"Census Case Tests Statistical Method," by Charles Seife. Science, 1 February 2002, page 783.
As a result of the 2000 U.S. census, the state of Utah lost a congressional seat. Utah is challenging a statistical technique used in the census, claiming that it violates the 1999 Supreme Court decision that barred sampling in the census. In 2001, three Utah judges ruled against the state and the U.S. Supreme Court has agreed to hear Utah's case. The technique in question is called "hot deck imputation"---a name that originated when computer punch cards were used in the census. Hot deck imputation is used when census officials are confronted with a questionnaire that has either bad or missing data. When that is the case, officials find a household in the current census as similar as possible to the one with the inaccurate or incomplete data and use that household's data to make an educated guess. While Utah claims that imputation amounts to sampling, the Census Bureau and other defendants have claimed that it is consistent with enumeration. Those who are familiar with the case wonder that if imputation is not allowed, what is to be done with households that claim zero residents or claim a resident of age 179?