FAS | Government Secrecy | November 2000 News ||| Index | Search | Join FAS


The Baltimore Sun
November 3, 2000

Are polygraph tests lying to us?

Tests: Mixed reading of Lee's nuclear secret data,
federal employee opposition to taking lie detectors
'reignite' 80-year-old controversy.

By Michael Stroh

When physicist Wen Ho Lee first denied leaking U.S. nuclear secrets to the Chinese, authorities from the Department of Energy in 1998 wired him to a polygraph to see if he was lying.

The Los Alamos National Laboratory scientist passed.

But when a polygraph expert from the FBI looked at the same test results later, he concluded that Lee had not told the truth.

How could the same lie detector test lead investigators to exactly opposite conclusions?

The case of Lee, who eventually pleaded guilty to one felony count of mishandling classified information, has left law enforcement experts trying to answer the same fundamental questions that have existed since the invention of the lie detector 80 years ago: Does the polygraph actually work? And is it fair?

"It's reignited this smoldering controversy," says Steven Aftergood, a senior research analyst with the Federation of American Scientists in Washington. In an essay being published today in the journal Science, Aftergood argues that a new federal policy requiring nearly 20,000 employees of the national nuclear weapons laboratories to take lie detector tests is having undesirable effects.

The policy has lowered morale, Aftergood writes, and caused some of the nation's most gifted scientists to leave, and made it harder for the labs to recruit talented young researchers for the weapons programs. The use of the polygraph, he writes, "symbolizes the defeat of reason by the national security state."

Despite such criticisms, the use of the polygraph test is on the rise.

Congress banned private industry's use of lie detectors as a condition of employment in 1988, but they are routinely used for employee screening at the FBI, Central Intelligence Agency, National Security Agency and local police departments around the country. The percentage of law enforcement agencies using polygraphs for this purpose rose from 16 percent in 1962 to 62 percent in 1999, according to a survey by Michigan State University's School of Criminal Justice.

There's also a growing market for polygraphs outside law enforcement. The American Polygraph Association, the largest polygraph accrediting and licensing organization in the country, reports that its membership has risen past 2,000 and is continuing to grow.

Private polygraph examiners handle everything from fishing tournaments to divorce cases. Winners of the annual Big Rock Blue Marlin Tournament in Morehead City, N.C., for example, must submit to a polygraph before collecting any prize money (to make sure they haven't stuffed rocks in the gut of their prize catch).

Lie detectors aren't designed to detect lies as much as the subtle physical changes that may occur when a person tells a lie. The word "polygraph" means "many writings," and that is what the polygraph machine produces: lots of squiggly lines on a scrolling piece of paper.

The test works like this: A subject is seated in a chair. Two rubber belts are wrapped around his chest and stomach to measure breathing patterns. A blood pressure cuff is wrapped around an arm. A metal plate attached to the fingers measures sweat gland activity.

The polygraph examiner then asks the person a series of questions. Some of the queries are "control" questions unrelated to the matter under investigation but establish a base line of the person's blood pressure, respiration and perspiration. Other questions directly address the actions under scrutiny.

The examiner interprets the person's physiological response to each of the questions, as recorded on scrolling paper, to judge whether the person is lying. And thus the uncertainty about polygraph results: they are a matter of judgment. "There's no red light or siren that comes on when the person lies," says Milton O. "Skip" Webb Jr., president of the American Polygraph Association.

The roots of the modern lie detector stretch back to antiquity. Like modern methods, early techniques to ferret out lies often relied on the behavior exhibited by liars -- sweaty palms, dry mouth, shifting gaze, racing pulse.

In China, for example, suspected liars were fed a handful of dry rice. If they could spit it out, the thinking went, they were telling the truth. If the rice stuck to their tongue, they must have something to hide.

The modern quest to detect liars using technology began with Cesare Lombroso, an Italian criminologist who in 1895 published a book called "The Criminal Man" in which he described his efforts using an early instrument to measure changes in blood pressure to determine whether several criminal suspects had lied.

In 1915, Harvard psychologist William Moulton Marston picked up on these early studies and devised a primitive lie detector based on blood pressure. According to psychologist and polygraph historian David Lykken, it was Marston, a colorful P.T. Barnum-like character, who was among the first to realize the lie detector's commercial possibilities.

In 1938, Look magazine described how Marston sometimes used his lie detection techniques in marital counseling. He also showed up in full-page ads testfying to the close shave offered by Gillette razors: "New Facts about Shaving Revealed by Lie Detector!" (Using the pen name "Charles Moulton," Marston would also invent the comic strip character Wonder Woman, whose magic lasso could force those held to tell the truth. )

But John A. Larson, a Berkeley, Calif., police officer, is the person generally credited with inventing the modern polygraph machine. In 1921, Larson, who eventually became a doctor, devised an instrument that could simultaneously record blood pressure, pulse and respiration. Later tinkerers improved upon Larson's design by adding sensors to measure perspiration.

Over the years scientists have tried to determine whether the polygraph actually works. But accurate studies are hard to do. "The science is not solid," says Aftergood, in part because investigators can rarely learn independently whether a subject who passed a polygraph test was indeed telling the truth.

In some studies, volunteers are recruited to be pretend criminals and then subjected to a lie detector test. But the results of such work, critics argue, don't mimic reality. "It's impossible to make the stakes as high in an experiment as they are in real life," says Aftergood.

Still, proponents of the polygraph argue the device is accurate in better than 90 percent of cases, and note that it's not uncommon for other types of test results to be open to interpretation.

"Your doctor can have you take a chest X-ray and say, 'I don't see anything.' Then he sends it over to a radiologist and the radiologist finds something the first doctor doesn't see," says Webb. "Happens all the time."

But enough guilty people have slipped past the polygraph to have given law enforcement officials pause. Most federal and state courts do not allow polygraph results to be entered as evidence.

CIA employee Aldrich Ames, for example, passed lie detector tests despite selling U.S. secrets to the Russians for more than eight years. There's also a mini-industry of Internet sites and books such as "Deception Detection: Winning the Polygraph Game" that purport to teach people how to beat the test.

"College students with 15 minutes of explanation can beat the lie detector," says David Lykken, a retired psychologist from the University of Minnesota. "Anybody who is working as a spy has been taught how to beat the polygraph." The advertised techniques range from curling one's toes to biting one's tongue during control questions to mislead the examiner.

Still, even critics of the polygraph acknowledge that it has led to admissions of guilt that they might not otherwise have gotten.

"The polygraph itself functions as a prop more than anything else," says Aftergood. "Yet, there are cases every year in which the prop works."




FAS | Government Secrecy | November 2000 News ||| Index | Search | Join FAS