ArticlePDF Available

Detecting Deceit via Analysis of Verbal and Nonverbal Behavior

Authors:

Abstract

We examined the hypotheses that (1) a systematic analysis of nonverbal behavior could be useful in the detection of deceit and (2) that lie detection would be most accurate if both verbal and nonverbal indicators of deception are taken into account. Seventy-three nursing students participated in a study about telling lies and either told the truth or lied about a film they had just seen. The interviews were videotaped and audiotaped, and the nonverbal behavior (NVB) and speech content of the liars and truth tellers were analyzed, the latter with the Criteria-Based Content Analysis technique (CBCA) and the Reality Monitoring technique (RM). Results revealed several nonverbal and verbal indicators of deception. On the basis of nonverbal behavior alone, 78% of the lies and truths could be correctly classified. An even higher percentage could be correctly classified when all three detection techniques (i.e., NVB, CBCA, RM) were taken into account.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Detecting deceit via analysis of verbal and nonverbal behavior
Aldert Vrij; Katherine Edward; Kim P Roberts; Ray Bull
Journal of Nonverbal Behavior; Winter 2000; 24, 4; Research Library
pg. 239
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
... Por otro lado, Vrij et al. (2000) parten de un modelo alternativo, conformado por información sensorial, información contextual (temporal y espacial), información afectiva y operaciones cognitivas. Además, introducen una división en la categoría de información sensorial, clasificándola en función de su procedencia, de modo que distinguen entre detalles visuales y detalles de sonido, olfato, gusto, tacto y sensaciones físicas. ...
... Esta variable se refiere al canal por medio del cual es percibida la información que posteriormente constituirá la base de los relatos de memorias experienciadas. Así, se ha observado que algunos estudios emplean experiencias vividas, ya sean acontecimientos vitales (e.g., Willén y Strömwall, 2011) o eventos diseñados en el contexto del experimento (e.g., Elnitb et al., 2015); mientras que otros estudios utilizan visualizaciones de eventos ocurridos a otras personas (e.g., Vrij et al., 2000). En este sentido, los distintos estudios realizados concuerdan en mostrar que el RM es más eficaz cuando se aplica a memorias autobiográficas que cuando se toma como base un evento visionado (Masip et al., 2005). ...
... Para el análisis de los criterios del RM, se tomó como base el listado propuesto por Sporer y Küpper (2005), que contiene un total de 8 criterios, los cuales se pueden dividir en función de si su origen es externo (claridad, información sensorial, información espacial, información temporal, afecto, reconstrucción de la historia y realismo) o interno (operaciones cognitivas). Asimismo, se ha tenido en cuenta la distinción propuesta por Vrij et al. (2000), relativa a la información sensorial, la cual define subcategorías para este criterio: detalles visuales, detalles auditivos, olfato, gusto y sensaciones físicas. ...
