Submitted by Charlton Stanley, (aka Otteray Scribe) guest blogger

Polygraph tests are 20th-century witchcraft.
 -Sen. Sam Ervin (1896-1995)

As the Obama administration and the Department of Justice ramps up the crackdown on security violators and leakers, the whole thing has taken a bizarre turn. There is an ongoing criminal investigation of instructors who claim they can teach job applicants how to pass lie-detector tests. The two men are Doug Williams, who operates Polygraph.com, and Chad Dixon, who had a website called “PolygraphExpert.net” which has been taken down. Chad Dixon has entered a guilty plea, but the charges against him are being kept secret under seal. Dixon faces a maximum sentence of up to 25 years in prison; however, prosecutors are asking for a two-year sentence. Williams has not been charged with any crime; at least not yet, but is said to be under investigation. His only publiccomment was to say he has done nothing wrong.

The criminal investigation has not been acknowledged publicly. What little news that has come out is the claim it is meant to discourage criminals and spies from infiltrating the U.S. government by using so-called polygraph-beating techniques. Several current or former polygraph examiners are alleged to have been providing training materials and classes on how lie detector devices work and how to “beat” them.

Doug Williams and Chad Dixon’s business records were seized. The records are believed to include the names of as many as 5,000 persons who sought advice from the two men. The government claims about twenty of those people applied for positions with the government or government contractors. About half of that group was hired, including one or more getting jobs with the National Security Agency (NSA).

Federal officials have adopted a unique and controversial legal theory that teaching clients how lie detectors work and how to pass the test is a crime, and not protected under the First Amendment.

I find this more than curious. By way of full disclosure, I own a voice stress analysis machine and several biofeedback devices. I first became interested in the detection of malingering, dissimulation and outright lying when I was still in graduate school, and have maintained that interest ever since. Some people lie to look good, and some lie to look bad. Some lie and don’t even know they are lying. Some lie when the truth would serve them better.

In this piece, we will take a look at exactly what it is the Feds are talking about. And we will puzzle about why they want to make it a crime for anyone to teach people how the machines work. Or more accurately, don’t work.

Everyone is familiar with anxiety. Hands sweat, voice trembles, breathing may become more rapid, and the heart races. Many times trembling is visible to the naked eye. Anxiety is a fear reaction. Both the polygraph and voice stress analysis take advantage of these physiological reactions to fear, and take measurements of them. The theory behind both machines is that an anxious person will react. Practitioners of polygraphy and voice stress analysis operate on the assumption that telling a lie will result in a predictable and measurable physiological reaction.

Voice stress analysis and the polygraph: A tale of two different machines with a common purpose.

There is undisputed rivalry between both technologies, and even among different manufacturers of the machines about which is “best.” One thing is undisputed among scientists. Both technologies measure stress, and do it rather well. ‘Stress’ is a term borrowed from physics, as is another physics term, ‘tension.’ Mental health professions describe the psychophysiological reaction to anxiety-producing situations as a stress reaction.

If both machines measure stress, then the next question is about detection of lies or truth. This is the problem. None of them does. All these machines are based on the unproven assumption that a lying subject will always be stressed, whereas little or no stress will be present if they are telling the truth.

One of the foremost authorities on the subject of lie detection was Dr. David Lykken, a psychologist and neuropsychiatrist at the University of Minnesota for 43 years. Lykken and his staff did pioneering research in human physiology and behavioral genetics.

In 1991, Dr. Lykken gave the Invited Address to the 1991 Convention of the American Psychological Association, acknowledging the Award for Distinguished Contribution to Psychology in the Public Interest.¹ In that address, he pointed to research over the previous forty years, including his own work and that of others. Dr. Lykken emphasized that science had proven conclusively there is no such thing as a “lie detector.” Dr. Lykken summarized much of the research in his book, A Tremor in the Blood.² The media interviewed him many times. He told reporters the same thing each time he was asked the question: This technology is junk science.

This exchange with a reporter from People Magazine³ was typical of his exasperation with the continued use of the polygraph and voice stress analysis by law enforcement and private industry:

People magazine reporter: “How reliable are the tests?”
Dr. David Lykken: “Half of innocent people fail them. You’d do as well flipping a coin. In particular, people with strong consciences and religious beliefs can be easily made to feel guilt and anxiety.”

One of the problems with any so-called lie detector machine is that no two people have the same stress reaction to any given stressor. Extensive research on psychopaths shows they have little or no anticipatory anxiety. They do not react to threat in the same manner as a normal person. By the same token, some people respond to almost any stimulus as a threat, and everything makes them anxious. One of the things emphasized during my training, was that any form of stress detection system, whether polygraph or voice stress, was useless in cases involving sex. As one experienced polygrapher instructor put it in his lecture, “In our culture, any discussion of sex makes everybody anxious.”

