This paper presents and analyzes an object-oriented analysis and design course that has been given in three different configurations for students who are already familiar with object-oriented programming. The results show that the course configurations have not had a major impact on the students' performances.