Thesis
Full-text available
El informe pericial psicológico se ha convertido en una prueba determinante en los procedimientos referentes a delitos del ámbito privado, demandándose especialmente la evaluación de la credibilidad de las declaraciones y el estudio de la huella psicológica del delito, el Trastorno de Estrés Postraumático. Para la primera tarea, la aproximación que más eficacia ha mostrado es el análisis del contenido de las declaraciones. Entre los instrumentos empleados para ello, se encuentra el Reality Monitoring (RM). Por su parte, para el estudio de la huella, es necesario adoptar una aproximación multimétodo, que combine entrevista e instrumentación psicométrica; y controlar la simulación. El instrumento psicométrico más empleado es el MMPI-2, que permite evaluar diversos trastornos, así como detectar la simulación mediante las escalas de validez. El objetivo de la presente tesis es estudiar la eficacia de las herramientas empleadas en el ámbito forense, concretamente el RM y el MMPI-2. Para ello, se llevaron a cabo revisiones meta-analíticas sobre ambas herramientas. Para el MMPI-2, se estudió el desempeño de las escalas F, Fptsd y Ds para detectar la simulación. Además, se comparó la escala F con su versión revisada (F-r) y se analizaron distintos moderadores: tipo de diseño, grupo de contraste y tipo de población. Para el RM, se examinó la eficacia de la puntuación total de la herramienta y de cada criterio de forma individual. Asimismo, se profundizó en el efecto de las principales variables moderadoras: edad de los participantes, tipo de evocación y forma de puntuación. Los resultados muestran que las escalas F, Fptsd y Ds del MMPI-2 detectan eficazmente la simulación; que la escala F discrimina con mayor precisión que la F-r; y que las variables estudiadas presentan efectos moderadores sobre los resultados obtenidos. Para el RM, la puntuación total presenta problemas debido a la falta de consenso en los modelos de medida; y sólo un criterio se ha mostrado consistentemente eficaz conforme al modelo, encontrándose afectados por los moderadores estudiados. Además, se ha encontrado que el instrumento no discrimina entre memorias vividas y visionadas, presentándose así una importante limitación para la práctica. A la luz de estos resultados, se puede concluir que las escalas estudiadas del MMPI-2 permiten distinguir entre respuestas honestas y simuladas, si bien se han de emplear en conjunción con otros indicadores y herramientas. De cara a futuros estudios, los efectos de las variables moderadoras deben ser tenidos en cuenta. Además, es necesario realizar revisiones meta-analíticas sobre otras escalas del instrumento. En cuanto al RM, los resultados advierten de sus limitaciones y de la carencia de validez judicial. La investigación futura ha de dirigirse a estandarizar la herramienta y optimizar los criterios contenidos en ella. Además, es necesario llevar a cabo revisiones meta-analíticas sobre otras variables del diseño.
... Objective approaches to measuring cues to deception could be preferable to selfreport. Indeed, cues to deception have frequently been assessed by having two or more observers independently make judgments about a participant's behaviors (e.g., Strömwall et al., 2006;Vrij et al., 2000). Observers have typically been trained to identify specific cues by researchers, and range in lie detection experience from university students (e.g., Mann et al., 2002) to police officers (e.g., . ...
... In two other studies, they were asked to rate nine cognitive (e.g., number of words spoken) and arousal (e.g., pitch, speech errors) cues (Sporer & Schwandt, 2006), and 13 cognitive (e.g., pauses) and arousal (e.g., blinking, head nods) cues . Each of these cues has also been assessed using various approaches, such as having observers assess the frequency (e.g., number of pauses; Vrij et al., 2000), duration (e.g., length of pauses; Vrij et al., 2000), and/or degree of particular behaviors (e.g., plausibility; Leal et al., 2010). The variability in the type, number, and measurement of cues has complicated the establishment of reliability and validity of trained coders' assessments across studies. ...
... In two other studies, they were asked to rate nine cognitive (e.g., number of words spoken) and arousal (e.g., pitch, speech errors) cues (Sporer & Schwandt, 2006), and 13 cognitive (e.g., pauses) and arousal (e.g., blinking, head nods) cues . Each of these cues has also been assessed using various approaches, such as having observers assess the frequency (e.g., number of pauses; Vrij et al., 2000), duration (e.g., length of pauses; Vrij et al., 2000), and/or degree of particular behaviors (e.g., plausibility; Leal et al., 2010). The variability in the type, number, and measurement of cues has complicated the establishment of reliability and validity of trained coders' assessments across studies. ...
Thesis
Cognitive load and arousal are constructs typically included in theories of deception, but they are often measured using a range of unvalidated techniques. Using a multitrait-multimethod analysis, I assessed the reliability and construct validity of common measures of cognitive load and arousal – self-report, trained coders’ observations, and behavioral measures – across three studies as secondary data. All measures showed good reliability, but achieved differing levels of validation. Measures of cognitive load (i.e., self-reported cognitive load, trained coders’ observations of thinking hard, and average response latency) showed some evidence of construct validity. In contrast, measures of arousal (i.e., self-reported arousal, trained coders’ observations of nervousness, and average skin conductance) did not achieve sufficiently high levels of validity. These findings suggest that researchers may not be assessing constructs of interest. Thus, researchers should exercise caution when using unvalidated measures to evaluate theories of, and the diagnosticity of cognitive and arousal-based cues to, deception.
