Census 2010 - No Sampling

From what I've read, the US 2010 census will not use sampling.

In 1990, the Census didn't count 8 million people and counted 4 million people twice. (Did they do a better job in 2000? I don't know.)

As the US population is around 300 million, this means that the Census has an error of anywhere from 1%-4% at the national level.

The people who aren't counted are poor, people of color, and undocumented immigrants.

This impacts communities at least at two levels - federal allocations of funds and the size of electoral districts. I'm unsure what federal funds are allocated based on population versus which are allocated by need (or by patronage), but I suspect the amount of dollars that some communities are missing out on is substantial.

In addition, the undercounting and overcounting of people leads to Democrat-leaning electoral districts that are larger than they should be, and Republican-leaning ones that are smaller. For instance, if you could have a Democrat-leaning district with 102,000 people and a Republican-leaning one with 99,000 where the US census counts the same (100,000) people. This would give the Republican-leaning voters 3% more voting power per person.

The impact of this effect would be seen most strongly on the local level, as the smaller the district, the more likely the Census is to get the count wrong. For instance, the Census will dramatically undercount a poverty stricken inner city neighborhood, but if you look at an entire US state - the rich and poor areas will be more in balance. So it'd be more of an issue for state senate and legislative elections, city council districts, and congressional house seats then for the US senate or president.

Ultimately this is yet another bias in the US electoral process which favors the Republicans (See my earlier post estimating that Obama needed 55% support of US residents over 16 to win the election - http://www.campusactivism.org/blog/node/271)

Fair sampling would mean smaller and thus more inner-city districts, or combining parts of existing inner-city districts into suburban districts giving Democrats a better chance of winning those. It could also help a progressive Third Party like the Greens make a breakthrough (into state legislatives, city councils, and ultimately congress).