A method was developed for assessing the extent to which a test reflects the topics addressed in the problems presented in a problem-based learning environment (curriculum validity). The intended curriculum for a unit from the second year of medical school at the University of Limburg in Maastricht (Netherlands) was analyzed, and an 132-item topic list was constructed to serve as a blueprint of unit content. At the end of the unit, students took an achievement test. The overlap between the intended curricular content, as specified in the topic list, and the information required to answer achievement test items correctly was assessed by four teachers (content specialists). Some curricular content domains were underrepresented on the test or totally absent, and some were overrepresented. It is concluded that the test does not accurately reflect curricular content as specified in the topic list. The method described seems to be a valuable approach for assessing the curricular validity of a test and could be adapted to use in test construction. Two figures and two tables highlight study concepts, and there is a seven-item list of references. (SLD)