Polygraph Is Poor Tool for Screening Employees
By WILLIAM J. BROAD
October 8, 2002
http://www.stat.cmu.edu/~fienberg/Polygraph_News/nyt-10-8-02.html In a report to the government, a panel of leading scientists said today that polygraph testing is too flawed to use for security screening. The panel said lie detector tests do a poor job of identifying spies or other national-security risks and are likely to produce false accusations of innocent people.
The 245-page report, by experts convened by the National Research Council, an arm of the National Academy of Sciences, said the scientific basis for polygraph testing is weak and much of the research supporting its use lacks scientific rigor.
The report is not the first time experts have questioned the reliability of lie detector testing, which is known to have limitations and whose admissibility in court, for example, is sharply limited. But it is the first by the academy, and private security experts said its findings could erode support for polygraph testing inside the federal government. Defense and intelligence agencies use it tens of thousands of times each year to screen prospective and current employees for espionage.
"National security is too important to be left to such a blunt instrument," said Stephen E. Fienberg, a professor of statistics and computer science, Carnegie Mellon University, Pittsburgh, and the panel's head. "The polygraph's serious limitations in employee security screening underscore the need to look more broadly for effective, alternative methods."
The report said lie detector tests, which measure pulse and breathing rates, sweating and blood pressure, have some usefulness in the investigation of particular crimes, though they are far from completely reliable. But in routine security screening they often flag innocent people as lying, while missing actual security risks, the panel said.
"No spy has ever been caught using the polygraph," Kathryn B. Laskey, a professor of systems engineering at George Mason University, and a member of the panel, told reporters yesterday.
The council is the research arm of the National Academy of Sciences, an organization of eminent scientists chartered by Congress to advise the government on scientific matters. The report is the academy's first formal assessment of polygraph testing. The panel performed its evaluation by reviewing previous research on lie detector tests and by visiting centers where the tests are performed and developed.
White House officials said the academy report would be studied carefully. "It's important to note," added Gordon D. Johndroe, spokesman for the White House Office of Homeland Security, "that polygraph examinations are one small part of a very comprehensive background investigation" for people in the government's most sensitive programs.
Recently, worries over possible atomic espionage prompted widespread use of polygraph screening at the nation's nuclear weapons laboratories, which are run by the Energy Department. The department requested the 19-month study and financed it.
"It could be decisive in the outcome of the long-standing debate" over the polygraph's reliability, said Steven Aftergood, a security expert at the Federation of American Scientists, a private group in Washington. "People will still want to use it but they'll no longer be able to say experts disagree" on its usefulness. "They don't."
Paul Gianelli, a law professor at Case Western Reserve who studies scientific evidence, said the panel's report was sure to stir wide debate and would probably reduce polygraph usage.
"This is going to be tremendously controversial because of the institutional investments in the polygraph," he said. But he said the report would be influential because its attack on polygraph testing was so fundamental. "It's saying you've not done the kind of science research that is required," he said.
Congress instructed the Energy Department to adopt polygraph screening in reaction to the case of Wen Ho Lee, a scientist at the Los Alamos weapons laboratory in New Mexico who was suspected of being a spy but freed from jail after admitting to a security violation.
An Energy Department official said that about 16,000 people in sensitive programs were targeted for polygraph testing and that examiners were doing the tests at a rate of about 2,000 people a year.
Laboratory workers have denounced the tests as demoralizing because of false positives and say some scientists have quit in protest and that others have avoided taking jobs there, potentially weakening the laboratories.
Linton Brooks, acting administrator of National Nuclear Security Administration, which runs the labs for the Energy Department, said it would carefully review the academy report and "consider the findings of the study in developing a new polygraph program over the next several months."
Congress has called on the department to take the academy's polygraph study into account in redesigning its program, and to come up with a new plan in six months.
The academy panel, Mr. Brooks said in a statement, "identified the fundamental conflict that we in the national security community must address: how to administer a program that is maximally effective in weeding out security risks while minimizing damage to the vast majority of loyal, patriotic employees. There is no easy answer, but it is a question that we will examine very seriously in the coming months."
To detect lies, polygraphs measure physiological changes as a window into the emotional state of liars. The theory is that their physical responses are unique. But critics say all kinds of emotions - hate, anxiety, depression - can cause a truthful individual to fail a polygraph exam.
Previously, the most thorough study of their reliability was published in 1983 by the Congressional Office of Technology Assessment, which has since been disbanded. It raised the prospect that spies "may well be the most motivated and perhaps the best trained to avoid detection" by developing skills at deceiving polygraphs.
The academy study, said Mr. Gianelli of Case Western Reserve, "is the most important report since then, and a lot more research has been done in those 19 years."
Still, the 14-person academy panel ended up faulting the evidence in favor of polygraphs as scientifically weak. It did so after examining large numbers of reports and visiting polygraph centers at several government agencies, including the Department of Defense's Polygraph Institute at Fort Jackson, S.C.
It illustrated the "unpleasant tradeoffs" in a chart showing a hypothetical population of 10,000 government employees that includes 10 spies.
If the tests were sensitive enough to detect 80 percent of the traitors, 1,606 employees would fail the test - 8 of them spies and 1,598 of them loyal. And if the sensitivity of the tests were lowered to avoid such imprecision, the number of false alarms would fall to 41 but 8 of the 10 spies would also escape detection.
The "unacceptable choice" for the Energy Department, the report said, lies between "too many loyal employees falsely judged deceptive and too many major security threats left undetected."
Panel members said the goal of accurate lie detection was not impossible but that polygraph science tended to avoid innovation and rely on antiquated physiological signs.
"We have newer measures that have not been applied" Marcus E. Raichle, a panel member from the Washington University medical school in St. Louis, told reporters yesterday at a news conference. He cited such methods as monitoring brains waves and heat changes on a person's face.
The panel, he added, found none of these newer methods "ready for prime time."
The panel called for more federal research on polygraph alternatives, saying a few million dollars a year could go a long way.
"What this committee does not want to see happen," said Dr. Fienberg of Carnegie Mellon, is a future scientific panel saying a decade from now that "the polygraph is not the right instrument and we still don't have a substitute."
triple x
