Sampling the 2000 Census

Sampling the 2000 Census

By Sarah E. Torian

Vol. 20, No. 3, 1998 pp. 24-25

In the upcoming months the Supreme Court will be making an important decision that will affect what has been termed the “most important civil rights issue for the remainder of the 105th Congress.” The issue is the Decennial Census and whether or not scientific sampling methods will be used to improve the accuracy of this population count in the year 2000.

The Decennial Census, a vital part of a fair and functioning democracy, is becoming less effective. For the first time since the 1940 census, the 1990 Census was less accurate than its predecessor. The 1990 net undercount or percentage of people who are missed by the Census count, was 50 percent greater than its 1980 counterpart. According to the Census Bureau, the 1990 Census missed 8.4 million people and double-counted 4.4 million, resulting in an undercount of 4 million. Of that 4 million, over 1.5 million (39 percent) were in the eleven Southern states.

Even more disturbing is the “differential undercount,” the difference between the undercount of whites and the undercount of ethnic minority groups. The “differential undercount” for the 1990 Census was the highest ever recorded since the Census Bureau began conducting post-Census evaluations in 1940. Most of those overlooked in the Census are children, people of color, and the rural and urban poor. Children under the age of 18, a group that represents 26 percent of the total national population, accounted for an incredible 52 percent of the undercount.

Considering the influence of the Census, these statistics are quite disturbing. The Census figures are used to allocate seats in the House of Representatives, to distribute tens of billions of dollars annually in federal, state and local program funds, to apportion electoral college votes, and to carry out congressional, state, and local redistricting, among other things. Since poor, minority, and central city populations and children, those most frequently undercounted, are also communities that are often most in need of social services and economic development programs and are frequently poorly represented as a constituency, it becomes increasingly important to ensure as fair and accurate a count as possible. Otherwise, these groups are denied important services and are denied their right to fair and equal representation.

Factors such as poverty, illiteracy, lack of education, fear or mistrust of government, language barriers, geography, and changing family structure all contribute to these lower rates of coverage.

In an effort to address the limitations of present enumerations methods, the Census Bureau, with the help of

Page 25

National Academy of Sciences, General Accounting Office and Commerce Department’s Office of Inspector General panels, produced “The Plan for Census 2000.” This plan combines a more aggressive direct enumeration effort, including several mailings to every household and multiple response options, with modern scientific sampling techniques to complete the count of the final non-responding households.

This plan is currently under attack by a Republican-led House that wants scientific sampling methods banned from use in collecting Census data and filed a lawsuit, arguing that the use of sampling methods would violate federal law. The Clinton administration, which supports the use of scientific sampling, is challenging a three-judge federal panel’s August 25 decision in that case that invalidated the Census Bureau’s plan without ruling on the constitutionality of sampling.

The Supreme Court will hear arguments November 30 and will likely decide by March whether scientific sampling can be used in the 2000 Census. A decision by March is vital. Since the lower court’s ruling, administration officials have been preparing for the upcoming Census on a “dual track.” one using scientific sampling and the other using only direct headcount methods. lawyers for both the House and the administration said, “If the current uncertainty continues beyond March 1999 [the] ability to conduct the most accurate census possible will be seriously threatened.”

For more information on scientific sampling in the 2000 Census, visit and

Sarah E. Torian is program assistant in communications at the Southern Regional Council