U.S. House Seat Hinges On Ruling

The battle began when Utah sued the U.S. Census Bureau over census data-gathering techniques that cost Utah a congressional seat, handing it to North Carolina.

The showdown -- Utah v. Evans -- made its way to the U.S. Supreme Court Wednesday and had the justices questioning the constitutionality of "hot deck" imputation -- a practice used to gather census data.

Hot deck imputation is the process of assigning occupancy to a residence based on the data of the nearest neighbor.

The U.S. Census Bureau has used hot deck imputation to determine population in past censuses, but the standards for the 2000 census were less stringent than in previous years. The looser standards are constitutionally suspect as they might qualify as a form of sampling, which the Supreme Court has previously ruled unconstitutional.

In 1999, the Supreme Court ruled 5-4 to prohibit "the use of statistical sampling in calculating a population for purposes of apportionment" but did not strike down the use of any statistical methods whatsoever.

During the hearing Wednesday, Utah counsel Thomas Lee argued that inferring the population of a home based on that of a neighboring residence is unconstitutional because making such an inference goes beyond using a mere statistical method for gathering data.

"That's sampling," he said. "Sampling is taking information about a part to make an inference about the whole."

But U.S. Solicitor General Theodore Olson, who represented the federal government, argued that estimating occupancy in a select few cases is not a form of sampling.

"Sampling technique is completely discrete from the imputation technique," Olson said. "In my thinking of it, it is drawing logical inference from the data available."

But Justice Sandra Day O'Connor held that the key was not to determine whether hot deck imputation was a statistical method but whether all statistical methods can be ruled unconstitutional sampling. "We have to establish whether this hot deck imputation is a type of sampling, which it seems to me it very well may be," she said.

Justice David Souter said a distinction should be made between making an inference for a small group and random sampling like that previously struck down by the court.

Souter held that it is logical to make an educated guess about occupancy based on who lives next door. "Birds of a feather flock together," he said. "People who live near one another tend to be alike."

But Souter said that no matter whether inference is determined to be unconstitutional sampling, Utah's grievances might not be redressable because more than two years have passed since the transgression occurred.

"It's a real issue in this case whether a statute that is geared to information at a particular time can reverse itself," he said.

O'Connor concurred that both states have progressed since the census was taken and that population likely has shifted. "Things have happened since that census was taken," she said.

But Lee claimed Utah could not have filed suit earlier than it did and that revised population numbers could still be legally gathered and used for apportionment.

The Supreme Court is expected to rule some time before its summer recess.

Even if the Supreme Court were to rule in favor of Utah and new population data were collected, President Bush might not submit those numbers to Congress and would be within his legal right not to do so, stated several justices, including Justice Antonin Scalia.

"I sort of wouldn't want to take away a representative from the people of North Carolina," he said. "If I were the president, I might not want to do it and risk making the people of North Carolina mad."

The State & National Editor can be reached at stntdesk@unc.edu.