- The Washington Times - Tuesday, March 13, 2001

In a bold move, acting U.S. Census Bureau Director William Barron took out the only leg left standing for proponents of statistical sampling. Rep. Carolyn Maloney, the self-appointed spokesperson for the failed movement, has argued for years that sampled census numbers would be more accurate, based on projections from some career professionals at the bureau.
Those same experts, and the actual count results they achieved, now disagree.
U.S. Commerce Secretary Don Evans recommended recently that the federal government release only the actual head count census numbers for Census 2000, based on Census Bureau Director William Barron's advice that "[we] are unable, based on data and other information currently available, to conclude that the adjusted data are more accurate for use in redistricting."
Clearly, the credit for achieving an improved census head count goes to the U.S. Census Bureau, which executed the largest civilian exercise of government action in our nation's history, and to the Congress, which approved the increased budget necessary to accomplish the task. Census 2000 is the most controversial and, by early measures, the most accurate census in U.S. history. In 1999, the U.S. Supreme Court in Glavin, et al vs. Clinton, et al disposed of the question of whether statistically sampled census numbers could be used for apportioning Congress by holding that only actual numbers would suffice. Now, with the law clear and the accuracy question answered by the Census Bureau's top professionals, where do the sampling proponents go? Probably back to court.
But prior experience with challenges to the 1999 Census shows court challenges to be a feckless endeavor.
Last week, a U.S. District Court in California denied a request by the city of Los Angeles and others for a temporary restraining order to stop a revised federal rule granting the commerce secretary the power to determine whether to release the statistically sampled census numbers to the states. Although currently languishing in the minutia of administrative law, the California federal lawsuit appears headed for familiar ground.
Following the 1990 Census, a series of lawsuits sought to compel the federal government to release so-called "adjusted" census numbers for purposes of apportionment and redistricting. In a key decision, the U.S. Supreme Court closed the door on the possibility that a local jurisdiction might have the power to compel the release of adjusted numbers (Wisconsin vs. New York, 1996). The court ruled that the commerce secretary's decision need only bear a rational relationship to the accomplishment of the actual count, effectively silencing pro-sampling local jurisdictions.
Although designed as the constitutional mechanism for apportioning congressional districts among the states, the census has also become the measure for appropriating hundreds of billions in federal spending. In terms of apportionment and its legal sibling, redistricting, the U.S. Constitution and federal law speak clearly the census shall be conducted by "counting the whole number of persons in each state." But when it comes to the myriad of non-constitutional purposes for the census, including the distribution of federal dollars, it has been the federal government's practice for at least two decades to update and use adjusted census numbers. In fact, federal law encourages it.
Despite the bright line drawn between sampled and actual count census numbers, sampling advocates continue to support the statistical count for both purposes. Following the 1990 Census, then-Commerce Secretary Robert Mosbacher declined to authorize the adjusted numbers. He cited the "slippery slope" of statistics, and the potential for partisan manipulation of the census data, as overriding concerns. At the time, some congressional Republicans urged the use of the adjusted data to compensate for undercounts among rural Americans, those with multiple residences, and others, who were deemed likely GOP voters. The Bush administration resisted the temptation and stood on constitutional principles.
Today, the new Bush administration faces a similar choice, and is apparently headed in the right direction to meet the April 1 deadline. Sampling advocates are pushing for inclusion of statistical numbers to "account for" persons deemed likely Democratic voters. The difference is that the U.S. Census Bureau has done a tremendous job improving the actual head count, to the point where the bureau's own experts concede that actual count is more accurate than sampling.
The decennial census is indeed political. It is a cornerstone of representative democracy, and the constitutionally approved mechanism for determining the apportionment of political power. Its political aspect must come only after the completion of an honest, accurate head count. Then, and only then, should the political games begin.

L. Lynn Hogue is president of the Southeastern Legal Foundation, the public interest law firm which won a U.S. Supreme Court decision on statistical sampling. He is a constitutional law professor at Georgia State University College of Law.

Sign up for Daily Newsletters

Manage Newsletters

Copyright © 2019 The Washington Times, LLC. Click here for reprint permission.

Please read our comment policy before commenting.


Click to Read More and View Comments

Click to Hide