Some of the apparent underreporting of Medicaid enrollment may be due to survey undercoverage--that is, the exclusion of complete households and individual members from the sample frame. Undercoverage affects the uninsured as well as Medicaid participants--and probably more so. The impact of undercoverage is difficult to gauge because the Census Bureau adjusts its survey estimates for undercoverage within categories defined by age, sex, race, and Hispanic ethnicity. That is, the missing households and people are included in the population estimates to which the survey is weighted, but their characteristics may be misrepresented. For example, if the undercoverage is concentrated in a subset of an adjustment category--such as the lowest income members--the adjustment will compensate only partially because it will increase the relative frequency of higher income people by the same amount that it increases the relative frequency of lower income people.