Purpose: The aim of this study was to analyze the instruction evaluation questionnaires that have been used by the Department of Radiology, Seoul National University College of Medicine, 2001, and to determine a means to improve them. Methods: This study used (1) a factor analytic technique to identify the instructional factors that determine the correlation between all pairs of the evaluation items. (2) A procedure known as Cronbach’s alpha, which allows the internal consistency of the instruction evaluation items to be estimated. (3) A correlation method to compare the instruction evaluation score results among the students, residents, peer faculties and the faculty itself. Results: The results were summarized as follows: First, the Instruction Evaluation Questionnaire included 12 items, 6 comprising a ‘teaching method connected with the lecturer’ factor and the other 6 a ‘teaching method connected with the teaching resource’ factor. Second, the Cronbach’s α index was found to be 0.91s. This indicates the high reliability of the items of the questionnaire for the instruction evaluation. Third, it was found that the correlation between the evaluators was very low (r=.345). In particular, the average score of the peer faculties was 3.33, which was lower when compared with the average score of other evaluators. Conclusion: There is no concordant opinion regarding the evaluation items that should be used in an instruction evaluation in a Medical College. However, the instruction evaluation items should consider various factors connected with the teaching and learning activity. There is a need to elaborate on the contents of the evaluation items in order that the instruction evaluation items be more reliable and have a greater validity.