The MCBS is a national probability sample of approximately 12,000 noninstitutionalized Medicare beneficiaries who are interviewed every 4 months for up to 4 years. Very old and disabled beneficiaries under age 65 are oversampled for some rounds of the survey. Proxy respondents answer questions for survey respondents who are unable to do so for themselves because of physical or cognitive impairment. Each year, approximately one-fourth of the sample is rotated out of the survey and replaced with new members, so that each annual MCBS data set represents a cross-section of the Medicare population enrolled in the program continuously since January 1 of that year, as well as members of a longitudinal beneficiary panel.
Each year, beginning in the winter of 1998, just prior to the first national distribution of the Medicare & You handbook, and continuing annually through the winter of 2006, one of the MCBS survey rounds has included a beneficiary knowledge and information needs supplement that is administered to community-dwelling beneficiaries. Through this supplement, the MCBS tracks national trends in (1)beneficiary knowledge and sources of information about Medicare through the periods before and after the education campaign activities were implemented, and (2)beneficiaries’ use and preferences for a variety of sources of information to stay informed about changes in the Medicare program.
A limitation of using the MCBS data to answer specific evaluation questions is that some questions change over time or are not included in some years. Although the exclusion of some questions allows for the collection of new information, it may limit the ability to track changes over time. A second limitation is that some questions are not asked of the entire sample in every year. Therefore, while weights allow for tracking trends in cohorts over time, longitudinal analysis measuring changes over time in individual beneficiaries is not always possible. A final limitation is that there is often a significant time lag between when questions are asked of beneficiaries and when the survey data are available for analysis. Thus, the MCBS is not particularly useful in identifying potential knowledge and information issues in “real time.”