WASHINGTON — She was one of the brightest students at a leading university when the Central Intelligence Agency offered her a job as a counter-terrorism analyst. But first, the 19-year-old was warned, she had to undergo a polygraph test to determine whether she could be trusted.
Instead of scrutinizing her ability to guard government secrets, polygraphers asked about her reported rape and miscarriage, the woman recalled. Over at least eight hours in three separate sessions, polygraphers repeatedy demanded to know her innermost thoughts, even after she started sobbing in shame.
“At one point, one of the polygraphers said to me, ‘Turn on the light inside so I can see,’ ” said the woman, who asked that her name be withheld. “I was amazed at how creepy and invasive the whole process was.”
Last year, more than 73,000 Americans across the country submitted to polygraph tests to get or keep jobs with the federal government, although such screening is mostly banned in the private sector and widely denounced by scientists. Many of the screenings probably aren’t as harsh as the CIA applicant described, but polygraphers at a growing number of U.S. agencies are asking employees and applicants questions about their personal lives and private thoughts in the name of protecting the country from spies, terrorists or corrupt law enforcement officers.
The federal government describes polygraph testing as an imperfect but effective way of preventing its secrets from being leaked at a time when almost 5 million people have been approved to access classified information. Many people who undergo polygraph tests describe them as one of the most emotional, terrifying and shameful experiences in their lives. Polygraphers routinely coax people into revealing secrets or experiences they haven’t told their friends, relatives or therapists. The polygraphers record the sessions and keep details of the results, sharing them across the government when someone applies to different agencies.
Scientists, however, don’t know whether polygraph machines can tell whether someone is lying or even withholding information. Some independent studies have concluded that polygraph testing is no more accurate than a coin toss.
Despite such doubts about the tests, Congress and the courts no longer aggressively scrutinize the usually secretive federal polygraph programs. People who undergo the tests often can’t get access to information about their interrogations, and most are barred from filing complaints in federal court.
The National Academies urged federal agencies in 2003 to stop using the tests as a screening technique. The organization, which advises the U.S. government on scientific matters, examined thousands of polygraph studies and concluded that the risk of innocent people failing the test, and spies passing it, was too high.
Since then, 15 agencies – from the National Security Agency to the FBI to the Postal Inspection Service – either have continued or expanded their polygraph screenings, McClatchy found. Many of the agencies now target a growing number of private contractors as well. Only the Department of Energy dramatically scaled back on screening after its own scientists protested.
“The federal government obviously has ignored the scientific consensus,” Stephen Fienberg, the chair of the National Academies’ polygraph panel, told McClatchy. “What we showed, without equivocation, is that the polygraph machine is too blunt an instrument to be relied on for screening.”
The federal government itself hasn’t reached a consensus on the best approach, or even the ethical limits, of screening more than 70 years after adopting the practice. McClatchy interviewed dozens of polygraphers, national security experts and people who’ve been screened and found vast differences in how the tests are conducted.
Six agencies, including the Department of Energy, try to stick to national security questions, such as whether someone has leaked classified information or has inappropriate relationships with foreigners, McClatchy found. These polygraphers are supposed to avoid delving into other personal matters, such as sexual conduct and psychological issues.
Bruce Held, the director of intelligence and counterintelligence for the Department of Energy, said he wanted to avoid relying on draconian security measures that might unintentionally encourage spying by alienating his employees.
“What we’re looking for is whether you are a spy, terrorist or saboteur, not whether you have some peccadillo in your life,” said Held, a retired CIA officer.
The nine other agencies that still use polygraph screening, however, see it as crucial in rooting out applicants or employees who are hiding crimes or deviant or unstable behavior that should bar them from certain jobs. These agencies delve into personal conduct such as past drug use, sexual perversions, undisclosed crimes and financial problems.
The Naval Criminal Investigative Service once asked only national security questions but recently decided to start asking its special agent applicants more personal questions. David Bogue, the head of the service’s polygraph program, said polygraph screening helped the agency root out “potential insider threats” and hire people who were “trustworthy and suitable.”
