Expectancy Effects

Given the operation of expectancy effects in many social interactions (see discussion in Chapter 3), one might expect that examiner expectancies of examinee guilt might influence not only examiners’ judgments of charts, but also examinees’ physiological responses during polygraph tests. However, we could find very little research on these issues. In one study, expectancies affected examiners’ scoring of charts that had previously been judged inconclusive, but not of charts with conclusive results (Elaad, Ginton, and Ben-Shakhar, 1994). We found only one small study (28 polygraph examinations) that considered the effects of examiners’ expectancies that were induced in advance of the polygraph examination (Elaad, Ginton, and Ben-Shakhar, 1998): The expectancy manipulation produced no discernible effect on test results. This evidence is too limited to draw any strong conclusions about whether examiners’ expectancies affect polygraph test accuracy.

There is a small body of research on the effects of examinees’ expectancies, conducted in part to test the hypothesis that so-called stimulation tests, which are intended to convince examinees of the polygraph’s ability to detect deception, improve detection accuracy. Although the results are mixed, the research provides some support for the hypothesis (e.g., Bradley and Janisse, 1981; Kircher et al., 2001).

Drug Effects

The potential effect of drugs on polygraph outcomes has received scant attention in the experimental literature. An early report examined the possible effect of the anxiolytic meprobamate (sometimes prescribed under brand names including Equanil and Miltown) on a concealed information task (Waid et al., 1981). This experiment was performed on a small sample of undergraduates and found that meprobamate in doses that were not detectable by the examiner significantly impaired the detection of deception in a concealed information analogue task. In a replication and extension of this study, Iacono and colleagues (Iacono et al., 1992) compared the effects of meprobamate, diazepam (a benzodiazepine) and propranolol (a beta-blocker) on detection of guilt with a concealed information task. Contrary to the findings of Waid et al. (1981), this study found that none of the drugs evaluated had a significant effect on the detection of deception, nor was there even a trend in the direction reported by Waid et al. The nature of the mock crimes was different in these studies, though drug dose was identical. Using diazepam and methylphenidate, a stimulant, in separate groups of subjects, Iacono, Boisvenu, and Fleming (1984) evaluated the effect of these drugs and a



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