Countermeasures and control questions

[Red]

I have a question in regards to this statement,

"The polygrapher assumes that if your physiological responses as
measured by the polygraph are stronger when answering a relevant
question (e.g. "Have you violated this agency's guidelines concerning
the use of illegal drugs?") than when answering the "control" questions
(e.g. "Have you ever lied to a loved one?"), then you must
have been deceptive in answering the relevant question. If your
physiological responses while answering the "control" questions are
greater, then you must be telling the truth in answering the relevant
question. And if your physiological responses while answering the
relevant and "control" questions are about the same, then the outcome
will be deemed inconclusive. If these assumptions seem overly
simplistic to you, you're right. As we stated at the beginning of
Chapter 1, polygraphy is not science: it is codified conjecture masquerading
as science."

If you are using countermeasures during control questions and answering with lies to relevant questions, would this not produce results that are equal?

Further, if the participant is choosing to use countermeasures does he/she answer yes to known control questions or only respond "no", knowing that no is the desired response?

Thanks

James.

[George W. Maschke]

James,

It is certainly possible that a person using countermeasures during "control" questions might fail to produce reactions as strong as those he/she may produce without countermeasures to the relevant questions. But there is no "Pinnochio response" that the polygraph measures each time a person lies, and there is no basis for assuming that a person who lies in response to relevant questions while self-stimulating in response to "control" questions will end up producing equal responses to both. In studies by Dr. Charles R. Honts and collaborators (cited in The Lie Behind the Lie Detector), about half of subjects provided a maximum of 30 minutes of instruction on polygraph procedure and countermeasures succeeded in beating the "test" (i.e., producing stronger reactions to the "control" questions than to the relevant questions).

As for the expected response to probable-lie "control," questions, in most cases it's "no." For example, "Did you ever lie to a loved one? Did you ever intentionally hurt someone? Did you ever take something that did not belong to you?" However, there are some "control" questions to which the expected answer is "yes." For example, "Are you a really honest person?

[Marty]

If you are using countermeasures during control questions and answering with lies to relevant questions, would this not produce results that are equal?

This is in fact an assumption made by Matte in specific incident poly. His scoring "quadra zone" methodology has a non-linearity when the responses are the same between comparison and relevants to produce a deceptive outcome. This is described as a counter-countermeasure. Highly debatable IMO. One possible side effect is that the DI and NDI distribution curves favor a DI outcome which is the case with his published distributions.

-Marty

[Commander]

Marty/George,

All that being said, is there empirical data to suggest how reactions to relevant questions may be reduced? If reactions to both controls and relevants are equal, then it seems to negate the effectiveness of the countermeasure.

I have heard several different schools of thought on this matter. Some suggest meditative thought, or thinking pleasant thoughts not related to the exam. Still, I have heard others say that there is no effective way to control the autonomous physiological responses to relevant questions.

Doug Williams indicates that it is familiarity with both the polygraph process as a whole as well as the types of relevant questions likely to be asked that creates a comfort level sufficient for minimizing reactions.

I am curious what techniques seem to be the most effective by people who have successfully employed them. Would it be feasible, perhaps, to show reactions to a combination of both irrelevant and known-lie controls?