... Earlier research in the field of non-verbal analysis relied solely on meticulous observation and analysis of video data, such as viewing hours of recorded videos in order to interpret social situations (Vrij et al., 2000;Hart et al., 2016). This method of behaviour analysis is prone to subjectivity, is time consuming and does not scale with large amounts of data. ...
... While non-verbal cues are generally accepted to be an important element within media interviews and are typically included within training, the accuracy of trainers in detecting these signals is uncertain as inference of emotions is subjective by nature (Vrij et al., 2000). ...
Article
Full-text available
It is well recognised that social signals play an important role in communication effectiveness. Observation of videos to understand non-verbal behaviour is time-consuming and limits the potential to incorporate detailed and accurate feedback of this behaviour in practical applications such as communication skills training or performance evaluation. The aim of the current research is twofold: (1) to investigate whether off-the-shelf emotion recognition technology can detect social signals in media interviews and (2) to identify which combinations of social signals are most promising for evaluating trainees’ performance in a media interview. To investigate this, non-verbal signals were automatically recognised from practice on-camera media interviews conducted within a media training setting with a sample size of 34. Automated non-verbal signal detection consists of multimodal features including facial expression, hand gestures, vocal behaviour and ‘honest’ signals. The on-camera interviews were categorised into effective and poor communication exemplars based on communication skills ratings provided by trainers and neutral observers which served as a ground truth. A correlation-based feature selection method was used to select signals associated with performance. To assess the accuracy of the selected features, a number of machine learning classification techniques were used. Naive Bayes analysis produced the best results with an F-measure of 0.76 and prediction accuracy of 78%. Results revealed that a combination of body movements, hand movements and facial expression are relevant for establishing communication effectiveness in the context of media interviews. The results of the current study have implications for the automatic evaluation of media interviews with a number of potential application areas including enhancing communication training including current media skills training.
... To accomplish deception and avoid detection in this twofold task, deceivers must invest cognitive resources into fabricating a story that did not actually take place, without contradicting themselves or otherwise arousing suspicion. For this reason, deception is assumed to be more cognitively demanding than telling the truth [41]. Two categories of functional words have been specifically identified as markers of cognitive complexity: exclusive and motion words. ...
... • kłamcy rzadziej mówią o jedzeniu, • kłamcy częściej mówią o wypoczynku i zajęciach relaksacyjnych (źródło:Zuckerman, DePaulo, Rosenthal 1981; DePaulo, Lindsay, Malone, Muhlenbruck, Charlon, Cooper 2003;Hauch, Blandon-Gitlin, Masip i Sporer 2014;Newman, Pennebaker, Berry, Richards 2003;Vrij, Edward, Roberts, Bull 2000). Próby replikacji nie wytrzymały niżej wymienione kryteria, w przypadku których nie stwierdzono istotnych statystycznie różnic:• zwiększony okres latencji w komunikatach kłamliwych, • zwiększona liczba błędów językowych w komunikatach kłamliwych, • brak koherencji w komunikatach kłamliwych, • dłuższe odpowiedzi w komunikatach prawdziwych, • większa liczba stanowczych stwierdzeń (leksemów), typu "zawsze" lub "nigdy" w komunikatach kłamliwych, • częstsze użycie czasowników modalnych w komunikatach kłamliwych, • częstsze użycie słów dotyczących negatywnych emocji w komunikatach kłamliwych, • częstsze użycie słów dotyczących pozytywnych emocji w komunikatach prawdziwych, • częstsze użycie liczby mnogiej w komunikatach kłamliwych, ...
