The Volokh Conspiracy

Mostly law professors | Sometimes contrarian | Often libertarian | Always independent


Rights and Wrongs of the Trump Administration's Revised Citizenship Test

Some of the changes are reasonable. But many of the new questions are badly designed and incorporate serious errors. Moreover, such tests raise the deeper issue of why immigrants are required to pass a test to get the right to vote, but natives are not.


The Trump administration recently introduced a new and somewhat more difficult citizenship test for immigrants. The 128 questions that can be used on the new test are listed here.

Some of the changes are reasonable, and arguably improvements over the old version. On the other hand, many of the new questions are badly drafted, incorporating various serious errors. Moreover, the very existence of this test (both the new and old versions) raises the question of why immigrants should be required to pass a test to get voting rights, but not natives.

The changes to the immigration test are far less objectionable than most of the Trump administration's other changes to immigration policy, most of which are calculated to keep immigrants out of the United States altogether, and have collectively made the US more closed to migration than ever before. By contrast, making the citizenship test harder doesn't deny people the right to live and work in the US, because all or nearly all of those eligible to take it in the first place already have the right to live and work in the US as permanent residents, and will not be deported if they fail the test. Indeed, they can even take the test again the next time it is administered. The main rights that turn on passing the test are the right to vote, and the right to certain types of welfare benefits.

Unlike the right to live and work in a given location, the right to vote is not just a personal liberty, but also, as John Stuart Mill put it, the right to exercise "power over others." Thus, there is some potential justification for restricting the franchise to those with at least a minimal level of political knowledge. That is, at least in theory, what the citizenship test is supposed to do. And it is also the reason why we deny the suffrage to children, among others.

Some of the changes from the old test to the new one are reasonable. For example, it makes sense to get rid of questions about geography, while adding more questions about history, law, and political institutions. The latter types of knowledge are far more likely to be useful for voting purposes.

It also makes sense to expand the number of questions on the test from ten to twenty. More questions make it less likely that someone will fail or pass simply because of a fluke related to the questions that just happen to be chosen for that administration of the test. The lower the number of questions, the more likely it is that a knowledgeable applicant will fail because she was unlucky, in that the questions selected for that particular administration just happened to be chosen from among the small number she didn't know the answer to. Conversely, a badly prepared applicant might get "lucky" if the ten questions chosen just happened to be among the few she is familiar with.

It is also worth noting that, while the new test is more difficult than the old, it still isn't that hard. The vast majority of the questions are fairly basic in nature, and applicants need only get 12 of 20 right (60%) in order to pass. That is the same percentage as on the old test (6 of 10). Thus, I'm skeptical of predictions by some immigrant rights advocates that the new test will lead to a great increase in the failure rate, though they may well be right to suspect that this is what the Trump administration is hoping for.

While some of the changes made on the new test are defensible, many of the questions are badly designed or based on factual errors. It is particularly  ironic that a test designed to weed out bad potential voters includes some egregious errors about the history of voting rights. Two of the questions ask "When did all men get the right to vote?" and "When did all women get the right to vote?" In reality it is simply not true that either"all men" or "all women" have the right to vote—even to this day.

Even if the relevant men and women are only adult citizens, current law continues to deny the franchise to large numbers of both men and women. Most states deny it to substantial categories of mentally ill people, and to many convicted felons. These are not marginal exceptions. They affect millions of people. And, of course, the franchise is also denied to millions of non-citizen residents of the US. That's why the citizenship test exists in the first place!

Another question asks respondents to name a power that is "only for the states." Among the possible correct answers are "Provide schooling and education," "Provide protection (police)," and "Approve zoning and land use." In reality, the federal government has a major role in each of these policy areas. The federal Department of Education funds and controls a wide range of education programs, there are numerous federal law enforcement agencies, and the federal government imposes a variety of restrictions on land use—especially in the many places where it actually owns large amounts of land!

I am sympathetic to the argument that the federal government has intruded into these areas far more than the original meaning of the Constitution allows. But the citizenship test should not conflate the normative debate over this issue with a factual question about the current distribution of power. And even under relatively narrow interpretations of federal power under the Constitution, the federal government would still have a role in each of these areas when it comes to military education, the federal territories, the District of Columbia and federally owned land.

