Comparison of Alternatives for Controlling Group Quarters Person Estimates in the American Community Survey

• Activity Number: 293
• Type: Contributed
• Date/Time: Tuesday, August 8, 2006 : 10:30 AM to 12:20 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Lynn Weidman*+ and Michael Ikeda and Julie Tsay
• Companies: U.S. Census Bureau and U.S. Census Bureau and U.S. Census Bureau
• Keywords: American Community Survey ; group quarters ; estimation ; controls

Abstract

Estimation for persons residing in group quarters (GQ) has been carried out only once with American Community Survey data. In 1999 there were 36 counties in sample and GQ stratification, sampling, and weighting was done separately for each county. For the full GQ implementation that began in 2006, a new GQ sort by type within state and selection across the whole state was used, making it possible to also weight GQ persons by state. A simulation study using Census 2000 data evaluated five alternatives for controlling estimates of GQ persons in the weighting: (1)no control, controlling by demographics for (2) GQ persons by state and for GQ and HU persons combined by (3) estimation area (large county or group of smaller counties) and (4) state, and (5) controlling by major GQ types for state. The simulation is described and analyses of the results are presented.