Article
Full-text available
This article attempts to investigate the legitimacy of including the linguistic perspective to the content analysis of witness testimony, which at the present time is being primarily developed in the area of forensic psychology. If we consider the fact that at least part of the nineteen CBCA criteria included in the Statement Validity Assessment method seems to be rather inaccurate and ineffective, it cannot be treated as a credible and objective tool to be applied in the Polish forensic practice. The article presents the former achievements of linguists in creating criteria for testimony content analysis and their definitions to possibly be applied in practice. It also presents the results of a diagnostic accuracy survey carried out on a Polish language material on the currently existing linguistic criteria and author’s new proposals.
... However, despite its extensive usage in our dialogues, lying is not a straightforward task. It requires a higher effort than telling the truth [14] [15], since, in order to inhibit cues that could expose it, the liar needs to mentally keep track of the lies in parallel with the real events [19]. Moreover, little is known about the cognition of deception [6]. ...
Conference Paper
Full-text available
Humans lie every day, from the least harmful lies to the most impactful ones. Therefore, in an attempt to design virtual agents endowed with advanced decision-making abilities , researchers not only focused their effort in designing cooperative and truthful agents but also deceptive and lying ones. In this paper we propose a model capable of engaging an agent in an uncooperative misleading dialogue with a user. This model gives to an agent the ability to reason about its knowledge and then autonomously adjust the story it tells depending on what its interlocutors might know and on how sensitive it considers the conversation topic to be. Such a model allows a story's author to focus on the main narrative, letting the model handle the generation of alternatives. We implemented the model in an agent called the Deceptive Virtual Suspect and conducted some preliminary experiments using an Interrogation Game.
Article
Full-text available
Deception detection can be of great value during the juristic investigation. Although the neural signatures of deception have been widely documented, most prior studies were biased by difficulty levels. That is, deceptive behavior typically required more effort, making deception detection possibly effort detection. Furthermore, no study has examined the generalizability across instructed and spontaneous responses and across participants. To explore these issues, we used a dual-task paradigm, where the difficulty level was balanced between truth-telling and lying, and the instructed and spontaneous truth-telling and lying were collected independently. Using Multivoxel pattern analysis, we were able to decode truth-telling versus lying with a balanced difficulty level. Results showed that the angular gyrus (AG), inferior frontal gyrus (IFG), and postcentral gyrus could differentiate lying from truth-telling. Critically, linear classifiers trained to distinguish instructed truthful and deceptive responses could correctly differentiate spontaneous truthful and deceptive responses in AG and IFG with above-chance accuracy. In addition, with a leave-one-participant-out analysis, multivoxel neural patterns from AG could classify if the left-out participant was lying or not in a trial. These results indicate the commonality of neural responses subserved instructed and spontaneous deceptive behavior as well as the feasibility of cross-participant deception validation.
Article
Groups often make better judgments than individuals, and recent research suggests that this phenomenon extends to the deception detection domain. The present research investigated whether the influence of groups enhances the accuracy of judgments, and whether group size influences deception detection accuracy. 250 participants evaluated written statements with a pre‐established detection accuracy rate of 60% in terms of veracity before viewing either the judgments and rationales of several other group members or a short summary of the written statement and revising or restating their own judgments accordingly. Participants’ second responses were significantly more accurate than their first, suggesting a small positive effect of structured groups on deception detection accuracy. Group size did not have a significant effect on detection accuracy. The present work extends our understanding of the utility of group deception detection, suggesting that asynchronous, structured groups outperform individuals at detecting deception. This article is protected by copyright. All rights reserved.
