The shaky science of lie detectors

A report about the test, given to CNN by Daniels' attorney Michael Avenatti, states the chance that Daniels lied is "less than 1%."

What does science say about the accuracy of polygraphs? Is it possible that people can fool the test? Those questions have been debated for years.

How polygraphs work

Polygraphs are based on the scientific assumption that lying causes stress. Studies show that when we deceive another person, the part of the brain that regulates emotion, called the amygdala, lights up. We feel shame and remorse. We get nervous.

Though we think we might be hiding that response from the person we are lying to, inside, our bodies betray us. Our breathing is likely to change; our heartbeat and blood pressure can rise; our legs or arms may twitch. We may even start to sweat.

A polygraph is designed to measure those physiological responses. First, the examiner hooks you up to a series of respiratory tubes, leg and arm monitors, a blood pressure cuff and a fingertip sweat detector. Then, a record is made of your baseline vital signs by asking common questions designed to measure truth: "Are you human?" "Have you ever lied to someone?" (It's assumed that we have all told white lies in our lives.)

Then the real questions begin, with enough repetition and variety to try to catch you in a falsehood. Each question, examiners say, can then be measured against the baseline to see whether it's a lie.

The National Polygraph Association says that "scientific evidence supports the validity of polygraph examinations" as long as they are conducted and interpreted with validated procedures. The association points to a meta-analysis of all peer-reviewed studies on polygraph testing that found an accuracy rate of 87%.

That's good enough for many police departments and federal agencies such as the FBI and US and district attorneys. Probation officers frequently use polygraphs to "prove or disprove" a person's statement; they are often used in sex offender parole cases, for example.

Critics cry foul

The validity of a polygraph reading has been widely criticized. One problem, critics say, is the tendency of humans to lie. Research shows that children begin fibbing before age 3, and they can master a white lie by age 6, which is also the age most children are busy telling whoppers.

As we grow, some of us become habitual liars, and that can affect how we respond to a polygraph. A 2016 study found that as a person lies more and more, the brain becomes desensitize, and is less likely to trigger an autonomic response.

"It is likely the brain's blunted response to repeated acts of dishonesty reflects a reduced emotional response to these acts," lead study author and neuroscience researcher Neil Garrett said in a statement.

Another problem, critics say, is that there could be many other reasons for a person to breathe more rapidly or experience a rise in blood pressure, heart rate and sweat.

"Fear, rage, embarrassment at having been asked a personal question, pain from the cardio cuff, even the tone of the examiner's voice can all cause the exact same reaction that the polygraph examiner would brand as a lie," Doug Williams writes in his book "How to Sting the Polygraph."

Williams is a former Oklahoma City police officer who says he gave countless polygraphs during his career before losing faith in their validity. He started a website and taught clients how to "fool" a polygraph and appeared on numerous television shows showing his techniques, including a demonstration for Diane Sawyer on CBS's "60 Minutes."

"The polygraph is not a lie detector, and it is not a truth verifier, it is simply a crude reaction recorder, and the reactions it records can be indicative of just about anything except deception," Williams wrote in his book. "I can even teach you how to duplicate this reaction by a simple breathing and muscle exercise. In fact, when you finish reading this manual, you will be able to control every tracing on the polygraph chart at will."

After an FBI sting caught Williams agreeing to help a confessed drug smuggler and another confessed child molester beat their polygraphs, he was convicted in 2015 on two counts of mail fraud and three counts of witness tampering and sentenced to two years in prison.

The widespread belief that it is easy to fool a polygraph hasn't died. The internet is filled with articles and tips on just how to do so.

The Employee Polygraph Protection Act

For years, many employers required applicants to undergo lie detector tests to order to land a job. By the early 1980s, more than a million Americans a year were taking polygraphs to either to get or keep a job. Some employers went over the line, asking embarrassing questions about private matters such as sexual preferences and toilet habits.

In 1988, the federal government stepped in and banned all private employers -- which is just about everyone -- from using polygraphs in the hiring process or as a reason to keep a job. The Employee Polygraph Protection Act allows polygraphs for hiring if the position in question includes handling drugs or security. A test can also be given when investigating a theft or crime -- but only after a written notice.

Federal, state and local government employees are not covered by the act but are often protected by civil service rules.

Not always admissible in court

Long before that report, the Supreme Court issued guidance on whether a polygraph could be admitted as evidence in court. In the 1993 case Daubert v. Merrell Dow Pharmaceuticals, the justices set the Federal Rules of Evidence as the standard for admitting expert testimony in any federal trial and left it up to local judges to apply those standards. After doing so, some judges allow the admission of a polygraph report, but others do not. Military courts, however, have banned polygraphs altogether.

Then there's the role of states. In Texas, for example, lie detectors are not admissible in criminal trials. In Florida, Georgia and Nevada, the test can be used if everyone agrees to it.

Not reliable enough for security screening

In 2002, the National Research Council of the National Academies of Sciences, Engineering and Medicine released a review of available evidence on the validity of polygraphs. The research was conducted at the request of the US Department of Energy, which wanted to use "polygraph testing to identify personnel who may jeopardize national security."

The council concluded that the US government should not use polygraphs to screen or clear employees or to identify spies because the evidence showing lie detectors work "is scanty and scientifically weak."

"Almost a century of research in scientific psychology and physiology provides little basis for the expectation that a polygraph test could have extremely high accuracy," the council stated.

Furthermore, it concluded that "the inherent ambiguity of the physiological measures used in the polygraph suggest that further investments in improving polygraph technique and interpretation will bring only modest improvements in accuracy."