The author describes a research methodology for college course improvement, and applies the results to education courses. In Japan, it is usually difficult to carry out research on college course improvement, because faculty cannot introduce experimental design approaches based on control and treatment groupings of students in actual classroom settings. In this study, however, the author followed the Design-Based Research (DBR) approach proposed by Reeves and other researchers, revising the DBR and then applying it to college course levels with special conditions. The revised DBR was applied to both "Instructional Methodology" and "Educational Technology" courses that are compulsory in teacher-training programs in Japan. This resulted in the identification of effective and ineffective instructional activities, which were then classified into common activities seen in both courses, and particular activities seen in only one course, respectively. The author concludes that this revised DBR is a practical and powerful method for improving existing college courses. (Contains 5 figures and 1 table.)