This article compares two measures of the extent of physician participation in Medicaid programs. The first, which has been used in most research to date on the subject, is based on physician estimates of the proportion of their patients who are Medicaid patients. The second derives from encounter forms for a sample of visits to the interviewed physicians. The comparison shows that physicians in the sample tended to overestimate by 40 percent the extent of their Medicaid participation. Because the two measures are highly correlated, the analysis of the determinants of Medicaid participation was not affected by the measure used. However, since physicians tended to overstate the proportion of Medicaid patients in their practices, interview data should not be used to measure the amount of physician participation or to calculate elasticities for the effects of policy changes on the extent of participation.