New estimates sharpen partisan conflict over use of sampling
By Genaro C. Armas/AP
Republicans now point to estimates, showing the 2000 census accounted for a higher percentage of Americans than 1990, to fuel their arguments that raw population numbers do not need to be adjusted to make up for an undercount.
The figures also disclosed that the percentages of undercounted minorities and children declined this time around. Still, Democrats argued the undercount proves millions of Americans could still be missed if the Bush administration declines to statiscally adjust the actual census count using a method known as “sampling.”
The heightened political tensions over the use of sampling surfaced on Feb. 14 during a hearing of the House Government Reform Committee’s census panel. The Census Bureau must decide by early March whether sampling is more accurate than the raw count, and whether the adjusted figures should be used as the basis for redrawing political district lines and redistributing over $185 billion in federal funds.
The Bush administration, though, could reverse the Census Bureau’s decision. Bush has said an actual “head count” is most accurate, but has not disclosed whether he supports sampling.
White House spokeswoman Clair Bucham said on Feb. 14, the issue was still under consideration by the Commerce Department and did not offer a timetable for any decision. Commerce oversees the Census Bureau.
“I could not be more pleased with the strength of these numbers,” Commerce Secretary Don Evans said in a statement. “While I can’t take credit, let me offer my praise.”
However, Rep. Martin Frost, D-Texas, said the sampling decision posed “an important test for President Bush, and for his purported effort to make the Republican Party more moderate and more inclusive.”
“How many millions of Americans are Republicans willing to disenfranchise, in order to preserve their political power?” asked Frost, third-ranking Democrat in the House.
The net undercount of all Americans in the census ranged from 0.96 percent to 1.4 percent of the population, according to the preliminary national estimates based on a follow-up survey.
The survey, which used sampling methods, is performed regularly after the census to measure the count’s accuracy. But it’s also drawn increased political attention this year as the Census Bureau and the Bush administration determine how to proceed with the Clinton-era sampling plan.
The net undercount in 1990 was 1.6 percent of the population — roughly 4 million people.
Most Democrats and civil rights groups support sampling. Republicans, in general, contend the follow-up survey should be used as an accuracy check only, and not to adjust raw results of the census that could be used for redistricting.
The Census Bureau dramatically increased its outreach efforts in 2000, and reversed a decades-long decline in the percentage of forms mailed back to the bureau. The census was deemed an operational success by all sides.
The follow-up survey found that the undercount of American Indians and Alaska Natives living on reservations may have dropped from 12.22 percent, in 1990, to between 2.77 percent and 6.71 percent in 2000.
Other preliminary figures released:
- The undercount for blacks in 1990 was 4.57 percent. The undercount for black, non-Latinos in 2000 was between 1.6 percent and 2.73 percent.
- The undercount for Latinos declined from 4.99 percent in 1990 to between 2.22 percent and 3.48 percent in 2000.
- For the Asian and Pacific Islander population, the undercount in 1990 was 2.36 percent. Census officials estimate that non-Latino Asians in 2000 may have been slightly over-counted by 0.09 percent, or could have been undercounted by as much as 2.01 percent.
- The undercount for whites changed from 0.68 percent in 1990, to between 0.44 percent and 0.9 percent in 2000.
- For Americans under 18, the undercount declined from 3.18 percent in 1990 to between 1.23 percent and 1.85 percent in 2000.
Though percentages may be lower, the actual number of people missed in some categories could be higher, because the overall population has grown.
At the hearing Rep. Carolyn Maloney, D-N.Y., a subcommitee member, urged the administration to allow the career statisticians at the Census Bureau to make the final decision.
But the subcommittee’s chairman, Rep. Dan Miller, R-Fla., called adjustment a “Pandora’s box, filled with unintended consequences, legal uncertainty and inaccuracy.
Statistics on the state and local levels are scheduled for release next month. |
