Doug Williams used to give polygraph exams. Now he’s going to prison for teaching people how to beat them
There was something odd, Doug Williams recalls, about the clean-cut young man who came to see him on Feb. 21, 2013. When Brian Luley had called two weeks earlier, he’d introduced himself as a deputy sheriff in Virginia applying for a job with U.S. Customs and Border Protection. To get the job, Luley needed to pass a polygraph test, and there were “a couple of reasons” he thought that might be a problem.
“I will get you ready,” Williams had promised. A former police detective and a man congenitally unencumbered by doubt, he claimed he’d already helped thousands of people beat the polygraph. In pictures on his website, polygraph.com, the goateed 67-year-old sat at a desk in a suit with his hands clasped, his silver hair slicked back from a tanned and furrowed brow. “Nervous or not, lying or not, no matter what,” he assured his clients, they would pass. The site had a manual and an instructional DVD for purchase, but if you really wanted to be sure, Williams would teach you one-on-one. Luley wanted to be sure.
Many of the people who sought out Williams over the years had secrets: marital indiscretions or professional lapses, drug busts or sex crimes. Williams never asked for details—those weren’t his concern. He has no affection for crooked cops or sexual predators, but what he hates above all else is the polygraph machine, an “insidious Orwellian instrument of torture,” as he calls it, that sows fear and mistrust, ruining careers by tarring truthful people as liars. “It is no more accurate than the toss of a coin,” he likes to say. When he’s feeling less generous, he’ll say a coin works better.
Williams’s office was a small basement room next to a massage studio in Norman, Okla. When Luley arrived, he didn’t act like other clients. He didn’t seem nervous, and he barely looked at the machine on Williams’s desk—an Axciton five-channel sensor box with blood pressure cuff, finger electrodes, motion-sensor pad, and two pneumographs to measure breathing. Williams described how a typical examiner would monitor Luley’s breathing, blood pressure, heart rate, and sweating. The examiner would ask “control” questions, assuming Luley would answer dishonestly—“Have you ever lied to anyone in authority to keep from getting in trouble?”—along with the questions the examiner really wanted answers to: whether Luley had committed and not disclosed serious crimes. The “controls” are intended to reveal a subject’s response to telling a lie; if Luley’s reaction to the relevant questions was stronger than his reaction to the controls, he’d be judged to be deceptive. The trick, Williams explained as he strapped the sensors onto his client, was to exaggerate the reaction to the controls and minimize the reaction to the “relevants.”
Little of this, however, appeared to get through. Luley displayed a striking compulsion for detailed self-incrimination. “If I tell them that I sold drugs in the jail when I was a jailer, can they use that against me?” he asked at one point. He also mentioned that he’d “messed around” sexually with a 14-year-old drug suspect after interviewing her.
“Keep that s--- to yourself,” Williams snapped, unaware that he was being surreptitiously filmed. “I mean, why would you ever say something like that anyway? If you’re gonna do that, save yourself the trouble. Don’t even go, if you’re going to sabotage your own damn self.”
Nonetheless, by the end of the session, Luley was able to produce what Williams judged to be a truthful, if not exemplary, polygraph “chart.” It was with some relief that Williams saw the young man out the door. Then the federal agents who had been waiting outside burst in.
The quest to defeat lying is as old as humanity. In Bronze Age China and India, suspects had to chew uncooked rice and spit it out to reveal if their mouths were dry. Medieval Europe had trial by fire or water. In the 1950s and ’60s, the CIA experimented with LSD as a truth serum. Then there’s torture, formalized in ancient Greece as a method to compel honesty and recast for the 21st century as “enhanced interrogation.”
The polygraph, invented in 1921, is today’s most widely trusted lie-detection device. It’s used to determine who gets hired by the CIA, the FBI, the Drug Enforcement Administration, and police departments all over the country. It helps decide who gets security clearances. Police detectives use it as an investigative tool, intelligence officers use it to assess the credibility of sources, and exams are commonly required as a condition of parole and probation for sex offenders.
This reliance continues despite the lack of scientific agreement about how well—or even whether—the polygraph works. Defenders point to studies that show a 90 percent accuracy rate. The National Academy of Sciences, however, in a 2002 report, found a “lack of understanding of the processes that underlie polygraph responses” and concluded that the quality of polygraph research “falls far short of what is desirable.” These criticisms, according to many psychologists, remain true. Some of the nation’s most notorious spies, including Aldrich Ames and Ana Montes, passed polygraphs. In a 1998 Supreme Court decision, Justice Clarence Thomas wrote, “There is simply no consensus that polygraph evidence is reliable,” and most judges have agreed.
