現代教育評価の諸問題

The Chairman first called the attention of the participants to the fact that the very appropriate theme was chosen for the symposium arranged by the University of Hiroshima, where most active research on factor analysis in Japan had been conducted.<BR>Discussion by each discussant was as follows:<BR>Osaka pointed out that since “the 8 year study”, for about 30 years, there had been no significant progress in the field of educational measurement and evalution he insisted that the educational goals should be subdivided systematically into subgoals, and that guidance for the individual student in the educational setting would be effective when as many sub-tests as possible to meet the needs of the sub-goals were administered. Furthermore, he emphasized that collaboration of the related fields, such as education, sociology, and psychology were indispensable to effective evaluation.<BR>Enumerating the following illustrations, Higuchi warned that educational measurement had limitation in its usability and was not to be misused: 1) Both the highest and lowest values of the phi-coefficient are determined by the proportion of marginal frequencies, 2)The reliability coefficient measured by test-retest is affected by means and standard deviations of the test, 3) Not so high coefficient of validity is expected even when scores of two tests are combined.<BR>Kitao regarded the role of educational evaluation as a regulator of learning effect of the individual learner and saw the points as follows: 1) Effective evaluation should meet to the information processing ability of the learner, 2) Timely evaluation which meets information retention ability is to be given to the individual learner, 3) Reinforcement as the result of evaluation is different due to the different needs in each individual. Accordingly it is possible to modify the need system of the learner by the evaluation, 4) It is necessary to give the learner information about his learning progress. With those viewpoints, Kitao insisted that both intra-individual and longitudinal evaluations were necessary in education.<BR>Shimizu stated that differences among junior high schools regarding their academic standards were to be taken into consideration when senior high schools use the academic record of applicants for admissiom.<BR>The following comments were added to their reports<BR>Akagi (Center for Science Education, Osaka) stated that objective tests were necessary for screening in admission, but subjective evaluation, at the same time, was indispensable for individual guidance in the school setting. Osaka stated that school differences in terms of academic standards among junior high schools would even be waved if senior high schools became accustomed to selecting applicants on the basis of academic records. Yanai (University of Tokyo), contrary to Higuchi's negative attitude toward the coefficient of validity of composite scores, stated that it would be possible to get higher validity by applying multivariate analysis.Arguing about that point, Higuchi said that it was necessary for cost and effect to be taken into consideration at the time of administration. Yasuda (Tottori University) saw inappropriateness in abusing objective measurement by other persons in predicting whether the individual learner would be successful or not in attaining the goal, because Yasuda thought evaluation was to be used as helpful to the learner for his self-understanding and self-actualization.<BR>Tanaka (Osaka University of Education) stated that experienced teachers would use not only intra-individual and longitudinal evaluation but also inter-individual evaluation for individual guidance.<BR>Tsuzuki (Nagoya University) said that teachers should not use any test unless they were very familiar with its use. Also he insisted that along with the efforts on the part of educational psychologists, teachers' cooperation was necessary for effective evaluation.