Medical errors are a major cause of morbidity and mortality, and cognitive errors account for many of these events. This study examined the basic science of the cognitive performance of trainees. We created a low-intensity medical simulator to perform a preliminary study of the ability of residents to recall and process patient information presented verbally. The subjects were separated into three categories based on critical care experience: novice (< or =8 weeks of critical care experience), intermediate (8-16 weeks of critical care experience), and expert (>16 weeks of critical care experience). The subjects were presented with three clinical cases. In the first case, the presentation contained 55 separate data points and subject recall was analyzed. In the second and third cases, a patient report was given, and the subjects were asked by a "medical student" to outline and explain their treatment decisions. Fifteen subjects completed the experiment (five novices, six intermediates, and four experts). Case 1 (recall): No significant differences among groups with regard to errors or total data points recalled (however, subjects who chose not to take notes had significantly poorer recall and committed more errors). Cases 2 and 3 (cognition and decision making): Intermediates and experts made significantly fewer errors. More importantly, the reasoning process (forward hypothesis based) of the more experienced residents differed from novices. This preliminary study demonstrates that the cognitive processes used by residents experienced in critical care are quantitatively and qualitatively different from those used by novices. These processes were also associated with far fewer cognitive errors in clinical decision making.