A Good Use For Sampling?

As you will recall, there was a huge debate over the use of statistical sampling in the 2000 census as a possible way of correcting what is presumed to be a large undercount of minorities. That debate resurfaced with President Obama’s now abandoned intention of moving the Census Bureau out of the Commerce Department and putting it under the control of political operatives in the White House, and now it has resurfaced again with his recent nomination of University of Michigan sociologist Robert Groves to head the Census Bureau.

Groves is a former Census Bureau associate director of statistical design, who served from 1990-92….

When he was the bureau’s associate director, Groves recommended that the 1990 census be statistically adjusted to make up for an undercount of roughly 5 million people, many of them minorities in dense urban areas who tend to vote for Democrats.

But in a fierce political dispute that prompted White House staff to call advisers to the bureau and express opposition, the Census Bureau was overruled by Republican Commerce Secretary Robert Mosbacher, who called the proposed statistical adjustment “political tampering.”

The Supreme Court later ruled in 1999 that the use of statistical sampling cannot be used to apportion House seats, but indicated that adjustments could be made to the population count when redrawing congressional boundaries.

So far, the debate over sampling has concerned its use to prevent an undercount of minorities, but there’s no reason it couldn’t also be used to correct an overcount as well, and Slate’s Mickey Kaus has pointed to a new problem that just may call for sampling as the solution. Last Thursday he asked: “Will Obama’s New Legalization Push Screw American Workers? Or Has It Already Screwed Them?”

Both supporters and opponents of illegal immigrant legalization think trying to pass it “while the U.S. economy is mired in economic turmoil” might be difficult. The fear/expectation is that Americans will see the new law as a plan to allow foreigners who aren’t supposed to be here to compete for the few jobs that are left, bidding down wages in the process. But here’s the thing: Just by re-opening the legalization issue, without passing any new law, Obama has already encouraged foreigners who aren’t supposed to be here to come and compete for the few jobs that are left, bidding down wages in the process. What better way to encourage more illegal immigration than by promising a possible amnesty in the next few years?

Kaus returned to that issue today, asking “Why might the Obama administration want to raise the topic of possible immigrant mass legalization even if it’s not going to happen this year?” His answer, based on the CNS article just linked:

… a state harboring more illegal aliens can gain more House seats as long as the Census Bureau finds the illegal aliens and counts them. This also means that the illegal alien population resident in the United States during a census year has the potential to alter the regional and philosophical balance of power in Congress.

Isn’t the fact that illegal aliens are notoriously hard to count exactly the sort of problem that sampling is supposed to solve?

There are, of course, a couple of obstacles to this solution. One is that pesky Supreme Court opinion barring the use of sampling in Congressional apportionment. But courts change their minds, or change when they change, and if not now, there will always be future censuses.

A bigger problem may be that many Americans would be furious at the thought that a Democratic president has appointed a Census director skilled in a technique whose use would increase Democratic majorities in Congress by counting large numbers of illegal aliens who can’t even vote (legally).

Always eager to avoid conflict and promote comity and conciliation, I have a compromise to suggest that might mollify both sides of this controversy: use sampling in areas that have high concentrations of Hispanics, but then use some smaller percentage of that total (60% worked for a similar compromise in Philadelphia in 1787) for the purpose of apportioning representatives, with the percentage used to be determined by sampling that produces an estimate of the typical ratio of illegal aliens to citizens and legal immigrants in such communities.

Oh, wait. I guess that compromise didn’t work so well when it was tried before.