Decennial Census

Lessons from the 1990 Census, which produced data less reliable than in 1980 despite being the costliest in history, led the Census Bureau to redesign key features for the 2000 Census. These changes included developing procedures for compiling a complete and accurate address list, increasing the mail response rate through outreach efforts, staffing census operations with a more capable workforce, and reducing costs and improving accuracy through sampling and statistical estimation. Congress, however, has not endorsed the overall design because of concerns about sampling and estimation. Although the dress rehearsal, now underway in three localities around the country, should be testing the operations and procedures planned for use in 2000, it will leave several design questions unresolved and some procedures untested. The longer the impasse continues between the Administration and Congress, the greater the risk that the 2000 Census will fail.

GAO noted that: (1) the dress rehearsal for the 2000 Census is currently under way at three sites: (a) Sacramento, California; (b) 11 counties in the Columbia, South Carolina, area; and (c) Menominee County in Wisconsin, including the Menominee American Indian Reservation; (2) although it was originally intended to demonstrate the Census Bureau's plans for the 2000 Census, the dress rehearsal will instead leave a number of design and operational issues unresolved; (3) these unresolved issues led GAO in 1997 to raise concerns about the high risk of a failed census in 2000; (4) accurate address lists and associated maps are the building blocks for successful census; (5) however, the Bureau has concluded that its original procedures for building the 2000 Census address list might not meet its goals of being 99-percent complete; (6) although the Bureau has since revised its address list development procedures, they will not be tested during the dress rehearsal, thus it will not be known until the 2000 Census whether they will meet the Bureau's goal; (7) the Bureau's outreach and promotion initiatives are designed to boost mail response rates and thus avoid costly followups to nonresponding households; (8) while the Bureau is to rely on partnerships with the local governments and organizations to raise public awareness of the census, the level of participation in these efforts has been inconsistent during the dress rehearsal, suggesting their impact on response in 2000 may be limited; (9) uncertainties surround the Bureau's ability to staff the 295,000 mostly temporary office and field positions necessary to conduct the census; (10) census jobs may not be as attractive as other positions, and, if current trends continue, the Bureau could find itself competing for workers in a tight labor market; (11) the Bureau's sampling and statistical estimation procedures, while they could reduce costs and improve accuracy if properly implemented, face methodological, technological, and quality control challenges; (12) in addition to these operational challenges, the Bureau has not finalized its plans for evaluating the dress rehearsal, thus it is not known whether the evaluations will provide needed data to assess the feasibility of the Bureau's plans for the 2000 Census; (13) further, Congress has not endorsed the Bureau's overall design of the 2000 Census because of its concerns over the Bureau's plans to use statistical sampling and estimation procedures; and (14) the longer this impasse continues, the greater likelihood of a failed census.