Get on With the Count

The Supreme Court has spoken on the volatile issue of how a census should be taken. Actually, the court spoke only in part. The issue remains charged, and fireworks could start anew when Congress deals with additional funding for the national head count this spring.

Counting every head, so to speak, lies at the heart of this controversy. As you probably recall, one side of the aisle in Congress took great umbrage at the idea of using scientific sampling methods to fill in census gaps. (And gaps there are: Some 4 million people were missed in 1990's count, mostly urban and mostly minority.)

Republicans argued that sampling techniques - no matter how praised by statisticians - violate the Constitution's requirement for an "actual enumeration." Above all, they weren't about to let numbers so derived be used to apportion states' representation in Congress.

On that, the court agreed. The five-justice majority found that federal law prohibits sampling for that crucial political purpose. But the justices left the door open to other uses of sampling: to set population figures for distributing federal spending, or even to guide state redistricting.

Its GOP critics, however, want sampling all the way out. Most Democrats, on the other hand, want to be sure that poorer Americans, who tend to vote Democratic, are counted. In the middle stands an already frustrated US Census Bureau, which needs quickly to set a steady course toward the 2000 count. Knocking on the front door of every household that doesn't mail back the census form is not a feasible alternative.

Justice Stephen Breyer, in dissent, hit it on the head. Sampling is a useful, statistically sound supplement to (not substitute for) an actual enumeration. As he said, it would achieve the "very accuracy that the census seeks and the Census Act itself demands."