Lawprof Steven Lubet argues that there are some possible errors in Salon; the test itself is here.
The article overstates matters somewhat, it seems to me; for instance, Prof. Lubet writes:
Everyone has the right to bear arms. (This is basically wrong, and probably ideologically motivated; the Second Amendment makes it clear that the right to bear arms is connected to a "well regulated militia," and the Supreme Court has held that this right does not belong to individuals — and in any event, it is an "alienable" right, as in the case of convicted felons.)
Actually, the Supreme Court has never held that the right to bear arms doesn't belong to individuals (see here for a list of all Supreme Court cases mentioning the Second Amendment [UPDATE: Prof. Lubet tells me that he'll try to correct this in the Salon piece]); but more broadly, the question (#78) just asks people to "Name two rights of everyone living in the U.S.," and lists among the acceptable answers "Freedom of expression," "Freedom of speech," "Freedom of assembly," "Freedom to petition the government," "Freedom of worship," and "The right to bear arms." It would be basically wrong and ideologically motivated to reject the answer "The right to bear arms," given that this is an answer endorsed by a federal court of appeals, several state Supreme Courts, Congress, the Justice Department, and many serious scholars.
Similarly, it seems to me it would be rather too picky to mark a layperson down because he read "everyone" as meaning nonfelon adults. After all, even other rights aren't entirely open to everyone; for instance, prison inmates have vastly reduced free speech rights, and next to no freedom of assembly rights or Fourth Amendment rights.
It's possible to argue that, since the preceding two questions refer expressly to the rights of citizens, and this one speaks of "everyone," people should recognize that this means "everyone, including noncitizens." But the Court hasn't even made it clear that the Second Amendment is limited to citizens. (I think that it, unlike some state constitutional right to bear arms provisions, may well be limited to citizens, but this is far from open and shut.) And an exam-taker won't even see the rights-of-citizens questions next to the rights-of-everyone question, since, as Prof. Lubet says, "The actual test for any applicant includes only 10 questions, selected at random, of which six have to be answered correctly."
Maybe an exam for lawyers or law students might demand this level of care, and knowledge of which constitutional questions are open and which are more settled. But I'd say that the more ideologically neutral thing is to accept a range of acceptable opinion about the Second Amendment, and not to be hyper-picky in this context, given that I doubt that we'd endorse hyper-pickiness in other contexts.
One could argue that the flaw isn't in the government's willingness to accept the right to bear arms as an answer to the question, but rather in its listing the right to bear arms as one of the selected set of accepted answers. After all, the list of answers to that question can't be comprehensive: The right to jury trial, the right to freedom from unreasonable searches and seizures, and several other rights have to be acceptable answers, too. Perhaps the sample answers should have focused on the rights that are most clearly recognized, or had some notation noting that there's controversy about the right to bear arms.
But that's a different and milder criticism than the article makes. Moreover, it's not unreasonable, it seems to me, that the government may want to promote even a controversial view that has been repeatedly (including recently) endorsed by Congress, and that is the view of the current Administration. It may not promote this view by grading down those who disagree with it — since the contrary view is also credible, and people who take it aren't legally wrong — but as I noted, the test doesn't seem to do it.
The same can be said, I think, of some of Prof. Lubet's other criticisms (though some are more sound). For instance, he writes "A member of Congress represents all citizens in that representative's district (wrong; he or she represents all people in the district, including noncitizens)." Again, the question (#26) is "Who does a U.S. Representative represent?," and the sample answer given is "All citizens in that Representative’s district (each state is divided into districts)." I wouldn't mark down someone who said "all citizens"; while a Representative is indeed generally seen as representing all his constituents, it seems to me that a test-taker — especially a layperson — can see representation as referring to those people to whom the popularly elected representative is actually accountable, and not just those whom he should in theory look out for.
Perhaps the sample answer should have said "all persons," but it doesn't follow that the test itself is wrong, unless the administrators mark down those who say "all persons" (which I highly doubt that they do). Again, if there is a quarrel to be had here, it's with the sample answer list, not with the test itself. And unless the sample answers are the only ones permissible — which would be clearly bad, but which I doubt is the case — I don't see evidence that the administration of the new test will "penaliz[e] applicants who actually understand the Constitution."