The 2nd year student evaluation of didactic lectures by one instrustor was carried out anonymously in 1991 in order to establish the most appropriate set of questionnaire items (including 11 types of questions with 5 choices of responses each and space for a free answer). The study was repeated 11 times, with a total of 286 responses collected. There was a certain amount of variability in the means and standard deviations in the data. Mutivariate analysis, such as multiple regression, was appled and three principal components (students' concern and lecturer's attitude, utilization of visual aids, and preparation and clarity of lectures) were extracted out 11 items by means of principal component analysis. In concluding, the author stresses the necessity of the further studies.

As the second phase of a trial of student evaluations of didactic lectures, a modified questionnaire (B5 type, consisting of seven items, each having five multiple-choice responses and more space for written responses) was filled out anonymously by all students receiving didactic lectures of one instructor in a preventive medicine and community health course from 1992 through 1994. The frequency of written responses was 51.8% higher than in our previous trial (1990-91). Multivariate analysis was also applied to investigate the items most related to the summative assessment. The minimal essential questionnaire items and the necessity of space for written responses in frequent trials is discussed.

A new method of evaluating learning, combining the 2 categories of knowledge/skills and attitude, has been introduced into the course on preventive medicine and community health. The goal was to improve student attitudes towards learning. In this article, our five years' experience from 1989 to 1993 is reported and follows a previous similar study from 1986 to 1988. This time, the records of 3 of 490 students were judged to be inadequate for promotion to the next grade. Although one-half of the students agreed with this method of evaluation, 15 % disagreed. The usefulness ofthis new method is discussed, based on our results.