Census and Sampling:

What Statistics tells us about it

William Hooper - SUNY Binghamton

Every ten years the U.S. government attempts to count every inhabitant of the country. So far, they have never succeeded. Why not? What could be done differently? Is sampling, as is currently being suggested by the Democratic party, the answer? Why do the Democrats favor sampling, while the Republicans are against it? If sampling would not correct the errors, what changes would it make?

The Supreme Court ruled earlier this summer that using sampling to determine representation in the House of Representatives was unconstitutional. The President then ordered two census reports:

1. A traditional census to determine representation in the House of Representatives.
2. A census utilizing sampling procedures to be used for all other applications of the census.

What will be the difference between these two reports? Which will be more accurate? Will the second report really favor the Democratic party, or the first favor the Republican party? How?

This talk is accessible at all levels.

William Hooper is currently studying for a PhD at SUNY Binghamton with research focused on Statistics. He received his B.A. in Mathematics and his M.A. from SUNY Binghamton in 1994 and 1996 respectively. He is a fellow in the "Preparing Future Faculty" Program in the Department of Mathematical Sciences.

 
• About this document ...