Conjoint analysis was used to assess resident preference for teaching hospitals that offer postgraduate clinical training programs. Data for the analysis were collected from 46 clinical residents who participated in an emergency room seminar held in July 2005. Categories that received high utility scores in the attributes of teaching hospitals were “hospital has a renowned teaching doctor”, “hospital is partnered with overseas institutions”, “conferences/study sessions are held more than 8 times per month”, “bedside teaching takes place daily”, “a monthly salary of 500, 000 yen”, “80 working hours per week”, “the hospital is located in a nearby city”, and “the hospital is not affiliated with a university” . Conjoint analysis is a useful analytical method for developing education systems and programs, as it can predict the degree of change in resident preference made by improving these categories.

1)The factors Japanese residents consider most important in postgraduate clinical training are: "relationship with supervising physicians," "well-developed training programs," and "guidance of supervising physicians."2)Items showing a large discrepancy between their importance to residents and residents' satisfaction were: "well-developed training program," "experiencing a large number of cases,""guidance of supervising physicians," and "guidance of senior residents."3)Multiple regression analysis showed that factors significantly influencing residents' satisfaction with training were "quality of the medical service" (r=0.59) and "consideration for accepted residents" (r=0.42).