The decennial census of the population is one of the most important tasks undertaken by the federal government–and one of the hardest. A complete count is impossible, because there are so many people in the United States, some of them hard to find. Experts say the last census missed about 4 million people, including 2.4 percent of those in Chicago.
The Clinton administration wanted to address this problem by using statistical methods known as “sampling” to arrive at estimates of people who are omitted by the traditional head count.
But in January, the Supreme Court ruled that federal law does not permit sampling for purposes of congressional apportionment. It’s not clear that, if obliged to decide, the justices would conclude that the Constitution does either.
The most noteworthy consequence of the verdict is that when it comes time to divvy up seats in Congress, some states may be shortchanged. That can’t be helped. What can be avoided is using a plainly faulty tabulation for other purposes.
The court held that sampling was forbidden for apportionment. For all other purposes, though, it not only is permissible but may be required. So the administration plans for the Census Bureau to come up with two numbers in 2000–one based on traditional door-to-door methods for parceling out House seats and another using state-of-the-art techniques for such purposes as distribution of federal money and state legislative redistricting.
That Solomonic proposal is imperfect, but not as imperfect as the alternative, which is to use the less accurate tally for everything.
Republicans object to spending any extra funds to supplement the conventional census, and warn the public will be confused. But it’s hard to see the sense in refusing to allocate government aid in accordance with where the intended beneficiaries actually are.
The Constitution may bar the use of estimates when the sacred matter of voting is involved, but that principle doesn’t apply when it comes to social welfare programs.
It has not escaped the notice of either party that the people who are missed in the old-fashioned census tend to be the kind of people (poor, minority, urban) who generally vote Democratic. But pretending they don’t exist is not likely to work to the long-run advantage of the GOP. Now that they’ve won on apportionment, fairness and political wisdom argue that Republicans should compromise on the other census battle.
