gling of these separate contributions; however, few of these concepts and methods have been used in polygraph research. Moreover, applied polygraph research has not for the most part taken advantage of advances in the psychophysiology and neuroscience of emotion, motivation, attention, and other processes that can affect the measures taken in polygraph testing (see, e.g., Coles, Donchin, and Porges, 1986; Cacioppo and Tassinary, 1990b; Cacioppo et al., 2000).

Polygraph research has not paid sufficient attention to advances in inductive inference in psychophysiology that have underscored the need to examine the specificity as well as the sensitivity of the mapping between a psychological state and a physiological manifestation (Strube, 1990; Cacioppo and Tassinary, 1990a; Sarter, Berntson, and Cacioppo, 1996). Specificity of the polygraph is threatened by any physiological process unrelated to deception that can systematically affect polygraph test scores.17 We have found very little research on ways that conditions other than deceptiveness might produce records that are judged deceptive and no evidence of any systematic attention to threats to specificity. As discussed in more detail in Chapter 5, empirical validation studies of the polygraph continue to emphasize the ability to make physiological differentiation between known lying and known truth-telling.

A particularly important gap is the absence of any theoretical consideration of the social (e.g., interpersonal) and physical context of the polygraph test. As already noted, an extensive basic scientific literature in social psychology and sociology details the myriad effects of perceptible personal features (e.g., status, race, gender), dispositions (e.g., traits), and histories (e.g., examinee expectancies, cultural norms, and values) on social perception (e.g., examiner expectancies) and on psychological and physiological processes within individuals (e.g., Shapiro and Crider, 1969; Waid, 1983; Cacioppo and Petty, 1983; Gardner, Gabriel, and Diekman, 2000; Hicks, Keller, and Miller, 2000; Blascovich et al., 2001b). We found no study of the mechanisms by which such variables might affect polygraph test outcomes: for instance, of the effects they might have on the selection of comparison questions, on the examinee’s understanding of the questions and the examination, or on the examiner’s behavior, subtle and otherwise, during the examination.

In short, the bulk of polygraph research, including almost all the research conducted by federal agencies that use the polygraph, can be accurately characterized as atheoretical. Studies report on efforts to improve accuracy by changing methods of test administration, physiological measurement, data transformation, and the like, but they rarely address the underlying psychological and physiological processes and mechanisms that determine how much accuracy might be achieved. Thus,

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