The announcement today by Sen. Judd Gregg that he is withdrawing as the president's nominee to be Secretary of Commerce--and citing the Census issue as one reason--is going to increase scrutiny of White House plans to try to change the way the 2010 Decennial Census is conducted. Clearly the president has done little to alleviate concerns of Gregg, a Republican, that the Census might be conducted in a politicized manner.

It is now unavoidable that the White House statements claiming oversight of the Census preparations were not flukes caused by ignorance or naivete. They must have been serious or Sen. Gregg would not have decided to withdraw. One previously could give the president the benefit of the doubt. He was too busy with other matters to make his position clear. Now he has lost another cabinet nominee--not because of faults in the nominee, but because the nominee didn't want to be part of the Administration that ignored the chain of command and tried to micro-manage a function traditionally left to career scientists.

It is also possible, however, that Gregg didn't quit, but was shoved. The West Wing officials by now may have figured out that there are legal as well as political risks if they try to change the plans for the Census from the White House instead of the Commerce Department. Lawsuits were threatened today by House Republicans.

It would be more expedient for the White House to have a pliable Secretary of Commerce in place if the aim is to "re-evaluate" the conduct of the 2010 Census in order to introduce adjustment of results through sampling and computer modeling. Gregg presumably would not have gone along--and would have been hard to run over.

But the legal issues will remain even if a willing partisan is nominated and confirmed as Commerce Secretary. There is a 1999 Supreme Court ruling that would make sampling-based adjustment difficult in the absence of compelling evidence that the customary hard count would be less credible. And that evidence not only is lacking, but a three year statistical study that was finished in 2003 to respond to this issue concluded just the opposite: adjusting the Census numbers through sampling and computer models could lead to a less credible Census result. A hard count has always been legally defensible. A fuzzy "adjusted" Census--where figures at the Census tract and block level would be demonstrably erroneous in many cases--could invite endless litigation and bad will.

Another problem for the Obama White House if it wants to change the Census approach: planning for the 2010 Census has been underway for years and now is in preparation for testing. The disruptions caused by an Administration decision to change those plans would cause great problems and probably agitate the resistance of career statisticians charged with responsibility for conducting the Census.

Finally, one wonders if the President understands that the Census is a function of government that requires not only integrity in fact but also the appearance of integrity. The reputation of the Census should not be compromised. It is hard enough to get people to cooperate in the conduct of the Census without creating a reputation for politicization.

Those on Capitol Hill that want White House direction--and an adjustment plan--for the Census say they worry about the Census Bureau's ability to reach out to minorities and urban dwellers and to "count everybody." In truth, the Census Bureau's record on reducing the "undercount" has improved with each new Census. Huge efforts are made to reach out to all groups.

However, what might well cause the undercount to grow would be efforts to politicize the Census, thereby generating doubt in large parts of the citizenry that normally are quite willing to take part. If that is the case, and the Administration refuses to back off, Sen. Gregg's withdrawal will prove to be only one in a series of likely developments that will trouble the 2010 Census.

Various kinds of individuals, myself included, have served as Directors of the U.S. Census Bureau. One thing I think we all have in common is a desire to have the statisticians' work respected, without politicization. As far a the general public, I felt a bit alone on this subject a week ago when I first blogged on this, but awareness has grown daily in a startling way. The Gregg withdrawal really makes the issue unavoidable.