John Espenschied writes for Computerworld in “Beyond Polygraph Testing (The DOE Goes Over a Cliff)” (17 June 2007), criticizing the Energy Department’s move to embrace random polygraph screening and suggesting more meaningful alternatives for enhancing security. Excerpt:
June 17, 2007 (Computerworld) — The National Academy of Sciences conducted an analysis of the Department of Energy’s security screening policy and published the report “The Polygraph and Lie Detection” (NAS Press) in 2003. The review indicated polygraph tests are ineffective because of high rates of false positives and susceptibility to behavioral or physical manipulation. The DOE is concerned with preventing actions by malcontent scientific researchers, computer analysts and others with access to sensitive data, and for years it has used lie detectors in the background check and risk assessment process for job applicants. Those are reasonable concerns, but the DOE is right to review its policy in light of newer data.
However, instead of doing away with this nonsense entirely, the new DOE policy actually expands the use of random polygraph assessments while reducing periodic assessments. This is precisely the wrong way to go.
We’re still betting that a galvanic skin change is an accurate predictor of malice? Does anyone really think decreasing blood pressure differentiates the inner placidity of someone with nothing to hide from a perfectly content psychopath? A polygraph may be a useful prop for a skilled interrogator developing a subjective opinion about a person, but when it comes to predicting human risk or truth of statements, the results of polygraph testing are statistically indistinguishable from coin-tossing.