Sporer, S., & Schwandt, B. (2007). Moderators of nonverbal indicators of deception: A meta-analytic synthesis. Psychology, Public Policy, and Law, 13 (1), 1-34 DOI: 10.1037/1076-8922.214.171.124
Caution: this is another post where I deviate a little from what's been covered in class. Make sure you can differentiate what you learn from the book from what you read here before you continue, because you're not going to be tested on this stuff. I'm just presenting it as a different viewpoint from which to consider these kinds of psychological issues. That said, caveat emptor and continue on.
Dr. Geiselman has talked about the Behavioral Assessment Interview in class, which seeks to provide interrogators with nonverbal cues for investigators to look for to figure out when someone is lying to them: averted gaze, body turned, legs or arms crossed. However, he also acknowledged that this isn't always diagnostic or predictive of deception. The interogatee could just be nervous, for instance. And even beyond this, the hit rate for this technique is not all that great.
The plain truth is that it's really hard, maybe impossible, to detect deception just looking at nonverbal cues. This is what Siegfried Sporer and Barbara Schwandt (of the Justus-Liebig-University of Giessen in Germany) found out. They conducted a meta-analysis on the research that has been done on deception and failed to find any consistent nonverbal indicators.
For those of you who don't know, a meta-analysis is an analysis looking at the results of published studies on a given topic - it takes all the data that has been collected on a question and aggregates it, looking for trends and consistencies. These are generally good ways to find the effect size of a given construct, even though they're susceptible to the so-called "file drawer" problem: many studies are done but never published, due to the investigator finding something that doesn't agree with his or her hypothesis, or finding no result at all. These have to be left out of meta-analyses, resulting in an over-reporting of a given phenomenon.
However, given these limitations, it's remarkable that the authors of this analysis were unable to find any gestures that consistently betray deception. In fact, what they did find was that deceivers show less nodding and foot movement than truth-tellers; people move more when they tell the truth than when they lie. The most troubling finding comes when they compare their results with the opinions of average joes and professionals: everyone predicts that liars will do things like avert their gaze, shift in their seats, and gesticulate, and these expectations are quite strong. Unfortunately, they're entirely unfounded.