CARLTON, B. L. and SMITH, B. J. The effects of aural versus visual presentation of questions during a detection of deception task. January 1991, Report No. DoDPI91-R-0002. Department of Defense Polygraph Institute, Fort McClellan, AL 36205.
The purpose of this research was to investigate the relationship between accuracy of a detection of deception task and the stimulus mode of the question presentation. That is, will the presentation of questions on a computer screen change the accuracy rate when compared to exams conducted, more traditionally, in a verbal mode? Eightly subjects were assigned to either a guilty or innocent condition. Guilty subjects were shown a video of a mock crime scenario, while innocent subjects viewed a clip from a training video. Half of the innocent and half of the guilty groups were given the exams aurally using a tape recorder and the other half were shown the questions on a computer terminal. Subjects were then given a guilty knowledge test by the experimenter using a Coulbourn polygraph.
While the polygraph exam was being administered, a second experimenter sat across from the subject. This second experimenter was responsible for programming the subject, while the experimenter running the exam was blind to the subject's guilt/innocent status. During the exam the subject was required to respond to the experimenter with 'no' to every item. The charts were scored by the following: 1) the original examiner; 2) a blind evaluator; and 3) using a scoring system introduced by Lykken. Overall accuracy of the decisions of the original examiner was 78%, 74% for the blind examiner and 76% for the Lykken system. Accuracy rates for subjects in the visual condition were 83% for the original examiner, 78% for the blind evaluator and 70% for the Lykken system. The decisions for the aural condition were 73% accurate for the blind examiner, 70% accurate for the blind evaluator and 83% accurate for Lykken scoring system. There was no significant association between an accurate decision and the stimulus mode condition for the original examiner, the blind evaluator or the Lykken scoring decision. (chi-square = .6091; p <.4351 and chi-square="2.0378;" p < .1534; chi-square="1.065," p < .3020). There was no significant association between the type of error and the stimulus mode for the original examiner (Fisher's exact p < .14) or the decision rendered by the Lykken system (Fisher's exact p < .25) whereas the type of error was associated with stimulus mode for the blind examiner (Fisher's exact p < .0075). This may be due to an artifact associated with the use of the experimenter as a confronter during the exam.
Key-words: detection of deception, stimulus mode, aural presentation, visual presentation, guilty knowledge test (GKT)
This report describes a unique study designed to determine if the accuracy of psychophysiological detection of deception (PDD) examination decisions are affected by the manner of question presentation. The authors interpret the results as indicating that the question presentation mode has very little influence on detection accuracy rates. Examination results were scored by the original examiner, a blind examiner, and using an objective scoring system. No differences were found among the three scoring system accuracy rates.
The results of this study clearly suggest that the accuracy of PDD examination decisions is not influenced by whether examinees hear or read the questions. This conclusion should however, be interpreted with some caution. The project was designed, implemented, and reported with above average scientific rigor. However, the non-standard procedure of having an collaborator within the subject's visual field throughout testing could have unduly influenced the results. In addition, negative results are always questionable in the absence of a statistical power analysis to indicate whether the data were sufficient to test the hypothesis. This project does, however, provide a basis for resolving the question as to the influence of visual versus aural question presentation on PDD examination decision accuracy.