No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Credibility Assessment at Portals
Abstract
In the post-9-11 world, assessing credibility at portals that control entry to countries, public transportation, and large public venues has become a major concern for governments. In response to this concern, the US Government has invested heavily in developing and applying new technology for credibility assessment at portals. The Government's efforts have included human- and machine-based approaches. Our chapter provides a critical review of technologies currently in application and under development. We found the science for these technologies lacking in both theoretical foundation and empirical validation. We provide a scientific perspective on the deceptive context of portal credibility assessment, and on existing scientific theory and research that may be relevant for addressing that context. Finally, we outline a scientific approach to theory development and scientific validation for future research in this area.
... One fundamental problem of deception research is that the accuracy of correctly identifying liars and truth-tellers on average exceeds the guessing level only by about 20 accuracy points (see Kleinberg et al., 2019a;Oberlader et al., 2016) both for manual coding procedures (Vrij et al., 2017) as well as for fully automated deception detection (e.g., Mihalcea & Strapparava, 2009;Pérez-Rosas & Mihalcea, 2014;Soldner et al., 2019). In particular, for large-scale settings where the base rate of persons of interest is often low, these accuracy rates are not satisfactorily (see Honts & Hartwig, 2014;Kleinberg et al., 2019aKleinberg et al., , 2019b. In this paper, we test whether deception detection performance can be augmented by combining two distinct modes of decision-making: automated classification and human judgment. ...
... Third, in our experiment, the base rate of deception was arguably higher than in many real-life settings where humans are predominantly honest (see Honts & Hartwig, 2014;Kleinberg et al., 2019b). A realistic base rate for security screening at boarders, for example, might be 1 deceptive passenger to 10,000 truthful ones (from Honts & Hartwig, 2014). ...
... Third, in our experiment, the base rate of deception was arguably higher than in many real-life settings where humans are predominantly honest (see Honts & Hartwig, 2014;Kleinberg et al., 2019b). A realistic base rate for security screening at boarders, for example, might be 1 deceptive passenger to 10,000 truthful ones (from Honts & Hartwig, 2014). Future steps on (integrated) machine learning detection approaches would ideally incorporate these considerations. ...
Background
Deception detection is a prevalent problem for security practitioners. With a need for more large-scale approaches, automated methods using machine learning have gained traction. However, detection performance still implies considerable error rates. Findings from different domains suggest that hybrid human-machine integrations could offer a viable path in detection tasks.
Method
We collected a corpus of truthful and deceptive answers about participants' autobiographical intentions (n = 1640) and tested whether a combination of supervised machine learning and human judgment could improve deception detection accuracy. Human judges were presented with the outcome of the automated credibility judgment of truthful or deceptive statements. They could either fully overrule it (hybrid-overrule condition) or adjust it within a given boundary (hybrid-adjust condition).
Results
The data suggest that in neither of the hybrid conditions did the human judgment add a meaningful contribution. Machine learning in isolation identified truth-tellers and liars with an overall accuracy of 69%. Human involvement through hybrid-overrule decisions brought the accuracy back to chance level. The hybrid-adjust condition did not improve deception detection performance. The decision-making strategies of humans suggest that the truth bias - the tendency to assume the other is telling the truth - could explain the detrimental effect.
Conclusions
The current study does not support the notion that humans can meaningfully add the deception detection performance of a machine learning system. All data are available at https://osf.io/45z7e/.
... One of the core assumptions of SPOT is that security-threatening passengers will behave differently compared to ordinary travelers and display suspicious behavior. That behavior is claimed to result from experiencing emotional states such as stress or fear (see also Honts & Hartwig, 2014). Central to SPOT are so-called Behavior Detection Officers (BDOs): employees trained in identifying passengers that show behaviors allegedly indicative of deception (Government Accountability Office, 2015). ...
... The key challenge for such methods is the time needed to gather reliable physiological, linguistic, and thermal data. On a large scale, time is a premium (see Honts & Hartwig, 2014;Kleinberg et al., 2017) and a short procedure is one of the key requirements. At the same time, a multimodal approach brings about new challenges, including the handling of false positives that come with each method and the generalizability of machine-learning models to various contexts. ...
There is an increasing demand for deception detection at scale. In situations in which larger numbers of people need to be tested, traditional deception-detection methods are limited because they often require extensive testing sessions or are limited in their flexibility to novel contexts. The aim of this chapter is to discuss the potential for large-scale applications of several deception-detection methods. Specifically, we evaluate the theoretical foundations of the methods, the potential for quick data-collection procedures, and the flexibility of implementing the methods in different contexts. Each method (behavioral observation, reaction times, speech analysis, verbal content, thermal imaging) is outlined, evaluated, and discussed on the suitability for large-scale purposes. There is no substantial evidence to support the validity of behavioral observation nor of speech analysis. Reaction times and verbal content analysis are promising for future of large-scale applications. To reach their full potential, for both methods the key challenges are a fast data-collection process (e.g., remote online testing) and a near real-time assessment of the data. We close this chapter with an outlook on possible new applications including the integration of methods (i.e., multimodal approaches).
... Its government funding has reached almost US$1 billion. Despite its continued implementation and increases in funding, the program is controversial because prominent scientists, the public, and government offices are concerned about its scientific validity and effectiveness [13]. ...
... His and colleagues' research applies because it involves high-stake situations with great consequences for individuals. However, independent research has not successfully replicated Ekman's findings and a recent published study showed that his approach was not as effective as claimed [13,14]. ...
