Editorial: “The truth about lie detection“
Can you tell when someone’s lying? Dead giveaways are scarcer than you might think, but some well-chosen mind games can turn you into a better lie detector
HOW DO you spot a liar? The truth is, it’s very difficult, even for those who work in law enforcement and security. Contrary to popular belief, liars don’t readily give themselves away with their facial expressions or body language and most of us are easily duped by a determined liar. But is it possible to improve our lie-detection abilities?
Advertisement
If you ask the man or woman in the street how they know someone is deceiving them, they will probably tell you that a liar cannot look you in the eye. This was the most popular answer given in 50 countries in a 2006 survey. Unfortunately, it’s not true: when engaged in conversation, liars and truth-tellers avert their gaze more or less equally. Other popular stereotypes about liars turn out to be just as unreliable. Liars don’t necessarily fidget a lot, shift their weight, scratch or hum and haw. So perhaps we shouldn’t be surprised that in lab studies of lie detection most people only perform fractionally better than by chance. “People typically believe that they are poor liars and good lie detectors,” says psychologist Aldert Vrij from the University of Portsmouth, UK. “In fact it is the other way around.”
For most of us, most of the time, poor lie detection doesn’t matter too much, because the majority of lies are trivial – false compliments, face-savers and the like. Occasionally, the stakes are higher: Did that man steal my purse? Is my wife having an affair? Does my boss secretly want to get rid of me? And sometimes lying can have devastating consequences. Take the Al-Qaida double agent who infiltrated an American camp in Afghanistan in 2009, blowing up seven intelligence officers. For law enforcers and security specialists, in particular, spotting a lie can be a matter of life or death – and the shocking truth is that they are no better at it than the rest of us.
Why we are so easily duped has been the subject of much psychological research in recent years. Now Vrij and fellow psychologists Pär Anders Granhag of the University of Gothenburg, Sweden, and Stephen Porter of the University of British Columbia, Canada, have brought the findings together (). Their paper is not merely a catalogue of ineptitude, they argue that we can all become better lie detectors. For professional interrogators their insights should prove invaluable, but the rest of us can pick up some handy tips too.
The psychologists begin by highlighting common errors that people make when trying to decide if someone is lying. At the heart of the problem is the widespread belief that liars give themselves away by their facial expressions and body language. Yet when Bella DePaulo of the University of California, Santa Barbara, and colleagues pulled together the findings of more than 100 studies looking for cues to deception, they found none that consistently stood out (). “There are no behaviours that always occur when people are lying and never occur when they are telling the truth,” says DePaulo. So non-verbal cues – the averted gaze, fidgeting and so on – can be misleading.
The idea that liars leak information about their true emotional state through so-called “micro-expressions” is not very helpful either. American psychologist Paul Ekman coined this term in the mid-1980s to describe emotional expressions that involve the whole face and are suppressed between 0.04 and 0.2 seconds later. It’s not that liars don’t exhibit these telltale expressions – in fact, Porter’s group found that a liar’s micro-expression of disgust, sadness or happiness can last up to a second. “But it’s usually more subtle, in the upper or the lower face and sometimes only on one side,” he says.
The trick is to know what you are looking for, and not to attach too much importance to it when you find it.
In practice, we tend to pay more attention to non-verbal than to verbal information, and if the two appear to conflict, we favour the former (). Porter’s and Vrij’s groups have both reported that people are best at detecting lies when they give the two channels equal weight.
Part of the problem with focusing on facial expression and body language is that it assumes only liars act nervously under questioning. In fact, truth-tellers can be just as anxious. Simply being grilled alters the way a person acts, whether or not they are lying. So, if possible, you should compare your suspect’s behaviour under interrogation with a similar situation in the past when you know they were telling the truth. In other words, establish a baseline of honest behaviour against which to judge a suspect.
When trained to take such steps, people do get better at spotting liars. But the improvement is not overwhelming – accuracy increases from around 50 per cent, or just above chance, to between 60 and 70 per cent.
The big problem is that detecting deceit is a cat-and-mouse game in which liars always have the upper hand. Every time someone tells a lie, they get feedback on how well they are doing, enabling them to refine their performance. The other person, by contrast, may never find out whether they were being duped or not, so rarely gets a chance to learn from their mistakes.
Vrij and his colleagues argue that we need to adopt a whole new strategy to substantially improve our lie-detecting ability, not just put less emphasis on interpreting body language.
Over the past decade, researchers have come to realise that liars stand out more because of the way they think than because of the way they act. Lying is more cognitively demanding than telling the truth, so the alternative strategy aims to exacerbate cognitive differences between liars and non-liars. “It’s almost impossible to ask questions that will make the liar more nervous than the truth-teller,” says Vrij, “whereas it is possible to ask questions that are more difficult for a liar to answer.”
“Liars stand out more because of the way they think than because of the way they act”
One way to do this is to ask your suspect to tell their story in reverse. That should prove harder if they are hawking a fictional scenario rather than retelling events they lived through and can remember. Another is to ask for a drawing of the scene in question. Liars find this harder, since they rarely rehearse the visuospatial details of an invented scene. Or you could draw the person out on details of timing. A common trick is to invent a false alibi by transposing an event that really happened to the time when the incident occurred. By asking questions on the exact chronology of the supposed event, you may get details that expose the truth.
Research by Vrij’s group suggests a particularly effective line of questioning to find out a person’s true opinion – what a suspected Al-Qaida double agent really thinks about the US, for example. First you ask them to argue in favour of their professed position, then to play the devil’s advocate and argue against it. The giveaway is that people are usually able to present more and stronger arguments to support their real opinion than one they have fabricated ().
