OBERLIN, M. Verbal conditioning of the galvanic skin response to deception. June 1994, Report No. DoDPI94-R-0018. Department of Defense Polygraph Institute, Ft. McClellan, AL 36205.
A "noise instruction" procedure designed to produce a consistent and marked anticipatory autonomic response in subjects during deception was examined. Sixty college students were randomly assigned to a "no instruction" control group or a "noise instruction" experimental group; after all had independently participated in a mock theft of one of two items from a drawer and all were instructed to lie about what they had taken. All subjects were tested twice. In Phase I, none of them receive the noise instruction. In Phase II, the experimental group subjects were told they would hear 1 or 2 very intense blasts of sound sometime during testing: if they lied. A Lafayette field polygraph and CODAS data acquisition hardware/software were used to interface with a computer to record electrodermal responses. Instructions and test questions were delivered by tape recorder. Independent t-tests were used to assess group differences, and paired t-test were computed for within group comparison. There were no significant differences between Phase I and Phase II measured responses of either the relevant or control questions for the control group. There were significant differences between responses to relevant questions recorded from the noise instruction group and the control group during Phase II. The noise instruction would appear to contribute little to the detection of deception during initial testing because both groups exhibited differential responding to control and relevant questions during Phase I testing. However, during Phase II the GSR of subjects exposed to the noise instructions remained high for the relevant questions while that of control subjects showed a substantial decrease; probably a result of habituation.
Key-words: conditioning, verbal conditioning, detection of deception, autonomic response, galvanic skin response, GSR, control question test
This study was designed to explore the use of an instructional parameter on human conditioning that was designed to produce consistent and marked anticipatory autonomic response during deception. If effective and reliable, it would directly impact the accuracy of most psychophysiological detection of deception (PDD) examinations which are currently dependent on deception resulting in automonic nervous system arousal. While other studies have shown deception can serve as a reliable stimulus in a classical conditioning paradigm, differential conditioning by instruction has not been adequately explored, even though it could easily be incorporated into standard (PDD) testing paradigms.
During this study, some of the deceptive subjects were provided instructions to the effect that during their PDD exam, lies would be followed by a loud blast of sound. Based on electrodermal responding, the "noise instruction" was not necessary to discriminate between relevant and control questions during initial testing of deceptive subjects. However, during subsequent testing, the noise instruction had the effect of countering subjects habituation to questions about the mock crime they committed and of maintaining subjects differential responses to relevant and control questions.
While these findings appear promising, the procedure did not work with all subjects. Personality variables, one being subjects' latent anxiety, appeared to have an influence on whether they responded strongly. Research should be undertaken to determine any correlation between latent anxiety and responsiveness to noise instructions. This may lead to noise or other conditioning instructions being incorporated into selected PDD procedures and a possible reduction in problems associated with psychophysiological habituation.