2001 Volume 39 Issue 1 Pages 15-27
The purpose of this research was to assess the validity of the Guilty Knowledge Test (GKT)- a widely used method of psychophysiological detection of deception in Japan. Eleven hundred and forty-six samples of GKT questions were collected from 271 field examination cases conducted in actual criminal investigations.
Each sample consisted of descriptions about the concrete content of the question, the original examiner's decision about physiological responses to the question, and the content of the examinee's memory about the item which was the subject matter of the question.
In 92.6% of the samples, examinees had actually memorized the item. The length of the interval between the date of the crime and the examination was not related to memory performance. Remarkable responses had been observed by examiners in 906 of 1032 samples, in which items had been actually memorized (the hit rate was 87.8%). On the other hand, ambiguous or no responses had been observed in 68 of 105 samples, in which items had not been memorized (the correct rejection rate was 64.8%). As a result of classifying questions by their content, questions about "number" and "behavior" had a lower hit rate than other sorts of questions, and questions about "behavior" and "date or time" tended to have a rather lower correct rejection rate than the others. Moreover, it was revealed that the Searching Peak of Tension Test was inferior to the GKT, especially in terms of the correct rejection rate.
These results suggest that the examiner's item selections from the crime information are appropriate, and Japanese field GKT practice has satisfactory validity. It is also suggested that examinee's physiological responses are not caused only by the recognition of the item.
