One solution might be to silo the people who evaluate the students from those that teach the materials. Treat the test more about evaluating the effectiveness of instructors and less about the students.