Census 'can of worms' makes experts squirm
Seattle Times staff reporter
Like trying to determine the motives of Florida voters who left chads hanging on presidential ballots, we may never know what Americans were trying to say when they answered race and Hispanic-origin questions on the 2000 census form.
"It's a can of worms and everyone knows it," says Richard Morrill, a University of Washington professor emeritus in geography. "But we're stuck with what we have because race is such an obsession in our society."
The U.S. Census Bureau seems acutely aware of the murky nature of the results. As officials trumpet the statistics as significant, they toss out enough caveats and equivocations to raise suspicions about the entire exercise.
"We have the fortunate job of just collecting the information, and we don't always have to give interpretation," said Jorge del Pinal, the bureau's chief of special population statistics, yesterday in Washington.
Matt Kelley, president and CEO of the Seattle-based Mavin Foundation, which publishes a magazine for mixed-race young people, said census numbers on race amount to "estimations based on fabricated racial categories."
"What really irritates me is the Census Bureau is basically saying, `We're giving you this information even though we realize it is inherently corrupt,' " Kelley said. "I would bet that behind closed doors, they are throwing their hands up in the air."
Race categories for the 2000 census changed from 1990, making it difficult to directly compare with past censuses. Yet such comparisons are being made.
In 1990, people marked one of five boxes: white; black; American Indian, Eskimo or Aleutian; Asian or Pacific Islander; and "some other race." In 2000, "Native Hawaiian and Other Pacific Islander" were given a separate box.
Under those confines, a Korean American and a Pakistani immigrant are considered the same - Asian. An African American whose great-grandparents were born in America is the same as someone who came to this country last year from Somalia.
But compounding the confusion is the fact that the 2000 census allowed respondents to mark more than one racial category for the first time. In effect, there are 63 possible racial combinations in 2000 compared with five in 1990. When Hispanic origin is added to the mix, there are 126 combinations compared with 10.
About 2.4 percent identified themselves as mixed race.
Mark Ellis, a University of Washington geography professor who researches race population projections, said precise counts of minority groups remain important because they affect voting rights and civil-rights policies.
"Power for a group usually is based on the size of its population," Ellis said. "There will be all kinds of lawsuits filed out of fear that people marked as mixed won't be counted wholly as a minority."
Although few people overall marked more than one race, they make a huge difference in the counts of certain minority populations. For example, the number of those who identified themselves as a Hawaiian or other Pacific Islander more than doubles when the count includes people who identified themselves as Pacific Islander and at least one other race.
Similarly, 2.5 million people identified themselves as American Indian or Alaskan Native, but the number increases to 4.1 million when the count includes those who marked Native American and at least one additional race.
The practice of treating Hispanic origin separate from race, though not new to the 2000 census, also appears to distort results. The census form asks separate questions on race and whether someone is of Hispanic origin, the theory being that Hispanic is an ethnicity and not a race.
The bureau reported yesterday that 97 percent of the 15.4 million Americans who marked the "some other race" box identified themselves as having Hispanic origin, which seems to indicate that many Americans identify their Hispanic heritage as a race.
"If nearly 15.4 million people found the Hispanic question so confusing to them that they rejected the race question outright, that is very telling," Kelley said.
The Washington Post last week wrote a story based on census numbers that said the nation's Hispanic population is roughly equal to that of African Americans - failing to acknowledge that 2 percent of the 35.3 million Hispanics identify their race as black. That means some of the people counted within the black total also are counted within the Hispanic total.
Bureau official Claudette Bennett cautioned those interpreting the census to take into account its most basic tenet: The survey is based on people's self-identification.
"We really don't know what individuals are trying to tell us about their heritage," Bennett said.