Let's be clear about the sampling issue.
The Census Bureau has been using sampling for years. Detailed socioeconomic data collected on the long form have always come from a representative sample of households. The plan for Census 2000 is a proposal to augment the traditional census with three samples:

1. A non-response sample of the estimated 34 million households that will not return their forms.
2. A "UAA Vacant" sample of households identified by the U.S. Postal Service that are "undeliverable-as-addressed/vacant," but which may, in fact, be occupied.
3. A random nationwide sample of 750,000 households that would be used to check for accuracy of the census and especially to adjust for an undercount.

Undercounts were a problem with the 1990 Census and they have been getting larger with each census. In 1990, some 8.4 million people may not have been counted. Many of the people who were missed in 1990 were members of racial and ethnic minorities. In response, the bureau designed its Census 2000 plan to obtain a more accurate count.

The bulk of the Census 2000 data would still be collected in the traditional way, which is contacting and enumerating every resident in the United States. The initial phase calls for multiple mail contact, a toll-free telephone number, blank forms at convenient locations, and a strong advertising and community-based publicity program.

Sampling would be used to bring Census 2000 response rates up to 90 percent in each of the country's 61,258 census tracts. Instead of sending enumerators to a set percentage of non-responding households (as was done in 1990), tracts with lower initial response rates will have a larger portion of housing units sampled. Characteristics of the remaining 10 percent of households that were missed would then be estimated on the basis of the sample. The final census would include the combined results of enumeration and the new sampling procedure.

"The Census 2000 plan, which includes a sampling component, is both more cost effective and more accurate than the traditional enumeration method. It would save as much as a billion dollars. We believe very strongly in our plan," says Tom Jurkovich, Associate Director of Communications for the Census Bureau.

So, why all the controversy?
Enter House Speaker Newt Gingrich, who is among those who literally elevated the sampling issue to a federal case. Mr. Gingrich and Republican Congressman Bob Barr of Georgia brought suits challenging the legality of sampling against the Census Bureau and the Commerce Department. In August and September of 1998, the United States Court of Appeals for the District of Columbia Circuit and a three-judge panel of the United States Court of Appeals for the Fourth Circuit in Roanoke, Virginia, both ruled that the proposed sampling violated Federal law.

The legal issue that was considered in federal court centered on apportionment of election districts and determined that previous law did not authorize sampling when apportionment was at stake. But there are other important issues tied up with Census 2000. Census data are also used to determine the distribution of federal funds and to draw boundaries for congressional districts.

The nation's cities are particularly concerned about an accurate population count since they want to get their fair share of tax dollars. House Republicans would like to keep their majority safe from reapportionment or redistricting plans that might favor the Democrats.

Now the sampling question goes to the Supreme Court, which has agreed to review the lower court decisions. Oral arguments set for November 30. A final decision is expected by the spring of 1999.

In the interim, the Census Bureau will proceed with its two-track planning process, preparing for both a census using an element of statistical sampling and a census that would not use those methods.

What started out a well-planned effort to improve the accuracy and cost effectiveness of Census 2000 has become a political quagmire. The National Academy of Sciences and other experts condoned this sampling process. Now it will be examined by the High Court. The outcome is likely to favor the old system and Census 2000 will most likely count those Americans who dutifully mail back their forms.