Community Groups Push to Adjust U.S. Census for Minority Undercount (in National News) Help Rico: Eight-year-old Leukemia Patient Needs Bone Marrow Donor (in Bay Area News) Forecasting Asia's Economy in 2001 (in Business) The Wonderful World of Jason Shiga (in A&E) Emil Amok: Bush's First Days (in Opinion)

Making Minorities Count

By Neela Banerjee

Using unadjusted data from the 1990 U.S. Census, minorities comprise 43% of California's total population.

However, with adjustments to the Census using sampling data, the minority population jumps to 76% of California. Source: Census Monitoring Board.

Whether or not to adjust for the damaging undercount of racial and ethnic minorities in the 2000 census is quickly shaping up to be one of the major civil rights issues of the new millennium.

Statistical agencies, civil rights organizations and community groups are all pushing for the use of sampling, along with the raw data from the head count, to give a more accurate picture of the American population.

President George W. Bush must decide whether or not his will be the first administration ever that would allow for this data to be used to redraw Congressional district boundaries and distribute $185 billion in federal funds.

Historically, the census undercount has adversely affected racial and ethnic minorities, as well as poor people and immigrants. The 1990 raw census tally excluded some 63,000 Asian Americans in California and 22,000 in New York.

“Sampling techniques were integrated into the 1990 census, but they did not use the numbers to adjust for the undercount of racial and ethnic minorities or the double count of traditionally white, affluent communities,” New York-based Asian American Legal Defense and Education Fund lawyer Glen Magpanty said.

In California, the unadjusted data listed the minority population at 43 percent, while the adjusted data reported that the actual minority population was closer to 76 percent. The state stands to lose some $5 billion over the next decade if the unadjusted numbers are released.

“Conservatives argue that the Constitution only allows a strict enumeration, or a strict head count,” Magpanty said. “So if you are going to get counted, you have to fill out a form or answer the questions of a census taker.”

But AALDEF and other civil rights organizations argue that the Constitution says that everyone has to be counted, and therefore the numbers need to be adjusted.

The undercount happens for a variety of reasons, but especially because people are fearful of repercussions due to their immigration status, trouble with the law or language barriers.

“If census takers go to the door and they knock, and then they hear ‘shhhh’ and the lights get turned off, they know there is someone in there. But regulations say they must put down zero,” Magpanty said.

Ted Wang, of Chinese for Affirmative Action in San Francisco, said that places such as Chinatown in San Francisco are full of hidden housing that the census board almost always misses.

“There are a lot of dormitories in Chinatown that do not have a real address, but many people live in them,” Wang said. “Also, there are many multiple family homes, which only get one form, and then only one of the families there will fill it out.”

Politically, most Republicans are against the use of statistical sampling to adjust the census because they say the method would create Democrat-friendly districts, and could cost the GOP its slim majority in the House.

“The court very clearly directed the federal government to rely on the actual numerical count,” Kevin Nguyen, executive director of the Sacramento-based American Civil Rights Institute, said. “Until there is an amendment, nothing can be done.”

Nguyen said that Republicans and Democrats have to move beyond the fight over adjusted figures and work together to make sure there is a way to have a more accurate count.

A recent report by American University history professor Allan Lichtman looks at the 10 most undercounted states in the 1990 census. The report states that “the use of corrected rather than uncorrected data had the potential to significantly affect opportunities for minority voters to participate fully in the political process and elect candidates of their choice.”

The 10 states are California, Texas, Florida, Georgia, Virginia, North Carolina, Illinois, Maryland and Louisiana.

Lichtman’s report shows that the use of corrected data would have increased the numbers of majority-minority districts. For all 10 states, the use of the corrected data would have had the potential to affect minority voter opportunities in more than 40 state legislative seats.

“New York has the second largest Asian American population in the country, and we have never seen an Asian American elected to city council, state legislature or Congress,” Magpanty said. “Part of that is that our community has been misrepresented because we haven’t been counted in the census.”

Magpanty explained that when people know they are the majority in a certain area, they are more likely to get involved in the political process and vote.

“When you know your numbers are strong enough to really affect things – then you would certainly get involved,” Magpanty said.

Lichtman’s report states that even though the adjustment would have a significant effect on minority voter opportunities, it does not appear to have serious partisan effects as the minorities would almost always be added to districts that were already predominately Democratic.

Magpanty pointed out that the overcount is also a very significant problem.

“It might be in the conservatives’ interest to double count the suburbs, the people who have two homes and who have kids in boarding schools,” Magpanty said. “The fewer people of color there are, poor people, people in cities, the better it might be for distributing resources to those upper classes. That is unconstitutional and wrong.”

The Census Bureau would use the Accuracy and Coverage Evaluation (ACE) sampling method. Over the summer, the Bureau sent out enumerators to interview 314,000 households throughout the country. The results of the ACE survey will provide the base for the adjustment.

“The 11 top professionals at the Census will look at all the data, including the ACE and decide whether or not to release the data,” explained Census Monitoring Board spokesperson John Chambers.

The Census Monitoring Board is a bi-partisan board established in 1997 to monitor all preparations and conduct of the Census Bureau. Four Democratic members were appointed by the President, and four Republican leaders were appointed by House and Congressional leadership.

“Now that the counting is done, our role has shifted to make sure the census data is the most accurate data available,” Chambers said.

Former Census Bureau director Ken Prewit resigned earlier this month, when he was not asked to stay on by President Bush. The decision may fall into the hands of new Commerce Secretary Don Evans, who has not yet come out with his opinion on the issue.

The undercount in San Francisco was over 20,000 in the 1990 census count. The Department of Human Services estimated that approximately $53 million was lost in federal and state funding because of this undercount.

But with huge education and outreach efforts by a coalition of community organizations, return rates were up this year, according to Wang.

San Francisco is also a site for a census bureau pilot program called the American Community Survey. This survey is slated to replace the census long form by the year 2010.

“The long form is the one that asks all the detailed questions,” Wang explained. “But by the time the data is released, it is already outdated. We are still using 1990 data results to make policy decisions.”

Wang said that this makes no sense, especially in a place such as San Francisco where some parts of the city have a completely different population than 10 years ago.

“The idea behind this survey is the Bureau will sample three to five percent of the population, so the census data is constantly updated,” Wang said. “So even if the actual adjusted results are not released, there will be a constant check to show what is really going on.”