A current focus in deception research is on developing cognitive-load approaches (CLAs) to detect deception. The aim is to improve lie detection with evidence-based and ecologically valid procedures. Although these approaches show great potential, research on cognitive processes or mechanisms explaining how they operate is lacking. Potential mechanisms underlying the most popular techniques advocated for field application are highlighted. Cognitive scientists are encouraged to conduct basic research that qualifies the ‘cognitive’ in these new approaches.
... KEYWORDS: forensic science, computational linguistics, deception detection, named entity recognition, linguistic inquiry and word count, reality monitoring, criteria-based content analysis With an increased demand for security systems like airport border control, researchers and practitioners alike have identified the need for applications to detect deception on a large scale (1). For example, airport security settings preclude many tools used in deception research (e.g., polygraphy, extensive interviewing) due to their limited applicability (e.g., real-time data analysis; scalability). ...
There is an increasing demand for automated verbal deception detection systems. We propose named entity recognition (NER; i.e., the automatic identification and extraction of information from text) based on three established theoretical principles: (i) truth-tellers provide accounts that are richer in detail, (ii) contain more contextual references (specific persons, locations, and times), and (iii) liars tend to withhold potentially checkable information. We test whether NER captures these theoretical concepts and can automatically identify truthful vs. deceptive hotel reviews. We extracted the proportion of unique named entities and compared its discriminative ability to a lexicon word count approach (LIWC) and a competing measure of sentence specificity (speciteller). Named entities discriminated truthful from deceptive hotel reviews and outperformed the lexicon approach by LIWC and sentence specificity by speciteller in diagnostic efficiency. This investigation suggests that named entities may be a useful addition to existing automated verbal deception detection approaches.
... KEYWORDS: forensic science, computational linguistics, deception detection, named entity recognition, linguistic inquiry and word count, reality monitoring, criteria-based content analysis With an increased demand for security systems like airport border control, researchers and practitioners alike have identified the need for applications to detect deception on a large scale (1). For example, airport security settings preclude many tools used in deception research (e.g., polygraphy, extensive interviewing) due to their limited applicability (e.g., real-time data analysis; scalability). ...
There is an increasing demand for automated verbal deception detection systems. We propose named entity recognition (NER; i.e., the automatic identification and extraction of information from text) to model three established theoretical principles: (i) truth tellers provide accounts that are richer in detail, (ii) contain more contextual references (specific persons, locations, and times), and (iii) deceivers tend to withhold potentially checkable information. We test whether NER captures these theoretical concepts and can automatically identify truthful versus deceptive hotel reviews. We extracted the proportion of named entities with two NER tools (spaCy and Stanford's NER) and compared the discriminative ability to a lexicon word count approach (LIWC) and a measure of sentence specificity (speciteller). Named entities discriminated truthful from deceptive hotel reviews above chance level, and outperformed the lexicon approach and sentence specificity. This investigation suggests that named entities may be a useful addition to existing automated verbal deception detection approaches.
Nahari, Vrij, and Fischer [(2014b), Applied Cognitive Psychology, 28, 122–128] found that, when participants were forewarned that their statements would be checked for verifiable details, truth tellers gave much more verifiable details than liars. In this direct replication (n = 72), participants wrote a statement claiming they had carried out their regular campus activities, whereas liars had actually stolen an exam. Statements were coded for verifiable details. Our primary prediction was confirmed: Truth tellers provided significantly more verifiable details than liars. Of note, the replication effect size (d = 0.49) was less than half that of the original (d = 1.14), and – like in the original study – was smaller than the lie‐truth effect size for total details (verifiable and unverifiable details combined; d = 0.80). We hope this will stimulate other independent investigations of VA to tell whether or not coding for verifiability will pass Ockham's razor test.
Background: Deception detection is a prevalent problem for security practitioners. With a need for more large-scale approaches, automated methods using machine learning have gained traction. However, detection performance still implies considerable error rates. Findings from other domains suggest that hybrid human-machine integrations could offer a viable path in deception detection tasks. Method: We collected a corpus of truthful and deceptive answers about participants' autobiographical intentions (n=1640) and tested whether a combination of supervised machine learning and human judgment could improve deception detection accuracy. Human judges were presented with the outcome of the automated credibility judgment of truthful and deceptive statements. They could either fully overrule it (hybrid-overrule condition) or adjust it within a given boundary (hybrid-adjust condition). Results: The data suggest that in neither of the hybrid conditions did the human judgment add a meaningful contribution. Machine learning in isolation identified truth-tellers and liars with an overall accuracy of 69%. Human involvement through hybrid-overrule decisions brought the accuracy back to the chance level. The hybrid-adjust condition did not deception detection performance. The decision-making strategies of humans suggest that the truth bias - the tendency to assume the other is telling the truth - could explain the detrimental effect. Conclusion: The current study does not support the notion that humans can meaningfully add to the deception detection performance of a machine learning system.
Background: Academic research on deception detection has largely focused on the detection of past events. For many applied purposes, however, the detection of false reports about someone's intention merits attention. Based on the verbal deception detection paradigm, we explored whether true statements on intentions were more detailed and more specific than false statements on intentions, particularly when instructed to be as specific as possible.Method: Participants (n = 222) lied or told the truth about their upcoming travel plans either providing 'as much information as possible’ (standard instructions) or being 'as specific as possible’ (i.e., mentioning times, locations, places; specific instructions), resulting in four conditions (truthful vs. deceptive intention by standard instructions vs. specific instructions). We collected data via a custom-made web app and performed automated verbal content analysis of participants’ written answers.Findings: We did not find a significant difference in participants’ statements specificity. The instruction to be as specific as possible promoted more specific information but did not help to discern honest from deceptive flying intentions. Conclusion: The experiment reported here attempted to demonstrate automated detection of verbal deception of intentions. The difficulty in capturing genuine, intentions, and the non-intrusive, non-interactive questioning approach might explain the null findings and raise questions for further research towards large-scale applicability. We conclude with suggestions for a novel framework on semi-interactive information elicitation.
