The instructor of the educational computing course at the University of Houston used action research to address the following questions and improve the course: "How could I formalize the process of action, reflection, and feedback?" and "Do certain types of learners do better with certain types of activities and certain structures for activities?" As part of the midterm, students were given two surveys and the following reflective questions: (1) "What was your most successful previous learning experience? Why?" (2) "What was your least successful previous learning experience? Why?" (3) "What do you consider to be your major blockages to learning?"; and (4) "Reflect back on one of your learning sessions in this class and describe what techniques or patterns of behavior you might be using to learn." The first survey, the Gregorc Style Delineator, indicated dominant learning styles (i.e., Abstract Sequential, Abstract Random, Concrete Sequential, Concrete Random). The Productivity Environmental Preference Survey showed environmental, emotional, sociological, and psychological learning preferences. The reflective questions were analyzed by the four learning styles categories. Use of the midterm assignment was found to be an effective way to formalize action, reflection, and feedback. Similarities were found within each of the four learning style groups in terms of types/structures of activities. (AEF)