Some critics may be too excited in their alarums about the danger of the Obama Administration's prospective politicization of the conduct of the 2010 Census. It is a serious threat that needs to be faced head on, but it is not the end of democracy as we know it.
Moreover, members of the U.S. Senate are going to get a chance soon to ask questions of the next nominee to head the Department of Commerce (the president's third choice), and then to vote on whether to confirm that nominee. After that they will get a chance to do the same with a a prospective new Deputy Secretary and then a new Director of the Census Bureau. A conscientious senator and his staff can get all the information they need through normal committee processes.
Meanwhile, members of the House will have several opportunities to conduct reviews of Census 2010 planning and find out what the Administration intends. Crucially, will the White House let the Census Bureau stick with the present plan that has been years in the making on a non-partisan basis, or does it intend to impose (or inveigle) a last minute, politically-motivated "adjustment" based on sampling?
Liberal Democrats and their allies in the media and the sciences have been lamentably quiet about this subject since it surfaced two weeks ago. There hasn't been a peep out of the National Academy of Sciences that I have heard. (I'll be happy to be shown that there has been one if I have missed it.) Some of these folks have strong opinions--which they present as facts--on other matters bearing on science policy, but they apparently have trouble standing up for the scientific discipline of statistics when it is being given a political arm-twist.
It makes one wonder how far politics reaches.
Meanwhile, the danger for Republicans is that they merely will vent and gesture when the confirmation and oversight hearings come. If they don't prepare themselves they are likely to be rendered ineffectual in the end. To put it bluntly, the Senate and House staffs need to be doing their homework now.
To begin with, they need to read and digest the 2003 report of statisticians who dealt authoritatively with the adjustment/sampling issue. It is readily available through the ever-cooperative Census Bureau. Legislators would be wise to consult with statisticians personally who know this subject cold and also with project managers who have handled a Decennial Census in the past. These experts know all the difficulties in both theory and practice.
Legislators also need to study the 1999 Supreme Court case on adjustment (United States Department of Commerce v. House of Representatives), a road-block to sampling for adjustment. And talk to some law professors. Among other things, they need to be alert to Administration attempts to use cagey language to dodge the Court ruling.
They also need to acquaint themselves with the outreach programs the Census Bureau already has underway for 2010 so that they can blunt interest group claims that adjustment by sampling and computer models is the only or best response to an undercount. (Thanks to such efforts in past decades, by the way, the undercount has substantially diminished.)
They also have to respond to the proposed legislative alternative of making the Census Bureau an independent agency, or, as another option, increasing the prominence and relative independence of the Bureau within the Commerce Department.
The defenders of a Decennial Census that relies on an actual enumeration and eschews the mistake of adjustment by sampling can prevail on this issue. But they will do so only if they do adequate preparatory work. Otherwise, they may get rolled. And that would be a shame for the country as well as for them.