For three and a half decades the polygraph has had no critic more dedicated or more obstreperous than Williams. He’s been a White House aide, a private investigator, a construction worker, and a member of the clergy. In the 1980s he was a barnstorming activist living out of a truck, giving talks and training sessions, going on talk radio and TV, testifying in courtrooms and before Congress. He eventually settled down and began charging for his instructional materials and tutoring. Over the past few years he’s learned, painfully, what can happen when an obsession becomes a business—especially when that business is spreading detailed information about a law enforcement technology that works best when its subjects know only what they’re told.
The first time Williams saw a polygraph was in 1966, after his sophomore year at Bethany Nazarene College, when he left the small Christian school outside Oklahoma City to enlist in the Air Force. High scores on his aptitude tests got him sent not to Vietnam but to Washington, to serve in the White House Communications Agency. The last requirement for getting the posting was an appointment with the machine.
For two and a half years, he sent and received encrypted messages from the White House Situation Room, earning letters of commendation from military and civilian superiors. It was a heady but ultimately disenchanting experience. “The intelligence community was just a web of intrigue and paranoia,” he says. “Ninety-nine percent of the s--- that’s classified should never be classified. Really what it’s all about is to cover up corruption and ineptness and outright fraud.”
Williams is sitting at the kitchen table of the single-story orange-brick house in Norman he shares with Kathy, his wife of 18 years. It’s early May, a week before the start of his trial in federal court. When I ask if he has any mementos from his White House tour, he goes into his study, a blast zone of papers and boxes, and eventually returns with a goodbye card signed by Henry Kissinger.
When his deployment was up, Williams returned to Oklahoma City and got a job as a police officer. Chris Eulberg joined the force the same year, and the two remain close friends. As a patrolman, Eulberg recalls, Williams got his picture in the local newspaper for single-handedly apprehending an armed robber. He also infuriated his fellow officers by reporting one of them for stealing a TV while investigating a break-in. In his off hours, Williams took psychology and criminology classes at Oklahoma City University.
One day in 1972 he saw a job posting for a polygraph examiner at sergeant’s pay, a significant bump up for him. He applied, got accepted, and was sent for a three-month course in New York. To his surprise, the bulk of the training focused on interrogation techniques. Many of the confessions Williams would elicit, the instructor explained, wouldn’t be during the test itself, but in the pretest interview. Subjects, certain they were about to be exposed, would simply break down. The trick was to take advantage of that.
For seven years, Williams ran polygraph exams for his department’s internal affairs bureau. Off duty, he freelanced for local businesses. At the time, store clerks, bank tellers, route salesmen, and other workers who handled cash were regularly polygraphed. If money or equipment went missing, the boss would often call in a polygrapher to question the staff. By some estimates, Americans underwent more than a million exams a year.
The argument for the polygraph has always been that it’s the best tool available for a difficult task. The device may not be infallible, its proponents concede, but humans are far more fallible: Even trained interrogators, studies suggest, are barely better than chance at spotting a liar. David Raskin is perhaps the leading pro-polygraph researcher. A psychologist and University of Utah professor, now retired, he helped develop the modern computerized polygraph. “Properly administered and interpreted,” he says, the exams are more accurate “than virtually any psychological test used as evidence in court—and better than a large number of the medical tests used as evidence, too.”
As an examiner, Williams saw the polygraph simply as a prop to scare confessions out of suspects, “a psychological billy club.” And over time, he came to believe that a properly trained subject could game the machine. The key realization occurred when a friend on the force came into the station describing a high-speed chase from the night before. With all the fear and adrenaline, he told Williams, the “pucker factor” had been high—he’d been “pinching doughnuts” out of the seat of his patrol car. That vivid image made Williams wonder if the same process could run in reverse, if an intentional “pucker” could trigger the physiological stress responses a polygraph traced. That would allow a subject to manipulate his response to the control questions and rig the test. For a lark, Williams hooked himself up to the machine in his office and clenched. “Lo and behold,” he recalls, “the most beautiful blood pressure increase you ever saw, accompanied by a corresponding increase in sweat activity in my hand.” The effect was even more convincing if he simultaneously altered his breathing pattern to fool the pneumographs.
About that time, Williams realized his job was destroying his life. “The thing about interrogation—you know, fear has a smell,” he says. “That smell would greet me in the morning, lingering in my office from the previous day. And it would be compounded day after day after day.” Hooking his subjects up to his machine to measure their blood pressure, his own would climb into dangerous territory. He took to drinking a nightly pint of whiskey, usually alone—as an internal affairs investigator, he had become a pariah in the department. In 1979, miserable and exhausted, he quit.
