A quasiexperimental research design involving a nonequivalent control group was used to examine the efficacy of a brief educational intervention in teaching medical students to select optimal diagnostic data consistent with the competing-hypotheses heuristic and Bayes' theorem when solving clinical problems. There was no significant difference between intervention (n = 119) and control (n = 89) groups in performance at baseline, but the intervention group performed significantly better (p less than .001) than the control group after training. The results suggest that some problem-solving skills, such as learning to use the competing-hypotheses heuristic, can be enhanced or learned independent of the acquisition of content knowledge.