Saturday, January 10, 2009

TV show depicts deception detection science as too exact

FOX TV is heavily running previews for a new mid-season drama called "Lie to Me" that purports to portray a scientist with perpetual five o'clock shadow and his hotter-than-thou research team who are experts in the "science of deception detection" - i.e., telling whether someone is lying.

Says a promo: "Dr. Cal Lightman and his team are effectively human polygraph machines, and no truth can be concealed from them." Of course, as human "polygraph" machines that would make their work discredited, pseudoscientific and too unreliable to use in court, but I'm sure that's not what the producers meant by that line.

The lead character is based on the work of behavioral scientist Paul Ekman whose landmark research cataloging and interpreting facial expressions has led him to create a lucrative consulting business teaching lie detection techniques to various official and corporate entities. FOX TV, naturally, has turned Ekman into a crime fighting detective type. Even more of a stretch, the show takes to absurdist extremes a concept from Ekman and Maureen O'Sullivan's work hypothesizing a small number of people are "truth wizards" who are exceptionally accurate at deciphering lies. In fact, the story goes so far as to pretend the lead character, Dr. Cal Lightman, himself enjoys such abilities.

Inarguably, deception detection using "microexpressions" is a big part of how humans identify deception in an interactive environment. But it can never be 100% accurate, or even close. In practice it's more art than science.

To use a Shakespearean analogy from Ekman himself at a recent MIT conference on security and human behavior,
Othello’s error was to read Desdemona’s fear correctly but to misunderstand its cause. We may indeed pick up on these cues, but that doesn't mean they will be interpreted correctly.

Reading micro-facial expressions to identify deception is a core method used in the "Reid technique" of police interrogtion, an approach that provides the basis for most police interrogation training in America. But with the possible exception of a handful of "wizards," there's little evidence police are better at detecting deception than the rest of us. According to the (highly recommended) recent book by Richard Leo, Police Interrogations and American Justice (pp. 98-99):
Numerous controlled studies have shown that people are not good intuitive judges of truth and deception, typically performing at no better than chance levels of accuracy. Controlled studies have also shown that even investigators and other supposed experts who routinely evaluate deceptive behavior are highly prone to error. Moreover, Kassin and Fong have shown that police interrogators and others specifically trained in the [Reid technique] not only fail to discriminate accurately between true and false statements much of th time, but also that behavior analysis training actually lowers the ability of police interrogators to discriminate accurately between true and false denials. Further, such training inflates their confidence in their judgments. (citations omitted)
Indeed, researcher Aldert Vrij thinks overreliance on Ekman-esque micro-expressions ignores more probative cues to deception and may lead to false accusations. According to this abstract from a recent paper:
deception research has revealed that many verbal cues are more diagnostic cues to deceit than nonverbal cues. Paying attention to nonverbal cues results in being less accurate in truth/lie discrimination, particularly when only visual nonverbal cues are taken into account. Also, paying attention to visual nonverbal cues leads to a stronger lie bias (i.e., indicating that someone is lying). The author recommends a change in police practice and argues that for lie detection purposes it may be better to listen carefully to what suspects say
"Lie to Me" premiers on Jan. 14 in FOX's sweet spot right after American Idol, and I'll be watching with interest if also a slightly jaundiced eye. Until then, for those interested in the actual science of deception detection, the always excellent (when the author gets around to posting!) Deception Blog supplied a superb end of the year roundup of deception-related research:

12 comments:

Soronel Haetir said...

I'm not sure about this particular show but I was very troubled by the couple lie-detection game shows that ran last year. I don't recall the name(s) but the basic idea was to force the player into revealing extremely mbarrassing things about their life, potentially job and marriage ruining things.

The sorts of questions that were asked seemed designed to toss out whatever reliability that polygraph techniques might enjoy, where my understanding is that the more specific the question the more accurate the result is likely to be with regard to truthfulness. The questions asked were extremely open-ended, things like "Have you ever touched a client inappropriately?" asked of a personal trainer.

