Polygraph Testing: Liar, Liar

Polygraph testing may be an effective plot device in spy movies, but a recent report issued by the National Research Council (an arm of The National Academies) found that when it comes to employee security screenings, polygraphs are deeply flawed.

Polygraph testing may be an effective plot device in spy movies, but a recent report issued by the National Research Council (an arm of The National Academies) found that when it comes to employee security screenings, polygraphs are deeply flawed. The 245-page report is the result of 19 months of study by a committee of statisticians, psychologists and mathematicians. The Department of Energy commissioned the report after embarrassing allegations of theft and espionage at its labs.

According to the report, one of the problems with using polygraphs to screen large numbers of employees is balance. Test administrators face a dilemma between two unappealing choices. "If you set the threshold for measuring deception too low, then many truthful people will get labeled as deceptive," says Stephen Fienberg, a professor of statistics and social science at Carnegie Mellon University and chairman of the panel. "If, on the other hand, you raise the bar because you don't want as many false positives, then too many deceptive people will be labeled as truthful."

This problem doesn't even factor in the use of countermeasures by subjects to disguise false responses as truth. Studies suggest that mental exercises like silently counting backward can alter the physiological response to a question.

While private companies may not use the polygraph, the government is free to do so, and employees of companies with government contracts often have to go through a rigorous government screening.

Unfortunately, there is no viable alternative to the polygraph at present. Researchers are looking at technologies like fMRI (functional magnetic resonance imaging) to monitor brain patterns, voice tremor detection and thermal imaging, but they won't be ready to replace the polygraph for years, if ever.

However, Fienberg and his committee still believe that some good can come from releasing their findings. "People have an overconfidence in the use of the polygraph, and their belief in its accuracy is perhaps worse than the risks of not doing [this kind of testing]," says Fienberg. He hopes that the report will shake the confidence that some government departments have placed in the polygraph and that it will lead to a greater interest in researching alternatives.

Recommended
Join the discussion
Be the first to comment on this article. Our Commenting Policies