Verified by 
I know, the TV show Lie to Me is toast, but with DVDs and such, we can all have our toast and eat it, too. But will it make us big and strong lie detectors?
When FOX first promoted the show in 2009, it claimed that it was "inspired by the scientific discoveries of Dr. Paul Ekman, a real-life specialist who can read clues embedded in the human face, body, and voice to expose both the truth and lies in criminal investigations." So would viewers who watched the show learn to become more accurate at distinguishing lies from truths?
That's what deception researcher Tim Levine and his colleagues set out to determine when the show had just begun to air. (Full reference is at the end.) They randomly assigned college students to watch either the series premier of Lie to Me or of Numb3rs (another crime drama in which Stanford math whizzes are the hero crime-solvers). In a third condition (the control), students did not watch any particular TV show as part of the study.
All of the participants then watched 12 videotapes of people who were tempted to cheat at a trivia game and thereby get a cash prize. Some did - and then lied about it in a subsequent interview. The others did not cheat and told the truth during the interview. On the 12 tapes, half of the people were lying and the others were telling the truth. Participants watched each clip, then recorded their judgment of whether or not the person was lying. By chance, participants would be right 50% of the time. Here are their actual accuracy scores:
60% Lie to Me
62% Numb3rs
65% Control
The differences were not statistically significant, so we can't conclude that people who watched Lie to Me actually did worse at labeling lies as lies and truths as truths, but they surely did no better than anyone else.
Many studies of human lie detection show that people have a "truth bias" - they see other people as truthful more often than they should. (Arguably, this particular bias is not necessarily a bad thing - maybe it is better to err on the side of believing people than disbelieving them.) In studies such as Levine's, truth bias is measured simply by computing the percentage of clips that participants judged to be truths. In fact, half of the clips were truths, so an unbiased score would be 50%. Here are the actual 'truth bias' scores for the three conditions:
51% Lie to Me
58% Numb3rs
59% Control
The difference between the viewers of Lie to Me and the participants in the other condition was a real one - the Lie to Me viewers were less inclined to judge the people on the videotapes as truthful. Those truth bias scores sum across judgments of both truths and lies. The bigger difference, though, was for the truths - Lie to Me viewers were much less likely than the other participants to see the actual truths as truthful.
So why didn't viewers of a show supposedly based on a real scientist and actual scientific research do better at detecting deception? Maybe they needed to watch more than one episode - but the differences were in the direction of worse accuracy for the Lie to Me viewers.
It is not surprising that promotions for TV shows engage in hype. Levine and his colleagues suggest that it is not just the show that has been oversold but the science of lie-detection as well. That's what Charlie Bond and I concluded, too, on the basis of our series of papers on lie-detection accuracy. In our review of hundreds of studies of skill at detecting deception, we found an average accuracy rate of only 54% (when a chance level would have been 50%). One of the papers in that collection is a study in which my colleagues and I compared the lie-detection judgments of experienced law enforcement officers to those of college students. The law enforcement officers were no more accurate - they only thought they were. (They were more confident.)
OK, so this has nothing to do with singles and I started my own blog in part so I'd have a separate place to discuss non-singles topics. But I think I still have more readers here than there, and I want to ask your help with something. I'm giving a workshop on deception-detection next week to about 70 polygraphers, and I want to make a point about the importance of a person's demeanor in whether they are judged to be lying. Some people just look a certain way, regardless of whether they are lying or telling the truth. Some people come across as honest most of the time, while other less fortunate ones often look at bit untrustworthy.
So here's my question: Can you think of any celebrities or other public figures who characteristically have a negative demeanor? Specifically, can you name any people who typically look contemptuous, or disgusted, or fearful? Thanks! I can describe the overall findings, of course, but the points will be more compelling if I can illustrate them with some good visual images.
Reference:
Levine, T. R., Serota, K. B., & Shulman, H. C. (2010). The impact of Lie to Me on viewers' actual ability to detect deception. Communication Research, 37, 847-856.