For more than a year, Williams drifted. He lived for a little while with his parents in Houston, where his father was a Methodist pastor. Moving back to Oklahoma City, he did investigative work for Eulberg, who’d left the force to practice law. He learned carpentry. One day, his sister told him she needed to take a polygraph for a job at a nightclub. Williams said she should prepare, that innocent people can fail the test just from nerves. He wrote up a set of instructions, and after she passed, her co-workers started asking for copies. Williams, with a clergyman for a father, had been raised to believe in the idea of a calling. He decided he’d found his. “It just kind of took over his life,” Eulberg says. “He decided that he had hurt people as a polygraph operator, and so he started this crusade.”
Williams bought a white 1967 Chevrolet panel truck and installed a fold-down bed in the back, along with a portable toilet, a hot plate, and a makeshift desk. For the next five years he drove around Texas and Oklahoma, sleeping at the building sites where he found carpentry work. After work and on weekends he’d present himself at union offices and churches to speak about the psychological billy club. He took his Stoelting polygraph, with its skittering ink needles and scrolling graph paper, and offered to show people how to beat it. The machine, he’d say, was effective only as a way to browbeat confessions out of subjects, whether they were true or not. He pestered radio bookers and TV producers over the phone and by mail, and he made a point of befriending clerks at convenience stores, so he could ask them to answer the pay phones there and pretend to be his answering service. They were usually happy to help—many had been polygraphed by their employers. “You talk about the spectrum of people who were at one time distributing my manuals,” Williams recalls. “They ranged from the American Civil Liberties Union on one end to the John Birch Society on the other.”
In the small world of polygraph examiners and antipolygraph activists, Williams became a celebrity. On July 30, 1985, he testified before a House Education and Labor subcommittee about a bill to ban employment-related polygraph testing. “My name is Douglas Gene Williams, and I plead guilty to crimes against humanity,” he announced to the assembled members of Congress. He appeared in a 60 Minutes segment with Diane Sawyer in 1986, helping the show orchestrate an elaborate setup. The producers brought three real-life polygraphers to a magazine office and told them a camera had been stolen. The polygraphers, thinking the crime was real, were told that one employee—a different person for each examiner—was the likely culprit. Each confirmed that, yes, their particular “suspect” was lying about being innocent.
Photographer: Thomas Prior for Bloomberg Businessweek
At this point, Williams was no longer living out of his truck. His father had helped him get a job as a student pastor in the town of Fargo, Okla. He’d preach on Sundays and officiate at weddings and funerals. Otherwise, he was free to use his small church office and ecclesiastical gravitas to wage war on the polygraph.
In 1988, Congress passed, and President Reagan signed, the Employee Polygraph Protection Act, which forbade testing in the private-sector workplace. Liberals and conservatives alike had pushed for the bill: Orrin Hatch (R-Utah) and Edward Kennedy (D-Mass.) authored the Senate version. There were a few exemptions—armored car companies, for example—but not enough to sustain the industry of small businesses that had grown up to provide testing. Most went out of business.
It was a resounding victory—for unions and civil liberties organizations, for psychologists who questioned the machine’s validity, for lawyers whose clients had been fired or imprisoned because of it. And, Williams ambitiously argues, for him. “Why did that legislation finally succeed?” asks the prologue to his self-published memoir, From Cop to Crusader. “Many say it was primarily due to the efforts of Doug Williams.” Feeling he’d earned a sabbatical, he resigned his pastorship, bought a motor home, and drove down to Galveston, Texas. Parking on the beach, he built himself a deck and lived there for seven years, working odd jobs when he periodically ran out of money for food and beer.
As the polygraph disappeared from businesses, though, the government was coming to rely on it more. In the wake of the Aldrich Ames disaster in 1994, the CIA toughened its polygraph testing. The FBI dramatically expanded its testing in 2001 after agent Robert Hanssen was also outed as a Russian spy. In 2010, Congress passed a law requiring all applicants for law enforcement jobs with Customs and Border Protection (CBP) to take the test.
The psychologist Charles Honts has, with Raskin, done many of the studies that polygraph advocates cite. A Boise State University professor and former U.S. Department of Defense polygraph instructor and researcher, he takes issue with the widely cited National Academy of Sciences finding that most polygraph research is plagued by methodological problems. Yet the way the federal government uses the test, he says, “causes me discomfort.”