This setup bothers me more because I think people are more likely to believe that the methods behind a game show where large sums of money are involved are accurate and fair than those depicted on a detective drama.

Don Dickson said...

Equally troubling to me is a form a deception detection known as "statement analysis." This has been championed by a fellow named Mark McClish, now or formerly of the FBI Academy.

Mr. McClish is the author of a book titled "I Know You Are Lying," and now he's even got a software program which purports to assist you to tell when someone is lying in a written statement.

Mr. McClish lost all credibility with me when he asserted in his book that the number 3 is a "liar's number." Stop any motorist and ask how much he's had to drink, and see how many will tell you they've had three.

Unfortunately, our DPS has had a few of its folks go to the FBI Academy and come home with this "science" in their little notebooks, and I've seen them try to use it against Troopers in disciplinary cases. It is the purest form of hogwash, IMHO.

Don Dickson said...

Correction: (I lied LOL) McClish taught this nonsense at the Federal Law Enforcement Training Center in Georgia.

Anonymous said...

Grits,
You need to go to WOAI.com and read about the probation chief.
Too funny!

Anonymous said...

From what I know of deception, physical cues include pupil dilation when a person is telling a lie. But this pupil dilation is very rapid. It's not as though the person's eyes stay dilated after one lie. I don't know what happens in situations where a person may be on drugs or medication that would cause their pupils to dilate.

One technique that's supposed to be foolproof is a type of scanning of the brain to see if a memory center is activated when an obscure object from a crime scene is shown to a subject. If their brain reacts, that means they had to have at least seen the object because the brain registers that they recognize it. I don't know if it's truly foolproof. It was used in India recently to convict a woman of murder.

Another technique uses MRI to map whether you're thinking of a lie or not. Thinking deceptive thoughts creates brain activity that's different from the activity created when you're thinking of the truth.

Gritsforbreakfast said...

To 10:36 - check out the "New technologies" post from the Deception Blog for a roundup of recent research on using brain scans for deception detection. There's some interesting preliminary work, but the technology (or perhaps, our ability to interpret the results) apparently isn't quite there yet.

The only problem with the memory center approach you describe is that new things may remind us of things we've seen before and thus trigger memory. Thus if someone shows you an object you've never seen but it REMINDS them of something from their past, they could be falsely accused of lying. Ditto for faces - the big problem with eyewitness IDs in all these DNA exonerations is that, in a lineup, people pick the photo that looks most like the person they remember, even if they've never actually seen the person they choose before. So memory is involved, or can be, even in those instances.

roy said...

I fear that what "24" did to get the public to accept torture as reliable in revealing truth, "Lie to Me" will do to get the public to accept interrogation as reliable in revealing truth.

Gritsforbreakfast said...

7:37 - The missus actually chastised me upon seeing this post for not making that exact point. She wanted me to do more FOX bashing!

Anonymous said...

Ever hear of a paper polygraph?
It's a multiple choice test that determines if you are being truthful by comparing answers on multiple questions.
ie, If you said you smoke marijuana over 100 times, yet you stated that you never sold marijuana, you are lying.

Anonymous said...

Whoo Hoo, Maybe we can get a Scientology angle in on this too. I mean if we are going to talk quack practices, lets pull out all of the stops... Throw in some numerology, a dash of speaking in tongues, and heck we have ourselves a party.

Sorry, although there are certain signs of an uneducated person lying, there are many, many ways to pass all of these tests and feel confident enough so that all of the 'signs' are negated or managed so that a person can 'pass'.. Both the CIA and the old KGB were experts in training their personnel in the arts of deception. Both organizations could (and probably still do) train their people to lie and pass these tests with 100% responses.

Eyes said...

So you don't believe in truth wizards? How about checking one out in action? Watch for yourself and see what its all about.

Generic Cialis said...

It is an odd show, deception is a huge part of ourselves, almost everyone lies and deceives for their own good.