Purpose. To evaluate a review process for identifying marginal performers among students in a clerkship. Method. To better identify the marginal performers among the students participating in the medicine clerkship at the University of Minnesota Medical School-Minneapolis, the Medicine Clerkship Committee reviewed in 1990-91 and 1991-92 all students rated by faculty or housestaff as below expectations for any of nine areas of clinical performance (27 students of 890, 3%). (In the past, a student was considered to be a marginal performer only if he or she was assigned an unsatisfactory numerical grade, calculated from the nine ratings, or if written comments by housestaff and faculty and the opinions of the attending faculty and clerkship site coordinator indicated that the student should fail.) Chi-square analysis was used to compare the number of students judged to be marginal performers under the review process with the number of marginal performers in 1988-89 and 1989-90. The two groups were also compared based on their preclerkship performances on standardized examinations. Results. Ten of those reviewed (37%) were judged to have performed marginally. Although the study group’s performance on standardized examinations was not different from that of students during the previous two years, significantly fewer students were identified as marginal performers before the review process began than afterwards (2 of 867, 0.2%, versus 10 of 890,1.1%, p < .05). Conclusion. Without changing the way in which faculty and housestaff evaluated students, the review process improved the medicine clerkship evaluation system by identifying significantly more students who were marginal performers.