Objective:Methods of assessing humanism in internal medicine residents have not been completely designed or evaluated. This study used patient satisfaction as a measure of humanism, and assessed the validity of using faculty physicians to evaluate residents’ humanistic behavior. Residents’ ability to assess themselves was also evaluated. Setting:A university-affiliated internal medicine training program. Subjects:Forty-seven internal medicine residents were evaluated by patients, faculty, and themselves. Design:Faculty physicians were given standard faculty evaluation and patient satisfaction forms, and were asked to evaluate residents. These evaluations were compared with the patients’ responses on the same satisfaction forms. Residents performed self-assessment using identical forms; these responses were compared with those of the faculty and patients. Results:There was no correlation between patients’ responses and those of the faculty or residents. There was a significant inverse correlation between resident and faculty responses, especially for the female residents (r=0.71). Conclusion:These findings suggest the need for further study of the evaluation process, including what factors influence individuals to respond as they do. It appears that the use of one rating group is not sufficient to achieve an accurate assessment of residents’ humanistic skills. The present status of the process of evaluating humanism is discussed.