Article
Full-text available
Reality Monitoring (RM) criteria has been proposed as a forensic tool in order to discern between perceived and imagined memories. However, no systematic evidence has been provided on its validity for use in testimony evaluation. Thus, a meta-analytic review was designed to study its validity in forensic setting. A total of 40 primary studies were found, yielding 251 effect sizes. Random-effects meta-analyses correcting the effect size for sampling error and criterion unreliability were performed. The results showed that the total RM score discriminated, d = 0.542 (δ = 0.562), between imagined and perceived memories of events. In relation to individual criteria, the results showed support for the model's predictions (more external attributes in perceived memories) for clarity, d = 0.361 (δ = 0.399), sensory information, d = 0.359 (δ = 0.397), spatial information, d = 0.250 (δ = 0.277), time information, d = 0.509 (δ = 0.563), reconstructability of the story, d = 0.441 (δ = 0.488), and realism, d = 0.420 (δ = 0.464), but not for affective information, d = 0.024 [-0.081, 0.129]. Nevertheless, except for temporal information, the results are not generalized (negative effects may be found). For cognitive operations, the results corroborated, although the magnitude of the effect was lower than small, the hypothesis (more cognitive operations in imagined memories), d =-0.107 [-0.178,-0.036] (δ =-0.119). The moderating effects of age (more cognitive operations on imagined memories in adults, and on perceived memories in underage), evocation type (external attributes discern between imagined and perceived memories, in both self-experienced and non-experimented accounts), and criteria score (the results varied by score) moderators were studied. As conclusions, forensic implications for the validity of the RM technique in court proceedings are discussed.
Article
The Criteria-Based Content Analysis (CBCA) technique was developed to distinguish children's truthful from fabricated allegations. Research results suggest some use for the procedure, but many important theoretical and empirical issues remain unresolved, including wide differences in the apparent usefulness of individual CBCA criteria, inconsistencies in the number of criteria used, and the absence of decision rules for evaluating an individual child. Other issues include the effect of the child's age and cultural background, motivation to lie, being coached to lie, and belief in the validity of a false memory. Although the CBCA technique shows some promise in enabling raters to differentiate true from false statements, the authors conclude that the presentation of expert testimony derived from CBCA analyses of an individual child would be premature and unwarranted.
Chapter
The development of the Statement Reality Analysis (SRA) technique is described. The technique is employed to assess the credibility of witness evidence in criminal cases. An expert psychologist is appointed by the court in cases in which a child’s evidence is central in criminal proceedings. The expert interviews the child, other principals in the event, reviews the forensic evidence, attends the trial and then renders an opinion to the court of the credibility of the child’s evidence. Procedures like SRA are employed in both parts of Germany and in Sweden. The chapter begins with a critical examination of eyewitness research and its limited value in real forensic contexts. The origins of SRA are then traced from a court decision in Germany in the 1950s to its full elaboration in the 1970s. The assumptions underlying the SRA procedure are detailed and the procedure outlined.
Chapter
A central goal of each trial process is the reconstruction of past events. The available information, whether it be factual evidence or verbal statements of the accused or of eyewitnesses, has to be evaluated in the light of this basic task. In the case of factual evidence, the evaluation task is basically a question of interpretation of facts while the facts themselves are only seldom in question. Things are quite different, however, if the evaluation has to be based on witness statements. Witness statements cannot be taken at face value. There are, in fact, numerous factors which might lead to discrepancies between the actual course of past events and details of a statement. The degree of congruence between a statement and the original stimuli will be referred to as the reliability of that statement. The various reasons for unreliable witness accounts can be classified into two basically different groups of factors: First, witnesses, although trying to give a correct and complete report of an event or a precise description of a person, may be subject to unconscious and unintended errors, caused for example by simple forgetting, suboptimal perception conditions, misleading postevent information, or biased instructions (see, e.g., Clifford & Bull, 1978; Davies, Ellis, & Shepherd, 1981; Lloyd-Bostock & Clifford, 1983; Loftus, 1979; Shepherd, Ellis, & Davies, 1982; Wegener, Köhnken, & Steller, in press; Wells & Loftus, 1984, for reviews of this line of research). These factors affect an aspect of the reliability of a statement which will be referred to as the accuracy.
Chapter
This chapter provides a theoretical analysis of why some lies fail while others succeed and presents a discussion of its implications for the design and interpretation of research on deception. New findings on the measurement of facial expression, body movement, voice and verbal behavior are reported which show that the combination of facial and vocal measures allows highly accurate identification of deceptive behavior. Other data show, however, that most people do not make use of the behavioral cues most relevant to detecting deception, but instead they depend upon those behaviors least valuable in the detection of deception.