Conceptualizing the Future Attribute Screening Technology (FAST) as double future device.
Anticipations of functionality as future projection and aims to detect future terrorists as preemptive intervention.
FAST represents an ensemble of contrasting socio-technical futures.
Importance of the sociotechnical status and discursive role of technologies in socio-technical futures.
Background
Academic research on deception detection has largely focused on the detection of past events. For many applied purposes, however, the detection of false reports about someone’s intention merits attention. Based on the verbal deception detection paradigm, we explored whether true statements on intentions were more detailed and more specific than false statements on intentions, particularly when instructed to be as specific as possible.
Method
Participants (n = 222) lied or told the truth about their upcoming travel plans either providing ‘as much information as possible’ (standard instructions) or being ‘as specific as possible’ (i.e., mentioning times, locations, places; specific instructions), resulting in four conditions (truthful vs. deceptive intention by standard vs. specific instructions). We collected data via a custom-made web app and performed automated verbal content analysis of participants’ written answers.
Findings
We did not find a significant difference in the specificity of participants’ statements. The instruction to be as specific as possible promoted more specific information but did not help to discern honest from deceptive flying intentions.
Conclusion
The experiment reported here attempted to demonstrate automated verbal deception detection of intentions. The difficulty in capturing genuine intentions, and the non-intrusive, non-interactive questioning approach might explain the null findings and raise questions for further research. We conclude with suggestions for a novel framework on semi-interactive information elicitation.
Although lie detection procedures have been fundamentally criticized since their inception at the beginning of the 20th century, they are still in use around the world. In addition, they have created some remarkable appeal in the context of counterterrorism policies. Thereby, the links between science and fiction in this topic are quite tight and by no means arbitrary: In the progressive narrative of the lie detection devices, there is a promise of changing society for the better, which is entangled in a fictional narrative provided by many cinematic and literary examples. By drawing on the concept of “diegetic technologies” formulated by David Kirby, we want to highlight the role of science fictional narratives in the historical development and current application of lie detection procedures. We therefore aim to use this conceptual frame to analyze the role of prevention-oriented mind reading procedures, which are developed for or already used in the “war on terror” with reference to their fictional predecessors. One virulent factor has to be highlighted in this context: The fictional engagement with possible new lie detection practices is in itself creating a legitimating ground for future security technologies.
In this paper we argue that there is little need for more of the traditional deception detection research in which observers assess short video clips in which there are few (if any) cues to deception and truth. We argue that a change in direction is needed and that researchers should focus on the questions the interviewer needs to ask in order to elicit and enhance cues to deception. We discuss three strands of research into this new ‘interviewing to detect deception’ approach. We encourage practitioners to use the proposed techniques and encourage other researchers to join us in conducting more research in this area. We offer some guidelines for what researchers need to keep in mind when carrying out research in this new paradigm.
Despite the commonsense belief that people do not confess to crimes they did not commit, 20 to 25% of all DNA exonerations involve innocent prisoners who confessed. After distinguishing between voluntary, compliant, and internalized false confessions, this article suggests that a sequence of three processes is responsible for false confessions and their adverse consequences. First, police sometimes target innocent people for interrogation because of erroneous judgments of truth and deception. Second, innocent people sometimes confess as a function of certain interrogation tactics, dispositional suspect vulnerabilities, and the phenomenology of innocence. Third, jurors fail to discount even those confessions they see as coerced. At present, researchers are seeking ways to improve the accuracy of confession evidence and its evaluation in the courtroom.
The topic of true and false intent has been more or less ignored within the field of legal psychology. This is remarkable considering the frequency and importance of situations calling for assessments of whether a person is lying or telling the truth about his or her intentions (e.g., when crossing a border). There were four aims to the present paper. The first was to outline a psycho-legal research program on true and false intentions. The second aim was to highlight some conceptual issues which might be relevant for planning and conducting research in this domain. The third aim was to offer some stepping stones which might assist researchers launching investigations on true and false intentions. The final aim was to briefly summarize the first round of empirical psycho-legal studies on true and false intentions.
This article reviews principles of effective training (Beebe, Mottet, & Roach, 2004), explicates the strengths and weaknesses of deception detection training research, and proposes a research agenda for deception detection training within the law enforcement context. Overall, deception detection training studies to-date fail to follow principles of effective training. Trainee needs are not assessed, training content is lacking, and the trainer and trainee engage in passive roles. A model is proposed with the aim of provid- ing a conceptual framework for thinking about deception detection training and estab- lishing a research agenda for future studies. In particular, it extends previously suggested changes to deception detection training research (Frank & Feeley, 2003), applies princi- ples of effective training methods (Beebe et al., 2004), and calls for a context-dependent (Flyvbjerg, 2001) shift in deception detection training studies. Law enforcement officers are confronted with making judgments of truth and deception on a daily basis. They decide on the veracity of claims about the whereabouts of suspects, who physically assaulted whom, if someone is using or selling drugs, among hundreds of other situations. The decisions law enforce- ment officers make regarding the truth of a suspect's statement have potentially tremendous consequences. Whether or not an of- ficer believes a victim's statement at a crime scene may dictate if
The study investigated experienced police officers' (N = 30) lie detection accuracy. Each police officer conducted an interrogation of a college student acting as a suspect either guilty or innocent of a mock crime and made a veracity judgment of the suspect. The police officers had the opportunity to conduct the interrogation in the manner of their own choice. The lie detection accuracy of these police officers was compared to that of police officers judg-ing videotaped versions of the interrogations. The police officers failed to detect deception better than chance. There was no difference in accuracy between police officers interrogating live and observing video. The interro-gators reported to rely on verbal content more when interrogating than when watching video. It seems as though police officers have a difficult time detect-ing deception not only in passive contexts but also in active ones.
