Accuracy of polygraph techniques: Problems using confessions to determine ground truth

Mangan et al. [D.J. Mangan, T.E., Armitage, G.C., Adams: A field study on the validity of the Quadri-Track Zone Comparison Technique. Physiol Behav 2008] have carried out a field study of polygraph test accuracy in which they relied on confessions to determine guilt as well as to clear co-suspects in the same case as innocent. Using this criterion for ground truth, they estimate polygraph accuracy by determining how often confessions are matched by failed polygraph tests and how often those cleared by confession have passed polygraph tests. They conclude that the polygraph was "100% accurate in the identification of the innocent and guilty." However, their method contains a flaw, not discernible by reading their article, that invalidates this conclusion. The flaw arises because confessions were obtained by the polygraph examiner who interrogated the examinee after deciding the test was failed. Under these circumstances, the criterion (the confession) and the test outcome (deception indicated) are not independent. The method thus virtually guarantees that the two will match, ensuring 100% "accuracy." Although largely ignored by the polygraph profession, this flaw inherent to confession-based field studies of polygraph validity has been known to confound these studies for over two decades. Hence, contrary to Mangan et al., their study design does not provide for an adequate estimate of polygraph test accuracy. Moreover, reviews of polygraph testing carried out by scientists at arms length to the polygraph profession have repeatedly failed to support the accuracy proponents claim for the polygraph.