Decennial Census

The results of the 1990 census reveal that the American public has grown too diverse and dynamic to be accurately counted solely by the traditional "headcount" approach. For the first time in decades, the 1990 census was less accurate than that of its predecessor, even as census costs soared. GAO estimates that the net 1990 census undercount was about 4.7 million people and that the 1990 census had at least 14.1 million errors, including missed persons and those improperly included in the count. GAO submits that this is an opportune time to rigorously explore reform in three main areas: (1) improving the development of address lists; (2) responding to declining public response rates; and (3) reducing reliance on costly, lengthy, and error-prone follow-up efforts. The primary challenge for the Census Bureau is to build on its advantages, heed the lessons of the 1990 census, and plan for the year 2000 with discipline and inspiration. Strong and continuing congressional oversight will be needed to ensure that the momentum for change is maintained and the redesign effort is successful. GAO summarized this report in testimony before Congress; see: Decennial Census: Opportunities for Fundamental Reform, by L. Nye Stevens, Director of Government Business Operations Issues, before the Subcommittee on Census and Population, House Committee on Post Office and Civil Service. GAO/T-GGD-92-51, June 10, 1992 (12 pages).

GAO found that: (1) the Census Bureau managed the 1990 census on a 10-year cycle, beginning planning in 1984 and ending major data dissemination efforts in 1993; (2) the 1990 census had millions of errors and omissions, a lower accuracy rate than that of the 1980 census, and dramatically escalating costs; (3) the Bureau used redundant, labor-intensive, and costly procedures to develop a list of about 100 million addresses for the 1990 census, but the list's errors and omissions affected both census quality and cost; (4) the Bureau could improve its address list for the 2000 census by building on its 1990 list and increasing its reliance on the U.S. Postal Service; (5) the census mail response rate was only 67 percent, down from 75 percent in 1980, and the continued downward trend in public cooperation affects census costs and quality; (6) the Bureau may be able to increase public cooperation by simplifying the questionnaire and modifying the census in response to social changes; (7) census follow-up efforts were costly, time consuming, and error prone, due to the unanticipated workload and staffing and schedule problems, and the quality of enumerations declined with extended data collection efforts; (8) the Bureau could improve its follow-up efforts by preparing for a lower response rate, evaluating its use of sampling, assessing coverage improvement strategies, and developing alternative methodologies; and (9) the Bureau's census redesign effort appears to be open to a wide range of participants and ideas, but reform will require strong and thorough planning efforts.

Recommendations

Our recommendations from this work are listed below with a Contact for more information. Status will change from "In process" to "Open," "Closed - implemented," or "Closed - not implemented" based on our follow up work.

Director: Team: Phone: