cations. For instance, the system will need to estimate the prevalence of persons eligible for disability benefits as well as develop measures that predict application. Key to such a system will be sufficient data to understand macro- and micro-level factors that distinguish participating and nonparticipating eligible populations.
The design of a disability monitoring system must consider the informational needs of the system and the impact of alternative design options on meeting analytic goals as well as the impact on various sources of survey error (e.g., whether the design should include the use of household surveys). Alternative design components include the following:
Data source or sources: Among the various data sources that could be included, alone or in combination, in the design of a disability monitoring system are data obtained from household-based surveys, physical examination, and administrative records. Among the options with respect to household data are stand-alone surveys that permit rich and deep national data on the size of the disabled population (e.g., similar to the NSHA), survey modules administered as part of preexisting data collection efforts (e.g., a supplement to the Current Population Survey or the Survey of Income and Program Participation), or the incorporation of a limited number of questions on existing national surveys (e.g., the National Health Interview Survey or the Behavioral Risk Factor Surveillance System [BRFSS]). Each of these options has implications for the error properties of the resulting estimates, including coverage, sampling, nonresponse, and measurement error. The use of administrative record data potentially suffers from similar sources of error.
Periodicity of measurement: Decision on periodicity requires answers to several questions, such as: If survey data are collected, how often should the data collection occur? What are the ramifications of more frequent or less frequent data collection on the utility of the data? How is periodicity affected if one decides to use repeated cross-sectional data collection or a longitudinal design?
Mode of data collection: For survey data collection, a decision will need to be made as to the mode or modes of data collection—such as telephone or personal interviews, self-reports, or observation and examination. Little is known about the effect of mode of data collection on the measurement error properties of self-reports of disability and impairments. In addition, the choice of a single mode of data collection has potential implications for the coverage of the population and the potential for nonresponse bias.
Self and proxy response status: Questions that need to be resolved include: Should only self-response be accepted for household