Polygraph tests used by nearly every federal national-security agency as a screening tool will flag loyal workers as security risks and free actual spies from suspicion, a panel of top scientists reported Tuesday.

Gathered by the National Research Council, scientists said the theory and research supporting polygraphy is too weak and the accuracy of the test is "insufficient to justify reliance on its use in employee security screening."

"National security is too important to be left to such a blunt instrument," said panel chairman Stephen Fienberg, a statistics professor at Carnegie Mellon University in Pittsburgh.

Two lawmakers called on the U.S. Department of Energy to replace its polygraph screening program, targeting 16,000 employees mostly in California, New Mexico and Washington, D.C., with a testing program solely for interrogation of suspects.

Yet beyond the Energy Department and its national labs -- Livermore, Los Alamos and Sandia -- the polygraph is deeply embedded in the U.S. national-security apparatus, with an estimated 40,000 workers or applicants tested every year at the CIA, Defense Department, National Security Agency, Secret Service, DEA and -- in the wake of the Aldrich Ames and Robert Hanssen spy cases -- the FBI.

Thousands more are tested at state and local law-enforcement agencies. This summer, many in Congress

who voted to polygraph nuclear weapons scientists were themselves "put on the box" in an FBI search for leaks at the Senate and House intelligence committees.

Inventors such as psychologist and feminist theorist William Moulton Marston -- later known for creating Wonder Woman, whose lasso compelled truth telling -- devised polygraphy to interrogate World War I spies. The polygraph became hugely popular over the next 80 years, and no one has been more captivated by its mystique than Americans and their law officers.

Yet, said NRC panelist Kathryn Laskey, a professor of systems engineering at George Mason University, "We stress that no spy ever has been caught using the polygraph."

The conclusions of the 310-page report are not new. Scientists have criticized polygraphs as poorly grounded and researched since their creation.

The 310-page NRC report, however, is among the most comprehensive and authoritative on the subject, and the first to highlight the national security risks of growing federal reliance on a test that invariably clears the spies and saboteurs it was designed to catch.

Employees of the nation's three nuclear-weapons labs hailed the report as powerful vindication, in large measure because it echoed their attacks on the scientific foundations of polygraphy and found them equally weak or nonexistent.

"It's time to stop it, for everybody," said Lawrence Livermore National Laboratory physicist Jeff Colvin, president of the Society for Professional Scientists and Engineers, a labor union.

"It doesn't get any better than this. There's no wiggle room here," said Dr. Alan Zelicoff, a physicist and physician at Sandia National Laboratories in Albuquerque, N.M. "We've been spending millions of dollars on a test that is not worthless, but worse than worthless because it does more harm than good."

In 1999, Congress went into a lather over suspected Chinese thefts of U.S. nuclear secrets and instituted polygraph tests for thousands of career nuclear-weapons employees. Scientists denounced the tests as "voodoo" and "junk science" that insulted their dedication to national-security work.

"You're talking about people who for the most part are very loyal and find it terribly offensive that their loyalty is questioned," veteran Livermore weapons designer David Dearborn said Tuesday. "Then you have an undependable piece of electronic flimflammery, and someone pops up and says 'I think you're being deceptive,' and your clearance is pulled. ... What are we getting as a nation in return? We're getting political cover at best. Because if that's the best we can do to catch spies, we're in trouble. You're not catching the people who are spying, and yet you are having large numbers of people suffer as they're treated like criminals."

Rep. Ellen Tauscher, the Alamo Democrat whose district includes Livermore and Sandia labs, persuaded the Energy Department to limit the population of polygraphed scientists to a minimum. It now includes somewhere in excess of 700 at both labs, mainly workers who handle weapons parts or weapons materials or have access to highly classified intelligence information.

In the wake of the report, Tauscher said she still supports screening polygraphs, used with other tools such as background checks, for those scientists, as long as the Energy Department's National Nuclear Security Administration doesn't over-rely on them.

Polygraphs, Tauscher said, "have been a silver bullet, and this report tells us what we knew intuitively -- that they are not."

Contact Ian Hoffman at ihoffman@angnewspapers.com

New Mexico senators Jeff Bingaman and Pete V. Domenici, who represent employees of Los Alamos and Sandia in Albuquerque, called on NNSA to abolish the screening tests.

"Polygraph tests may have a role to play in law enforcement, but they don't work as a screening tool for our national laboratories," Bingaman said. "In the panic to protect classified information, Congress rushed to implement a policy that had the effect of treating prospective lab employees as suspects. From a practical standpoint, this policy never made sense to me. Now we have scientific evidence that it doesn't work. It's time to change this flawed policy."

NNSA acting administrator Linton Brooks said the NRC report pinpointed the problem: "How to administer a program that is maximally effective at 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 we will examine very seriously in the coming months."