The purpose of the study reported here was to explore the characteristics of premature diagnostic conclusions in a group of physicians, medical students, and residents. When the subjects were asked to construct complete, precise problem lists from three case abstracts, premature closure occurred frequently, it could be recognized with good interrater reliability, and it seemed to appear with equal frequency regardless of the level of training.