How people try (but fail) to spot a lie

Can you tell when someone is lying? What clues are you looking for that might give them away? At this year’s SpotOn event in London, Chris Street presented an interactive demonstration of just how difficult it is to spot lies and how the obvious visual clues we might associate with lying, simply don’t exist.

What a liar looks like

Before we start thinking about ways of detecting lies, we need to know a little about how the people around us actually lie. Researchers have given people diaries and asked them to record the lies and truths that they tell over a 24-hour period. What has been found across studies is that people tell the truth far more often than they lie.

It makes sense – the general purpose of language is to communicate what is in our head to another person, and a truthful account of what we are thinking does that best. Sometimes we fudge the details or outright lie, but that’s the exception, not the rule. We need to keep that in mind.

The differences between liars and truth-tellers are incredibly small and rarely seen.

How about the behavior of a liar? When people think about liars, they tend to begin listing behaviors that might give a liar away – visible behaviors. But Pinocchio’s nose does not exist. We are simply too good at lying. Yes, even you. If (or when) you lie, you don’t hide your mouth or avoid looking that person in the eye.

Tales are a little less convincing, perhaps, but out of 100 lies you can expect around three to seven of them to be less convincing than a truth. The differences between liars and truth-tellers are incredibly small and rarely seen.

In fact, in our research we have found that people are quite skilled liars. In one study, participants believed they were taking part in research on how people’s bodies sway while solving math problems. This premise allowed us to put them into a motion capture suit and record their behavior.

Participants believed they were taking part in a study on how people’s bodies sway while solving math problems.
Chris Street

They were greeted by a rather mean and strict researcher (yours truly) and a friendly and cheerful research assistant. While the researcher was out of the room, the research assistant ‘accidentally’ knocked a laptop onto the floor, breaking it. When the researcher returned, he asked the participant whether they saw what had happened.

About half of the participants lied to protect the friendly research assistant from the potential wrath of the mean researcher (I’m quite a nice person really, I swear!). The other half told the truth and told us that it was broken accidentally by the assistant.

What we found initially is that liars moved less – not in any particular part of their body, but overall. Interestingly though, they did not appear to be ‘rabbits in headlights’. Although they moved less, the pattern of the behavior was less patterned and rhythmic compared to truth-tellers, who seemed to get into a repetitive pattern of movement.

It suggests that liars were controlling how they displayed their behavior to meet the demands of the task on the fly, while truth-tellers produced a more habitual and unconsidered set of behaviors.

If you are starting to think that spotting a lie is not going to be an easy task, you would be right.

Trying to spot a lie

If there is practically no information available, what do we do when have to decide if someone is lying or not? It turns out that we are highly inaccurate, with accuracy rates only slightly better than chance. Despite our poor performance, the Adaptive Lie Detector theory, or ALIED, argues that our strategies are informed and ‘smart’.

It begins by claiming we look for clues that are directly and causally related to the statement being made by the speaker, which are called ‘individuating cues’. For instance, if I claim to have been to Wales last week, an individuating cue could be the verbal or nonverbal behavior I display, but it could also be a claim from a third party who corroborates my story, or any other information that directly relates to this statement.

Click here to display content from Twitter.
Learn more in Twitter’s privacy policy.

Unfortunately, these cues are rarely present – remember that people are good liars and give little away. To compensate, ALIED argues that people use information that generalizes across statements, called ‘context-general information’. As an example, you may recall that people tend to tell the truth more often than they lie. This does not tell you that my specific statement was the truth, but it gives an indication of how likely a truth is to be observed in general. This allows us to fill in for the lack of more reliable individuating cues and to make an ‘educated guess’. It is this flexibility in the information being used and the functionality of adapting to the situation that gives the theory its name.

ALIED takes the position that people are not necessarily error-prone, then, but rather are using sensible strategies. The low accuracy results from the lack of reliable clues available in the environment. If we want to increase lie detection accuracy, we need to develop methods that generate these individuating cues. That may mean encouraging speakers to include details that can be verified with documents (e.g., submitting receipts when claiming expenses).

View the latest posts on the Research in progress blog homepage