Chapter
Criteria-based content analysis (CBCA) and statement validity assessment (SVA) are semi-standardized methods for the credibility assessment of children’s statements in cases of sexual abuse. CBCA consists of a systematic analysis of the content of children’s statements using a set of defined criteria, while SVA incorporates additional information from other sources than the statement itself. This chapter provides condensed descriptions of CBCA and SVA and summarizes recent simulation and field studies on the validity of CBCA. The results of these studies demonstrate the usefulness of CBCA for the purpose of credibility assessment of children’s statements about sexual abuse.
Article
The Criteria-Based Content Analysis (CBCA) technique is the core component of the Statement Validity Assessment (SVA), an interviewing process that attempts to judge allegations of child sexual abuse for truthfulness. As the nucleus of the SVA, the CBCA has been promoted as a way to detect truth by analyzing the verbal content of such an allegation. However, all conclusions of the CBCA's usefulness as an indicator of truth have been based on European and/or White perspectives, and this is the first study to address the implications of probable differences of verbal content meaning across ethnic groups. The results failed to support the hypotheses that Black speakers' truthful statements contained fewer criteria than Whites' statements, or that CBCA training would only enhance truth detection of White speakers. First, in contrast to earlier work, the present study found that, despite the fact that true and false statements contained criteria to differing degrees, the total number of CBCA criteria in an adult's statement was uncorrelated with truthfulness for statements made by Blacks or by Whites. A discriminant analysis showed that CBCA verbal content criteria did, in fact, have different significance for Whites than for Blacks. Certain CBCA criteria were stronger predictors of truth for one ethnic group than for the other. It is suggested that more research is needed before the CBCA is relied upon in criminal investigations.
Article
Accuracy in the ability to detect truths and lies is important in a legal setting. It might be used as a tool in police investigations to eliminate potential suspects, to check the truthfulness of informants or to examine contradictory statements of witnesses and suspects in the same case. A consistent finding in the detection of deception literature is the truth bias: People's accuracy at detecting truths is usually higher than their accuracy at detecting lies. The present article examines whether the existence of a truth bias depends on the type of lie. It is argued that a truth bias may occur when people judge extensive statements (e.g. elaborations), but that a lie bias may occur when people judge statements which do not provide much verbal information (e.g. denials). Fifty participants (college students) were exposed to 20 video fragments of 20 people telling elaborations (10) or denials (10). Half of the elaborations and denials were truthful, the other half were deceptive. After each fragment, the participants were asked to indicate whether the person was lying or telling the truth and how confident they were in their decision making. As predicted, with regard to elaborations a truth bias was found and with regard to denials a lie bias was found. In other words, people have difficulty in accurately judging deceptive elaborations and truthful denials. The study further revealed individual differences in participants' confidence at detecting deceit. The more socially anxious/shy the participants reported themselves to be, the less confident they were in their ability to detect deceit. Also, the more extraverted they themselves reported to be, the more confident they were in their ability to detect deceit. The importance of confidence on improving people's ability to detect deceit will be discussed.
Article
Humans lack the ability to detect deceptive communication when it is present. This review examined several explanations for this state of affairs. Twenty years of research in deception has shown that there is not a reliable set of nonverbal or verbal indicators of deceptive communication. Moreover, human lie detectors’ veracity judgments are often affected by cognitive biases and erroneous stereotypical information about how a prototypical liar should look. The current review also suggests that the inability to distinguish lies from truths may be a function of the decoding task presented to receivers in interpersonal communication. Receivers of deceptive communication must reject information they have already accepted, must draw inferences of another's underlying intent, and rarely receive any feedback with regard to their inferences. Additional factors that mediate lie detection accuracy are also reviewed. Finally, suggestions were made regarding how scholars should proceed with future research efforts in lie detection.