user education materials, web site design, table content, and other aspects of the data products and education and outreach program for the ACS.
A major continuing survey, such as the ACS, requires continued monitoring to ensure that data collection and production processes are performing well, to identify problem areas for investigation and development of improved processes, and to provide information to users about sampling and nonsampling errors of which they should be cognizant. For these purposes, it is essential to develop and track an appropriate set of performance measures.
Some performance measures are for use by survey managers to ensure that survey data collection and processing operations are being carried out as specified and within quality control tolerances and to flag problems for investigation. Such measures may track timely completion of check-in and data capture of mailed-back questionnaires, interviewer productivity, and the like. The panel did not review what measures the Census Bureau uses for quality control of the ACS; we trust that the Census Bureau has developed a set of appropriate measures and periodically reviews them for relevance and usefulness in identifying problems on a timely basis.
Other performance measures are useful not only to survey managers, but also to inform users of the quality of the data across areas and population groups. The Census Bureau has long experience with monitoring and maintaining the quality of its survey operations. For the ACS, it has taken a further step to put up on the ACS web site basic indicators of sampling and nonsampling errors in the data.
The Census Bureau currently provides four indicators of nonsampling errors, which can be accessed from the main ACS web site under “Using the Data” (http://www.census.gov/acs/www, “Quality Measures”). The four measures are sample size, survey coverage rates relative to the 2000 census-based population estimates, survey response rates (unit response), and item nonresponse rates (refer back to Box 2-4). At this time, all four indicators are available for the Census 2000 Supplementary Survey, the ACS 2001–2004 test surveys, and the 2005 ACS for the nation and states.
The panel commends the Census Bureau for providing quality measures for ACS estimates on its web site. For these measures to be useful, it is important that users of the data access them and interpret them correctly. The Census Bureau’s data user advisory group and network of user educa-