Lie to Me

From A Test of the Micro Expressions Training Tool:

Image result for lie to meThe theory behind micro‐expressions posits that when people attempt to mask their true emotional state, expressions consistent with their actual state will appear briefly on their face. Thus, while people are generally good at hiding their emotions, some facial muscles are more difficult to control than others and automatic displays of emotion will produce briefly detectable emotional “leakage” or micro‐expressions (Ekman, 1985). When a person does not wish to display his or her true feelings s/he will quickly suppress these expressions. Yet, there will be an extremely short time between the automatic display of the emotion and the conscious attempt to conceal it, resulting in the micro‐expression(s) that can betray a true feeling and according to theory, aid in detecting deception.

…The METT Advanced programme, marketed by the Paul Ekman Group (2011), coined an “online training to increase emotional awareness and detect deception” and promoted with claims that it “… enables you to better spot lies” and “is meant for those whose work requires them to evaluate truthfulness and detect deception—such as police and security personnel” (Paul Ekman Group, METT Advanced‐Online only, para. 2). The idea that micro‐expression recognition improves lie detection has also been put forth in the scientific literature (Ekman, 2009; Ekman & Matsumoto, 2011; Kassin, Redlich, Alceste, & Luke, 2018) and promoted in the wider culture. One example of this is its use as a focal plot device in the crime drama television series Lie to Me, which ran for three seasons (Baum, 2009). Though a fictional show, Lie to Me was promoted as being based on the research of Ekman. Ekman himself had a blog for the show in which he discussed the science of each episode (Ekman, 2010). Micro‐expression recognition training is not only marketed for deception detection but, more problematically, is actually used for this purpose by the United States government. Training in recognising micro‐expressions is part of the behavioural screening programme, known as Screening Passengers by Observation Technique (SPOT) used in airport security (Higginbotham, 2013; Smith, 2011; Weinberger, 2010). The SPOT programme deploys so‐called behaviour detection officers who receive various training in detecting deception from nonverbal behaviour, including training using the METT (the specific content of this programme is classified, Higginbotham, 2013). Evidently, preventing terrorists from entering the country’s borders and airports is an important mission. However, to our knowledge, there is no research on the effectiveness of METT in improving lie detection accuracy or security screening efficacy.

…Our findings do not support the use of METT as a lie detection tool. The METT did not improve accuracy any more than a bogus training protocol or even no training at all. The METT also did not improve accuracy beyond the level associated with guessing. This is problematic to say the least given that training in the recognition of micro‐expressions comprises a large part of a screening system that has become ever more pervasive in our aviation security (Higginbotham, 2013; Weinberger, 2010).

Note that the online training failed but micro-expressions are real and better, more intensive training or maybe an AI could do better though on that last I wouldn’t accept the hype.

Hat tip the excellent Rolf Degen on twitter.


I'm not sure this is a fair test. I spent a little bit of time with Dr. ekman back in the early 2000s and he had fairly specific guidance.

His micro expression training tool was one thing... available to the public. His lie detection thing was something entirely different that he taught to police forces only.

Claiming that METT doesnt solve lie detection goes along very well with the Ekman materials.

I suspect that a shy person would appear to be lying when they are not using this technique. AND that a bi-polar person would always appear to be telling the truth even when they lie.

I worked in this area a dozen years ago and we felt at the time that Ekman and, more specifically, some of his colleagues were really overselling what could be accomplished. We tried to partner with them on some research, but they wanted us to pay licensing fees for working together on experiments. Definitely not typical behavior for academics working to improve the state of knowledge in the field.
A colleague did a review of SPOT training for DHS and found it to be quite flawed.

I'm shocked, SHOCKED to learn that aviation security in the US is based on ineffective, debunked theories and is more concerned with following the latest hype train than actually accomplishing its stated mandate.

Sure, sure. What we need now is the ability to detect subtle lies. And not the ability to do something about
that jump up and stare us all in the face.

