For once, Brett Kavanaugh was right when he told a Senate committee he hadn’t yet volunteered to take a lie detector test because “they’re not reliable.”
As Kavanaugh observed, polygraph test results are not legally admissible in federal courts, though a few states still allow them. This is precisely because the results contain so many false positives (suggesting the subject is lying when he’s telling the truth) and so many false negatives (failing to detect actual lies) that overall they are about as trustworthy as a coin toss.
Based on “junk science,” polygraph tests should play no role in determining who’s fit to serve in the federal government. Yet AG Jeff Sessions plans to require National Security Council staff to take polygraphs, to confront the Trump administration’s foremost issue: insiders keep “leaking” embarrassing true facts about behind-the-scenes confusion and paranoia in the White House.
Unfortunately, the very mention of “lie detector” testing lends public credibility to a scientifically discredited practice.
During a polygraph test, a subject is questioned while a machine tracks his blood pressure, pulse, breathing, and skin conductivity (an indirect measure of sweating), all regulated by the autonomic (involuntary) nervous system. Swings in these readings are supposed to reveal deliberate deception — lying.
However, studies over three decades indicate that the tests have an average error rate of over 50 percent. Throughout the history of medical science, no scientific studies have shown that the emotional response linked to lying could be measured. Some people stay remarkably calm under stress, while others show exaggerated responses to the slightest challenge.
One can even purchase instruction on how to “pass” the test; “cramming for the exam,” learning to control reactions using “countermeasures,” is fully legal.
As physician and physicist Alan Zelicoff wrote, “the truth is this: the polygraph is a ruse, carefully constructed as a tool of intimidation, and used as an excuse to conduct an illegal inquisition under psychologically and physically unpleasant circumstances.”
The usual format compares answers to “relevant” questions with those of “control” questions. Control questions are supposed to compensate for the effects of the generally threatening nature of relevant questions. Anyone might grow apprehensive when asked “have you ever molested a child?” even if he’s innocent, especially when a child abuse investigation is underway.” He thinks “I didn’t do anything like that, but what if the machine doesn’t believe me?”
Predictably, the American Polygraph Association, a professional organization for polygraph examiners, has faith in the test’s accuracy, not to mention a concern for the income derived from administering it. They provide licenses for practitioners in 28 states. Their trade journal reports both scientifically dubious studies and anecdotes of their accuracy.
The majority of these practitioners complete a six-week to six-month post-high school training course in polygraphy. They’re not required to have formal training in medicine, psychology, physiology, or behavior — yet those are the very disciplines upon which the testing is based. The majority of polygraphers cater primarily to the legal system upon which their income depends, creating a potential conflict of interest.
Some smaller police departments designate a police officer to be their department examiner instead of using an outside “professional.” The officer can have little or no formal training, other than the limited instruction provided by the company that sells the machine.
This can create a bias based on an officer’s predisposition to believe that most suspects are generally guilty. His colleagues aren’t in the business of arresting innocent people, right? If the officer knows about the circumstances of the charges, and the evidence already collected, his bias makes him more likely to interpret an inconclusive response as a lie.
Although the test can provisionally identify some lies, the National Academy of Sciences’ comprehensive review in 2003 concluded that the test produced too many false positives.
A man accused of murder was “offered” a polygraph (refusal per se is often regarded as evidence of guilt) during an investigation. He agreed — and “flunked” the test.
On the basis of other evidence, he was acquitted. But his community, convinced the test was accurate, shunned him and threatened him with bodily harm. Months later, the actual murderer was apprehended and convicted. A test failure can ruin lives.
An instructive example of a false negative is the case of Aldrich Ames. In 1995 Ames was arrested and convicted of spying, despite having “passed” five polygraphs during his long career in U. S. intelligence. Ames himself subsequently wrote a letter from prison calling polygraphs “junk science,” “a superstition,” and “a refuge from responsibility.” “Bureaucrats can abandon their duties and responsibilities to junk scientists and interrogators masquerading as technicians.”
Suspects are often offered the test prior to criminal proceedings; declining can foster a presumption of guilt. They can also be told they flunked when they actually passed, hoping to provoke a confession.
If the accused accepts the test, a false positive can be used to support intensifying the investigation of the suspect, while the real perpetrator evades detection. In some states, it can lead to a guilty verdict in court.
By federal law, it’s illegal for most private companies to require or invoke polygraphy as a condition of employment. It can be used in jobs involving security, or drug handling, or while investigating a specific crime.
Federal agencies like FBI, CIA and NSA are exempt, and continue to use polygraphy in employment decisions.
Isn’t that reassuring?
Jon Hauxwell, MD, is a retired family physician who grew up in Stockton and
lives outside Hays.