In an experiment, nonverbal indicators of deception in police interrogations of mock crimes were examined. Both vocal and nonvocal behaviours were scored. Thirty participants were subjected to long interrogations (over 9 minutes) conducted by 30 experienced police officers, asked to interrogate as they normally do. Although the liars reported being significantly more nervous, and found the task more strenuous than the truth tellers, no differences in the nonverbal behaviours scored were found. In an analysis of the strategies employed, both truth tellers and liars were found to try to not make excess movements. The principal verbal content strategy for the liars was to keep the story simple, and for truth tellers to keep the story real. The reasons why, in this demanding situation, the truth tellers and liars could not be distinguished by their nonverbal behaviour are discussed.
In many legal proceedings, fact finders scrutinize the demeanor of a defendant or witness, particularly his or her nonverbal behavior, for indicators of deception. This meta-analysis investigated directly observable nonverbal correlates of deception as a function of different moderator variables. Although lay people and professionals alike assume that many nonverbal behaviors are displayed more frequently while lying, of 11 different behaviors observable in the head and body area, only 3 were reliably associated with deception. Nodding, foot and leg movements, and hand movements were negatively related to deception in the overall analyses weighted by sample size. Most people assume that nonverbal behaviors increase while lying; however, these behaviors decreased, whereas others showed no change. There was no evidence that people avoid eye contact while lying, although around the world, gaze aversion is deemed the most important signal of deception. Most effect sizes were found to be heterogeneous. Analyses of moderator variables revealed that many of the observed relationships varied as a function of content, motivation, preparation, sanctioning of the lie, experimental design, and operationalization. Existing theories cannot readily account for the heterogeneity in findings. Thus, practitioners are cautioned against using these indicators in assessing the truthfulness of oral reports. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Since 1988 with the Employee Polygraph Protection Act, the US federal government has become the primary user of polygraph tests. The government uses the polygraph in 3 applications: as an investigative tool, as a preemployment screening device, and as a periodic screening device. Research and analyses conducted on the Department of Defense's counterintelligence scope polygraph (CSP) screening program in 1986–1989 indicate that the polygraph tests used in that program were unable to discriminate truthtellers from deceivers. It appears that the CSP polygraph examinations correctly classified only about 2% of guilty Ss. There is little evidence to support the use of polygraph as a preemployment screening device. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
15 males (mean age 29.2 yrs) participated in an experiment to evaluate the polygraph in a real-life situation; 2 Ss had actually cheated on a test. All Ss went through a standard polygraph test using the control questions method. This procedure guarantees the objective identification of "liars" without jeopardizing the real-life appearance of the experimental situation. Each S was evaluated by 3 polygraphers: One had access to the polygraph charts only, one observed the S's behavior but not his charts, and a third had both kinds of information. The evaluations of all 3 polygraphers were compared with the criterion. Evaluations based on both behavior observation and physiological charts were superior to those based on either type of information alone. However, evaluations based on the physiological information alone were not superior to those based on the behavioral information alone. (13 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
People tend to overestimate their capacity to detect lying in others and to underestimate their own ability to tell lies. These biases were demonstrated in a sample of 60 police officers. In a lie-detection task, the officers evaluated their accuracy as high and were overconfident in their judgements. In fact, their performance was below chance level. Participants also received false feedback about their performance. When the feedback suggested that they had performed better than they thought, this further enhanced their perceived lie-detection capacity and also increased their belief in their ability to conceal their own lies. When the feedback suggested they had performed worse than they thought, their ratings of both lie detection and their lie-telling abilities were lowered. Results are discussed in terms of anchoring, availability, and the self-assessment bias. On a practical level, the tendency of police interrogators to overestimate their ability to detect deception could change suspicion into certainty and increase the risk of a false confession. Copyright © 2003 John Wiley & Sons, Ltd.
A study was conducted to examine Spanish police officers’ and nonofficers’ lie- and truth-detection accuracy, as well as their estimated detection ability. The participants were 121 police officers and 146 undergraduates who watched videotaped truthful and deceptive statements. They had to indicate: (1) whether each statement was truthful or deceptive, and (2) how good police officers were, in comparison with the general population, at detecting the truthfulness or deceptiveness of a statement. Results indicate that police officers’ accuracy was not higher than that of nonofficers, rather, while the officers reached an accuracy rate close to chance probability, the undergraduates surpassed that probability. Officers had a very strong tendency to judge the statements as deceptive; this made them less accurate than the students in judging the truthful accounts, while both groups reached a similar accuracy when judging the deceptive ones. Both occupational samples considered that the police are more capable of identifying truths and lies than the general population. However, this belief was stronger among the officers themselves than among the nonofficers. No significant correlation between estimated ability and accuracy was found for either sample. The results are explained in terms of the participants’ wrong beliefs about the cues to deceit and the socialization process that police officers undergo, which would increase their confidence and perceived ability while hindering their learning of the actual indicators of deceit. The need for officers to receive training is emphasized, and some directions are given on how this training should be carried out.
We hypothesised that the responses of pairs of liars would correspond less with each other than would responses of pairs of truth tellers, but only when the responses are given to unanticipated questions. Liars and truth tellers were interviewed individually about having had lunch together in a restaurant. The interviewer asked typical opening questions which we expected the liars to anticipate, followed by questions about spatial and/or temporal information which we expected suspects not to anticipate, and also a request to draw the layout of the restaurant. The results supported the hypothesis, and based on correspondence in responses to the unanticipated questions, up to 80% of liars and truth tellers could be correctly classified, particularly when assessing drawings.
