In 2015, activists hold signs during a news conference in front of the U.S. Supreme Court during a case about whether voting districts should continue to be drawn by using census population data. (Photo by Alex Wong/Getty Images)

Should the 2020 Census revive an old question about citizenship status — or is doing so too glaringly political? That’s been debated hotly this week by pundits, state attorneys general, and more. Many worry that such a question will suppress minority responses — leading to an undercount of the population in “blue” states, with serious consequences for apportioning members of Congress and allocating federal funding for social programs.

But for better and for worse, the census has been political for a long time.

The U.S. Constitution mandates that the federal government run a census every decade to enumerate the country’s residents. Though the census is primarily designed to count the population, the Supreme Court has recognized the government’s ability to ask questions far beyond this scope.

That’s where politics comes in.

Census politics about race could be overt or subtle — with serious consequences

Let’s start with an overtly political question asked after the Civil War. To check on the progress of Reconstruction in the former Confederate South, the U.S. government included a question in 1870 asking whether a male citizen age 21 or above had had his “right to vote … denied or abridged on other grounds than rebellion or other crime.” The goal was to ensure that former slaves were being accepted as free citizens.

As the political will for Reconstruction quickly subsided, so too did the will to use this information for monitoring the progress of the 14th and 15th Amendments. The 1870 Census was the first and only one to ask about voting restrictions.

But that wasn’t the end of the census’s politicization of racial and ethnic identity and citizenship, which simply became more subtle. Jennifer Hochschild and Brenna Powell’s research has shown how the census shapes and is shaped by our understanding of race, ethnicity and social identity.

Starting in 1850, the census began asking individuals about their country of birth — a move toward using the census to create racial and ethnic categories. Before 1930, the census experimented with a variety of options in which individuals could characterize themselves racially and ethnically, including identifying as “Hindoos” or as a member of the “Mexican race.” All that shifted over time.

Take the case of Mexican Americans. For much of history, the U.S. census treated Mexicans as white. But in 1930, the Census Bureau began including a new option under race that was explicitly labeled “Mexican” — thereby separating Mexicans from whiteness. This was loudly protested by Mexican Americans and the Mexican diplomatic mission.

As the census director in 1936, William Lane Austin, commented, “The classification by race or color of individuals, or even entire populations, is not only very difficult, but is a very delicate matter to the United States Government.” Recognizing the fraught intricacies of racial classification and its relationship to racial hierarchy, future censuses took the separate classification of Mexican off the table for several decades.

Such categories could have extremely serious consequences. For instance, during World War II, the Census Bureau gave the Secret Service the names and residences of people classified as Japanese or of Japanese descent, so they could be rounded up into camps.

The politics of asking about citizenship

Asking about citizenship in tandem with national origin has been debated before. During the major immigration wave between 1850 and 1930, the United States had another heated debate about the pluses and minuses of the influx of “foreign” ethnicities, like the Irish, Italians and Slavs. Beginning in 1850 and lasting until 1950, in response to congressional and public demand for tracking the entrance of “desirable” races, the census asked where each individual was born, as well as his or her naturalization status.

The government used these censuses to track how well immigrants had assimilated, largely by comparing English literacy rates, employment status, and whether immigrants continued to live in enclaves. The Dillingham Commission — a House commission specifically appointed to investigate U.S. immigration in 1911 — specifically drew on this data in concluding that Congress should restrict immigration. From this recommendation, Congress wrote and passed the 1924 Emergency Quota Acts, which severely restricted “undesirable” immigration from Southern and Eastern European countries, based on statistics computed from the results of the census.

Asking about race and ethnicity always has political consequences

Beginning in 2000, the census began allowing individuals to mark more than one race, in response to increasing rates of immigration and intermarriage. Civil rights groups such as the NAACP opposed the change, fearing that this would weaken enforcement of voting rights and segregation policies, which were based on single-race population counts.

From these debates we get the prediction that the United States will be “majority minority” as of 2044 — but based, of course, on previous generations’ definitions, which will continue to change.

Not everyone agrees that measuring multiculturalism by enabling individuals to check as many boxes as they feel applies is sufficient. Some, including the American Anthropological Association (AAA), have repeatedly and unsuccessfully lobbied to eliminate the term “race” entirely. The idea is that not enumerating racial differences would help push American society into a post-racial world — that the definition is itself the division.

This is part of a broader argument advanced by the AAA that race itself is a social construct — a categorization without inherent meaning.

So what’s at stake now?

There may be something to the AAA’s argument. Work by political scientists Evan Lieberman and Prerna Singh shows that simply enumerating new groups in a census can create conflict. Looking at data across countries, and particularly examining India, they find that countries that enumerate new groups are more likely to have ethnic violence.

What’s more, the questions asked on surveys can influence who responds and how they respond. For instance, an internal research report by the Census Bureau noted that, when asked about citizenship and national origin, respondents in a pilot survey “provided incomplete or incorrect information about household members … tried to break off the interview … and seemed visibly nervous.”

Of course, that’s the issue currently up for debate: Would merely asking about citizenship status reduce the number of people who fill out their census forms, thus artificially lowering the count of ethnic groups with large numbers of undocumented immigrants, and reducing the power of blue states? We shall see.

Shom Mazumder (@shom_mazumder) is a PhD candidate in government at Harvard University studying the politics of social identity and violence.

CORRECTION: The piece has been changed to clarify the American Anthropological Association’s position in this statement. First, the piece no longer suggests that the AAA was opposed to allowing individuals to mark more than one racial category on the Census. Second, the piece has been changed to note that the AAA’s statement called for the elimination of the term “race” but did not call to “exclude race and ethnicity questions entirely.”