OBJECTIVES: Health locus of control (HLC) is associated with health-related behaviors such as adherence and participation in health screening. However, HLC among physicians may be different from that among the general public. It is important to understand the potential gap in HLC between physicians and the general public. The aim of the study included two steps: 1) to evaluate item bias of the HLC scale between physicians and the general public, and 2) to characterize HLC among physicians compared to the general public. METHODS: Data for the general public were obtained from the health diary study that involved a random sample from a nationally representative group of households in Japan. Physicians’ data were collected from a web-based survey of Japanese physicians. Multi-group structural equation modeling was used for examining item bias in the Japanese version of the HLC scale (HLCS-J) between the two groups. Differential item functioning (DIF), including uniform and non-uniform types, were used for measuring item bias. Dimensions with no uniform and non-uniform DIFs were then compared using multiple linear regressions. RESULTS: Data on the HLCS-J of 2194 people from the general public and 895 physicians were available. Uniform DIF was recognized for the dimensions of internal, professional control and control by spiritual powers. Chance and family control dimensions had no DIF. Mean score for chance control (17.2) among physicians was greater than that (14.9) among the general public (adjusted p<0.001), while mean score for family control (21.7) among physicians was lower than that (22.1) among the general public (adjusted p<0.001). CONCLUSIONS: Our psychometric evaluation of the HLCS-J indicates item bias in the dimensions of internal, professional control and control by spiritual powers. Physicians believe that chance has a greater impact but family control has a lesser impact on health than do members of the general public.