Background. The surgical literature suggests that collaborative learning using peers may be a valid way to teach surgical skills and there is a growing interest in the use of computer-assisted learning for this purpose. Combining this evolving technology with this type of teaching would theoretically offer a number of advantages including a reduction in the amount of faculty time devoted to this task. In this study, we evaluate the efficacy of a type of collaborative learning in a computer-assisted learning environment. Materials and methods. We designed a prospective, randomized study comparing novice learners who were allowed to work in pairs with those who worked independently in a specially equipped computer-assisted learning classroom. Both pretest and posttest assessments were performed by videotaping this skill. Three experts then evaluated the videotapes, in a blinded fashion. Three different outcomes were assessed. Results. Seventy- seven subjects were enrolled in and completed the study. Comparison of the outcome measures demonstrated no between group difference in the average performance scores or posttest times. The proportion of subjects who correctly tied a square knot was significantly lower in the computer-assisted peer teaching group when compared with the computer-assisted learning alone group (P = 0.04). Conclusions. Collaborative learning in a computer-assisted learning environment is not an effective combination for teaching surgical skills to novices. (C) 2000 Academic Press.