Decades of research has shown that people are poor at detecting lies. Two explanations for this finding have been proposed. First, it has been suggested that lie detection is inaccurate because people rely on invalid cues when judging deception. Second, it has been suggested that lack of valid cues to deception limits accuracy. A series of 4 meta-analyses tested these hypotheses with the framework of Brunswik's (1952) lens model. Meta-Analysis 1 investigated perceived cues to deception by correlating 66 behavioral cues in 153 samples with deception judgments. People strongly associate deception with impressions of incompetence (r = .59) and ambivalence (r = .49). Contrary to self-reports, eye contact is only weakly correlated with deception judgments (r = -.15). Cues to perceived deception were then compared with cues to actual deception. The results show a substantial covariation between the 2 sets of cues (r = .59 in Meta-Analysis 2, r = .72 in Meta-Analysis 3). Finally, in Meta-Analysis 4, a lens model analysis revealed a very strong matching between behaviorally based predictions of deception and behaviorally based predictions of perceived deception. In conclusion, contrary to previous assumptions, people rarely rely on the wrong cues. Instead, limitations in lie detection accuracy are mainly attributable to weaknesses in behavioral cues to deception. The results suggest that intuitive notions about deception are more accurate than explicit knowledge and that lie detection is more readily improved by increasing behavioral differences between liars and truth tellers than by informing lie-catchers of valid cues to deception.
In 2 diary studies of lying, 77 college students reported telling 2 lies a day, and 70 community members told 1. Participants told more self-centered lies than other-oriented lies, except in dyads involving only women, in which other-oriented lies were as common as self-centered ones. Participants told relatively more self-centered lies to men and relatively more other-oriented lies to women. Consistent with the view of lying as an everyday social interaction process, participants said that they did not regard their lies as serious and did not plan them much or worry about being caught. Still, social interactions in which lies were told were less pleasant and less intimate than those in which no lies were told.
An information-gain approach to the analysis and interpretation of eyewitness identification data is described. The information-gain analysis is grounded in Bayesian statistics, permitting the important role of prior probabilities to be explored. This approach also forces a more complete treatment of the data and reveals important patterns that have escaped previous attention in the eyewitness identification literature. Particularly important is the ability of information-gain analyses to make salient the exonerating value of eyewitness behaviors rather than just their incriminating value. Analyses of sample data sets show how the exonerating value of filler identifications and "not there" responses can actually exceed the incriminating value of identifications of the suspect at certain points in the distribution of prior probabilities.
Detecting deception is an inherently difficult task, but one that plays a critical role for law enforcement investigators in the interrogation room. In general, research has failed to indicate that performance in this domain is improved by training or prior experience. A signal detection framework is applied to the paradigm to better conceptualize the influence of these two factors. We found that although neither factor influenced discrimination accuracy, there was an effect on response bias such that training and prior experience appeared to increase the likelihood of responding "deceit" as opposed to "truth." This "investigator bias" was observed both in a review of the literature and in this study of North American law enforcement investigators who took part in a forensically based deception-detection task. Possible theoretical mechanisms and practical implications of these findings are discussed.
Do people behave differently when they are lying compared with when they are telling the truth? The combined results of 1,338 estimates of 158 cues to deception are reported. Results show that in some ways, liars are less forthcoming than truth tellers, and they tell less compelling tales. They also make a more negative impression and are more tense. Their stories include fewer ordinary imperfections and unusual contents. However, many behaviors showed no discernible links, or only weak links, to deceit. Cues to deception were more pronounced when people were motivated to succeed, especially when the motivations were identity relevant rather than monetary or material. Cues to deception were also stronger when lies were about transgressions.
Ninety-nine police officers, not identified in previous research as belonging to groups that are superior in lie detection, attempted to detect truths and lies told by suspects during their videotaped police interviews. Accuracy rates were higher than those typically found in deception research and reached levels similar to those obtained by specialized lie detectors in previous research. Accuracy was positively correlated with perceived experience in interviewing suspects and with mentioning cues to detecting deceit that relate to a suspect's story. Accuracy was negatively correlated with popular stereotypical cues such as gaze aversion and fidgeting. As in previous research, accuracy and confidence were not significantly correlated, but the level of confidence was dependent on whether officers judged actual truths or actual lies and on the method by which confidence was measured.
Deception detection research has largely neglected an important aspect of many investigations, namely that there often exists evidence against a suspect. This study examined the potentials of timing of evidence disclosure as a deception detection tool. The main prediction was that observers (N = 116) would obtain higher accuracy rates if the evidence against the suspects (N = 58) was presented in a late rather than early stage of the interrogation. This prediction was based on the idea that late evidence disclosure would trigger lack of consistencies between the liars' stories and the evidence; this could be used as a cue to deception. The main prediction received support. Late disclosure observers obtained an overall accuracy of 61.7%, compared to 42.9% of Early disclosure observers. Deceptive statements were identified with high accuracy (67.6%) in Late disclosure, indicating that the technique in this form is beneficial mainly for pinpointing lies.
This Project was sponsored by a grant from the Economic and Social Research Council (RES-000-23–0292). In Experiment 1, we examined whether three interview styles used by the police, accusatory, information-gathering and behaviour analysis, reveal verbal cues to deceit, measured with the Criteria-Based Content Analysis (CBCA) and Reality Monitoring (RM) methods. A total of 120 mock suspects told the truth or lied about a staged event and were interviewed by a police officer employing one of these three interview styles. The results showed that accusatory interviews, which typically result in suspects making short denials, contained the fewest verbal cues to deceit. Moreover, RM distinguished between truth tellers and liars better than CBCA. Finally, manual RM coding resulted in more verbal cues to deception than automatic coding of the RM criteria utilising the Linguistic Inquiry and Word Count (LIWC) software programme.