Details about how the U.S. government conducts polygraph screening are rarely discussed publicly, because many polygraphers cite the need to protect national security and many people who’ve been screened fear being identified. Many federal agencies, including the CIA and FBI, declined to grant McClatchy interviews or respond to basic questions, such as how many people they polygraph.
To prevent abuses in such a secretive culture, inspectors from the government-run National Center for Credibility Assessment routinely scrutinize polygraph programs. William Norris, the director of the center, said inspectors interviewed top officials and reviewed a sampling of test results to ensure that “ethical, professional and technical standards” were being met. Federal polygraphers also receive more than three months of training at the center.
A McClatchy reporter, however, spoke to veteran polygraphers from a wide array of agencies who described how they often rely on their own instincts and experiences to determine the relevance of a topic, comparable to a skillful police interrogation of a criminal suspect. Some polygraphers, for instance, think that asking someone about being raped could be legitimate in certain circumstances. Others disagreed.
“Where is the line? That depends on the polygrapher and the agency,” said John Sullivan, a retired CIA polygrapher of more than three decades. “It can be a slippery slope. At a certain point, the government can justify almost anything.”
One Defense Intelligence Agency employee accused a polygrapher last year of jumping to the wrong conclusions during her screening because she’s a U.S. citizen of Pakistani origin, according to a copy of her complaint. The polygrapher accused her of trying to deceive him when she countered that she was telling him the truth.
The woman, an intelligence officer, previously worked with polygraphers who told her that even truthful people can be seen as having deceptive responses because they second-guess themselves. As a result, she said in the complaint, she concentrated on providing honest responses. However, the polygrapher said he thought she was using what are known as “countermeasures” to prevent him from reading her test results.
The woman no longer works for the Defense Intelligence Agency but she asked not to be named because she still works for the federal government. When she appealed and asked the Defense Intelligence Agency to view the tape, the agency restored her national security clearance.
“Polygraphers aren’t trained scientists,” said Fienberg, who’s a nationally respected statistician. “They haven’t a clue what impact an interrogation has on people’s likely responses.”
One computer systems analyst said she’d failed the National Security Agency’s polygraph screening four times because she was seen as deceptive when she was asked about drug use. The 57-year-old said she’d never tried drugs – even in her 20s in the 1970s.
“I’ve never been a crowd follower,” she said. “When everyone else might have been smoking pot, I was playing the oboe.”
In the effort to pass, the woman said, she offered up habits such as drinking wine every day and responded to questions about whether cheating had led to her two divorces. It had not. One reason she doesn’t want to be named is that she’s held a security clearance for more than a decade with Defense Department agencies that didn’t ask applicants about past drug use during polygraph screening.
The woman speculates that she may be reacting to the drug question because her adoptive sister is a longtime drug addict with whom she’d had to cut off contact.
“But I have no idea,” she said. “Sometimes I feel like I’m not really thinking about anything at all other than the fact that I’m being polygraphed.”
Based on a National Academies statistical model, thousands of people screened last year could have been labeled as failing the test when they were telling the truth. Most courts won’t accept polygraph results as evidence in criminal cases. Even polygraphers acknowledge that their calls could be wrong at times.
“I would say there are people who may flunk the polygraph test because they’re extremely nervous,” said Kevin Boyle, a veteran federal polygrapher who’s overseen more than 1,600 polygraph tests. “As a polygrapher, you have to be aware that it’s never going to be 100 percent accurate.”
For many people – even polygraphers who undergo the screening themselves – the technique has undeniable psychological power. Customs and Border Protection applicants have admitted to cooperating with drug cartels. Counterintelligence officials reveal that they’ve spilled government secrets. One of the more serious criminal confessions is downloading child pornography.
“We have received admissions from people who have had sex with a dog,” said Boyle, who’s a polygrapher with the Postal Inspection Service. “You would be amazed what people admit to. What the research doesn’t show is that it really does work.”
