Utah v. Evans

LOCATION: Los Angeles City Hall

DOCKET NO.: 01-714
DECIDED BY: Rehnquist Court (1986-2005)

CITATION: 536 US 452 (2002)
ARGUED: Mar 27, 2002
DECIDED: Jun 20, 2002

Thomas R. Lee - Argued the cause for the appellants
Theodore B. Olson - Solicitor General, Department of Justice, for the federal appellees
Walter E. Dellinger, III - Argued the cause for the North Carolina Appellees

Facts of the case

In conducting the 2000 census, the Census Bureau used "hot-deck imputation" to fill in certain gaps in its information and resolved certain conflicts in the data. Under this methodology, the Bureau imputes the relevant information by inferring that the address or unit about which it is uncertain has the same population characteristics as those of its geographically closest neighbor of the same type. Hot-deck Imputation increased North Carolina's population by 0.4% while increasing Utah's population by only 0.2% such that North Carolina will receive one more Representative and Utah one less than if the Bureau had simply filled relevant informational gaps by counting the related number of individuals as zero. Utah brought suit against the officials charged with conducting the census, claiming that the Bureau's use of hot-deck imputation violates 13 USC section 195, which prohibits use of "the statistical method known as 'sampling,'" and is inconsistent with Article 1, section 2, clause 3 of the Constitution, which states that an "actual Enumeration be made." Utah sought an injunction compelling a change of the official census results. The District Court found for the Bureau.


Does the Census Bureau's use of "hot-deck imputation," in the 2000 census, violate the statutory provision forbidding use of the statistical method known as sampling? Is this methodology inconsistent with the Constitution's statement that an "actual Enumeration be made?

Media for Utah v. Evans

Audio Transcription for Oral Argument - March 27, 2002 in Utah v. Evans

Audio Transcription for Opinion Announcement - June 20, 2002 in Utah v. Evans

William H. Rehnquist:

The opinion of the Court in No. 01-714 Utah against Evans will be announced by Justice Breyer.

Stephen G. Breyer:

When the Census Bureau counts population, it first sends out forms by mail to each address on a long list of addresses that it has.

It then contacts people who do not respond by sending employees to make personal visits, and they may make up to six such visits, but if they all fail to elicit the information, then the Bureau uses what it calls imputation, which is the subject matter of this case.

The imputation takes an address on its millions of addresses list, an address where there is inadequate information or maybe conflicting information, and it imputes the status say whether the address is a vacant lot or a house, or it imputes the number of occupants, how many people live there, and it does it by assuming that this unknown address has the same characteristics as the closest nearby similar unit.

In the year 2000 the Census imputation accounted for about 0.4% -- it is four tenths of one percent of the total count -- but because the imputation counts were spread in an uneven way across the country, that tiny difference makes a difference.

It gave North Carolina an additional seat in Congress, and if imputation had not been used, the seat would have gone to Utah.

Well, Utah found this out and it brought a suit, and in this suit it asks the Court to set aside imputation as violating both statute and the constitution.

Now, the lower court in this case, which is a three-judge Federal Court in Utah, held that the Bureau could use imputation; it was lawful, and that the seat belongs to North Carolina.

We now review that decision and we affirm.

We first have to decide, as North Carolina argues, whether Utah maybe could not even bring this suit because it would lack standing, but we find that Utah does have standing focusing on the relevant issue because in all likelihood, if Utah had won the case, which it does not, but if it had won, the Census Bureau, the President, and relevant congressional officials would in all likelihood of taken the steps that would have been necessary to provide Utah with that extra seat.

Therefore, this is a real case.

Then second, we turn to the statutory question.

The statute forbids the Bureau to use “the statistical method known as sampling”.

Is imputation that kind of statistical method?

Is it sampling?

The Solicitor General, in argument here, used an example to help explain why he thought it was not.

He said imagine a librarian who wants to find the number of books in a library, suppose the librarian uses a statistically sound way to take a sample, say count the books on every 10th shelf, then suppose the librarian uses a statistically sound method to extrapolate from the sample to the whole, say he multiplies by 10.

That is what is the SG said was the statistical method called sampling.

Now suppose instead that the librarian tries to count each book one by one, that is not sampling, and it does not become sampling just because the librarian finding a few empty spaces on some shelves, say the readers have the books out, imputes to those empty spaces a number of books, not even if he does this imputation in a rather technical way, say by measuring the size of nearby books and dividing the size of the empty space by the average size of a book.

Now, we think that example patches the basic intuitive distinction between sampling and imputation.

Now, we support that intuitive conclusion by elaborating the nature of the differences by referring to technical literature, by looking at statutory history and reduce fill a certain amount of ink along the way, but we end up where we started.

When the statute says statistical sampling, it does not mean to include imputation.

So, the statute does not forbid its use.

Third, we ask whether the constitution forbids imputation.

The relevant part of the constitution says “representatives and direct taxes shall be apportioned among the several States according to their respective numbers counting the whole number of persons in each State.

The actual enumeration shall be made within three years after the first meeting of the Congress in such manner as they shall by law direct.

Now, you may have noticed there were two words in there, actual enumeration.

But we think those two words being there do not require say, an individual contact between a counter and each and every person counted.

Nor do those words otherwise restrict Census taking method so stringently that they would forbid the imputation used here.

In reaching this conclusion, we explore the context, the deligation of authority to Congress, a relevant history, contemporary linguistic usage, and Census practices both in the 18th century and more recent times.