In Experiment 2, we examined the effects of the three police interview styles on the ability to detect deception. Sixty-eight police officers watched some of the videotaped interviews of Experiment 1 and made veracity and confidence judgements. Accuracy scores did not differ between the three interview styles; however, watching accusatory interviews resulted in more false accusations (accusing truth tellers of lying) than watching information-gathering interviews. Furthermore, only in accusatory interviews, judgements of mendacity were associated with higher confidence. We discuss the possible danger of conducting accusatory interviews.
By questionnaire, 631 police investigators reported on their interrogation beliefs and practices-the first such survey ever conducted. Overall, participants estimated that they were 77% accurate at truth and lie detection, that 81% of suspects waive Miranda rights, that the mean length of interrogation is 1.6 hours, and that they elicit self-incriminating statements from 68% of suspects, 4.78% from innocents. Overall, 81% felt that interrogations should be recorded. As for self-reported usage of various interrogation tactics, the most common were to physically isolate suspects, identify contradictions in suspects' accounts, establish rapport, confront suspects with evidence of their guilt, and appeal to self-interests. Results were discussed for their consistency with prior research, policy implications, and methodological shortcomings.
In two experiments, we tested the hypotheses that (a) the difference between liars and truth tellers will be greater when interviewees report their stories in reverse order than in chronological order, and (b) instructing interviewees to recall their stories in reverse order will facilitate detecting deception. In Experiment 1, 80 mock suspects told the truth or lied about a staged event and did or did not report their stories in reverse order. The reverse order interviews contained many more cues to deceit than the control interviews. In Experiment 2, 55 police officers watched a selection of the videotaped interviews of Experiment 1 and made veracity judgements. Requesting suspects to convey their stories in reverse order improved police observers' ability to detect deception and did not result in a response bias.
The authors report a meta-analysis of individual differences in detecting deception, confining attention to occasions when people judge strangers' veracity in real-time with no special aids. The authors have developed a statistical technique to correct nominal individual differences for differences introduced by random measurement error. Although researchers have suggested that people differ in the ability to detect lies, psychometric analyses of 247 samples reveal that these ability differences are minute. In terms of the percentage of lies detected, measurement-corrected standard deviations in judge ability are less than 1%. In accuracy, judges range no more widely than would be expected by chance, and the best judges are no more accurate than a stochastic mechanism would produce. When judging deception, people differ less in ability than in the inclination to regard others' statements as truthful. People also differ from one another as lie- and truth-tellers. They vary in the detectability of their lies. Moreover, some people are more credible than others whether lying or truth-telling. Results reveal that the outcome of a deception judgment depends more on the liar's credibility than any other individual difference.
In recent years, DNA exoneration cases have shed light on the problem of false confessions and the wrongful convictions that result. Drawing on basic psychological principles and methods, an extensive body of research has focused on the psychology of confessions. This article describes the processes of interrogation by which police assess whether a suspect is lying or telling the truth and the techniques used to elicit confessions from those deemed deceptive. The problem of false confessions emphasizes personal and situational factors that put innocent people at risk in the interrogation room. Turning from the causes of false confessions to their consequences, research shows that confession evidence can bias juries, judges, lay witnesses, and forensic examiners. Finally, empirically based proposals for the reform of policy and practice include a call for the mandatory video recording of interrogations, blind testing in forensic crime labs, and use of confession experts in court.
People are generally poor at detecting deceit when observing someone’s behaviour or listening to their speech. In this chapter I will discuss the major factors (pitfalls) that lead to failures in catching liars: the sixteen reasons I will present are clustered into three categories: (i) a lack of motivation to detect lies; (ii) difficulties associated with lie detection; and (iii) common errors made by lie detectors. Discussing pitfalls provides insight into how lie detectors can improve their performance (for example, by recognising common biases and avoiding common judgment errors). The second section of this chapter discusses 11 ways (opportunities) to improve lie detection skills. Within this section, I first provide five recommendations for avoiding common errors in detecting lies. Next, I discuss recent lie detection research that introduces novel interview styles aimed at eliciting and enhancing verbal and nonverbal differences between liars and truth tellers. The recommendations are relevant in various settings, from the individual level (e.g., “Is my partner really working late?”) to the societal level (e.g., “Can we trust this suspect when he claims that he is not the serial rapist the police are searching for?”).
ifteen years ago at the beginning of a book chapter on training to detect deception (Bull, 1989) I noted that: An advertisement urging people to join a British police force recently appeared in a Sunday newspaper. Part of it stated that, ‘Most people speak the truth most of the time. When they lie they experience stress and it usually shows’. Accompanying the words in the advertisement were a number of photographs of people. One of these showed a man touching the side of his nose with the index finger of his left hand. The caption for this stated that, ‘The man with his finger to his nose is showing one of the signals associated with lying’. The advertisement continued by saying that, ‘After training, you'll register the particular things a person does when conversing normally. When a change of topic brings about significant changes in their actions you'll notice that too’. Another similar advertisement stated, ‘If you're interviewing a suspect, how do you know if he's telling the truth? You'll be taught the rudiments of body language, gesticulation and body movements that indicate stress and nervousness’. In closing that chapter I made the point that: Until a number of publications in refereed journals appear demonstrating that training enhances the detection of deception, it seems that some police recruitment advertisements and police training books are deceiving their readers.