Another intriguing tactic turns the gaze-aversion myth on its head. Though liars don’t look away excessively, anyone trying to recall an event will do so from time to time, because locking gazes is distracting. By asking a suspected liar to maintain eye contact, you can increase the cognitive load. Since this is already higher for a liar, the chances are that they will betray themselves. In a study published last year, Vrij’s group found that liars told to maintain eye contact displayed more verbal and non-verbal cues to deceit than those telling the truth, making it easier for observers to identify them.
Then there’s the strategic use of evidence, SUE for short, which exploits the fact that liars possess different information from non-liars. Police interrogators already use SUE, and a high-profile murder case in Sweden in 2009 provides an example of how it works. After a young woman called Nancy Tavsan was killed in Gothenburg, the police arrested a suspect nicknamed Skuggan (“Shadow” in Swedish). They had some information about his activities on the night of the murder, but nothing to link him with the murder weapon, a glass bottle. Instead of revealing everything they knew immediately, they waited for Skuggan to dissemble during questioning and then confronted him with their evidence.
Gradually, he learned that the police always have more information than they are willing to reveal. When the interrogators finally brought up the subject of the bottle, Skuggan tried to play them at their own game, offering details they didn’t have, and that he couldn’t have known if he were innocent.
SUE counteracts our natural inclination to immediately confront a suspect with the evidence we have. And we can all benefit from it. In a mock scenario involving a stolen wallet, psychologist Maria Hartwig of the John Jay College of Criminal Justice in New York and colleagues found that volunteers who had received just 3 hours of training in SUE obtained an impressive 85 per cent accuracy rate in detecting lying, compared with 56 per cent for interrogators without training (Law and Human Behavior, vol 30, p 603).
Results like this have led Vrij, Granhag and Porter to urge would-be interrogators to actively exploit the psychology of liars rather than just passively look for cues to deception in body language and facial expressions.
This approach is endorsed by psychologist Tim Levine of Michigan State University in East Lansing, whose work has contributed to the change in emphasis. “What makes the expert [lie detector] an expert isn’t the ability to watch non-verbal behaviour,” he says, “it’s the ability to ask the right question.”
Other psychologists are not convinced that we should downplay the role of emotional cues. Ekman points out that in most studies volunteers are just play-acting and can’t replicate the motivation felt by a person telling a high-stakes lie in real life. It is the threat of serious punishment that determines the emotional and cognitive burden on a liar, he says, and without it behavioural “hotspots” or cues to deception vanish. Ekman is trying to overcome this problem by introducing more jeopardy into his lab-based scenarios.
Soon-to-be-published research by Porter’s group cleverly sidesteps the shortcomings of volunteer liars. They used “family pleader” videos – televised pleas for information made by individuals whose relative or spouse has gone missing – where there was reliable, independent information about whether the pleader was responsible for the disappearance of the missing person or not. When people trained to look for both verbal and non-verbal cues scrutinised around 80 such videos, they managed to distinguish the liars from the truth-tellers with more than 90 per cent accuracy. Among the giveaways was the tendency for liars to use a tentative word such as “if” where a truth-teller would use a more concrete one such as “when”. Liars also displayed brief expressions of surprise on their upper face – which is what happens when a person tries and fails to look distressed.
Admittedly, the analysts were able to view the videos at a resolution of 30 frames per second, a level of scrutiny not available in most situations. Still, the impressive detection rates show that when the stakes are high liars expose themselves, and that anyone can spot the signs of deceit if they just know what to look for. “Traditionally liars had the advantage,” Porter says. “Now we’re moving towards a situation where the lie detector will have the advantage.”
The interrogator’s little helper
Most people are hopeless at detecting liars, so it would be handy if a machine could help out.
The first mechanised lie detector, the polygraph, was invented a century ago. It measures a range of physiological indicators while a person is being questioned, exploiting the notion that a liar is more likely to be nervous than a truth-teller. Hence it looks for signs of stress, including an accelerated pulse and changes in skin conductivity due to increased sweating.
In recent years, using the same logic, video-based eye-trackers have been used to measure pupil dilation, and speech analysers to detect signs of stress in the voice.
Tomorrow’s lie detectors will take the nervous-liar logic a step further. The FAST programme being developed by the US government’s Department of Homeland Security aims to detect signs of future “malintent” in airline passengers, using non-invasive sensors such as cameras and remote heart-rate detectors.
It sounds promising, but all these technologies are open to the same criticism, namely that the nervous-liar logic is flawed. Civil-rights activists who oppose FAST, for example, point out that truthful people can be just as stressed as liars, especially when they have no alibi or risk going to prison if they are not believed. Psychologist Mark Frank of the State University of New York at Buffalo, a consultant on FAST, does not deny this but he defends the technique saying that it will be used merely as a screening tool and not to determine malintent per se.
Prospects do not look good for anyone seeking an alternative technological fix for the problem of lie detection. Another approach is based on the principle that lying is associated with a different pattern of brain activity from telling the truth. Evidence is lacking that a particular brain region or network is consistently associated with deception, but that hasn’t stopped several US companies from offering lie detection services based on functional MRI scans of the brain.
American psychologist Paul Ekman is sceptical that neuroscientists will ever discover a “lie area” in the brain, although he applauds the hunt. “I’m willing to mortgage my house [to bet]there isn’t one,” he says.
Perhaps it’s just as well that psychologists are now finding ways to improve the abilities of human lie detectors.
