How to spot lies

When talking to anyone but Pinocchio most people trust their instincts and unconscious knowledge of body language to determine whether someone is lying.

Chris Street, University of Huddersfield, and Daniel Richardson, University College London, have challenged this assumption in a breakthrough study. Rather than looking at the whole package, people are actually better off focusing consciously on a single behavior that can show deception, such as whether or not a person is thinking hard or if they appear tense, to determine if they are lying.

ResearchGate: How do people fair as lie detectors?

Chris Street: People make for poor lie detectors. Two researchers in 2006 looked at past studies and found that accuracy is quite reliably around 54 percent - where chance is 50 percent. So people are only slightly better than flipping a coin. It turns out that people are better at spotting truths than they are at spotting lies. People are no better than chance when judging lies, but hit around 61 percent accuracy when judging truthful statements. One likely reason for the difference in truth versus lie detection is that people have a bias to judge people as telling the truth, known as the “truth bias.”

RG: Can you explain what your study was testing?

Street: The “indirect lie detection” method has been used for over 30 years to show that people have unconscious knowledge about deception that they can’t seem to access consciously. The indirect method does not tell participants about deception. Instead, they judge whether, say, the speaker is thinking hard or not. If we were to put all the “not thinking hard” people into one pile and the “thinking hard” people into another, we would find more liars in the “thinking hard pile”. That is, this rating seems to indirectly separate out liars and truth-tellers. What is particularly interesting is that this type of indirect judgment is often more accurate than making a direct lie or truth judgment.

However, this task is unlikely to tell us anything about unconscious thinking. All it requires is that I, as the experimenter, pick a clue that I know is related to deception – such as thinking hard. I ask you to separate out speakers who are or are not thinking hard. Because I already know that clue is indicative of deception, then provided you can actually spot when people are thinking hard, you will, without any conscious or unconscious knowledge, separate out liars and truth-tellers.

RG: Why can’t people detect liars with accuracy when asked to do so directly as opposed to indirectly?

Street: In the indirect method, people focus only on a single cue – for example, thinking hard. When making a lie or truth judgment directly, people use a variety of behaviors – for example, “Is this plausible? Does the speaker look nervous?” and so on. These behaviors can all conflict - for instance, I might tell you a plausible tale (which sounds truthful) but I look super nervous (which suggests I’m lying). This conflict causes confusion – should you judge me to be lying or telling the truth? The focal account that we proposed argues that this confusion reduces accuracy. By only focusing on a single cue that we know is useful in distinguishing liars and truth-tellers, we remove this potential for conflict.

So the focusing effect can explain why indirect judgments boost accuracy, and we now know that this focusing happens even when the judgment is being made consciously.

RG: In your study you had a really interesting way of gathering the lies that people could watch. Can explain it to us? How did you think of it?

Street: We were looking for a task that would get people to lie or tell the truth. A lie can be thought of as a statement that you know to be false, and that you intentionally try to make me believe is true. The intention to mislead me is quite important. If we brought people into the laboratory and said “lie to me now”, then the speaker will be aware that I know they are lying - after all, I asked them to lie. So can they really try to convince me that they are lying, when I know what they are about to say is false? Probably not.

There is also another problem with asking them to lie: the social repercussions around lying are no longer present. If you caught me in a real-life lie, then there is some social embarrassment and perhaps other consequences. So we wanted to collect a set of lies that were natural, spontaneously generated by them, and to lie to someone who they believed were unaware that they were being lied to. To do this, we came up with a task that involved lying on the instruction of a third party - often the case when someone asks us to keep a secret, for instance. The result was the Bloomsbury Deception Set, which we are keen to share with other researchers.

We had a research assistant (RA) pose as a film director’s assistant outside a London filming studio. He approached people on the street and asked them if they would like to take part in a documentary. If they agreed, the RA asked which countries on a list that they had visited. When we knew what countries they had and had not been to, the RA said that he was short on time, and asked if they would do him a favor and make up a story about visiting a country they had not been to, as well as tell the truth about a trip they had been on.

When they agreed, they were taken into the filming studio where there were multiple large television cameras positioned and were left alone with the film director (actually another research assistant). The director explained he was an anthropological researcher and interested in the types of people they met in various countries. He stressed that it was important that the accounts were true because they would contribute to his research. The speaker was then given a form from a fake television company stating that there could be consequences to their lying, and they then signed the form. Two people confessed to the director and did not tell a lie. But the remaining 20 both lied and told the truth.

RG: Is there any way people can boost their ability to detect lies?

Street: One new approach to boosting lie detection is to try to increase the clues to deception that the speaker gives off. For instance, if raters are telling a lie about an emotion they are experiencing, we might want to try to push some emotional buttons to evoke an emotional response. Then the truth-tellers should give signs of genuine emotional response and liars may struggle to successfully express such a strong emotion. But what is commonly done in the literature is to exploit the cognitive load theory, which claims liars think harder than truth-tellers. If this is the case, then we can make it even harder for both liars and truth-tellers to deliver their tale. For instance, I might ask you to recall it backwards, or ask you to do a math task while you’re telling me your tale. Because lying is difficult, according to the theory, and because the mental resources are being even further tasked by having to recall backwards, this might make the lie even more obvious and lead to more “leaking” of clues or tells to deception.

RG: During the upcoming debates do you have any suggestions for viewers to help them detect if the presidential candidates are lying?

The focal account would suggest that people become aware that many of the so-called “clues” to deception that are often touted in popular psychology books and online articles are not backed by research. For example, there is no link between deception and eye contact. So ignore eye contact altogether when making your judgment. In a large-scale study in 2003, Bella DePaulo and her colleagues found that most clues to deception are very weak and unreliable. Perhaps the better clue to deception in that study, albeit still not perfectly reliable indicators, was that liars are less forthcoming than truth-tellers. The focal account, then, would say pay attention to this clue and ignore less reliable behaviors like eye contact. What is really interesting for the focal account is that if you get indirect raters to decide whether a speaker is being cooperative or not, it seems to give better detection accuracy than making an explicit lie-truth judgment. Remember that the indirect method gets people to focus on only a single clue and ignore everything else. Focusing on the more reliable clues, and ignoring the less reliable ones, should boost your lie detection scores.

In about 90% of the time the “is this person forthcoming” clue will show up indiscriminately, but on 8% of the occasions it will show up more often when someone is lying. These are very rough estimates, but they give an idea of just how unreliable the better clues to deception really are. Clearly the focal account is not going to solve all our lie detection woes, but it goes some way to showing that reasoned, conscious judgments using only the more reliable clues should aid in detecting a lie.

Image courtesy of Kenny Louie