Do final grades reflect written qualitative evaluations of student performance?

An evaluation system should allow for early and continued identification of specific problems. A review of negative evaluations of students throughout a problem‐based program revealed that 66% of these occurred during the last third of the program. Was the evaluation system too insensitive to detect problem students earlier in the program, or was the faculty reluctant to give less‐than‐satisfactory grades regardless of identified problems? A blinded review of 608 written summaries of student performance revealed that the descriptive, non‐numerical system that was used did have the capacity to detect specific student behaviors. Independent reviewers identified student problems and agreed with the final grade given by the original evaluators in 92% of the cases; however, the reviewers identified more performance problems and labeled them earlier. Insensitivity of the evaluation system was of less concern than the faculty's reluctance in awarding negative decisions. Reasons for this reluctance are explored; suggestions are made for improving the evaluation process.