U.S. intelligence agencies nonetheless decline to identify spies who’ve been uncovered during polygraph tests. The CIA is said to have detected spying by at least two employees, including former CIA officer Harold Nicholson, who was convicted in the late 1990s of selling the names of covert officers to Russia.
Others seem to be able to get through the screenings without problems. CIA officer and Soviet spy Aldrich Ames cleared one. Polygraphers then came up with ways to catch people who try to beat the test. Despite their efforts, Bruce Ivins got through an FBI polygraph before being identified as the 2001 anthrax killer.
“Anybody who has been suspended from high school is not going to be intimidated by a polygrapher,” said Rob Caruso, a security consultant and cryptographer who was polygraphed.
Caruso, one of the rare applicants who allowed McClatchy to publish his name, said he was amused by his polygraph screening for the Naval Network Warfare Command because he saw it as such a crude way of detecting security risks.
During the test, Caruso said, the polygrapher seemed intent on getting him to confess to something, so he admitted getting caught spray-painting a drainage ditch as a teenager, “to give them something.” However, he didn’t admit to drug use although he’d experimented when he was younger. He passed.
“They were real squares,” Caruso said. “They were asking questions like I’m going to cop to things.”
Polygraphers said they weren’t trying to be intrusive when they asked about such matters. Many people initially show as reacting to questions even though they may not be lying. They may be thinking of something irrelevant or embarrassing that they don’t want to reveal. The polygraph machine, however, measures their blood pressure, sweat activity, respiration and heart rates as possibly indicating deception. To determine what’s making them react, polygraphers talk people into telling them what’s on their minds.
Boyle said he thought that polygraphers should “lightly probe” sensitive areas such as someone who professed to be the victim of a crime.
“I believe we should leave it up to the applicant to share the information,” he said.
Other polygraphers may not be as gentle, he acknowledged. “Everyone’s interrogation style is different.”
Once the troubling thought or behavior is revealed, the test is redone and many people pass.
Others don’t. The CIA applicant who says she was asked in 2005 about her rape recalled being pressed to reveal everything because the test showed that she was hiding something. At one point, she revealed she was bothered by a claim by another polygrapher that the agency hires alcoholics but not homosexuals. The polygrapher leaned in and asked, “Why? Are you a homosexual?” She said she wasn’t.
McClatchy was able to confirm the CIA’s interest in her reported rape and miscarriage by reviewing her heavily redacted security-clearance records, which were released under open records laws. The records show that she revealed before the polygraph that she’d smoked pot three times to deal with the pain of her miscarriage, records show. She also told them the miscarriage was the result of a rape.
The applicant wonders whether her state of mind at the time made her seem as if she were lying when she had nothing to hide. As the agency itself notes in records McClatchy obtained, the woman had been preliminarily diagnosed with post-traumatic stress disorder.
Her security-clearance request was rejected because of “personal conduct.” McClatchy couldn’t determine what the agency was most concerned about because of the redactions in the documents. Like most agencies, the CIA refused to give the applicant a copy of her recorded polygraph screening, citing policy.
The woman offered to sign a privacy waiver allowing the CIA to discuss her case but the agency said it couldn’t discuss individual cases. In a general statement about its program, the agency defended its polygraphers as “highly trained . . . among the best in their profession.”
Polygraphers who work at several agencies said asking someone about being sexually assaulted or raped could be legitimate in some cases, especially if the person brought up the topic and it seemed to be triggering a deceptive reading.
“A false rape report is a felony,” said one polygrapher, who wasn’t authorized to speak by name. “Would you want someone who files a false report of a crime to hold a sensitive government position?”
Sullivan agreed that polygraphers have a duty to follow up if someone implies that he or she has leveled any false criminal accusation. However, he said that pursuing details about someone’s rape “clearly crosses the line.”
“I can’t see how you can get anyone who is a victim of a rape to get through a polygraph once the emotional response is triggered,” he said. “I would question whether this is something that they really need to know about in the first place.”
Tish Wells contributed to this article.