The need to know ourselves is a fundamental requirement for our democracy. Government that purports to be "representative" must have knowledge of those it represents. This is so fundamental to our way of thinking that the taking of a census was written into the Constitution. The decennial census has evolved over the last two hundred years into much more than merely the tool by which congressional seats are reapportioned among the states. It has become the main provider of the country's profile, with the information it collects written into legislation that affects everyone. But the census is only taken every ten years, and we are a rapidly changing nation, increasing in number by at least 1 million housing units and 2 million people every year. A snapshot of us once a decade is no longer sufficient to determine programmatic needs. We need a better way than extrapolation, especially for small areas, to know ourselves over the decade so that efficient use can be made of our most scarce resource-- government funds. Continuous Measurement (CM) can provide the information critical to Federal and state program evaluation and formulation, as well as serve the needs of researchers and the private sector.

A successful Continuous Measurement program will depend on our ability to design and implement a monthly national sample survey large enough to produce reliable annual estimates of socio-economic characteristics for states and areas of 250,000 population or more, and annual rolling accumulated estimates for small geographic areas based on five years of data. Accomplishing this formidable task will require that monthly national samples of 400,000 housing units, decreasing to 250,000 housing units after 2001, be selected from a comprehensive, accurate, and constantly maintained list of all the nation's units--a Master Address File.

The design of the sample and data collection process for the housing unit and group quarters components of CM, as well as other special populations, will require quite different procedures. The housing unit component drives the overall design and produces the data -- operational and demographic -- on which the viability and benefits of a Continuous Measurement program will be judged.

This paper discusses some of the operational plans and issues related to the housing unit component of a Continuous Measurement program. It describes our initial understanding of the issues that have guided plans for the 1996 test of the entire CM system in selected sites. A more comprehensive version of this paper that includes reference citations and data on which the assumptions have been based is available from the authors on request.