The research proposed here is designed to test, evaluate and modify a series of methods designed to estimate total population, demographic characteristics and socioeconomic characteristics for sub-city and sub-county areas. The particular focus of this endeavor is the assistance of planners for and providers of mental health and related services in being able to monitor changes in their particular populations at risk. The areal focus here is on sub-county and sub-city populations for the simple reason that while research has demonstrated the importance of neighborhood characteristics in the presence of certain mental health conditions, the technology to produce estimates of these neighborhood characteristics is largely undeveloped. Census data are available once every ten years to meet this need, but these are prone to become dated quite quickly; to some degree population counts and characteristics are updated annually for cities and counties, but this is not true for smaller areas. Our intention is to modify the set of procedures developed for updating the demographic and socioeconomic data for larger areas to meet the particular small area data needs of mental health professionals and epidemiologists. Those procedures which we will deal with here include vital rates, housing unit, regression (ratio-correlation), modified projection techniques, and synthetic techniques. These will be compared and contrasted in terms of their comparative ability to accurately estimate population and its characteristics for the same set of geographic areas (census tracts in the Houston-Galveston SCSA); from this we will make statements regarding the impact of type of small area on accuracy of estimation for the various techniques. Finally, as part of the final report, we will prepare a manual for the production of estimates, including specific recommendations on the type of technique best suited for a particular estimation problem and particular type of area.
