Census Sensibility
of the Statistical Research Division of the US Bureau of Census, details how this is possible in the American Scientist. [1]
However, this is more than a scientific or mathematical question. It is a question about political process. It is a question about whether we ought to enact a process making it simple for less than scrupulous people to make minor adjustments to statistical assumptions in gathering the 750,000 household sample and potentially make significant differences in the apportionment of representation. Although 750,000 is a large number, that is less than 2,000 per Congressional District. It does require much imagination to realize that small changes in the samples at the margins could affect apportionment of key Congressional Districts.
The Clinton Administration is particularly notorious at bending the supposedly neutral rules of government. Michael Barone has pointed out that the Clinton Administration had little problem in ignoring criminal checks normally required by the Immigration and Naturalization Service to get more Democrats naturalized for the 1996 elections. (Note the Administration's assumption that avoiding criminal checks would admit more Democratic than Republican citizens.) It is not likely that this Administration "would suddenly develop scruples about adjusting census numbers for political purposes."
Democrats are rightly concerned that the Census in the past has not counted marginalized citizens, citizens they believe would vote Democratic. There are Republicans who, if they were in power, might exhibit Clinton-like ethics and display the moral restraint of a tomcat in fudging the numbers for political benefit. Imagine if a Richard Nixon had at his disposal the ability to alter Census numbers.
This ought not be a question of passing political advantage. If more money is required to make a more complete actual enumeration, then spend the money. It is a small price to pay for maintaining the integrity of the political process.
Latest Articles