The improvement of computing courses is a permanent need and is a goal established by any teacher. Suggestions of possible course improvements should be made by teachers and students. Computer project-based courses involving a significant number of people pose difficulties to listening to all their opinions. The purpose of our research is twofold: to develop a study to understand teachers' perceptions about possible course improvements by listening to students and teachers and to help introduce the Q methodology as a viable approach to be used in computing courses improvement. The focus is on project-based courses. Q methodology was used as the research methodology in an exploratory study of teachers' perceptions about possible improvements. The course investigated involves a software development project of median complexity in a reasonable simulation of a real-world environment. Suggestions of possible improvements were collected from the students of the course and used as a concourse. A sample of the concourse was presented to the teachers, who rank-ordered them, and then, data were analyzed and interpreted. A literature review about the use of the Q methodology on the improvement of courses was also done. The review shows that while Q methodology has been used frequently on the context of education, its use specifically for course improvement is much reduced. The empirical study produced two distinct factors relative to teachers' opinions about possible course improvements: one corresponding to a more experienced view of the course than the other. Some limitations were found and discussed. The study contributed positively for the computing project-based course improvement, in spite of some limitations, and confirmed the adequacy of Q for course improvement studies. (Contains 9 tables, 3 figures, and 6 notes.)