The purpose of this study was the development of a rating instrument to assess the use of humor in physician- patient interactions, and to compare humor use as a function of patients' socioeconomic status. The 46-item Physician-Patient Humor Rating Scale (PPHRS) was used to rate 246 audiotaped primary care interactions. Four subscales were reliable and valid, demonstrating correlations with patient satisfaction and reports of physician humor, with physician satisfaction and with separate affective communication ratings. There was a significant difference in use of humor as a function of patient socioeconomic status, such that there was greater mutual trust between physicians and high versus low income patients.