Skip to Navigation
University of Pittsburgh
Print This Page Print this pages

February 5, 2009

Census Bureau tracks changing face of U.S.

One of the things that President Barack Obama referenced in his inauguration remarks was this country’s “patchwork heritage.”

In his Jan. 22 Pitt talk, “Measuring Race and Ethnicity in a Changing, Multicultural America,” part of the Center for Race and Social Problems lecture series, a U.S. Census Bureau official discussed that patchwork heritage in relation to census questions over the decades.

Howard Hogan, who is associate director for demographic programs at the U.S. Census Bureau, was responsible for the statistical design of the 2000 census. He provided an overview of how questions on race and ethnicity have changed on the U.S. census, which has been completed every 10 years since 1790.

Hogan cautioned that his views did not necessarily reflect official Census Bureau positions. He also warned that certain terms appearing in past census questions — for example, colored, mulatto, orgoon — once were acceptable in society, but now are outmoded and even potentially offensive.

“Like all societies, America divides people into groups,” said Hogan, who also is an adjunct professor in the Department of Statistics at George Washington University, a fellow of the American Statistical Association and chair-elect of the association’s survey research methods section. “Unlike many other societies, in America these groups and their meaning have evolved. Official statistics, including the census, both reflect society’s constructs and helped shape those constructs. The election of Obama as president show that these constructs likely are going to be changing again. But I don’t think there are many serious people out there who believe we’re in a post-racial America.”

Measurements of race and ethnicity, Hogan said, should have three properties: The categories should be recognized by society and by individuals; they should be consistent, so that a person would tend to fall into the same group over a long period of time, and the measures should be predictive of social and economic opportunity.

“I use the concept of race in a very flexible context that over time has shifted just about every 20 or so years in response to social, political and demographic trends,” he said.

Hogan broke down the history of the U.S. census into five broad periods: 1790-1840, when the chief goal was to count the number of people of European descent and the number of slaves; 1850-1920, the age of immigration; 1930-1960, the age of consolidation of racial and ethnic categories; 1970-1990, the age of civil rights, and 2000-2010, the age of multi-racial and multi-cultural America.

“When the census started out in 1790, we thought of ourselves as being composed of English, Dutch, German, slaves and Indians,” Hogan said. “There was very little immigration in these earlier decades, and society was dominated by the slave-state, free-state struggle. This is where we began the practice of classifying non-whites in relation to whites. The race or colored category on the census is rooted in the U.S. Constitution, as was the practice of counting slaves as three-fifths of a person, which was a political compromise between the North and the South, and the exclusion in the census count of Indians who were not taxed, that is, tribal Indians.”

Article 1, Section 2 of the Constitution mandated the 10-year census primarily in order to apportion representatives to the U.S. House of Representatives, but also to apportion taxes, Hogan noted.

The Constitution states: “Representatives and direct taxes shall be apportioned among states according to their respective numbers, which shall be determined by adding to the whole number of free persons … excluding Indians not taxed, three-fifths of all other persons.” (The definition later was amended by the 14th Amendment.)

“First, the emphasis is on taxes. They were thinking about economic, military and political power, not one person-one vote,” Hogan maintained. “Secondly, although they’re talking about race, it’s white and everyone else.”

By 1820, the census was separating blacks — categorized as colored — from other non-whites.

“A low point was the 1840 census, where there were a couple questions geared toward social statistics in those days, including: ‘Are you insane?’ In tabulating the free blacks in New York and insanity, social scientists, pseudo-scientists, wanted to prove that freedom caused insanity. Even statisticians said the numbers were so small as to be meaningless, but it was still argued,” he said.

The 1850-1920 period, when the United States population grew from 23 million to 106 million, featured massive immigration.

Immigration created a challenge for the Census Bureau because up until 1850, white essentially meant English, Dutch or German, not, for example, Irish or Italian. The 1850 census added a differentiation among whites in the place-of-birth category to measure that other dimension, Hogan said.

“The 1860 census is a very important one. The race categories are white, black, mulatto, non-taxed Indian and Chinese (in California only),” he said. “It might be argued, I think with some persuasion, that the 1860 census contributed to the Civil War because a number of Southern states were losing population, losing ground in the number of representatives and that motivated some states to seek secession.”

In the 1870 census, the racial categories were unchanged, except Japanese was added to the California census questionnaire.

“But by 1870, for the first time we had second-generation immigrants: Those Irishmen who came in the 1840s and 1850s, their children were now native-born Americans, so by asking place of birth you could not separate them from anybody else,” Hogan said. “What’s the solution? The Census Bureau added one question: Was your father born somewhere else?”

That concept was expanded on the 1880 census to include the birthplace of the person, meaning the U.S. state or territory, as well as the place of birth of the father and of the mother.
The late-19th century also saw the rise of “race science theories” with the categorization of blacks based on “blood quantum.”

