The American Community Survey has implemented several improvements during the 2011 data collection period. The goal of these improvements is to increase the quality of survey estimates. They are being implemented as swiftly as resources allow.
Planned Improvements - housing unit address sample
- In January 2011, the ACS began conducting personal visit follow-up with every address selected in the initial sample in remote Alaska.
- Beginning with the personal visit follow up for the June monthly sample (which takes place in August), every non-responding and non-mailable address in the following areas will be followed up with a personal visit: American Indian areas with an estimated American Indian population greater than 10 percent, all Alaska Native Village Statistical areas, and all Hawaiian Home Lands. Due to the large number of non-mailable addresses in these areas, conducting 100 percent follow-up for approximately 27,000 additional sample addresses is the most effective way to improve the quality of estimates in these areas.
- In January 2011, the ACS modified the sample design to improve the reliability of estimates for small Census tracts. This improvement increases the sampling rates for smaller tracts and slightly decreases the sampling rates for larger tracts, thereby minimizing the differences in the reliability of our estimates. In addition, the sampling rates for the very smallest governmental units were increased, improving the reliability of estimates for these areas.
- Beginning with the June 2011 monthly sample, the initial sample will increase to 3.54 million addresses annually (approximately 295,000 addresses per month).
Purpose and Goals of Improvements
- To reduce sampling error, thereby enhancing the ability of users to determine trends and changes over time.
- To facilitate more efficient allocation of funds distributed on the basis of ACS estimates.
- To facilitate a more accurate portrait of community and small group characteristics.
Many of the benefits of these improvements will be realized with the release of 2011 ACS data in 2012. The full impact of the improvements will require 5 years of full data collection and will be best seen in the 2012-2016 ACS 5-year estimates.
The Census Bureau will provide additional details prior to the release of the 2011 ACS estimates in the fall of 2012.