Sample Representivity in the American Community Survey

• Activity Number: 418
• Type: Topic Contributed
• Date/Time: Tuesday, August 6, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Survey Research Methods Section
• Author(s): Don Keathley*+ and Steven P. Hefter
• Companies: US Census Bureau and U.S. Census Bureau
• Keywords: American Community Survey ; housing unit address sample ; sample representivity

Abstract

The American Community Survey (ACS) selects national housing unit address samples on a yearly basis. Each sample is selected systematically, using geography and estimated occupied housing unit counts within specific geographies as sort variables. Every housing unit address on the ACS frame is eligible for sample once every five years, with approximately one-fifth of the addresses being eligible in a given year. Weighted response rates for the yearly samples average above ninety-seven percent, so the ACS has respondent information on the vast majority of the sampled units. But, there is still a two-plus percent nonresponse rate, and it is uncertain as to whether these cases are systematically different from the respondents, for one or more estimation categories of interest, e.g., race. Sample representivity statistics attempt to quantify the representativeness of the responding units to the nonresponding units and, by extension, to the entire frame for these categories