While patient satisfaction has been studied extensively and considered an outcome measure of physician-patient encounters, physician satisfaction has received limited consideration. The reliability and validity of an "encounter-specific" physician satisfaction scale were examined. The Delphi technique was used to develop a 39-item questionnaire comprising four hypothesized domains of physician satisfaction: interactive, personal, professional, and contextual. Factor analysis supported two of the four hypothesized categories. Exclusion of all items not contributing to the reliability of the two subscales or differentiating between high and low scores in an item analysis allowed a condensation of the scale. This 16-item scale was tested in a second phase of the study, which supported the reliability and validity of the two-dimension encounter-specific physician satisfaction scale. Further use of the scale may prove useful in exploring this neglected component of physician-patient encounters.