The test works best, Honts argues, when the questions are about specific events. Job screening questions, however, tend to be broad: “Have you committed serious crimes that went undiscovered?” or “Have you had unauthorized contact with a foreign national?” The difficulty is compounded, he says, in looking for spies or aspiring Edward Snowdens. “There are hundreds of thousands of people who work in the federal government who need security clearance,” he says. “How many people working for foreign governments apply for those jobs? If you’re looking for something that only occurs one-tenth of 1 percent of the time, running a test that’s 90 percent accurate doesn’t help you.” Depending on where you set your threshold, you either miss most of the spies or you cast suspicion on tens of thousands of innocent people. Sometimes you do both.
“We don’t know how many good people we lose, and we don’t know how many bad people have gotten through and haven’t gotten caught,” Honts says. “And we don’t know whether the polygraph is at all predictive of either of those outcomes.”
Williams doesn’t like the term “polygraph countermeasures.” He believes it implies that the polygraph is inaccurate only when people are trying to fool it. Instead, he teaches people to “sting” the device, to trick the examiner the way undercover cops trick their quarry. Over the years, as examiners adapted their methods, he adapted his, too. He no longer teaches the pucker—polygraphs now have a seat pad to catch it—opting instead for mental visualization techniques.
In 1996, soon after Williams tired of life as a middle-aged beach bum and moved back to Oklahoma, he built a website and started charging $47.95 for digital downloads of his manual, at the time still an exotic technological process. Several years ago, he added private lessons: $1,000 if clients came to his office, $5,000 if he had to travel to them. That, and Kathy’s income as a real estate agent, allowed him to acquire some of the trappings of affluence—a Mercedes-Benz SUV, a vintage gold Rolex (bought on EBay), and a kitted-out Harley-Davidson he’d ride therapeutically at high speeds on the interstate.
He had competition. There was Polygraph Consultants of America, started by (and solely consisting of) Indiana electrical contractor Chad Dixon, who’d read Williams’s manual. Russell Ehlers, a small-town Wisconsin police chief, has offered online instruction on passing a preemployment test. A book by activists George Maschke and Gino Scalabrini, available for free on their website, describes exam countermeasures such as controlled breathing, mental arithmetic, and biting one’s tongue.
Williams, though, claimed to be the best: a certified polygraph examiner and expert interrogator with a decadeslong track record of successfully training people. The anonymity of Williams’s clientele makes it impossible to independently evaluate that track record, of course, but he guaranteed to one-on-one students that they’d produce three perfect truthful charts by the end of their lesson. “It is a proven fact that Doug Williams can teach you how to pass any polygraph test, given by anyone, for anything, anytime,” his website read. It was only natural that when federal investigators declared war on polygraph countermeasures, he became a target.
On Oct. 15, 2012, Williams got a call from a man named Javier Domingo Castillo. He told Williams he was a Department of Homeland Security inspector who’d helped a friend smuggle cocaine into the country. The friend got caught and confessed, and as part of the resulting investigation, Castillo was being polygraphed. He was concerned enough that he wanted Williams’s top-shelf option: $4,500 (after some bargaining), plus the cost of a first-class ticket and a hotel suite, for Williams to come to Virginia.
In reality, Castillo was an undercover investigator with CBP’s Office of Internal Affairs, and the sting he set in motion was part of a larger countermeasures crackdown called Operation Lie Busters. It’s perhaps surprising that Williams—a former internal affairs detective who named his polygraph-beating technique the “sting”—would fall for such a trap. What’s even more surprising is that, before he did, he saw through it.
“Frankly, Javier, I SMELL A RAT,” Williams e-mailed on Oct. 17. “You are doing what I told you not to do, and that is telling me that you are lying. ... DO NOT SEND THE MONEY—DO NOT MAKE THE RESERVATIONS—I AM NOT GOING TO HELP YOU!!!!” When Castillo called Williams to ask him to reconsider, Williams angrily rebuffed him: “Either you’re an agent trying to set me up or an idiot. I ain’t working with either one of those guys.” But in recordings later played in court, Castillo’s repeated pleas softened Williams somewhat. He hung up saying he’d think about it.
At that point, according to trial testimony by Douglas Robbins, the CBP agent in charge of the investigation, there were no plans to reach back out to Williams. The investigation was dead in the water, or would have been, had Williams himself not revived it less than two hours later. Castillo received an e-mail from an address he hadn’t seen before, firstname.lastname@example.org, offering a workaround. “If I get a call from a number that I don’t know,” it read, “and an e-mail from a person I don’t know, with a name I have never heard of, I could work with that person, especially if he is telling the truth and is just nervous because he knows it doesn’t work.”
The deal was back on. Castillo called Williams, who asked jocularly if the agent had heard from the paranoid chicken. Warming to Castillo further, Williams decided they didn’t need to bother with the fake e-mail and phone number. “You’re just a desperate man in need of help, am I right?” Williams asked. Castillo said he was.
