Gives examples of major sources of unreliability in multiple choice items in health professions classroom achievement tests (clues to response combinations, mutually exclusive alternatives, implausible distractors) and offers some suggestions for eliminating them when writing such tests. (MF)