Two Swedish researchers, Anders Eriksson and Francisco Lacerda, published an analysis of VRA (aka Voice Stress Analysis) in the International Journal of Speech Language and Law.⁴ Here is the abstract of their paper, verbatim:

 “A lie detector which can reveal lie and deception in some automatic and perfectly reliable way is an old idea we have often met with in science fiction books and comic strips. This is all very well. It is when machines claimed to be lie detectors appear in the context of criminal investigations or security applications that we need to be concerned. In the present paper we will describe two types of ‘deception’ or ‘stress detectors’ (euphemisms to refer to what quite clearly are known as ‘lie detectors’). Both types of detection are claimed to be based on voice analysis but we found no scientific evidence to support the manufacturers’ claims. Indeed, our review of scientific studies will show that these machines perform at chance level when tested for reliability. Given such results and the absence of scientific support for the underlying principles it is justified to view the use of these machines as charlatanry and we argue that there are serious ethical and security reasons to demand that responsible authorities and institutions should not get involved in such practices.”

That journal article in the International Journal of Speech Language and the Law pointed out “discrepancies between the claims the producers and vendors make and what their products are capable of delivering.” As a result of that journal article, the authors were threatened with a libel suit by one of the companies making these devices. The journal received similar threats and, rather disappointingly, took the article offline. But hey, this is the Internet age, and stuff has a way of hanging around forever. All the threatened lawsuit and ensuing caving in by the journal did was trigger the Streisand Effect. Here is the full journal article, preserved by the magic of the Internet. It has been published and republished many times, on many sites, including Wikileaks.

In summary, lie detector technology has no known statistical properties with regard to detecting deception of any kind. It has not been accepted as science in the scientific community. The only thing scientists seem to agree on is most of these machines measure stress reactions in humans, and to that extent, they can measure stress in people who feel stress—that’s it. There are too many variables in the psychology and physiology of deceptiveness for reliable data to be measured with any confidence at all.

Peer reviewed articles about detection of deception published in mainstream scientific journals state flatly that no technology invented thus far can be considered a reliable and valid detector of deception. Every “blind” research project has resulted in “inconclusive” findings; that is, the technology has no better results than chance. A caveat: One has to be careful to differentiate between rigorously controlled studies done by independent scientists and published in reputable peer-reviewed journals, and studies sponsored by manufacturers or groups with a financial interest in the outcome. Also, testimonials are not science.

A quick search of the Internet for information on how to beat lie detectors yields a lot of hits. Same thing on how to beat the MMPI-2, which is used for screening applicants for many types of jobs, including law enforcement. Because of the ongoing government investigation, several online web sites providing this type of educational information have shut down.

So let’s get this straight. There is a junk science out there that does not work, as established by research study after research study over more than a half century. Despite the scientific evidence, the government not only continues to use it, but wants to prosecute critics? One Wisconsin police chief has been quoted as saying he thinks he may be under investigation simply because he has been a public critic of this technology.

I like this observation by California attorney Gene Iredale, as quoted in a McClatchy news article:

Citing the scientific skepticism, [Mr. Iredale] compared the prosecution of polygraph instructors to indicting someone for practicing voodoo.

“If someone stabs a voodoo doll in the heart with a pin and the victim they intended to kill drops dead of a heart attack, are they guilty of murder?” asked Gene Iredale, a California attorney who often represents federal defendants. “What if the person who dropped dead believed in voodoo?

“These are the types of questions that are generally debated in law school, not inside a courtroom. The real question should be: Does the federal government want to use its resources to pursue this kind of case? I would argue it does not.”

What has been established by research is that most of the “positive” results of these devices have been due to the interrogation skills of the examiner and not the machine. Most experienced examiners are skilled at convincing the subject to confess. I have done that myself. This is where claims of success come from. It is called confirmation bias.

As Dr. Lykken said in his famous speech, when raw data from these “tests” are presented to experienced examiners in blind studies, they score no better than chance. There was one case where a suspect was brought into the police station. This was in the 1960s, when the first Xerox machines became available and a lot of people had never seen one. The detectives had some wires and a vacuum cleaner hose which they attached to the suspect. The other ends were tucked behind their brand new Xerox machine. Every time they asked him if he had committed the crime, an officer poked the button on the Xerox machine. Out would come a piece of paper with, “HE’S LYING” in big block letters. The suspect confessed. I suppose that meant Xerox makes a great lie detector machine.

There are several ongoing research studies to see whether the fMRI can detect deception with any reliability and validity. The research is still in its infancy, with no results that can be considered even remotely conclusive enough to meet any legal evidentiary standard.

The criminalization of imparting information that is widely available in this Internet age appears be a dangerous precedent. I would like to hear a clear and unambiguous explanation why teaching someone to “beat” a machine that hard science research over more than a half century has shown to not work, is illegal, and to prosecute them does not violate their First Amendment rights?

The floor is open for discussion.