On Oct. 27, two days early to avoid Hurricane Sandy, Williams flew to Washington. At a hotel in Alexandria, Va., while investigators listened from the next room, he gave Castillo an hour-and-a-half lesson. He explained the difference between relevant and control questions and went through the hypnosis script he used to get his students to relax when asked relevants. Castillo was to imagine being on a beach. “Wave after wave, gently rolling into the shore,” Williams repeated softly in his Oklahoma twang. Then he had Castillo think of something frightening to call to mind to spike his breathing, sweating, blood pressure, and heart rate when asked a control question. He went over potential questions Castillo might be asked and ran the agent through his practice tests, complimenting him on his performance. Then they were done. After Williams got back to Oklahoma City, he tweeted, “Another person is now confident and prepared to pass their polygraph test.”
Four months later, Williams trained CBP Special Agent Luley, the supposed Virginia sheriff’s deputy, and found himself being served with search warrants for his office and home.
On May 12, Williams went on trial in Oklahoma City’s federal courthouse. For training Castillo and Luley, he was charged with two counts of mail fraud and three counts of witness tampering, each charge carrying a maximum sentence of 20 years and a $250,000 fine. Two young lawyers from the Department of Justice’s Public Integrity Section flew from Washington to prosecute him. In the mostly empty gallery were CBP agents, lawyers, and the antipolygraph activist Maschke. A soft-spoken former U.S. Army intelligence officer, Maschke had traveled from The Hague, in the Netherlands, where he works as a translator. He wore a pin that said “I know the lie behind the lie detector” until the judge noticed it and asked him to keep it in his pocket while in her courtroom. Kathy Williams came on the second day of the trial, accompanied by her adult daughter, who wept steadily.
Midway through that day, Williams decided to plead guilty to all five counts. The jury had heard the Castillo recordings but had not seen the Luley video, with the undercover agent’s references to sex with a 14-year-old girl. Standing before the judge and speaking in a chastened voice, Williams read the plea aloud.
While it’s not always prosecuted, it’s a crime to knowingly help someone lie to a federal agent. In December 2012, Dixon, of Polygraph Consultants of America, pleaded guilty to charges similar to Williams’s; he was sentenced to eight months in prison. Training someone is different in the eyes of the First Amendment from writing a book or a blog post describing countermeasures, acts that have a broader civic purpose. “If there’s a bad use and a good use, then the speech is protected,” says Eugene Volokh, a law professor at the University of California at Los Angeles. “It’s hard to really think through whether polygraphs are a good idea without having a sense of whether they’re easily defeatable. But once you get this one-on-one interaction with somebody whose purposes you know, then it stops being dual-use and becomes only single-use. This person is not going to be making policy about polygraph examinations—this guy wants to defraud somebody.”
Still, using criminal prosecution to prevent people from learning how to fool a test doesn’t suggest great confidence in that test’s diagnostic power. And Operation Lie Busters can’t bust the Internet, which has a wealth of information on countermeasures—some of it anecdotal and anonymous, some of it peer-reviewed research. Mark Zaid, an attorney who represents federal law enforcement and intelligence employees, says he’s seen a sharp uptick in clients being denied jobs and clearances because they’re accused of trying to beat the exam. “It’s out of control,” he says. The easier the information is to find, he argues, the more paranoid polygraph examiners become.
Companies are working to bring new types of lie detectors to the market—and, they hope, to the courtroom. Some use electroencephalography or functional magnetic resonance imaging to trace brain activity; a startup called Converus has developed an exam based on eye movement and pupil dilation. Raskin, the polygraph researcher, is on the company’s science team. Focusing for now on Latin America, Converus markets its “preemployment and periodic employee screening solution” to government agencies and the sorts of businesses that in the U.S. have long been forbidden from using the polygraph.
Williams’s website is still online. Immediately after his guilty plea, he took it down, but a few weeks later, mercurial as ever, he reversed himself and put up a stripped-down version. “It’s my way of saying, look, you know, this is what I have done for 35 years,” he explains.
Sentencing could come as early as the end of summer, shortly before Williams turns 70. In the meantime, he studies the Bible, prays, watches TV, and occasionally takes a ride on his Harley. He’s been having panic attacks, and he was recently diagnosed with chronic obstructive pulmonary disease. Speaking on the phone from his house two months after his plea, he’s uncharacteristically subdued. He still rolls out the well-practiced applause lines and the old epithets for his Orwellian nemesis, but he can’t muster the same motivational anger. “It’s a lot easier to pick a fight when you’re 35,” he says, “than to finish it when you’re 70.”