Research round-up 1: Catching liars

I know I’ve really neglected this blog over the past few months (pressure of work and a doctorate to finish). Over the next few posts I’ll share with you all the articles and stories I hoped I’d have time to comment on this year but just didn’t. I’d like to promise to be better in 2009, but for the first half at least I’m going to struggle. Hang in there, eventually I’ll get back to being a better blogger!

The second post in this series deals with research and commentary on new technologies, like fMRI, for lie detection. But first I round up some recent research on deception detection methods which don’t require a $1 million giant magnet or wiring your subject up to a polygraph or brain scanner.

Who can catch a liar?

Let’s start with a bit of back and forth that began with an article by Charles Bond and Bella Depaulo which appeared in Psychological Bulletin this year. Bond and Depaulo’s analysis suggested that individual differences in lie detection ability are vanishingly small. Accuracy in lie detection, argue Bond and Depaulo, is more to do with how good the liar is at lying than individuals are at detecting deceit.

The authors report a meta-analysis of individual differences in detecting deception… 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.

This is a direct challenge to, in particular, the work of psychologists Maureen O’Sullivan and Paul Ekman, who have been investigating people they claim are extraordinarily accurate at deception detection – people they have dubbed ‘wizards’ of deception detection. So here comes O’Sullivan, right back at Bond and Depaulo:

…[Bond and Depaulo's] conclusions are based principally on studies with college students as lie detectors and lie scenarios of dubious ecological validity. When motivated professional groups have been shown either high stakes lie scenarios or scenarios involving appropriate liars and truth-tellers, average accuracies significantly above chance have been found for 7 different professional groups reported by 12 researchers in 3 countries. The replicated and predicted performance of extremely accurate individual lie detectors (“truth wizards”) also undermines the claim of no individual differences in lie detection accuracy…

Therese Pigott and Meng-Jia Wu also weigh in highlighting some methodological problems with Bond and Depaulo’s novel meta-analytic technique:

…[Bond and Depaulo] have presented a creative solution to the problem of estimating the standard deviation of deception judgments in the literature. Their article raises methodological questions about how to synthesize a measure of variation across studies. Although the standard deviation presents a number of problems as an effect size measure, more methodological research is needed to address directly the question raised by Bond and DePaulo (p.500).

Of course Bond and Depaulo get a right to reply. They repeat their analyses using the suggestions made by Piggott and Wu, and come to the same conclusion. They also take on O’Sullivan’s criticism and analyse data on experience to see what differences exist between college students and judges with ‘professional experience’. They conclude: “In moderator analyses, we looked separately at inexperienced and experienced judges. The individual differences in lie-detection accuracy were actually smaller among experienced judges than inexperienced ones” (p. 503).

Some empirical research that appears to support Bond and Depaulo’s claim was published online earlier this year, in advance of print in Law and Human Behaviour:

We examined whether individuals’ ability to detect deception remained stable over time. In two sessions, held one week apart, university students viewed video clips of individuals and attempted to differentiate between the lie-tellers and truth-tellers. Overall, participants had difficulty detecting all types of deception. When viewing children answering yes-no questions about a transgression (Experiments 1 and 5), participants’ performance was highly reliable. However, rating adults who provided truthful or fabricated accounts did not produce a significant alternate forms correlation (Experiment 2). This lack of reliability was not due to the types of deceivers (i.e., children versus adults) or interviews (i.e., closed-ended questions versus extended accounts) (Experiment 3). Finally, the type of deceptive scenario (naturalistic vs. experimentally-manipulated) could not account for differences in reliability (Experiment 4). Theoretical and legal implications are discussed.

But just in case you think it’s all over for the ‘wizards’, here’s a study from Gary Bond which suggests it can’t be dismissed quite yet. Out of more than 200 participants (law enforcement officers and college students) G. Bond discovered eleven (all LEOs) who could detect deception at greater than 80% chance, and from these, two potential ‘wizards’ who could maintain this accuracy rate over time. All of which indicates that there may indeed be ‘wizards’, but that they are (as O’Sullican has always recognised) very rare.

…Two experiments sought to (a) identify expert(s) in detection and assess them twice with four tests, and (b) study their detection behavior using eye tracking. Paroled felons produced videotaped statements that were presented to students and law enforcement personnel. Two experts were identified, both female Native American BIA correctional officers. Experts were over 80% accurate in the first assessment, and scored at 90% accuracy in the second assessment. In Signal Detection analyses, experts showed high discrimination, and did not evidence biased responding. They exploited nonverbal cues to make fast, accurate decisions. These highly-accurate individuals can be characterized as experts in deception detection.

How do you catch a liar?

What about the rest of us who aren’t ‘wizards’? Another discussion piece, which appeared earlier this year in a special issue of Criminal Justice and Behaviour focusing on scientific and psuedoscientific practices in law enforcement, was by Aldert Vrij who issued a challenge to law enforcement professionals to focus on verbal rather than non-verbal cues to lie detection:

…deception research has revealed that many verbal cues are more diagnostic cues to deceit than nonverbal cues. Paying attention to nonverbal cues results in being less accurate in truth/lie discrimination, particularly when only visual nonverbal cues are taken into account. Also, paying attention to visual nonverbal cues leads to a stronger lie bias (i.e., indicating that someone is lying). The author recommends a change in police practice and argues that for lie detection purposes it may be better to listen carefully to what suspects say.

Signs of lying

Continuing with their programme of research on the ‘cognitive load’ hypothesis Vrij, Sharon Leal and their colleagues published some psychophysiological evidence that lying does indeed increase mental effort, and that this increased effort can be detected by studying blink rates and skin conductance:

Previous research has shown that suspects in real-life interviews do not display stereotypical signs of nervous behaviours, even though they may be experiencing high detection anxiety. We hypothesised that these suspects may have experienced cognitive load when lying and that this cognitive load reduced their tonic arousal, which suppressed signs of nervousness. We conducted two experiments to test this hypothesis. Tonic electrodermal arousal and blink rate were examined during task-induced (Experiment 1) and deception-induced cognitive load (Experiment 2). Both increased cognitive difficulty and deception resulted in decreased tonic arousal and blinking. This demonstrated for the first time that when lying results in heightened levels of cognitive load, signs of nervousness are decreased. We discuss implications for detecting deception and more wide-ranging phenomena related to emotional behaviour.

And finally in this round-up, an article published in Psychology of Aging this year which tested older adults’ ability to detect deceit and whether any impairments were related to a lesser ability to recognise facial emotion expressed by the lie-teller.

Facial expressions of emotion are key cues to deceit (M. G. Frank & P. Ekman, 1997). Given that the literature on aging has shown an age-related decline in decoding emotions, we investigated (a) whether there are age differences in deceit detection and (b) if so, whether they are related to impairments in emotion recognition. Young and older adults (N = 364) were presented with 20 interviews (crime and opinion topics) and asked to decide whether each interview subject was lying or telling the truth. There were 3 presentation conditions: visual, audio, or audiovisual. In older adults, reduced emotion recognition was related to poor deceit detection in the visual condition for crime interviews only.

Next round-up will cover recent research on new technologies for lie detection!