Immediately before and after taking a course on the physician-patient relationship, 102 sophomore medical students were measured on their preference for using understanding responses in a hypothetical patient interaction simulated by on objective test. Sixteen of these students were also rated on their use of understanding responses in an actual patient interview. A comparison of the pre- and post-measurement showed that students could identify an understanding response on a paper-and-pencil-test as well as make understanding responses in an actual interview situation to a significantly greater degree after having taken the course on the physician-patient relationship. This response type was compared to evaluative, reassuring, hostile and probing responses. All of these decreased as a function of the course, except the probing response which showed an increase in quality level. Implications for the training of students in a variety of service professions are discussed. (Author)