Cite as: 536 U. S. 452 (2002)
Opinion of Thomas, J.
"actually" counted, its use to adjust the census for apportionment purposes runs afoul of the Constitution.
The Court's further reflection that "the Bureau's only choice is to disregard the information it has, using a figure of zero, or to use imputation in an effort to achieve greater accuracy," ante, at 478, makes no difference as to whether it is constitutionally permissible. Even if hot-deck imputation produces more accurate results (and we do not have the means to answer that question), the Framers well understood that some Americans would go uncounted. Accuracy is not the dispositive factor in the constitutional consideration. Despite their awareness that estimation techniques could be used to supplement data, the Framers chose instead to require an "actual Enumeration" or "counting of whole persons." Disappointment following the first census did not prompt a change in this view or in the text. A zero must remain a zero under the dictates of the Constitution.
The Court takes the position that "enumeration" may be incompatible with gross statistical estimates, but concludes that such gross estimates are not at stake here. See ante, at 476. I derive little comfort from the fact that the Court has drawn a constitutional line at " 'gross statistical estimates.' " Ibid. The Court neglects to explain the boundaries of such gross estimates, begging the question of how "gross" must "gross" be? The Court nonchalantly comments that the Census Bureau used the method "sparingly," see ante, at 477, and that the "inference involves a tiny percent of the population," ante, at 479. But the consequences are far from trivial. One State's representation in Congress is reduced while another's is fortified. If the use of hot-deck imputation in the next Census shifts the balance of power in "only" two or three seats, will the Court continue to defend the method? Today, we deal with hot-deck imputation. But if history is our guide, surely other statistical methods will be employed in future censuses and there will be similar challenges. By accepting one method of estimation as con-
509
Page: Index Previous 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 NextLast modified: October 4, 2007