Investigation of Data Release Rules for Medians and Zero Estimates for the American Community Survey

• Activity Number: 20
• Type: Topic Contributed
• Date/Time: Sunday, August 1, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Karen Ellen King*+
• Companies: U.S. Census Bureau
• Keywords: American Community Survey ; Data Quality/Reliability Indicator ; Data Release Rules

Abstract

The American Community Survey (ACS) uses various techniques to improve the quality of the 1-year and 3-year data products it releases based on a series of data release rules. The release rules use the size of an estimate's coefficient of variation (CV) as a quality/reliability indicator. Products with too many estimates having large CVs fail the release criteria and are not published. This research is in response to feedback from ACS users that the current release rules are too conservative, especially those applied to certain types of estimates. We look into alternatives to the current release rules as they are applied to ACS 1-year and 3-year medians, ratios and zero estimates.