Vetting deception detection

We should be wary of unsubstantiated alternative approaches when we’re trying to improve our deception detection anti-fraud practices. Make sure that any theory you hear is backed by evidence and peer-reviewed research.

We've all seen offers for courses that promise to teach us how to detect deception in subjects of an investigation. Some are legitimate and based upon peer-reviewed empirical science. However, many are based upon ideas or notions without any solid scientific support, which can result in serious consequences. We have to learn to distinguish unsubstantiated claims from those that are evidence-based. However, this isn’t always an easy task. The aim of this column is to assist anti-fraud professionals in making such a distinction.

Science and pseudoscience of deception detection

Although the practice of detecting deceit dates back to biblical stories, empirical research regarding the detection of deception — as it is known today — started in the 1960s. In one of the first theories linking demeanor and deception, the “leakage theory” advocated that nonverbal cues, or “tells” could (unknowingly) expose emotions experienced internally by liars. (See Nonverbal leakage and clues to deception, by P. Ekman and W. V. Friesen, Psychiatry, 1969.)

A little more than a decade later, the “four-factor theory” suggested that the behavior of liars could not only be influenced by the emotions they experience but also by their arousal plus mental efforts required to deceive successfully and attempts to control their behavior to appear credible. (See “Verbal and nonverbal communication of deception,” by M. Zuckerman, B. M. DePaulo and R. Rosenthal, in “Advances in Experimental Social Psychology,” 1981, edited by L. Berkowitz, Academic Press.)

In the following years, both the leakage and four-factor theories largely influenced deception detection research. Several attempts have been made to find valid nonverbal signs that someone is lying. However, in 2003, Bella DePaulo and colleagues demonstrated that most non-verbal behaviors typically associated with deception were, in fact, weakly correlated and, more than often, unreliable. (See Cues to deception, by B. M. DePaulo, J. J. Lindsay, B. E. Malone, L. Muhlenbruck, K. Charlton & H. Cooper, Psychological Bulletin, 2003.)

Deception scholars grew distant from traditional approaches. Their focus shifted from the nonverbal to the verbal responses of liars and truth tellers, and they developed different techniques leading to significant improvements in deception detection. However, so-called experts continued to place undue emphasis on non-verbal communication to identify liars. (See “Deception detection,” by V. Denault and L. Jupe, in “Psychology and Law Factbook 2,” 2017, edited by B. Baker, R. Minhas and L. Wilson, European Association of Psychology and Law Student Society.)

Deception detection accuracy

Detecting deceit merely by watching and listening to someone can be a difficult endeavor. Several peer-reviewed papers have highlighted that the deception detection accuracy of lay individuals and presumed lie experts (such as police officers) is generally similar to what one could expect from chance alone. In addition, testing has shown that training courses tend to often offer mixed results in terms of improved accuracy. (See Does training improve the detection of deception? A meta-analysis, by V. Hauch, S. L. Sporer, S. W., Michael and C. A. Meissner, Communication Research, 2014.)

However, although we’ve yet to find the silver bullet to detect deception, anti-fraud professionals can make preliminary assessments of approaches offered to improve their deception detection accuracy using current scientific consensus. In addition, empirical research provides important insights on what ought to be important for anti-fraud professionals in face-to-face interactions.

For example, according to research into the “strategic use of evidence” (SUE), when a suspect ignores the incriminating evidence an investigator possesses, the investigator’s deception detection accuracy can reach levels above 80 percent if the investigator asks specific open-ended and follow-up questions. (See Strategic use of evidence during police interviews: When training to detect deception works, by M. Hartwig, P. A. Granhag, L. A. Strömwall and O. Kronkvist, Law and Human Behavior, 2006.)

In addition to the SUE, anti-fraud professionals should also pay close attention to the “verifiability approach” (VA) in which the interviewer asks suspects to provide very detailed answers to specific questions. In practice, the interviewer asks the interviewee to provide evidence that the investigator can check. Liars know that a detailed account sounds more plausible, but they also know that if the interviewer investigates their stories further, they might uncover their deception. Therefore, liars’ simple strategy is to provide lots of unverifiable details. Because truth tellers are generally able to provide more verifiable details than liars, fraud examiners can use VA as a diagnostic tool to assist them when trying to differentiate liars from truth tellers. (See “The applicability of the verifiability approach to the real world,” by G. Nahari, in “Detecting concealed information and deception: Recent developments,” 2018, edited by J. P. Rosenfeld, Academic Press.)

Anti-fraud professionals should pay close attention to the 'verifiability approach' in which the interviewer asks suspects to provide very detailed answers to specific questions.

Importance of peer review: first step toward evidence-based practices

Anti-fraud professionals can better identify unsubstantiated (and substantiated) approaches by understanding the current state of empirical research about the detection of deception. We should ask questions if we hear or read extravagant claims without extravagant evidence and see faulty methodologies and no connectivity with other research. We should ask for the peer-reviewed papers that vetted the concepts or techniques.

In their scientific (peer-reviewed) papers, skilled researchers must thoroughly describe their results and the steps they took to get them — offering any other researchers the possibility to find flaws that could undermine their conclusions about detection deception — just like any other scientific theories. In other words, all evidence will be transparently available for criticism or approval. Without the scientific justification that comes with peer-reviewed papers, trusting so-called experts on deception detection might be deemed an act of blind faith.

However, the lack of peer review shouldn’t always be a criterion for absolute rejection. For example, experiential knowledge — knowledge gained in practice — is vitally important for anti-fraud professionals. But it’s also not free from limitations. Just because a particular deception detection tool worked in one particular situation, doesn’t mean that this can be generalized to other situations.

Possible consequences

We should be wary of unsubstantiated alternative approaches when we’re trying to improve our anti-fraud practices. For example, believing that someone is lying when in fact they’re telling the truth can harm them, distort a fraud examination, misuse resources and result in legal consequences. Of course, the scientific method doesn’t preclude any of these results. However, justifying evidence-based practices is not only easier but comes with fewer risks, irrespective of the persuasion that so-called experts can exhibit when presenting their approaches. Good faith is not a synonym of good practice.

Much of this article was adapted from a revised version of a session Vincent Denault presented at the 2018 ACFE Fraud Conference Canada. — ed.

Vincent Denault, LL.M., LL.B, CFE, is a lawyer and a lecturer in the department of communication at Université de Montréal in Canada. He’s also a research assistant at the university’s Communication and Health Research Center, a doctoral student and co-director of the Center for Studies in Nonverbal Communication Sciences of the Montreal Mental Health University Institute Research Center. Contact him at vincent.denault@umontreal.ca.

Louise Marie Jupe is a doctoral student and research associate in the department of psychology at the University of Portsmouth in the U.K. Contact her at louis.jupe@port.ac.uk.