Can you tell whether suspects are lying based on what they say, how they say it, how they sound when they say it, or how they look? Can you do so even without any special equipment, such as a polygraph to monitor physiological responses (Honts, this volume) or a camera to record interviews, which can then be studied in microscopic detail (O'Sullivan and Ekman, this volume)? These are the questions we will address in this chapter. Reading this chapter will not give anyone grounds for feeling smug about his or her deception detection prowess in forensic contexts (nor in any other contexts, for that matter). The study of verbal and non-verbal behavioural cues to deception is an inexact science, and probably always will be. We will offer hints, based on the current state of the science, about the kinds of behaviours that are more or less likely to intimate that a suspect may be lying. Our suggestions may be of some value in cases in which more definitive evidence, such as DNA, is not available or has not yet been uncovered. We hope our review will also prove useful in demonstrating why sweeping statements about perfect cues to deception are deserving of deep scepticism. There are two fundamental questions in the study of verbal and nonverbal cues to deception. First, are there any such cues? That is, are there any behaviours that indicate whether a person may be lying?
The role of on-the-job experience in fostering skill at detecting deception was examined. A deception-detection test was administered to three samples of more than 100 subjects each: a group of undergraduates with no special experiences at detecting deceit; a group of new recruits to a federal law enforcement training program, who had some limited on-the-job experience at detecting deceit; and a group of advanced federal law enforcement officers, with years of experience working at jobs in which the detection of deception is very important. Although the officer samples were more confident about their judgments of deceptiveness than were the students, they were no more accurate than the students. None of the three groups showed a significant improvement in deception-detection success from the first half to the second half of the test; however, the advanced officers felt increasingly confident about their performance as they progressed through the test. Correlational analyses of the relationship between accuracy and confidence provided further evidence that experience does not improve people's awareness oi the accuracy or inaccuracy of their judgments. The findings from this research are compared to the results of research on other kinds of professional decision-makers (e.g., clinical psychologists), and several theoretical perspectives on the role of experience in decision making are discussed.
All deception studies published to date have been laboratory studies. In such studies people lied only for the sake of the experiment, consequently the stakes were usually low. Although research has shown that most spontaneous lies told in real life are trivial, such studies tell us little about lies where the stakes are high (such as police/suspect interviews). In Study 1, we discuss the behaviour of an actual suspect while he was interviewed by the police in a murder case. Although the man initially denied knowing and killing the victim, substantial evidence obtained by the police showed that he was lying. On the basis of this evidence, the man confessed to killing the victim and was later convicted for murder. To our knowledge there has been no other study published that has analysed the behaviour of a liar in such a high-stake realistic setting. The analysis revealed several cues to deception. In Study 2, we exposed 65 police officers to six fragments (three truthful and three deceptive) of the interview with the murderer and asked them to indicate after each fragment whether the man was lying or not. The findings revealed that the participants were better at detecting truths (70% accuracy) than lies (57% accuracy). We also found individual differences among observers, with those holding popular stereotypical views on deceptive behaviour, such as ‘liars look away’ and ‘liars fidget’ performing least effectively as lie catchers. Copyright © 2001 John Wiley & Sons, Ltd.
The use of physiological measures to detect deception, known variously as the psychophysiological detection of deception (PDD), lie detection and the polygraph, has a history in psychological science that goes back to the end of the nineteenth century (Lombroso, 1895). Although applied interest in PDD was primarily an American phenomenon during most of the twentieth century, international application, research and interest has expanded rapidly in recent years (for reviews see Barland, 1988; Honts, Raskin and Kircher, 2002). PDD is applied in a variety of settings for a variety of purposes. In some jurisdictions in the United States the results of PDD tests are admissible as evidence in legal proceedings. In many countries law enforcement uses PDD as an investigative tool to check the veracity of suspects and informants. PDD can also be used as a pre-employment screening tool to verify a person's credibility as part of the vetting process. In the United States, PDD use in the national security system to screen for persons engaged in hostile actions against the government is ubiquitous. Finally, a recent phenomenon in the United States is to use the polygraph as part of the release programme for convicted sex offenders. As a condition of release, convicted sex offenders are required to take and pass periodic polygraph tests concerning new sex offences or other prohibited behaviour.
One of the most fascinating sub-divisions within the rapidly growing field of psychology and law is the area of deception detection. Traditionally this area has been characterised by a number of approaches which have analysed different aspects of deception such as verbal content, non-verbal behaviour, and polygraph testing. The last few years' intensive research has resulted in an impressive corpus of new knowledge about issues such as cross-cultural deception, the detection of simulated amnesia and false confessions, lie-catching expertise and how best to train professionals in detecting deception. This book provides a state-of-the-art account of current research and practice, written by an international team of experts and will be a valuable resource for academics, students, practitioners and all professionals within the legal domain who need to tackle questions of credibility and reliability.
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.
This volume provides a practical guide to interrogation, its effects and implications, based on scientific findings. The main theme is the 'how and why' of false confessions—which are shown to be more frequent than is generally believed. . . . This is the first book to raise the important issues of false confession from a scientific perspective, backed up by references to well-known contemporary cases (including the 'Guildford Four,' 'Birmingham Six' and 'Tottenham Three') in which the author has been involved.
This unique book considers, in depth, police interrogation methods and the processes by which confessions are elicited. Concerned with both theory and practice, the basic principles of interrogation are explained in detail, and supplemented by relevant psychological research data. Present practice is described from the author's extensive experience of investigating the reliability of confessions in actual cases.
