Redistricting Will Occur by the Numbers -- Literally

Article excerpt

In a stunning development, a committee of career-level professionals at the U.S. Bureau of the Census recommended on March 1 that the unadulterated actual count of Census 2000 be used as the basis for congressional redistricting. As a result, an alleged Bill Clinton scheme to tilt the political process for the next 10 years may have met its comeuppance.

Just before Clinton's commerce secretary, "Slow Count" William Daley, left to become the Gore/Lieberman campaign manager, he delegated the decision on census adjustment to Clinton appointee Kenneth Prewitt, then-director of the Census Bureau. It was expected that Prewitt would recommend a scheme to "adjust" or pad the numbers in historically Democratic districts using an estimating method called Accuracy and Coverage Evaluation (ACE).

But two things happened: George W. Bush was elected president and Prewitt was shown the door; and the actual numbers of Census 2000 turned out to be more accurate than ever before, within 1 percent of the error rate instead of the 1.6 percent of the 1990 Census. As a result, the Executive Steering Committee for ACE Policy said it was "unable to conclude that the adjusted Census 2000 data are more accurate for redistricting. Accordingly ... the unadjusted census data [should] be released as the Census Bureau's official redistricting data." Acting Director William G. Barron Jr. passed forward the recommendation to Commerce Secretary Don Evans, who will make the final decision in April as required by law.

Democrats had tried to make the alleged undercount into a civil-rights issue, much like the Florida vote, charging that the poor, the homeless and migrant

workers were least likely to be found or to respond to census-takers knocking on the door. Therefore, they asserted, an actual enumeration of persons as required by the U.S. Constitution would be inaccurate. Instead, the count should be "adjusted" to reflect estimates of the allegedly missing persons based on statistical sampling. …