Concerning the course construction project “Engineering Design”, a decision had to be made about the instructional procedures of the seminar stage of the course. In order to come to this decision in a more justifiable way an experiment was designed to compare four instructional procedures (viz. a teacher-centred egruleg-, a student-centred egruleg-, a teacher-centred ruleg-, and a student-centred ruleg procedure) on cognitive pupil gain and satisfaction. For testing the hypotheses an analysis of variance design was used. Twelve groups of ca. 20 students were divided into the four cells of a 2×2 factorial design with one completely nested factor. The results of the experiment have shown that there are no differences between the procedures as to pupil gain and satisfaction. However, differences were found between the separate student groups (nested factor), irrespective of the instructional procedures. A plausible explanation seems to be the effect of interaction between instructional procedures, teachers and students.