ON THE ISSUE OF ANXIETY DISORDERS AND THE POLYGRAPH’S FALSE POSITIVE RATE


Here’s a reply I sent on October 18 to an inquiry from an individual who made the following email inquiry:


---------- Forwarded message ----------
Date: Thu, 19 Oct 2006 00:11:28 -0400 (EDT)

I used to work as a psychotherapist and was asked this question today.
> In my search online I came up with your name and thought possibly you might know the answer.
> So here is the question: If a person has an anxiety disorder could that create a false positive on a polygraph test?



My email reply was as follows (individual’s name deleted): 

       If you go to http://www.psych.utoronto.ca/~furedy/polygraph.htm, and look especially at table 1 of the paper http://www.psych.utoronto.ca/%7Efuredy/Papers/ld/North%20American%20Polygraph%20...,
you'll see the background for what may seem a rather dogmatic answer to your question **below.  All the best, John



> 
**It's highly likely that it would, especially because the polygraph "test" is not a test at all but an unstandardizable dynamic
interview which is grossly biased by the level of anxiety that the examinee feels, not in general, but about the crime of which s/he issuspected of.

If you look at table 1, you'll see that the examinee's fate (being classified as "deceptive", "truthful", or "inconclusive") depends
entirely on the size of the autonomic response s/he gives to the relevant questions in comparison with the "control" questions.
Normally, even for an Innocent examinee, since polygraph cannot distinguish between the anxious innocent and the anxious guilty, it is
likely that the relevant question (eg., did you steal the money) will produce greater responding than the "control" questions (e.g., "did
you do anything you were ashamed of, aside form the issue under investigation), which, in the case of the innocent examinee, will lead to
a classification of "deceptive").  If an individual actually suffers from an anxiety disorder, it is at least likely that the strain
associated with being a suspect will increase the magnitude of the response to the relevant question.

There is also another danger with someone who has an anxiety disorder, of which they are ashamed.  In this case, as detailed in the
"polygrapher's dilemma paper (http://www.psych.utoronto.ca/~furedy/Papers/ld/CQT%20Polygraphers%27%20Dilemma.d...; see footnote detailing
FBI's attempt to suppress this paper), the cost of being classified as "truthful" may be damage to the psyche of the examineee, as the
only way to boost responding to the "control" question is to increase its relative significance to the examinee.  As a hypothetical
example, suppose the examinee is innocent (has not stolen the money), but his anxiety disorder stems from his conviction that has had
inappropriate sexual relations with his mother, a belief that he has not revealed even to his therapist.  A "control" question that asked
about this issue (recalled that in the pre-test interview, a polyrapher can investigate this issue, and then make up a "control" question
accordingly), could absolve the examinee of criminal guilty, but could cause permanent damage to his psychic economy.  In fact, this
could happened even in a criminally guilty examinee who is more concerned about being found guilty of theft than being found (morally)
guilty of having sexual relations with his mother.

       Sorry, I've gone on for longer than I intended to, but hope this answers your question about what I regard as America's 2nd most
serious (after AIDS) social disease (see also www.antipolygraph.org).

       All the best, John
------
Postcript to my email as of Nov. 9 to www.antipolygraph.org:

     More generally, the question of how any psychological or physiological factors affect the accuracy of a test can only be raised sensibly with a genuine, standardizable test.  The accuracy of an unstandardizable interview cannot be specified in the first place, because that accuracy (and the result) depends too much on the context, the beliefs of the examinee and examiner, and the rapport between the examinee and the examiner.  So while it is possible to determine the (approximate) accuracy of an IQ test given by a competent tester (e.g., an IQ test is less accurate than a blood test which, in turn, is less accurate than a DNA test), and also to determine (approximately) the deterioration in accuracy produced by such factors as sleep deprivation leading to fatigue in the testee, none of this is possible with the unstandardized and unstandardizable polygraph, which is not a genuine psychological test.