Time magazine is reporting that ‘behaviour detection officers’ have been introduced to US airports who have been trained to pick out potential terrorists by analysing, at least in part, facial expressions. Despite the enthusiasm of the authorities for this new approach, there’s no clear evidence that it will be effective.
America’s Transport Security Administration describes the job as involving “voluntary encounters with the public under the SPOT Program, to determine whether elevated behaviors indicate that the individual may be involved in a terrorist or criminal act or activity”.
This means the SPOT system is likely to be based on one of Ekman’s two commercial systems for analysing facial expressions: the Facial Actions Coding System (FACS) or the Micro-Expression / Subtle Expression Training Tool (METT/SETT).
Ekman’s Facial Actions Coding System (FACS), a well-researched method for coding the individual muscle movements or component parts that make up a facial expression.
The idea is that its hard to fully hide emotions as they often quickly or partially emerge on the face before we suppress them, so by being able to detect ‘microexpressions’ we can get a better idea if someone might be trying to hide inner stress.
It is designed by to be used by researchers, who have time to carefully examine video tape, but also by people who need to catch microexpressions as they happen – on the fly.
Research has shown that people who are better at detecting microexpressions are better as detecting lies [pdf], but so far, the (admittedly limited) evidence suggests that training people to detect microexpressions doesn’t make people better lie detectors.
There are only two small studies I know of that have tested this (I would be interested to hear of more), none of which inspire much hope. Both studies looked at whether FACS training improved clinicians ability to detect faked vs genuine pain.
The METT/SETT is even less well-researched. In fact, only one pilot study that I know of has used it at all.
Yet Ekman makes some grand claims for its effectiveness. In a recent article for the journal Behavioral Science and the Law he notes that:
This Micro Expression Training Tool (METT) (Ekman, 2002) includes feedback about the correct answers, morphed faces contrasting the most difficult to discriminate emotions, and a pre and post test. In two recent studies, Ekman & Frank (2005) provided training on detecting deception using METT. They obtained a very large
increase in accuracy with less than one hour of training with METT.
Rather than research published in a peer-reviewed journal the Ekman and Frank (2005) reference turns out to be to a document listed only as “Revealing concealed emotions. Retrieved from http://www.paulekman.com”, which, I’m damned if I can find. Anywhere.
However, it is likely these techniques are already being used in training law enforcement officers, security guards and the like. Ekman’s corporate website notes that he’s signed a three year contract with the UK Police’s Anti-Terrorist Unit at New Scotland Yard.
One possibility is that either Ekman’s company or the US authorities have done extensive unpublished research to show that training in these or similar methods are effective at helping staff detect potential terrorists or risky passengers.
One difficulty with all deception research is that participants told to lie in the lab are not necessarily good models for ‘real-life’ deception, with all its complex motivations and emotional force.
Lab-based lies are likely to be a poor substitute an actual covert terrorist situation.