We naturally detect lies all the time. It could be a dip in our partner’s voice alerting us to the fact that they’re concealing their emotions; a child’s eyes drifting back to the drawer containing a present they weren’t supposed to open; or an implausible story told by a colleague trying to explain why the company’s petty cash has gone missing.
Yet we also often fail to detect lies. Why? Researchers have been trying to answer this question for over a century and the liars are still slipping through our fingers. But the latest research may help shine a light on where we’ve been going wrong.
A notable recent study was conducted by Associate Prof Timothy Luke and colleagues at the University of Gothenburg. They looked at research published over the past five years by 50 international experts on lie detection to analyse how they go about telling when someone is lying.
But first they had to decide on exactly what a lie is. We might use the word ‘lie’ when referring to someone saying an outfit you’re not sure about looks good, or a partner you think is trying to hide an affair, or a murderer who claims to be innocent. But are these comparable? Surely some lies have greater significance than others? Luke prefers to separate ‘white’ lies from what he calls deception.
“The construct of deception is more complicated than a lot of people think,” he says. “There are many kinds of psychological processes that can underlie it. We’re not talking about the same thing. Even superficial things, such as the length and type of communication, matter.”
Whether you’re texting your lie or telling it straight to someone’s face, the core of deception is an intentional attempt to mislead another person, Luke says. But deciding what constitutes a lie is one thing; detecting it is quite another. Are there really any cues that reliably betray deception in others?
Can you spot a liar from their eyes alone?
One commonly held belief is that liars are reluctant to meet another person’s gaze. And yet, in the Gothenburg study, 82 per cent of experts agreed that liars are no more likely to avoid eye contact or look away than truth tellers.
“The empirical work on deception detection is massive,” says Pär-Anders Granhag, a professor of psychology at the University of Gothenburg and one of the study’s co-authors. “But the only single issue that a large majority of the experts agree upon is that gaze aversion is not a diagnostic cue for deception.”
Similarly, 70 per cent of the experts agreed that liars don’t seem more nervous than truth tellers. This might come as a surprise, as nervousness and gaze aversion are two of the four key behaviours supposedly displayed by liars.
The other traditional indicators are that a liar will continually shift posture or touch themselves more often, and will give an account that’s less plausible, logical or consistent than if they were telling the truth.
These beliefs also stand on shaky empirical ground. The researchers found that links between deception and fidgeting (body movement), the lengths of time it took for subjects to answer questions (response latency), and whether their accounts were consistent, made sense or were expressed easily (fluency) were not clear-cut. Some experts said that liars did these things more, some that they did them less, and others that there was no difference.
Read more:
Words matter
Prof Aldert Vrij, an expert in the psychology of deception at the University of Portsmouth, wasn’t involved with the Gothenburg survey, but says that the most prevalent misconception about deception is “the idea that non-verbal lie detection works.”
What he’s suggesting is that people who try to use non-verbal lie-detection methods should proceed with caution, even if those methods involve technology such as polygraphs, video analysis, taking brain ‘fingerprints’ using neuroimaging machines, or looking for changes in vocal pitch – all of which are controversial areas of deception-detection research.
So are there any effective methods for spotting a liar? According to Luke, one cue is promising: a lack of detail. Some 72 per cent of experts agreed that liars provide fewer details than truth tellers.
Vrij agrees, saying that instead of examining how people behave, we should examine what they say. He says that there are several verbal indicators, including the number of details and the ‘complications’ that appear in a subject’s statement.
Complications are things that go wrong or are unexpected. They add clusters of details that make a story more convoluted – for example, saying that you initially didn’t see someone you were due to meet because they were waiting at a different entrance from the one you expected them at.
Vrij also points out another tell. “Statement-evidence inconsistency is another cue,” he says. “A liar’s statements are less consistent with the available evidence than statements from truth tellers.”
Granhag agrees: “There are no reliable non-verbal cues, but there are reliable verbal cues,” he says.
“If there’s an inconsistency between what a person tells you and the facts that you hold, there is a high likelihood that the person is trying to deceive you,” he adds.
For example, if you have footage of someone committing a crime, but that person says that they didn’t do it, it’s quite likely that they are lying to you. This seems so obvious that it almost doesn’t need to be said. However, the benefit of making this explicit is that it steers an investigator away from guessing whether a potential culprit is lying based on how they’re behaving, and forces them to look at the available facts instead.
How to challenge discrepancies
Turning this into advice for those who need to separate lies from truth, including detectives, Luke and Granhag have proposed a ‘Shift-of-Strategy’ approach to gather information that suspects are intentionally concealing.
