In a forensic context, a lie is defined as a sender transmitting a message with the deliberate intention of misleading the receiver to cultivate a false belief or conclusion. The important element of this phenomenon is that the sender is cognizant of his or her intent to distort the truth, and the expectation is that the receiver does not anticipate or detect this. In criminal investigations, a primary aim of law enforcement agents and expert witnesses is to reliably and accurately distinguish truths from lies. This article reviews the tools developed to aid the detection of lies, explores lie detection accuracy, and examines stakes and response biases in lie detection.
Tools That Aid Deception Detection
Verbal
Several speech analysis tools have been developed to aid experts in distinguishing truthful from fabricated accounts of events in the forensic setting. These include the Statement Validity Assessment, which is a credibility assessment tool used for accounts relating to interviewees’ reporting of events and experiences. Criteria-Based Content Analysis, a component of Statement Validity Assessment, can be used by interviewers to focus on features considered to correspond with truthful accounts. Scientific Content Analysis is another credibility assessment tool, which helps appraise both quality and content of information given by interviewees. While accuracy rates using Criteria-Based Content Analysis in laboratory settings (i.e., accurately categorizing a truthful account as truthful and a deceptive account as deceptive) are found to be higher than chance level for detecting truth and lies, these procedures are generally not satisfactorily admissible as key evidence in criminal court cases.
Physiological
Physiological tools developed to aid deception detection have had varying degrees of success. Polygraph tests indirectly infer deception through the examination of physiological reactions in relation to a series of structured yet unstandardized questions. The autonomic arousals typically evaluated are respiration, sweating of the fingers, and blood pressure. The Comparison Question Test is one type of a polygraph procedure universally used in criminal investigations. It compares responses of relevant questions specific to the crime (e.g., “On June 21, did you stab your husband?”) with control questions not specifically related to the crime and which are generally broader in scope (e.g., “Have you ever hurt someone who trusted you?”). The underlying assumption of this test is that a truth teller will indicate a higher level of arousal for control questions in contrast to relevant ones since relevant questions inquire into a crime they are not culpable of. In contrast, a liar is anticipated to reveal higher arousal for relevant questions since these questions correspond to a more severe and current threat. There is insufficient evidence demonstrating that polygraph procedures are highly accurate in detecting deception, due to the underlying theoretical assumptions of the tests. Empirically, it cannot be verified that distinctive physiological responses diagnostically indicate that deception is present. Consequently, polygraphs are used by police or attorneys as a basis for interrogations and risk assessments rather than being presented as evidence of deception in criminal courts.
An equally controversial advancement in physiological procedures to aid lie detection is functional magnetic resonance imaging. Studies have shown potential in this technology, which reveals that certain regions of the brains are more engaged when an individual is lying compared to when he or she is telling the truth. The same cannot be determined when individuals are telling the truth, demonstrating that lying is more cognitively demanding than truth telling. Yet, these findings cannot systematically verify the underlying assumption that an individual is lying when certain regions consistently activate during a task, even across several trials. Any subsequent results displaying neural patterns can only be considered as correlational and not causational with associated cognitive activity or behavior of deception requiring additional substantiation. In its present form, the functional magnetic resonance imaging procedure fails to satisfy the reliability prong of the Daubert standard set by the U.S. Supreme Court, as well as its precedent, the Frye standard. Research has also cautioned that the accuracy of detecting false positives and true negatives, while higher than that of the polygraph procedure, may see to a dramatic decrease in higher stakes cases due to noncompliance and countermeasures that can be employed by suspects. Lie detection reports grounded on functional magnetic resonance imaging results are yet to be endorsed as admissible evidence into trial courts (i.e., United States v. Semrau, 2012).
Lie Detection Accuracy
Lie detection accuracy is usually quantified by the number of correct assessments relative to the total number of assessments made by receivers. Deception detection research has revealed that individuals are generally poor lie detectors, no better than chance.
Earlier explanations as to why we are poor include that we often draw invalid cues of deception from accounts given, such as the Othello error. Examples of this error include nervousness as a sign of lying, liars always do not give rich detail in their accounts, and eye-gaze aversion is always an indication of lying. Another prevalent belief is that nonverbal cues are more diagnostic and reliable than verbal ones, due to the belief that liars are more capable of regulating their verbal behaviors compared to their nonverbal ones.
For a while, it was believed that the two main reasons human judges are poor in detecting deception is because people tend to use undiagnostic cues and because there is a lack of diagnostic cues when it comes to deception. Conversely, it has been found that humans are poor judges of deception due primarily to the weaknesses of the behavioral cues themselves. In reality, then, people seldom rely on invalid cues.
The ability to detect lies is also subject to individual differences. An individual’s ability to detect lies and truths is not associated with confidence levels but with other characteristics (e.g., personality traits). Some groups of individuals also perform better than others. For instance, the Central Intelligence Agency’s federal officers, county sheriffs, and clinical psychologists were able to accurately judge lies and truths at a rate significantly higher than chance level. In addition to the issue of observer lie detection ability, certain senders simply appear more credible than other individuals whether or not they are telling the truth—a term coined sender detectability.
Stakes Do Not Matter
It was long posited that lie detectability in high-stakes conditions will be different from that of laboratory simulations (i.e., scenarios that include trivial lies or lies that are not of a critical nature with critical penalties, typically set within a laboratory condition or university setting and restricted by ethical concerns). It was thought that the cues that liars exhibit will be more discernible in a high-stakes situation than in a low-stakes one. Thus, the higher the stakes, the higher the chance more reliable deception cues will be elicited and thus deceit will be more likely to be detected.
Contrary to these earlier propositions, truth–lie detection in a higher-stakes scenario did not fare much better, or easier. Furthermore, there was little difference in how detectable lies were across different scenarios. These scenarios included statements moderated by the level of sender’s motivation (i.e., sender is unmotivated or highly motivated) and highly emotional statements.
Response Biases in Lie Detection
The truth bias hypothesis holds that veracity judgments will always be greater than 50% for truthful statements. Observers are more likely to ascertain honest messages accurately more often than lies due to a bias toward truthful messages. The truth default theory claims that the truth default passively draws assumptions about a statement or sender. It is a cognitive default hypothesized to occur without conscious reflection and can be empirically measured via the observation of a truth bias. Essentially, the truth bias can be made with or without conscious reflection. Truth bias needs not be a cognitive default, so a judge can arrive at a truth bias without requiring a truth default; however, a truth default helps explain why a truth bias is preselected and transpires.
Conversely, the lie bias has been shown to be exhibited more often by certain groups, including police officers and prison inmates. This bias posits that observers will achieve accuracy greater than chance level in correctly categorizing statements that are deceptive compared to truthful ones.
In contrast to a default in believing or disbelieving others, the Adaptive Lie Detector account posits there are no such cognitive defaults. Instead, response biases are indications of observers attempting to make an informed guess when little information is available or when diagnostic value of cues present is low. In these cases, individuals are theorized to adapt their decision-making according to the specific situation by using heuristics, which causes biases to surface. Contrary to the truth default theory, under the Adaptive Lie Detector account, both truth and lie biases are posited to arise from the same underlying process.
The accuracy of these theories rests on whether these response biases are inherent defaults or otherwise. Since these defaults are predominantly unconscious, it then becomes vital that the role of unconscious lie detection be revealed. To date, there is not much existing empirical evidence for unconscious but accurate lie detection.
Final Thoughts
While having the potential to aid the detection of deception, verbal and physiological tools in their current form do not meet scientific criteria in terms of relevance, probative value, and reliability. Furthermore, standing theories in lie detection research are subject to be further developed, supported, or challenged.
References:
- Bond, C. F., Jr., & DePaulo, B. M. (2008). Individual differences in judging deception: Accuracy and bias. Psychological Bulletin, 134(4), 477–492.
- Hartwig, M., & Bond, C. F. (2011). Why do lie-catchers fail? A lens model meta-analysis of human lie judgments. Psychological Bulletin, 137(4), 643–659. doi:10.1037/a0023589
- Hartwig, M., & Bond, C. F. (2014). Lie detection from multiple cues: A meta-analysis. Applied Cognitive Psychology, 28(5), 661–676. doi:10.1002/acp.3052
- Street, C. N. H. (2015). ALIED: Humans as adaptive lie detectors. Journal of Applied Research in Memory and Cognition. doi:10.1016/j. jarmac.2015.06.002
- Vrij, A., Granhag, P. A., & Porter, S. (2010). Pitfalls and opportunities in nonverbal and verbal lie detection. Psychological Science in the Public Interest, 11(3), 89–121. doi:10.1177/1529100610390861
Court Cases
- Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993).
- Frye v. United States, 54 App. D.C. 46, 293 F. 1013 (1923).
- United States v. Semrau, 693 F.3d 510 (6th Cir. 2012).