Formal evaluation of surgical simulators is essential before their introduction into training programs. We report our assessment of the Mentice Corp Procedicus shoulder arthroscopy simulator. This study tests the hypothesis of construct validity that experienced surgeons will score better on the simulator than individuals with minimal to no experience with the technique. Thirty-five subjects were stratified into three groups (novice, intermediate, and expert) based on their past 5 years' experience with shoulder arthroscopies. Each subject had an identical session on the simulator and completed anatomic identification, hook manipulations, and scope navigation exercises. We found statistically significant differences among the three groups in hook manipulation and scope navigation exercises, with the expert group performing the exercises more quickly (P =.013) and more accurately (P =.002) than the other two groups. No statistically significant differences were found among the groups in the identification of anatomic landmarks. Experts rated the simulator as an effective teaching tool, giving it a mean score of 4.22 and 4.44 (maximum, 5) for teaching instrument control and triangulation, respectively.