Abstract
The purpose of our research was to develop a final exam that was to be used as a standardized oral exam for the Intensive Japanese Language Preparatory Course. The text was to be developed with Japanese Government Graduate-level Scholarship students as the object of this research. In this presentation, we will introduce the development process of this oral final exam at the International Student Center, Okayama University. The test format is an interview with the instructor. There were 9 evaluation categories using OPI as reference, such as "function" and "content". We adopted the evaluation method of listening to the subject's recorded utterance and judging it in comparison with the guidelines which we prepared beforehand. After the exam, upon calculating the coefficient of interrator reliability, we found that while categories such as "function" and "content" had a comparatively high rating, "pronunciation" had an extremely low rating, and "grammar" as well as "sociolinguistic competence" were not able to get a sufficiently high rating. It is thereupon which we rearranged the evaluation categories, while at the same time completely adopting the Guttman scale, and upon refining the evaluation descriptions and rejudging the same sample, we were able to reduce the evaluation variance significantly.
