Purpose: There has been a trend in medical education toward development of procedural skills through sophisticated computer-based models prior to interaction with patients. The validity of such models, however, compared to bedside training is not fully understood. The aim of this validation study was to determine if performance on a computer-based colonoscopy simulator (CBCS) appropriately differentiated skill levels based on actual colonoscopy experience. Methods: GI fellows, GI faculty, and Internal Medicine (IM) faculty from our institution were recruited between July 2003 and January 2004. A CBCS (AccuTouch Endoscopy Simulator, Immersion Medical) requiring both technical and interpretive skill was used for the study. Each subject performed one case to familiarize himself with the simulator equipment, followed by three cases for the study. Parameters measured included total procedure time, insertion time, withdrawal time, percent of mucosa visualized, patient discomfort, volume of air insufflated, and identification of luminal pathology. Comparisons were performed between the groups and the least and most experienced gastroenterologist (first year fellows vs. GI faculty) using the Kruskal-Wallis one-way (ANOVA) followed by the Wilcoxon signed rank test for pairwise comparisons. Results: Performance parameters of fellows (n = 11), GI faculty (n = 10), and IM faculty (n = 10) were reviewed. Results from one case were excluded because of complications that resulted in early termination of the case for many subjects. Significant differences were seen when comparing IM faculty to GI fellows and GI faculty for total procedure time (p = 0.003; p <0.001), insertion time (p <0.001; p <0.001). When comparing the least and most experienced gastroenterologist, differences were noted for total procedure time (p = 0.003), withdrawal time (p = 0.002), patient discomfort (p = 0.03), and volume of air insufflated (p = 0.02). Conclusions: The CBCS performance differences were most apparent when comparing physicians with little or no endoscopic experience to those with intermediate to great experience levels. The CBCS did not differentiate between upper level GI fellows and GI faculty. Such differences suggest that the basic technical skills of endoscopy are employed while using the simulator. If additional validation studies support this finding, the simulator will have a role for learning skills required for technical competence, enabling trainees to acquire procedural skills prior to performing colonoscopy on actual patients.