Census Confusion

By Michael Scherer, Columbia Journalism Review. Posted February 6, 2003.

Imagine this headline: "Baptists Now Outnumber Blacks in Louisiana, Says New Study."

Doesn’t work right? The reason: Any such study would have to count black Baptists against themselves to compare overlapping categories of race and religious belief. It’s like comparing organic apples with red apples.

Now consider this headline, which The Associated Press ran Tuesday: "Hispanics Now Outnumber Blacks As Largest U.S. Minority Group." Similar versions ran in papers and on web sites all over the country. "Hispanics Have Edged Past Blacks As The Nation’s Largest Minority Group," said The New York Times.

Can that be true? Unlike blacks, Hispanics do not make up a racial group. They are a self-identified ethnicity, a group of people who generally trace their roots back to Latin America or Spain. Hispanics can be white, black, Native American, Asian, or a blend of all these racial categories. So when does an ethnicity really outnumber a racial group, particularly if both categories share some of the same members? The answer is not so clear.

While the importance of race and ethnicity remains a topic of vigorous debate, journalists too often overlook the nuanced differences between them and the boundaries that define them. Newspapers and local television stations still run stories that describe police suspects as "Hispanic males," a description not much more telling than "Anglican males" or "Atheist males." On Wednesday, the Chicago Tribune described a jury of twelve as "six blacks, four whites and two Hispanics." One can only guess at the skin color of the latter two, or the ethnicity of the first ten.

"The rules that people are making to stabilize these categories are coming unglued," explains Margo Anderson, who studies the Census at the University of Wisconsin. In fact, government approved racial and ethnic categories have never been much more than gross generalizations. Over the last two centuries, racial categories like "Mulatto," "Hindu," and "Octoroon," meaning one-eighth black, appeared on Census forms. The current categories are still just rough approximations, especially when it comes to the self-defined category of Hispanic. "It has now become so commonplace to think of Hispanic as a race," explains William H. Frey, a demographer at the University of Michigan, pointing to a common fallacy.

Such distinctions become crucial when it comes to reporting statistics like Census data, especially since the government began allowing people to describe themselves by marking multiple races. On Tuesday, the Census Bureau released voluminous new information on the racial and ethnic makeup of the U.S. The data came without any comment, leaving reporters with the task of interpretation.

At the AP, policy prescribed the coverage. Since April 2001, the AP has used "black" to mean non-Hispanic blacks as well as non-Hispanic blacks who also describe themselves as belonging to a second or third racial group, like Asian, white, or Native American. The AP uses "Hispanic" or "Latino" to mean Hispanics of all races. "The decision was made to ensure that our reporting minimized confusion," explains Jack Stokes, a spokesman for the AP. As a result, the AP read the new Census data in a very specific way: It counted 36.1 million blacks and 37 million Hispanics, and thus concluded that Hispanics had outnumbered blacks.

The problem is that this formulation counts black Hispanics as Hispanic, but does not count black Hispanics as black, effectively erasing 1.7 million people from the total black population. Considering that 1.7 million uncounted blacks are more than the 900,000 difference the AP describes, the story’s headline comes into question.

Other reporters, who did not share the AP’s view of the data, found themselves in a bind when the story moved across the wires. Paul Overberg, Census reporter for USA Today, had to explain to his editors why he could not follow the lead of the AP. "Who wanted to believe me when I said, ‘No, that’s not the way we count these things’?" says Overberg.

He chose to compare two other numbers, the total Hispanic population (37 million) and the total population of blacks, including those who also claimed another racial identity, regardless of Hispanic ethnicity (37.7 million). This effectively double counts the 1.7 million Hispanics who consider themselves at least partly black. In this interpretation, which assumes that race and ethnicity are comparable minority categories, blacks are still the larger minority group. In the end, Overberg’s story carried a markedly different headline: "Hispanics Inch Towards Outnumbering Blacks." The Miami Herald and The Washington Post were also cautious, emphasizing the rate of increase for Hispanics. "Hispanics Close to Becoming Largest Minority," read the Herald’s banner. "Hispanic Population Booming In U.S.; Census Finds Growth Outpacing Blacks'," said the Post. The San Francisco Chronicle chose to be explicit: "America’s Ethnic Shift; Latinos Pass Blacks Unless You Count Black Latinos."

But several other papers opted for a third interpretation that supported the most news-making interpretation of the data. Both The New York Times and the Atlanta Journal Constitution decided to double count black Hispanics, but unlike Overberg, they did not count those who considered themselves black in combination with some other race. Consequently, they found that the total number of Hispanics (37 million) surpassed the total number of people whose only racial identity is black (36.2 million) by roughly 800,000. Both the Times and the Journal-Constitution mentioned multiracial blacks later in their stories. Like the AP, these papers chose a more provocative path through the maze of numbers presented by the Census, leading assertively with a single interpretation. Without qualification, the Journal-Constitution’s headline read, "Latinos Surpass Blacks As Largest Minority."

News organizations that simply trumpeted this new milestone might have served their readers better with a more cautious approach. As one member of the journalism email listserve Census-L noted, "I’m worried that the AP is drawing black-and-white conclusions from gray data." If nothing else, these stories show that American demographics are no longer black or white. In truth, they are a blended, difficult-to-define shade of brown.

Michael Scherer is an assistant editor at CJR.

Liked this story? Get top stories in your inbox each week from AlterNet! Sign up now »