Accuracy of the Data (Multi-Year Estimates Study)
The data contained in these data products are based on the American Community Survey (ACS)
sample interviewed in 34 counties from January 1, 1999 through December 31, 2005. Data
profiles were produced for six sets of one-year estimates, five sets based on three consecutive
years, and three sets based on five consecutive years. In 2008, the Census Bureau will publish
the first full-sample multi-year data products, based on data collected in 2005, 2006, and 2007.
The first products based on five years worth of data will follow in 2010. Publication of these
multi-year estimates for these 34 counties is intended to allow users an early look at data similar
to what will be released in the coming years.
The purpose of this documentation is to provide users with a basic understanding of the ACS
sample design, estimation methodology, and accuracy of the ACS data used in these multi-year
estimates. The ACS is sponsored by the U.S. Census Bureau, and is an integral part of the plan
for the 2010 Census.
Thirty-six counties were selected for inclusion in the 1999, 2000, and 2001 ACS Comparison
Study. These counties would be sampled using the methods then planned for the full-sample
ACS, and at a high enough rate to allow tract-level estimates to be made using just those three
years’ worth of data. For more information on the ACS Comparison Study, see
http://www.census.gov/acs/www/AdvMeth/acs_census/index.htm. Due to budget constraints,
the sampling rates for two of the counties, Ft. Bend and Harris, TX, were too small to support
reliable estimates below the county level. Because of this, these two counties have been
excluded from the Multi-Year Estimates Study data products. The remaining 34 counties had
their sampling rates reduced in the ACS sample for 2002-2004 to levels similar to those planned
for full implementation of the ACS. The 34 counties are listed below.
Table 1. List of the 34