With the 2000 census fast approaching, the Clinton administration had pressed the court to permit demographic sampling and estimates to obtain a more accurate count of the population in likely Democratic areas. Past studies have shown that poorer people living in inner-city areas -- and presumed to vote Democratic -- have been harder to reach and thus often undercounted when direct methods, such as mail-in questionnaires or door-to-door inquiries, have been used. "Whatever the correctness of the legal interpretation," says Pooley, "the decision is wrong on the reality. Every responsible analyst has said there is a serious undercounting without sampling adjustments in certain areas of the country." As a result, the 2000 census is expected to produce fewer Democratic congressional districts, and thus fewer Democrats in Congress, than might have been possible with statistical estimates. Success in politics, as always, remains a numbers game.
If you think the conservative-liberal divide on the Supreme Court doesn't count for much, guess again: It does indeed count -- especially when it pertains to counting. By a 5 to 4 vote, the Justices ruled on Monday that, for the purpose of apportioning congressional seats, the Census Bureau can't use estimates to determine the population of areas where it's difficult to obtain the actual number of residents. Writing for the conservative majority of the court, Justice Sandra Day O'Connor wrote that the language and history of the federal law governing the census cannot be interpreted to permit statistical adjustments. "The clear arithmetic of the decision," says TIME senior writer Eric Pooley, "is that it will be a blow to Democratic interests."