and systematic, into the results of polygraph examinations. The implications of these errors for polygraph test interpretation depend on the nature of the error. If errors were known to be randomly distributed across individuals and physiological indicators, they would be reduced by multiple measurement across multiple channels—an approach commonly used in polygraph testing.

Of more serious concern are sources of error that may reflect consistent rather than random causes and that may lead guilty individuals to appear truthful on the test or innocent ones to appear deceptive, thus reducing the accuracy of the test. We have noted that one cannot rule out, on theoretical grounds, the possibility that polygraph responses vary systematically with characteristics of examiners, examinees, the test situation, the interview process, and so forth.14 Such factors may cause systematic error in polygraph interpretation and need careful consideration, especially if basic scientific knowledge suggests that a particular factor might systematically affect polygraph test results. It is convenient to distinguish two classes of potential sources of systematic error: those that derive from stable or transient characteristics of examinees or examiners (endogenous factors) and those that derive from factors in the social context of the polygraph examination.

Endogenous Factors

Among the characteristics of examinees and examiners that could threaten the validity of the polygraph are personality differences affecting physiological responsiveness; temporary physiological conditions, such as sleeplessness or the effects of legal or illegal drug use; individual differences between examiners in the ways they conduct tests; and countermeasures. For such conditions to threaten the validity of the test, they would have to differentially affect responsiveness to relevant and comparison questions (e.g., by reducing a guilty examinee’s responsiveness to relevant questions). Although there have been studies of the effects of some personality variables and some drugs on polygraph detection of deception (see Chapter 5), there have been few systematic efforts to ascertain whether and how any such relationships might vary across the particular indicators used in polygraph testing. We have not seen persuasive scientific arguments that any specific personality variable would influence polygraph accuracy. If such effects were found to exist, however, it would be possible in principle to use information on the personality variable to adjust polygraph test scores.

An example of an endogenous factor that could be imagined to decrease the specificity of the polygraph, mentioned at our visit to the U.S. Department of Energy (DOE), is what was termed the “guilty complex”—



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement