With the release of Christine Blasey Ford’s lie detector test record comes the revelation that proprietary (that is, secret) software developed at the Johns Hopkins Applied Physics Laboratory, called PolyScore, was used to grade her performance as “No Deception Indicated—Probability of Deception is Less Than .02.” The gaping problem behind this sciencey statement, whether you want to believe her or not, is politely alluded to in a 2003 report from the National Research Council, which found such algorithms to be untethered from the “psychophysiological processes underlying the signals being recorded, except in a heuristic fashion.” That is, the supposed validity of the assumption that measurements made by a polygraph machine, such as skin conductance or breathing rate, can indicate when someone is lying comes from the reported experience of polygraph examiners, not from scientific research–two fundamentally different sources of information. This has always been and always will be the crucial flaw in lie detector tests. Moreover, multiple references are made throughout the NRC study to APL’s refusal to provide data about PolyScore necessary to fully critique it. APL can write complex scoring algorithms until hell freezes over, which the federal government will be happy to fund for at least that length of time, but never escape the a priori error of positing that lying makes everybody sweat, pant, and throb. The Hopkins brand name is golden window dressing for the black-box in Laurel.
Either Brett Kavanaugh or Christine Blasey Ford may be lying, but a polygraph test is tangential at best to the matter. It seems unlikely, given Ford’s scientific literacy, that she could be unaware of the polygraph’s methodological weaknesses. She testified today before the Senate Judiciary Committee that when she was interviewing attorneys, they asked her if she would be willing to take the test. She said she took it “based on the advice of counsel.” She added that she had never taken such a test before, discussed how to take one, or given other people such advice. She said that she was “scared of the test itself.” That she submitted to it on August 7 should not be held against her. Rather, here is prominent evidence of its bizarre ubiquity in Washington political and legal culture.
10/1/18: And a friend at Harvard Law. Everybody does it, so should you.