only when the consequences were serious. The implications for measuring deception are discussed.