Sampling Errs
Republicans may have received a big boost in their crusade against census sampling last week when, of all things, the Vermont unemployment numbers came out. Describing a sizable drop in the unemployment rate between July and August, state officials said that a statistical sampling error by the U.S. Census Bureau potentially made the difference and that more accurate data will be available at the end of the year.

The leading opponent of sampling in Congress, Rep. Dan Miller (R., Fla.), immediately publicized the admission, which his staff had found in an Associated Press story. "This is why we cannot under any circumstances have sampling in the 2000 Census," he said. "If the Census Bureau made a statistical sampling error in calculating the unemployment rate of one of the least populated states in the nation, with a population of 589,000, how can we trust that they won't make a sampling error estimating 10 percent of the U.S. population--some 26 million people?"

Despite a federal court decision that says sampling for purposes of Congressional apportionment is illegal, the Clinton administration continues to forge ahead with a plan to sample in the next census. As long as the Supreme Court hasn't barred the practice, says "the most ethical administration in the history of the republic," then it's okay to spend tax dollars on sampling preparations.

The Supreme Court, for its part, probably will rule on sampling by early next year.