1. Problem: Berg & Rapaport studied response bias in an unstructured test situation, and found in such a situation the response bias appeared at a high level of statistical significance, and occurred at culturally valued options, and that different cultures would presumably cause different patterns of response bias. I have attempted to study to what extent such response bias appears in the same situation, and to reexamine their results of the study.
2. Subjects and Procedure: Subjects were 830 students in the first and second year of the Faculty of Education of the Kumamoto University and the Kumamoto Women's University (465 males, 365 females). Each one was requested to choose and mark, on the distributed sheet of paper, a statement that seemed to be correct for a question among various statements orally given. Then questions had alternative answers for each, other ten had four-optionchoices for each, and another ten three-option choices. This test was administered two times with interval of five months, and each time it was held in both regular and inverted orders of options.
3. Results: This study reveales that response bias appeared, in such an unstructured test situation as Berg & Rapaport studied, at a covincingly high level of statistical significance, reaching chi-square values above 200 in some cases. Although no occurence was observed at certain positions of options, the response bias appeared exceedingly at culturally high valued options and at the options that would meet the organic and physiological fundamental needs of human beings. As for the cultural factors, it seems that other cultures to some extent, would show different types of response bias.
And it could be stated that those who respond always abnormally in such a situation would be deviate in their personality from a group of normal subjects.