Bandura's self-efficacy theory was used a model for examining levels of confidence and clinical skills among undergraduate speech pathology students. Forty-four second-year students rated their confidence in their ability to perform a number of clinical tasks at the beginning and at the end of the academic year, which was their first clinic experience. These efficacy expectations were compared with clinical supervisors' assessments of the students' performance on the same tasks. An attempt was also made to assign students to pairs and small groups for training on the basis of their initial efficacy expectations Efficacy strength (but not level) increased significantly over the year, and efficacy level (but not strength) was moderately and significantly related to clinical supervisors' ratings of clinic performance. Practical problems arose with the grouping, but the results suggested that there may be some possible benefits from assigning students to pairs on the basis of their self-efficacy. These results show self-efficacy to be only moderately well related to clinic performance, but it is suggested that relatively high baselines may have attenuated the strength of the obtained relationship. The role of self-efficacy in clinical skills training warrants further investigation.