An official website of the United States government
Here’s how you know
Official websites use .gov
A .gov website belongs to an official government organization in the United States.
Secure .gov websites use HTTPS
A lock (
) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.
This study examines whether reasonable and consistent sub-annual estimates can be produced from the American Community Survey (ACS) without changes to the existing weighting methodology and, by extension, whether it would be advisable to include variables for month or quarter of interview on the ACS Public Use Microdata Sample (PUMS) files. The ACS was fully implemented for housing units in 2005 and for the total population, including the group quarters (GQ) population, in 2006. The ACS is a continuous monthly survey with an annual sample of roughly 3 million addresses, approximately 250,000 addresses per month, resulting in about 2 million annual interviews. Twelve months of sample interviews are then cumulated to produce an annual 1-year period estimate. Similarly, 36 and 60 months of sample interviews are cumulated to produce 3- and 5-year period estimates, respectively. The first 3-year period estimates, for 2005 through 2007, were released in 2008. The first 5-year period estimates, for 2005 through 2009, will be released in late 2010. A number of users have approached the Census Bureau to request sub-annual estimates through the ACS Custom Tabulation Program. Others have requested information on month or quarter of interview be added to the ACS PUMS. With a few exceptions, the Census Bureau has declined requests pending the results from research designed to understand the consistency and properties of sub-annual estimates using ACS data. There are some concerns about the potential impact of certain aspects of the current ACS design on estimates of sub-annual periods. One is the ACS multi-mode data collection operation. The sample selected for a given month can be interviewed at any time over three consecutive months by one of three possible collection modes (mail, telephone, or personal visit). This introduces subtle differences in the number of respondents and the distribution of their characteristics between months. Another is the current ACS weighting methodology. The weighting/estimation process produces weighted estimates for twelve-month periods, i.e., one or more calendar year. The process is not designed to produce weighted estimates for shorter periods, i.e., monthly, quarterly, or biannual. These design components will be discussed further in this document. The focus of this research is on two questions: A. How do monthly ACS estimates for race, sex, and Hispanic origin compare to monthly estimates from the Census Bureauâs Population Estimates Program (PEP)? B. Do ACS sub-annual estimates behave in a demographically logical and consistent way over time using the current weighting methodology?
Share
Some content on this site is available in several different electronic formats. Some of the files may require a plug-in or additional software to view.
Top