“The 1890 census is my favorite census because this is where the pseudo-scientists of the 19th century got their way,” Hogan said. The race categories were white, black, mulatto, quadroon, octoroon, Indian, Chinese and Japanese.”

Black meant more than three-quarters black; mulatto meant one half to three-quarters black; quadroon, one quarter to one half black, and octoroon, less than one quarter black, he said.

“There was a separate question for American Indians, all of whom were counted for the first time, and how much white they were,” Hogan said. “These questions are statistically meaningless. No one knew how to classify people into these categories, and they weren’t asked that again.”

In 1900, the term Negro first was used on the census as a subset of colored. Colored, which used to mean black, now meant all non-whites.

“The 1920 census was the only failed census. The main purpose of the census is to divide up representatives of Congress, based on population, but after the 1920 census, Congress did not re-apportion; they just said ‘We’re going to use the 1910 census,’” Hogan said. “Why? There were some technical errors, was their excuse.”

He dubbed the period 1930-1960, when the U.S. population grew from 123 million to 179 million, as one of consolidation and assimilation from the Census Bureau’s perspective.

“There was also a lessening of white ethnicity: What it meant to be Irish, for example, in 1960 is pretty different from what it meant in 1920,” he said. “The 1930 census dropped mulatto, and we were categorizing people by the ‘one drop rule’ to enforce segregation. Both black and mulatto persons were determined to be black without distinction.”

The one question on color/race on the 1930 census included the categories white, Negro, Indian, Japanese, Chinese, Filipino and Hindu, a lame attempt to differentiate between natives of India and Native Americans, Hogan said. “By 1960, racial categories added Hawaiian, Alaskan, Eskimo, Aleut and my favorite, ‘et cetera.’”

The 1960 census also was the first to measure interracial unions, Hogan noted. Only .04 percent of the population said they were in an interracial union in 1960. That percentage had grown to 5.7 percent by 2000. That also was the year the Census Bureau added the option for individuals to declare themselves to be “more than one race.”

By 1970, political pressure to better measure the Hispanic population was growing. The Census Bureau sent out a sample form, which went to about 5 percent of the population, with questions on Spanish origin. “But one of the questions was: Are you Central American or South American? The problem was that folks in Georgia were saying, ‘We’re Americans and we live in the South — Isn’t that was a South American is?’ So there was a tremendous response in the sample to South American population. We needed to work on that.”

In 1976, Congress mandated that Hispanic data be included in all federal surveys, including the census.

In 1997, the Office of Management and Budget issued Statistical Directive 15, which established minimum standards for survey questions on race and ethnicity: “The standards have five categories for data on race: American Indian or Alaska native; Asian; black or African American; native Hawaiian or other Pacific Islander, and white. There are two categories for data on ethnicity: ‘Hispanic or Latino,’ and ‘Not Hispanic or Latino.’”

The 2000 and 2010 censuses are built around that directive, “although we obviously are still learning. It makes it difficult when, for example, Latino means one thing in Los Angeles, [typically] Mexican American or Chicano, but a different thing in mid-town Manhattan, [typically] Puerto Rican,” Hogan said.

He said that of the 10 questions on the forthcoming 2010 census, only one is on ethnic origin and one is on race, although there are many options for answers.

The first of the two questions is: “Is this person of Hispanic, Latino or Spanish origin?” The census specifies that Hispanic origins are not races, Hogan said.

The choices offered are: No, not of Hispanic, Latino or Spanish origin; Yes, Mexican, Mexican American, Chicano; Yes, Puerto Rican; Yes, Cuban, or Yes, another Hispanic, Latino or Spanish origin.

The second question is: “What is this person’s race? Mark one or more boxes.”

Possible answers are: white; black, African American or Negro; American Indian or Alaskan Native; Asian Indian; Chinese; Filipino; Japanese; Korean; Vietnamese; other Asian; native Hawaiian; Guamanian or Chamorro; Samoan, other Pacific Islander.

The data are used for civil rights monitoring issues. If a person checks white plus one minority, he or she is counted as the minority for the purposes of civil rights issues only. If a person checks two minority races, and if civil rights complaints are in either one of these, the person is counted in the complaint category.

“So we now have a tremendous concentration on measuring race and ethnicity in the census. But there are challenges to measuring the multi-cultural society that we’ve become. When we say white, most people think of Europe, not Turkey or Afghanistan or the Middle East,” Hogan said. “And what of the person who says, ‘I’m not Hispanic, I’m Mexican American?’ What does that tell you? That they don’t think of themselves as part of a global Hispanic population. The point is that race and ethnicity are fluid concepts that reflect social, political and demographic environments.”

—Peter Hart


Leave a Reply