2007 年 80 巻 p. 165-186
The purpose of this paper is to reexamine the use of “Comprehensive and Multi-dimensional Evaluation” as the basis for University Entrance Examinations. Though the phrase “Comprehensive and Multi-dimensional Evaluation” itself was first articulated in the 1997 report of the Central Council for Education (Chuou Kyoiku Shingikai), the concept itself came into existence immediately after the postwar period. In fact, “comprehensive evaluation” was merely an excuse for avoiding having to add the score of Japanese Scholastic Aptitude Test (Shingaku Tekisei Kensa, used from 1947 to 1954) into the total score of the University Entrance Examination. Moreover, the term “multi-dimensional evaluation” appeared in the outline of the University Entrance Examination (Daigaku Nyugakusha Senbatsu Jisshi Youkou), as it is proposed in the first report of the National Council on Educational Reform (Rinji Kyoiku Shingikai) in 1985.
In fact, the report of the Central Council for Education (Chuo Kyoiku Shingikai) in 1971 stated that “Comprehensive and Multi-dimensional Evaluation” was scientifically valid as a basis for University Entrance Examinations. The report is famous as the only report based on evidence, and is generally known as the “1971 Report” (Yonroku Toushin). In the interim report, the Central Council for Education stated that follow-up surveys by the National Institute for Education and the Educational Test Research Institute (Nouryoku Kaihatu Kenkyujyo) had proven that a “Comprehensive and Multi-dimensional Evaluation” could be a valid selection method for predicting a good Grade Point Average after entrance to university.
However, the two surveys cited contained simple statistical errors. The first, survey by the National Institute for Education, failed to control for the “Selection Effect.” A “Selection Effect” is a “restriction in range problem,” caused by cutting off the distribution at the passing grade. As a result, there is a tendency to misunderstand the fact that, in actuality, academic achievement tests on University Entrance Examinations have little relationship with Grade Point Average after entering university. To tell the truth, this problem had been pointed out as early as 1924 by Japanese psychologists who were interested in Entrance Examinations. In the second survey, by the Educational Test Research Institute, the inevitable nature of multiple correlation coefficients was ignored. As the number of independent variable increases one by one, the multiple correlation coefficient necessarily reaches the maximum of 1. In this paper, the follow-up research data from the Educational Test Research Institute is recalculated using a multiple correlation coefficient adjusted for the degrees of freedom. The conclusion is different from that reached by the Central Council for Education.
This demonstrates that there is absolutely no scientific ground for the use of “Comprehensive and Multi-dimensional Evaluation.” In other words, it is not necessarily correct that putting a lot of effort into University Entrance Examinations and using anything more than academic achievement tests as reference for University Entrance Examination will lead to more students gaining good grades after entering university. If this mismeasure of academic achievement is not properly recognized, the number of university students who cannot achieve even low basic competence level will surely increase.