He’s right. It’s true, Vox, a national media company willfully lying to the public is a major problem.

Or as their star and part owner Matt Yglesias says, “Fighting dishonesty with dishonesty is sometimes the right thing for advocates to do, yes”

You have really poor skills at determining the truth, right? It isn't what the full FBI investigation finds. That's negated when someone you regard as left wing reports what the FBI finds.

“Fighting dishonesty with dishonesty is sometimes the right thing for advocates to do, yes” - Vox Media

So what is the lie? He found conclusively that there wasn't any collusion in the trump campaign. There were some instance of Russians doing things, including sourcing a bunch of rumours that has tied the political class in knots for two years, generating FISA warrants, etc.

I doubt Vox takes that angle though. They were fully invested in the conspiracy theories and usually conspiracy theorists don't change their mind when presented with reality. They prefer their lies.

"If you like your insurance, you can keep your insurance." - Barack Obama

"I did not have sex with that woman." - Bill Clinton

Always good to have a test data set.

Sure sure, those are exactly the same as welcoming Russian interference.

“Over the course of my career, I’ve seen a number of challenges to our democracy. The Russian government’s effort to interfere in our election is among the most serious,” Mueller said.

Again, what we have here is an inability to deal with very important lies central to the safety of our democracy.

Shouldn't you be harassing somebody who can actually do something? Perhaps one of your Senators?

You could call them once per hour until they do something.

That would be a lot more productive than pestering a bunch of people that think you are a bit loony and couldn't directly make a change in any case.

White nationalists such as yourself are not loony. They’re evil.

I’m proselytizing to sinners. If even one can be saved, it’s worth it.

"I’m proselytizing to sinners. If even one can be saved, it’s worth it."

Wow, Sowell's "Vision of the Anointed" was right.

You still mad at Obama for not responding when he found out?

Something that's been known, used, and talked about since forever, it's called being a "good judge of men", "sizing up", etc. Only it was a far more expansive and useable sense than what can ever result from AI (which does not exist and is not more than a SV/Wall Street hype term).

Maybe AI doesn't exist, but it can beat you at chess (and go, and starcraft, and poker...)

"A problem that proponents of AI regularly face is this: When we know how a machine does something 'intelligent,' it ceases to be regarded as intelligent."

" Note that the online training failed but micro-expressions are real... "

No, the referenced study allegedly proving that micro-expressions are "real" across human populations ... does no such thing.
That very narrow study is highly subjective and non-scientific.

A micro-expression perhaps fit for bogus, faux, and otherwise tentative and thoroughly incomplete "sciences" (especially those appropriated prematurely for avid dissemination by our corrupt and corrupting Media Establishment):


i heard something about a study of why some sheepdogs were ineffective, and they determined that rather than being fully assertive, those dogs rocked back on their “heels” for just the slightest moment. and the sheep noticed

Best part of that show was when they showed some public figure making the face they were describing. It's no doubt oversold, but those frowning, caught politicians were fun to watch...

Hi Alex. I'd be even more skeptical than you of any work involving micro-expressions. The paper you cite in support of their existence actually states: "The notion of the 1/25th–1/5th of a second full-face micro-expression long advocated by Ekman was not supported here." They find other types of micro-expressions, but they are rare and, in other work by ten Brinke and colleagues (e.g. ten Brinke & Porter 2012 Cry Me a River: Identifying the Behavioral Consequences of Extremely High-Stakes Interpersonal Deception), they show the micro-expressions not to be diagnostic at all. The whole thing should likely go down the bin.

The whole theory of Microexpression is discredited - only people deeply invested are keeping this going. There are no universal basic expression that show emotions : there are learned patterns, based on culture. The within group variation is so large that these cannot be categorized by facial muscle movements.
Read "how emotions are made" by Liza Barret - recommended on this site by Tyler !

Comments for this post are closed