|
Abstract:
|
In 2010, the U.S. Census Bureau will publish the first set of 5-year period estimates from the American Community Survey (ACS), based on data for 2005-2009. Published for small places, tracts, and other small areas, the 5-year ACS estimates will attempt to replace the long-form data from recent decennial censuses. Because the effective sample size for the ACS will be somewhat less than half that of previous censuses, users will face an increased challenge to distinguish true variation from sampling error. The concept of the false discovery rate has become increasingly useful in other disciplines confronted by large numbers of estimates, such as micro-array analysis in genetics and fMRI studies of the brain. The paper will review this concept and suggest its possible future application, based on a preliminary analysis of published data from the ACS Multiyear Estimates Study.
|
