Georgew, the Washington Time archive is 7 days. After that, all articles are for paid.

The truth is polygraphs lie
Steve Chapman
CREATORS SYNDICATE
Published 10/16/2002

     In May 1978, four men were arrested by Chicago police for murdering a suburban man and raping and murdering his fiancee. All the suspects claimed they were innocent, but there was no real doubt about their guilt: Three of them, after all, had failed a polygraph exam.
     Eventually, the Ford Heights Four, as they became known, were convicted for these brutal slayings, and two of the defendants were sentenced to death. But in 1996, DNA evidence exonerated all four. They had spent 18 years behind bars, partly because the lie detector lied.
     A report issued last week by the National Academy of Sciences recommended that the federal government stop using polygraphs to screen for security risks. Why? Because, in the words of the study, these devices are "intrinsically susceptible to producing erroneous results." That's academese for "I wouldn't trust one as far as I could throw it."
     The Energy Department adopted polygraph screening of employees in response to the case of Wen Ho Lee, a scientist accused of spying for China but convicted of only a minor security violation. DOE now tests about 2,000 people a year. But George Mason University systems engineering professor Kathryn Laskey, a member of the NAS committee, noted, "No spy has ever been caught using the polygraph."
     There are particular dangers in subjecting lots of people to polygraphs in the effort to find a few wrongdoers, because false positives will greatly outnumber "true" positives. Some employees who have done nothing wrong will nonetheless have physiological reactions that look suspicious. Some accomplished liars will be able to fool the machine.
     To nab 8 out of every 10 real spies, the NAS report found, the device would probably have to erroneously implicate nearly 1,600 people. If it were set to minimize false positives, 80 percent of the real spies would slip past. But even then, 20 innocent people would be flagged for every guilty one.
     The same fallibility that renders these machines unusable for employee monitoring makes them dangerous for criminal investigations as well. Police and prosecutors regard polygraph results as the closest thing to a dead-bang certainty. But that faith lacks any foundation. "Almost a century of research in scientific psychology and physiology provides little basis for the expectation that a polygraph test could have extremely high accuracy," concluded the panel.
     And there is no reason to think better technology will help. People simply don't respond in a clear and predictable way to questions about what they may have done wrong. The "inherent ambiguity of the physiological measures used in the polygraph suggest that further investments in improving polygraph technique and interpretation will bring only modest improvements in accuracy," said the report. Polygraphs are a crude instrument that can't be refined.
     The consequences of a misleading polygraph exam are bad enough in the employment arena, where someone can lose a job or not be hired. But they're much worse for criminal suspects, who can be locked away or even put to death because their pulse rate rose too much in a stressful situation.
     A polygraph result generally can't be used as evidence in court. But some states allow the information if both the prosecution and the defense concur. So prosecutors may offer suspects the opportunity to clear themselves. Innocent suspects sometimes feel they have nothing to lose and much to gain from going along — only to fail the test.
     A couple of weeks ago, one Jimmy Williams was officially cleared by an Ohio court after spending 10 years in prison for the alleged rape of a 12-year-old girl. In fact, the rape never happened, but the Akron man nonetheless managed to fail a polygraph exam. Because his lawyer had agreed in advance to admit the results, the jury was told the lie detector had implicated him.
     Other defendants have been victimized not only by the polygraph itself but by its aura of infallibility. Gary Gauger was sentenced to death for the murder of his parents on their McHenry County, Ill., farm but was eventually exonerated. He took a polygraph during his interrogation, and the results were inconclusive. But the police told him he had failed it.
     He was so rattled by the news that the cops were able to get him to speculate aloud how he might have killed his parents. Those statements were then used to convict him of a crime he never committed.
     Our medieval forebears had their own lie detector test: Suspected witches were dunked in water, on the theory that the guilty would float and the innocent would sink. Polygraphs aren't quite as preposterous, but they're bad enough.