The ways in which a rating scale developed for problem-based medical curricula provides information that is useful in the area of curriculum reform were demonstrated in a study involving 95 first-year medical students at the University of Sherbrooke in Quebec (Canada). The rating scale, based on the theory of problem-based learning of W. H. Gijselaers and H. G. Schmidt (1989), covers educationally important dimensions through 10 subscales, most with Likert items. A new problem-based, community-oriented curriculum was instituted at the university in 1987. After each curriculum unit, the scale was administered to all students. Results were used to evaluate three variables: (1) quality of problems used; (2) adequacy of lectures; and (3) tutor performance. The problem of the absence of absolute standards for what can be considered sufficient performance of a course is circumvented by applying a reliable measurement approach in which each course or teacher is compared to other courses and teachers, and in which the highest score constitutes the norm against which others are judged. Four graphs illustrate the evaluation. (SLD)