A number of questions on the test dress up normative preferences as factual issues. For example, one question asks who members of the House of Representatives represent, and there is a similar question about who senators represent. The "correct" answers are "citizens" in their respective states (for senators) or districts (for members of the House). Whether members of Congress are supposed to represent all residents of their constituency or only those who are citizens is a actually a contested normative issue. As a practical matter, members who represent areas with large immigrant populations often do take non-citizen interests into account in various ways. Presenting this as a factual question with an obvious right answer is wrong.

The same can be said of the question which asks why "[i]t is important for all men age 18 through 25 to register for the Selective Service." Among the "correct" answers are that it is a "[c]ivic duty," and that  it "[m]akes the draft fair, if needed." The framing of both the question and the list of correct answers overlooks the long history of opposition to the draft, which dates all the way back to the 19th century. There have long been those who oppose it on the grounds that the draft is a form of unjust forced labor, that it is unconstitutional, that many of the wars fought by draftees are immoral and unjust, or some combination of all three reasons. Opposition to the draft (and draft registration) is as much an American tradition as support for it—perhaps even more. The answer that "it makes the draft fair, if needed" overlooks the longstanding argument that the male-only draft is unjust sex discrimination. A recent court decision has even ruled that it is unconstitutional for that very reason, though the ruling was later reversed on appeal.

These are far from the only questions on the test that are either based on factual errors, smuggle in contestable normative preferences, or some combination of both. If the federal government is going to continue to have a civics test for new citizens, they should hire some people to draft it who actually know what they are doing—at least enough to avoid very basic errors.

Some of these flawed questions are carried over from the old test (though not the one about representation). But it is still a mistake to include them in the new one.

Despite the flaws in the current citizenship test, I am not on principle opposed to requiring would-be voters to pass a test of basic political knowledge. Voter ignorance is a serious problem, and such a requirement might potentially curtail it, at least at the margin. Even the current flawed test might still be better (or less bad) than no test at all.

But that, in turn raises the question of why it should be imposed on immigrants, but not native-born citizens. One possible answer is that we can reasonably assume that natives already know these things. But, sadly, that isn't true. Studies show that almost two-thirds of current American citizens would fail even the old citizenship test if they had to take it without studying. The percentage who would fail the new one may well be even higher.

Perhaps the answer is that voting is an inherent right of citizenship, so all citizens should be given the franchise, regardless of how ignorant they might be about politics and government. But this runs into the fact that we already deny the franchise to large numbers of citizens based on their real or imagined lack of competence to be good voters: children, many of the mentally ill, and numerous convicted felons. In combination, these groups add up to a third or more of the population. If it is morally permissible to deny the franchise to incompetent (or supposedly incompetent) children, mentally ill people, immigrants, and felons, why not to ignorant native-born adults? Indeed, some of these groups (most notably children) are denied the franchise based on crude generalizations about their competence that don't actually apply to many of them.

Discrimination between politically ignorant immigrants and similarly ignorant natives is another example of morally arbitrary discrimination based on parentage and place of birth, on which most immigration restrictions are based. Why not just impose a knowledge test that applies to all would-be voters regardless of whether they are immigrants or natives, children or adults, mentally ill or not? Any who pass are entitled to vote, and those who fail can try again later.

One possible answer to this question is that we cannot trust the government to come up with an objective test, as opposed to one calculated to weed out opponents of the party in power. This problem is already evident in some of the Trump administration test questions discussed above, which seem designed to privilege more conservative answers to questions (such as the one about who members of Congress represent). The incentive for the government to "rig" the test would be even greater if it applied to all potential voters, not just immigrants.  The history of voting tests is not a happy one; it includes, for example, "literacy tests" used to weed out black voters in the Jim Crow-era South. This is the main reason why I am skeptical of adopting knowledge tests as a general solution to the problem of political ignorance. The idea is likely to be  pernicious in practice, even if defensible in theory.

But even if they cannot be easily changed, we should at least acknowledge the morally questionable aspects of the citizenship test for immigrants. And if we are going to continue to have one, it should be more competently designed. Perhaps the people hired to design the citizenship test should be required to meet certain standards of civic knowledge themselves! To the classic problem of "who guards the guardians," we might add the issue of "who tests the testers."