The 2020 Census is set to take place at a time of political turmoil, when Americans are experiencing a crisis in confidence in federal institutions. Unfortunately, the census is likely to exacerbate that crisis, because the Trump administration has enlisted it in the work of maintaining Republican political control.

Signs of the administration’s strategy emerged in May, when John Thompson, director of the Census Bureau and a 27-year veteran of the agency, resigned over a congressional budget forecast he said was inadequate. The proposed cuts would undermine efforts to expand access — getting the word out to undercounted communities or experimenting with online responses.

Those warning bells rang louder in December when news broke that President Trump would appoint as deputy director Thomas Brunell, a political scientist who has defended Republican gerrymandering tactics in court. Then, two weeks ago, ProPublica reported that administration officials have asked to include a new question about citizenship status — an addition clearly aimed at scaring immigrants away from participating and being counted.

This should concern every American.

An accurate count of the American people is the foundation of much of our politics: Everything from the allocation of congressional seats to billions of dollars in federal funds is determined by how many heads are counted and where. What questions are asked can also have profound consequences. So perhaps it isn’t surprising that Trump, like his predecessors, is trying to game the census for political gain.

But such efforts to manipulate the census have not always worked. A massive bureaucratic endeavor, the census demonstrates the power of the national government — and highlights the limits of its power in the hundreds of minute decisions made by those doing the counting and the individuals to be counted.

The census has always been controversial. Who is counted matters. Mandated by the Constitution, the census resulted from a heated debate between big states and small ones about how to determine each state’s representation. Once the delegates to the Constitutional Convention reached a compromise in which each state would be represented equally in the Senate while seats in the House of Representative would be allocated proportionally based on population, they had to determine how to count populations.

It was this debate over who counted that pitted slave states against those without slaves. Article I, Section 2 of the Constitution, which mandates a count every 10 years to determine apportionment, initially instructed: “Representatives and direct Taxes shall be apportioned” among the states, “according to their respective Numbers, which shall be determined by adding to the whole Number of free Persons . . . three fifths of all other Persons.”

Before the Civil War, slaveholders such as Andrew Jackson, president when the 1830 Census was conducted, were keen to have their slaves counted, for according to the three-fifths compromise, slaves — “all other Persons” — counted toward each state’s share of congressional seats. This math also gave slaveholders more electoral votes. Because of that, they maintained a stranglehold on the office of the presidency: 10 of the first 16 presidents owned slaves and had a vested interest in protecting the three-fifths compromise that helped put them in office.

After the Civil War ended slavery, Republican leaders anticipated that, as the party of Lincoln, they would attract the loyal support of freed men who now counted as full citizens — and as voters. And they were right. Black men voted for the first time in the South in 1868, helping elect Ulysses S. Grant to the presidency and offering their support to other Republican candidates.

Like other Republicans, out of a mixture of benevolence and party politics, Grant embraced the proposed 15th Amendment, which enfranchised blacks. But, they worried, granting black men voting rights wasn’t enough to maintain political power.

So Republicans also tried to harness the power of the census. They began thinking about how they would go about the counting and what questions they would ask when they did. They found the answer they were looking for in the 14th Amendment.

The 14th Amendment affirmed a basic principle laid out in the Constitution: that the number of representatives granted to each state is based on their “respective numbers, counting the whole number of persons in each State.”  But while it eliminated the three-fifths clause, it also gave Congress the power to take representation away from states that denied citizens “the privileges or immunities” of citizenship, including, of course, the franchise.

The second clause of the 14th Amendment gives the federal government the power to reduce a state’s congressional representation according to the number of male citizens over 21 who are unlawfully disfranchised. (In 1870, some men who fought for the Confederacy continued to be legally disfranchised, although in dwindling numbers.)

So in December 1869, months before the 1870 census was to get underway, Republicans in the House of Representatives proposed adding two questions to the census. Under the heading “Constitutional Relations,” Question No. 19 asked whether the subject was a male citizen of voting age, and Question No. 20 asked whether the subject had been unlawfully disfranchised. These last-minute additions reflected Republican desires to use the census to enforce the 14th Amendment’s disfranchisement clause to punish states that denied black men the vote — and ultimately, to guarantee Republican control of Congress.

Congress ended up not using this disfranchisement data to determine apportionment. But even asking such questions mattered. As enumerators took to the streets and roads, knocking on doors and entering the homes of people throughout the nation, these men with pencils in their hands stood as striking reminders of the power of the federal government.

In communities still recovering from the Civil War, their questions likely elicited various responses. For historians, their activities generate important questions: Did posing these new questions among recently freed men in the South communicate to them that the federal government would protect them if they were threatened or attacked when they tried to vote? Did disfranchised former Confederates receive the questions as an insult? Or did the latter steer clear of enumerators, some of them federal soldiers, whom they believed would put the numbers to no good use?

While we don’t have all these answers, we do know that the efforts to manipulate the census for political gain failed. These last-minute additions to the 1870 Census did not fix the problem of black disfranchisement, nor did they allow for a congressional reapportionment that would play easily into Republican hands.

Gaming the census proved tricky, with the outcome decidedly mixed. Rather than providing a political boost, Republican management of the census provoked intense criticism. Amid these howls, President Grant authorized a recount in three cities that complained of gross undercounting and other errors. Two of these cities, Indianapolis and Philadelphia, voted reliably Republican.

The Trump administration’s attempt to game the census by adding a last-minute question about citizenship to the questionnaire may have the intended effect of intimidating immigrants who would otherwise turn out to be counted. This could help the party hold onto political control in communities where Latinos might otherwise tip the balance of power away from the  Republicans. Or it could end up reducing the counted population in reliably red states like Texas, thereby reducing their representation. As in 1870, there is no telling how the census will play out once the forms get into people’s hands and the counting begins.