How to conduct Census 2000 has become a highly contentious issue, and the delay in selecting an approach threatens the census's eventual effectiveness. In dispute is the proposed "census sampling" technique for addressing the problem of undercount. Because census data are used to apportion federal funds and to determine how congressional seats are divided among and within states, the ultimate decision will have an important political effect.

Media reports misleadingly suggest that sampling would heighten the accuracy of the count. But in Sampling and the Census, Kenneth Darga presents a compelling case against the proposed procedure. In an attempt to address an inaccuracy at the national level, the adjustment would seriously undermine the reliability of census data at the state and local levels.