A problem solving, laboratory experience was devised in which first-year medical students were given a case description and then required to make judgments about what microbiology specimens should be collected and to analyze the results of laboratory tests in terms of implications for patient care. Over a four-year period revisions were made in the instruction and evaluation methods of the microbiology course. The three design variables which were included in the study were number of laboratory hours spent on problem solving techniques, class credit awarded for completion of laboratory activities, and exposure to problem solving examination items. Variations in these instructional methods were considered across the four years. Two of the variables were eventually dropped from the statistical analysis because of multicollinearity. A regression equation predicted student achievement on problem solving test items from number of hours spent in problem solving laboratory exercises. It is concluded that hours spent in problem solving laboratories had the strongest relationship with problem solving achievement, although test item exposure and student accountability also had an effect. (Author/SW)