Questions on race and ethnicity should be eliminated from the 2010 census. Those questions produce unscientific and misleading data and can lead to bitter arguments over race.
It has been a year since the 2000 census was taken, and now, as the results tumble out, 281 million Americans have been divided, officially, into 126 racial and ethnic categories. The time has come to ask whether that is 125 too many. Should the 2010 census eliminate the questions on race and ethnicity?
It should. The data are getting worse, sillier, and perhaps dangerous. The proliferation of pigeonholes—which will continue unless it is stopped—is causing arguments that are sharp and may become bitter. Instead of slicing ourselves thin and thinner, we should say what most Americans believe, that we are all Americans, as diverse as we want to be, reveling in our unmeasured and immeasurable diversity, which nowadays should be none of the government’s business.
As shall be seen, this is also the general view of the distinguished political scientist Kenneth Prewitt, now the dean of the graduate faculty at the New School. In 1998, Mr. Prewitt was appointed director of the Census Bureau by President Clinton, a post in which he served with energy and distinction until January 20, 2001.
We don’t know how to measure “race.” A Texas resident lives in Texas. A sixty-eight-year-old is sixty-eight years old. I could continue such a display of demographic erudition. But race is different; you are what you say you are, as in a public opinion poll. Mr. Prewitt says, “The measurement of race is not grounded in science, as anthropologists and biologists have pointed out for years.”
Misreporting the Data
Not knowing how to do it, we do more of it. Under understandable pressure from Americans of mixed race ancestry, we have engaged in category creep. Once there were three basic racial classifications on census forms, then four, then five, now six: “white,” “black or African American,” “American Indian or Alaska Native,” “Asian,” “Native Hawaiian or other Pacific Islander,” plus “Some Other Race.” And for the first time the census invited respondents to “mark one OR MORE races.” This yields sixty-three potential flavors of Americans.
(Beneath these listings are eighteen additional “response categories” such as “Japanese,” “Samoan,” “Asian Indian,” and “Guamanian or Chamoro.” Just try multiplying all that out.)
Double the sixty-three. Every American is further categorized, nonracially, as “Hispanic or Latino” or “Not Hispanic or Latino,” offering a total of 126 categories.
The 2000 census made Hispanics the largest minority group in America, barely surpassing blacks. Hispanics now make up 12.5 percent of the population and nearly a third of the California total.
This got California state librarian Kevin Starr excited. Mr. Starr announced that “The Anglo hegemony was only an intermittent phase in California’s arc of identity. . . . This is a reassertion of the intrinsic demographic DNA of the longer pattern, which is part of a California-Mexico continuum.” Is Mr. Starr getting set to run for office, or for the hills?
The potential mischief of this complexity is apparent in the New York Times headline on the California story: “Census confirms whites are in minority.” Why mischief? Because most Hispanics declare themselves to be whites. Because Mr. Starr’s “intrinsic demographic DNA” neglects growing communities of Koreans, Japanese, Russian Jews, Chinese, Canadians, Thai, Iranians—and whites. Because fertility rates in Mexico are sinking rapidly. Because the 126 groups are intermarrying at high rates. Arc of identity, indeed.
Mr. Prewitt says pressure for more category classification is inevitable: “Proliferation begets proliferation. . . . To resist would be to invite an ever more intense politics of racial measurement.” He notes that while Arab-American leaders reluctantly accepted the government’s decision not to separately count Arab-Americans in 2000, they will surely make a more intense request in 2010. If successful, that would jump the number of combinations and permutations of Americans to 252.
What would be so wrong with dumping the race and ethnicity categories?
Not much. It is not racial data that keeps polls open to blacks in once-segregated states; it is political power and the U.S. Constitution. The courts are already (and properly) unwinding the skein of racial preferences. We have almost a decade to make whatever legal changes are necessary. There would little loss of scholarship, because the data are so weird in any event and because a simpler category of “ancestry” on the “long form” would give plenty of information.
And the census doesn’t ask everything. There is no census religion data, but various religious groups have pretty good estimates. We get estimates on voting behavior by group from commercial polling firms.
Mr. Prewitt says we ought to prepare ourselves to govern without a racial classification system. He’s right.
And we ought to make that change with gusto and grace. People get very nasty about race and ethnicity, all the time, everywhere. Counting that way is so very twentieth century. By stopping, we can serve ourselves well and offer a model to the world of what a twenty-first-century nation looks like—immeasurably diverse.
Ben J. Wattenberg is a senior fellow at AEI.