Substantial use of clinical peer teaching among students has been reported, but there is limited description of outcomes and no reports of the use of a theoretical framework. The purpose of this study was to investigate the effects of peer teaching on baccalaureate nursing students' clinical performance. It was hypothesized that students who were taught by peers will: (a) achieve significantly higher improvement scores than students taught by teachers alone; and (b) rate their preference for peer teaching equal to or higher than instructor teaching. Bandura's (1971) social learning theory provided the framework for the study. The experimental design involved 50 volunteer subjects on two surgical units, one for peer teaching and one for instructor teaching. Data were collected from pre- and post-psychomotor and cognitive tests of a surgical dressing procedure and from a Clinical Teaching Preference Questionnaire (CTPQ). Experimental subjects achieved significantly higher cognitive improvement scores (t = 1.67; P < 0.05) and moderately higher psychomotor improvement scores in support of hypothesis 1. Responses on the CTPQ showed support for hypothesis 2.