It involves drip-feeding evidence to a suspected liar to challenge discrepancies in their story without directly accusing them of lying. In practice, this involves asking someone what happened, then presenting them with evidence that contradicts their statement, and seeing how they try to accommodate it.
“If a person changes their story when you present parts of the background information that you hold, you’re on your way to catching a lie,” says Granhag.
This method isn’t perfect. Investigators who use it need to be conscious that what seems like a lie can sometimes be down to a simple error of memory, especially if the suspect is asked about an event that happened long ago. Differentiating between an intentional fabrication and an unintentional one (or a false memory) is often very difficult.
Despite the problems associated with purported behavioural tells, such as gaze aversion, Vrij says that many practitioners are reluctant to swap those for more useful cues based on what a suspect is saying. Old myths and methods die slowly.
“Most annoying is the assumption that comes from TV shows… that lead the general public [and] professionals, to think they can catch an individual liar,” says Prof Amina Memon of the University of London, a leading researcher on lie detection and investigative interviewing, and another of the Gothenburg study’s co-authors.
Police following a hunch about a suspect based on the stereotypical profile of a liar may use coercive tactics that cause innocent people to confess to crimes they didn’t commit. Because of this, Memon advocates for a neutral, fact-finding approach to interviewing, rather than trying to guess whether someone is lying.
But there’s a bigger issue lurking behind all of this. Perhaps the reason why we haven’t found universal cues to deception is because they simply don’t exist.
For the past century, researchers have almost exclusively taken what’s known as a nomothetic approach. This means that they’re looking for the ‘laws’ of deception – cues that are exhibited by everyone. But maybe this kind of one-size-fits-all approach hasn’t worked simply because everyone lies differently.
A poker player applies this logic when they look for another player’s ‘tells’ – behaviours that indicate whether that person is bluffing. Tells are unique to individuals, so one person might scratch their nose when they have a bad hand, another might cough more, while yet another looks more calm than usual.
If you throw these three people into a research setting, a nomothetic approach won’t get you far. These differences will simply look like noise.
Unique signs of lying
If we want to understand the cues, researchers need to adopt an ‘ideographic’ approach and focus on what makes each individual unique, argues Luke. This would involve creating a personal profile of how each person lies about the same kinds of things, and in similar settings.
“Testing the same person under varying conditions (so-called ‘repeated measures’ experimental design) is the way to go,” says Memon.
One example of this approach was published in a 2022 paper by Dr Sophie van der Zee and co-authors, who developed the first deception model specifically tailored to an individual.
Using a fact-checked database of tweets by Donald Trump while he was president, they found that the language he used when he lied was systematically different from his truthful tweets. Once they made a personalised profile, the scientists could predict whether his tweets were untrue with an accuracy of 74 per cent.
This kind of personalised deception-detection model can work for those who already have a large online presence in which they lie a lot. Artificial intelligence can help collate and examine these existing data. But what about people who are less present online, or who don’t lie in posts?
Some things you can fact-check, but most everyday posts and messages are so personal that it’s hard to even identify them as lies, so even AI models may struggle.
“There’s no guarantee that a machine-learning model is going to actually work in ... situations where you don’t know the right answer” Luke says.
Exactly how researchers are going to overcome logistical barriers remains to be seen, but it seems clear that a shift in the science of lie detection is underway. It’s time to move away from what Luke calls “crude averages”. “People are a little too fascinated by having a cool trick to catch someone in a lie,” he says.
The crux is that researchers studying deception have repeatedly found that evidence from controlled environments shows most people are bad at detecting lies. The liars can evade detection partly because they also know the stereotypes and play into them.
Our confirmation bias can also make us overconfident: we disproportionately remember the times when we caught liars and don’t realise all the other times when we didn’t.
In instances when we succeed, too, Luke isn’t convinced that the cues we think we employed really are the keys we used to unlock the truth.
“Think about the last time that you caught someone in a lie. How did you know?” he asks. “It’s probably not because they looked up and to the left. You probably had some evidence: a receipt, a text message, a witness. These are the ways people tend to actually tell whether someone is providing the truth.”
Even when you don’t have concrete external evidence, you may be able to assess situational factors. “In the real world, you often have some understanding of why someone would be motivated to lie,” says Luke.
The reason you’re better able to guess when someone you know is lying from subtle cues, such as glances, is because you know them. In these situations, it’s best to read the situation more than the person, says Luke, and try to figure out their motivations.
The take-home message is that, though behavioural cues for deception may exist, they’re likely to be highly personal. “It’s better to trust your own detective work and check what people say against evidence,” says Luke.
Stereotypical cues are not going to cut it – in fact, they might actually make you worse at catching a liar. And if you can’t find any evidence? Luke’s advice is simple: “proceed with caution.”
Read more:
