The Supreme Court heard arguments Monday on whether Census 2000 should be traditional or accurate. Although the court isn’t given to tipping its hand, the indications are that the Nine Wise Ones may sit this one out.

Steering clear of a fight about power and money that can only be settled by long-dead Founding Fathers — always wise.

Overall, recent censuses have been remarkably accurate — it is estimated that the 1990 count missed only 4 million Americans, not bad for a population so big and so mobile. The problem is that most of that undercount was in cities and in poor rural regions and mostly among minorities.

That’s where the power and money come in. Congressional districts are apportioned according to population, ditto some $182 billion in various forms of federal aid to local governments.

The solution proposed by the National Academy of Sciences (in a study ordered after the 1990 census by then-President George Bush) is to use statistical sampling to fill in the gaps. Statistical sampling is a proven technique; experts say its use in Census 2000 would cut error by roughly three-fourths and, by reducing the number of repeat visits by census-takers, would cut the cost from $4.8 billion to $4 billion.

A census that is more accurate census and $800 million cheaper probably sounds good to most Americans. It certainly is appealing to the Clinton administration and to House Democrats, who want as much of their traditional constituency counted as possible, and to cities and poor rural regions, who want and deserve their fair share.

It’s abhorrent to the majority of House Republicans, who have tied a ban on statistical sampling to a series of funding bills, ranging from disaster relief to transportation. While they’ve gotten some public-opinion black eyes for the tactics, they may well win in the end.

The GOP bases its objection on the Constitution, which calls for an “enumeration” every 10 years. The Framers, framing at a time when the population pretty much stayed put and out in the open, did not elaborate, they left no clue as to whether the point is to test the diligence of census-takers or to get the best possible census. The Constitution is a wonderfully clear and consise operator’s manual for a new nation; this is just one more example of where a little wordiness would have been welcome. At the very least, the authors could have sprinkled it with the occasional “use common sense.”

Actually, the Republicans’ best argument against sampling is one the court cannot consider: The Clinton administration could manipulate the census to its own advantage. At any other time and place, such an objection would seem paranoid. But this, after all, is the Clinton administration.

The GOP probably will prevail. The Supreme Court decided in 1996 that the 1990 census did not have to be corrected, even after sampling proved it was flawed. Two federal courts already have ruled against sampling in Census 2000, the high court may see no reason to overturn. And the Republican majority gets to set the final budget for the census bureau next spring.

Some of the undercounted are responding by getting active. Many black, Hispanic and Asian leaders already are urging their followers to stand up and be counted when the census-takers come knocking. This would be a good practice for community leaders everywhere to emulate. Unless the GOP has some constitutional objection to people answering the door.