Information on the Census Bureau's Plan for Conducting the 2000 Census Without Statistical Sampling
GGD-99-12R: Published: Dec 17, 1998. Publicly Released: Dec 17, 1998.
- Full Report:
Pursuant to a congressional request, GAO reviewed the Department of Commerce's Status Report on Planning for a Decennial Census in Year 2000 Without the Use of Scientific Sampling as of April 1998, focusing on: (1) the Bureau of the Census' plans and cost estimates--as contained in its April 1998 report--for preparing a decennial census that does not require the use of sampling; and (2) how the Bureau's fiscal year (FY) 1999 cost estimates for preparing such a design were developed.
GAO noted that: (1) Commerce's April 1998 status report presented a preliminary plan for doing the 2000 Decennial Census without the use of statistical sampling; (2) many of the elements of the Census Bureau's current sampling-based design, such as field office infrastructure and staffing, questionnaire development and delivery, special population and area counts, automated data processing, and marketing and local partnerships, could require modification or expansion if a nonsampling census is conducted in 2000; (3) Bureau officials said that the Bureau's efforts to revise its current design to exclude sampling procedures required additional research and operational analysis that the Bureau had not previously considered or conducted; (4) the Bureau expects to issue a plan for conducting a census without sampling, an associated master activity schedule, and FY 2000 cost estimates for a nonsampling census by the end of December 1998; (5) in Commerce's status report, the Census Bureau identified a need for $312 million in additional FY 1999 funding to prepare for a non-sampling effort; (6) this estimate included cost increases for seven of the eight decennial census program activities in the President's FY 1999 budget, such as $82 million for field data collection and support systems and $77 million for data content and products; (7) for the April 1998 status report, senior Bureau officials provided their best guess judgments, based on prior census experience, of the costs of potential components of a nonsampling plan; and (8) these officials said that their estimates were not supported by in-depth quantitative analysis and that, until the design for a nonsampling plan is completed, the Bureau could not estimate precisely what the total cost of such a census would be.