Using the American Community Survey to Improve Sample Design and Survey Weighting

• Activity Number: 539
• Type: Roundtables
• Date/Time: Wednesday, August 1, 2012 : 12:30 PM to 1:50 PM
• Sponsor: Social Statistics Section
• Author(s): Alfredo Navarro*
• Companies: U.S. Census Bureau

Abstract

Virtually all national surveys use complex sample designs that involve some level of stratification. The specific survey objectives determine how much stratification is required. As the demand for more detailed survey estimates increases, so does the requirement for finer levels of stratification. Weighting and estimation methods are also becoming increasingly more complex and adjustments to offset non-response and coverage errors are now standard in survey weighting. These adjustments for coverage require an independent data source. In the past decade, the Census 2000 long form sample estimates have provided the best source of data for stratification and weighting adjustments. The 2010 Census did not collect these detailed social, economic, and housing data. The American Community Survey (ACS) is the new source of this information, providing valuable information for survey designers. This round table will discuss how best to use ACS data in sample design (stratification) and survey weighting, and will cover advantages as well as issues and limitations.