There's a widespread belief that when people look up and to the right while speaking, they're probably lying. But results of three separate experiments performed by a team of British and Canadian researchers suggest that the belief is nothing more than a myth, like alligators in the sewers.

There are definitely differences in the behavior of liars and truth tellers. But the direction of their eye movements doesn't seem to be one of them.

How did the idea that liars can be detected by their eye movements gain acceptance? That's a story in itself.

The third experiment looked at a real life situation, since liars in the laboratory rarely have the motivation to lie well that people have in real life, where the stakes are much higher.

Neuro-Linguistic Programming (NLP) is a school of thought that arose in the 1970s. It's concerned with the relationship between brain, body and language and is keenly interested in how people can improve their communication skills. Its advocates soon began to claim that when right-handed people look up and to the right, they are likely to be visualizing an imagined event or sound, but when they look up and to the left they are likely to be visualizing a remembered event or sound, one that actually happened.

Over time, this idea became simplified to looking up and to the right indicated lying, while looking up and to the left indicated truthfulness. And like a viral video (though slower), it spread throughout the Internet and became part of the 21st century's Book of Common Wisdom. It's even taught in lie detection training courses.

Yet there's little evidence backing up this claim. So the British and Canadian team went looking for some. And came up empty three times.

In the first experiment, 32 participants, mostly college students and all right handed, were given a cell phone and sent on a short trip to an office, where they either placed the phone in their pocket or in a desk drawer. They then were sent to an interview room where they were filmed while being asked questions about what they did in the office. Half told the truth, half lied.

An analysis of the films showed no relationship between participants' eye movements and whether they were telling the truth or lying.

In the second experiment, 50 new participants watched the films from the first experiment. They were asked to judge whether the people in the films were lying or telling the truth. Half were told of the alleged relationship between eye movement and lying, half were not.

There was no difference in the accuracy of the two groups. Searching for liars by their eye movements did not make them easier to find.

The third experiment looked at a real life situation, since liars in the laboratory rarely have the motivation to lie well that people have in real life, where the stakes are much higher. Here the researchers looked at 52 videos of high-profile cases where people made a public plea for the return of a missing relative. Over time, there's been compelling evidence that half of these people were telling the truth while the other half knew very well where their relative was and were lying.

There was no correlation found between people's eye movements and whether or not they were telling the truth. In fact, the people who were lying actually glanced more often to their upper left than to their upper right, though that relationship was not statistically significant.

In short, the researchers found no evidence that liars can be detected by their eye movements. And without other evidence to back up the claim, they think that people and organizations should stop teaching that they can.

For more details of the study, see the article published in PLoS ONE on July 12, 2012.