raymond.nelson wrote on Jan 16^th, 2008 at 11:02pm:

Mr. Maschke, I don't believe its splitting hairs to make this point. This point determines whether we are having a credible discussion about the empirical basis for polygraph testing, or simply engaging in some straw-man discussion about a mythical and inaccurate understanding of the test. Words themselves cannot be measured (at least in the way you are implying regarding polygraph tests). Response to stimulus can be measured and understood. The term "lie detector" is simply a term of convenience, not an accurate or empirical description. Epistemological complications aside, discussions about speaking the truth are a distraction from measurable concerns about behavior. r

I don't expect to change any minds, but on the off chance the this writer and others really are interested in the empirical evidence, I invite you to review the analysis I have done on the no-nonsense, real-world, data published in the polygraphers' own literature which shows that reviews negative predictive value (that is calculating the probability that "if someone passes a polygraph, how likely is it that they are ACTUALLY telling the truth?" and positive predictive value (that is that is calculating the probability that "if someone FAILS a polygraph, how likely is it that they are ACTUALLY telling engaging in deception?".  

By the polygrapher's own literature, the answer is: about 55 - 60% for both questions, or, the equivalent of flipping a coin.

On this we base important national security, occasional legal and employment decisions?  The science could give no worse a grade to the theory behind and claims of the value of the CQT polygraph.

Respectfully,

Al Z.

---