Aimed primarily at clinical psychologists working in this field, "The Psychology of Interrogations, Confessions and Testimony" will also be invaluable reading for police officers, lawyers, judges, social workers, and academics and researchers in the field of psychology and the criminal justice system. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Research on the detection of deception, via non-verbal cues, has shown that people's ability to successfully discriminate between truth and deception is only slightly better than chance level. One of the reasons for these disappointing findings possibly lies in people's inappropriate beliefs regarding ‘lying behaviour’. A 64-item questionnaire originally used in Germany, which targets participants' beliefs regarding truthful and deceptive behaviour, was used. The present study differed from previous research in three ways: (i) instead of a student population, police officers and lay people were sampled, (ii) both people's beliefs regarding others' deceptive behaviour and their beliefs regarding their own deceptive behaviour were examined, and (iii) both non-verbal cues to, and content characteristics of, deceptive statements were examined. Results were consistent with previous studies, which found significant differences between people's beliefs regarding deceptive behaviour and experimental observations of actual deceptive behaviour. Further, police officers held as many false beliefs as did lay people and finally, participants were more accurate in their beliefs regarding their own deceptive behaviour than they were in their beliefs regarding others' behaviour.
In this paper we argue that the emerging concept of episodic future thought (EFT) is a good candidate for capturing the core mental processes at play when forming an intention. Furthermore, we argue that tapping essential EFT features can be helpful in understanding how statements on the forming of true and false intentions may differ. Specifically, we argue that the processes relevant for constructing a convincing cover-story (which a guilty suspect needs to mask his criminal intention) will be much less characterized by the typical EFT features. To test our assumption we devised a new experimental set-up accommodating the main characteristics of intent and allowing for episodic future thought. The combined empirical evidence strongly supports our assumption that EFT is a helpful concept for illuminating the differences that may occur when forming true and false intentions. Copyright © 2010 John Wiley & Sons, Ltd.
Can the science of deception detection help to catch terrorists? Sharon Weinberger takes a close look at the evidence for it.
We adapted and applied the Wells and Olson's (2002) Information Gain Analyses to examine the relative usefulness of a common psycho-physiological deception detection (PDD) technique, the Comparison Question Test, in forensic and screening settings as compared to unassisted lay and professional persons. We found that in forensic settings PDD provided substantial improvements in information gain over unassisted laypersons across nearly the complete range of the base rate of guilt. This was true for accuracy estimates based on laboratory and field data. At p(guilt) = 0.9, a benchmark set by critics of PDD, PDD provided 27 times the information gain of credibility decisions made by unassisted lay persons. Analyses of a screening PDD indicated that only deceptive outcomes provide useful information gain at relevant low base rates of guilt. These results strongly support the use of PDD in forensic settings and have implications for how screening PDD results are used.
Because of special characteristics of nonverbal behaviors (e.g., they can be difficult to suppress, they are more accessible to the people who observe them than to the people who produce them), the intention to produce a particular nonverbal expression for self-presentational purposes cannot always be successfully translated into the actual production of that expression. The literatures on people's skills at using their nonverbal behaviors to feign internal states and to deceive are reviewed as they pertain to the question of whether people can overcome the many constraints on the translation of their intentions into expressions. The issue of whether people's deliberate attempts to regulate their nonverbal behaviors can be detected by others is also considered.
A meta-analysis was conducted of research on the relation between judges' accuracy at detecting deception and their confidence in their judgments. A total of 18 independent samples revealed an average weighted accuracy-confidence correlation of .04, a relation not significantly different from zero. However, confidence was positively correlated with judges' tendency to perceive messages as truthful, regardless of the actual truthfulness of the messages. Judges were also more confident when they really were rating truths compared to when they were rating lies. Also, men were more confident than women, and judges who had a closer relationship to the message sender felt more confident in their judgments of truths and lies. Methodological and theoretical explanations for these findings are discussed.
Concern with crime and terrorism makes it increasingly important to be able to detect lying. Most lie detection tools used to date are arousal-based protocols. The majority of these protocols are based on the assumption that, because of their fear of being caught, liars will be more aroused when answering key relevant questions (‘Did you steal the money?’) than when answering comparison questions. According to the US National Research Council's well-documented report [1], however, this premise is theoretically weak. Liars do not necessarily reveal more signs of arousal when answering key questions. Conversely, truth tellers might be anxious and hence show signs of arousal when answering key questions.
Lying: Moral Choice in Public and Private Life Vintage Books Accuracy of deception judgments
- S F Bok
- Jr
- B M Depaulo
Bok, S., 1989. Lying: Moral Choice in Public and Private Life. Vintage Books, New York.
Bond, C.F., Jr. DePaulo, B.M., 2006. Accuracy of deception judgments. Personality and
Social Psychology Review 10, 214-234.
The screening of passengers by observation techniques programme: analysing the issues
- M Perry
- A Gibley
Perry, M., Gibley, A., 2011. The screening of passengers by observation techniques programme: analysing the issues. Aviation Security International 17, 12-13.
Social Cognition: From Brains to Culture
- S T Fiske
- S E Taylor
Fiske, S.T., Taylor, S.E., 2008. Social Cognition: From Brains to Culture, first ed.
McGraw-Hill Higher Education, Boston, MA.
Reading between the lies: identifying concealed and falsified emotions in universal facial expressions
- S Porter
- L Ten Brinke
Porter, S., ten Brinke, L., 2008. Reading between the lies: identifying concealed and falsified emotions in universal facial expressions. Psychological Science 19, 508-514.
Handbook of Polygraph Testing
- M Kleiner
Kleiner, M. (Ed.), 2002. Handbook of Polygraph Testing. Academic Press, San Diego, CA.
Searching passengers' faces for subtle cues to terror
- D Q Wilber
- E Nakashima
Wilber, D.Q., Nakashima, E., 2007. Searching passengers' faces for subtle cues to terror.
The Washington Post 19 September, D1.