The purpose of this project was to develop and evaluate a program to teach medical students how to order diagnostic tests in a cost-effective manner. The 1-month educational program included a seminar, a simulated patient-care exercise, special case presentations by students, newsletters about diagnostic tests, and concurrent review of patients' bills. Content analysis of answers to open-ended questions and pretests and posttests were used to measure differences in the study and control groups. Although students said the program was useful, no significant differences were found in students' knowledge, attitudes, or simulated test-ordering behavior. The authors conclude that the lack of improvement in objective measures limits the potential effectiveness of restricted efforts such as this one and that the discrepancy between the subjective and objective measures reinforces the need for more rigorous evaluations of programs that teach cost-effective diagnostic test use.