How the U.S. census has measured race over 230 years

On the first Monday in August 1790, just over a year after the inauguration of President George Washington, America’s first census marshals began knocking on doors. The new country’s constitution decreed that each state would be represented in Congress “according to their respective numbers.” A national enumeration was in order.

And so, marshals took to the streets with schedules, quill pens and ink in hand. The census intended to enumerate every person in the original 13 states, three districts (Kentucky, Maine and Vermont) and one western territory (Tennessee). The Northwest Territory and American Indian communities were left out. The marshals asked the head of each household to record his or her name and the size of the household, and then placed the residents into one of three racial categories: “white” people, “other free people” and “slaves.”

That first census tallied 3.9 million residents, approximately the number of people living in Los Angeles today.

The U.S. population has expanded and diversified a great deal since that first count. And the categories the census uses to describe people have diversified as well. Still, the underlying purpose of the procedure, which happens every 10 years, remains the same: to determine each area’s proportional representation in national and local politics and to distribute government funding for social services.

In that 1790 census, white people made up about 80 percent of the total population, enslaved black people represented 18 percent and other free people represented the remaining 2 percent. These three categories were the primary markers of racial difference in the population until 1860.

As retired census demographer Campbell Gibson notes in his online compilation, American Demographic History Chartbook: 1790–2010, racial categories used in the census tended to reflect social attitudes and political considerations. One such political consideration was built into the Constitution. To boost their representation in the House of Representatives, delegates from southern states wanted slaves to be counted with the free population. But northern states wanted the free population to be more heavily weighted. Delegates finally agreed on the Three-Fifths Compromise — each slave was counted as three-fifths of a person until the category “slave” was removed in 1860.

As additional race categories were added to the census in the 19th century, the trained enumerators who replaced the marshals determined race through observation. The power for defining the population was in the hands of the data collectors, rather than the people being counted.

The accuracy of historical census data fluctuated for minority groups as social power and political priorities changed. For example, the American Indian classification was first included in the 1860 census when California (1850), Minnesota (1858) and Oregon (1859) were entering the union and fighting for seats in the House of Representatives.

But the American bureaucracy was ill-equipped to count this population. In a local enumeration of New Mexico pueblos conducted from 1850 to 1870, for example, racial indications were often “unreliable,” according to a history on the U.S. Census Bureau’s website. One enumerator might use “non-White” to refer to a pueblo community member, another might refer to the same person as “Indian,” and another might write in “copper.” According to census data, the American Indian population (which, until 1890, did not include those living on reservations) went from 44,000 in 1860 down to 26,000 in 1870 and then up to 66,000 in 1880.

The census began to acknowledge members of other racial communities in the latter half of the 19th century as immigration from non-European countries increased. In 1860, the first year that a category for Chinese Americans was included, in California only, about 35,000 people were counted in that group. Many Chinese immigrants had arrived in California hoping to make their fortunes in the Gold Rush of 1849 or to help build the first transcontinental railroad. When “Chinese” appeared in the national census in 1870, the count was 63,000, which rose to 105,000 in 1880.

In 1882, however, Congress passed the Chinese Exclusion Act, prohibiting immigration from China for the next 10 years; further anti-Chinese policies and social norms led the Chinese-American population to drop below 100,000 until well into the 20th century. Immigration from Japan had fewer constraints in that time: The Japanese-American population grew from 148 people in 1880 to 138,834 in 1930.

Congress established national immigration quotas in 1921, limiting immigration from any given country to 3 percent of the population from that country already living in the United States. The next census after this policy was passed, in 1930, counted the country’s existing immigrant population with more demographic categories than ever before, including categories for “Filipino” and other Asian minorities. This census was the first and last to specify “Mexican” as a race, enumerating about 1.4 million people in that category. However, the authors of the report on the 1940 census revised the population figures from 1930 to classify the Mexican population as white.

Immigration largely stagnated from the 1920s to the 1960s. But a game changer was the Immigration and Nationality Act of 1965. This law removed immigration quotas and created a seven-category preference system that prioritized relatives of U.S. citizens and professionals with specialized skills. Immigration rose sharply in the second half of the 20th century, with major waves coming from Asia, the Pacific Islands, Mexico and the Caribbean. The Filipino population, for example, jumped about tenfold from 180,000 in 1960 to more than 1.8 million in 2000. Census categories expanded with increasing diversity in that same time period.

On the eve of the new millennium in 2000, the number of racial categories the census used to delineate Americans had multiplied from the original three to 16, and respondents could choose from four Hispanic/Latino ethnic origin options. And for the first time, Americans could self-identify with more than one race.

Immigrants have driven America’s population shifts, from Irish and Germans in the early 19th century to an ever-widening ethnic pool today. According to 2018 population estimates from the Census Bureau, there are 2.3 million foreign-born children and 15.9 million native-born children with immigrant parents living in the United States.

Thanks to this generation of new arrivals, first-generation immigrants and their children will represent 36 percent of the U.S. population in 2065, according to Pew Research Center projections reported in a 2015 study on the impacts of immigration on the American population.

The last few years have seen a shift in U.S. attitudes toward immigrants, with the current administration restricting immigration and creating a legal and political environment that could discourage immigrant families from participating in the census.

“If you’re an immigrant living in the U.S. or trying to come to the U.S., the institutions have turned against you,” says Austin Kocher, a faculty fellow at Transactional Records Access Clearinghouse, or TRAC, an organization based at Syracuse University in New York that compiles and disseminates data from the federal government.

According to TRAC’s analysis, in the last three years, there have been spikes in the numbers of asylum cases that are denied, deportation cases that are considered and people detained by Immigration and Customs Enforcement.

Throughout U.S. history, immigration has always been tied to politics. As the 2020 census begins, it remains to be seen how the current political climate will affect who participates in this enumeration and how they will be counted.