The battle continues to rage over whether or not the Census Bureau can restore a question about citizenship status to the 2020 census short form, despite the fact that the same question has been asked on the long form which roughly one in six Americans fill out every ten years since 1970. Several states have already filed lawsuits seeking to block the addition of the question, adding to the endless pile of suits currently clogging the courts which grows every time the President tweets anything.

Such questions don’t seem to be slowing down the Commerce Department (which oversees the Census Bureau). Under a new proposal currently being considered, the government would review the answers to the citizenship question and compare them to other data, looking for cases where someone may be lying about their citizenship status. If you’re thinking that sounds like a data management nightmare you’re probably correct, but they clearly think it’s possible. (Daily Mail)

The Trump administration – which is adding a citizenship question to the census – plans to mine other government records to ensure the 2020 surveys are answered accurately.

The Wall Street Journal reported Tuesday that the government may even change respondents’ answers.

In December, the Department of Justice requested that the Census Bureau, which falls under the Department of Commerce, reinstate the citizenship question in order to help the agency enforce the Voting Rights Act, which is supposed to protect minorities.

Last week, Commerce Secretary Wilbur Ross announced his decision to include the citizenship question on census forms.

I suppose there are other documents in the system which reflect current citizenship status, but how well are they organized? I’m certain we never hit 100% in terms of census responses, but most of the country does it so we’re talking about the records of at least 300 million people. Even after you’ve scanned them all into the system (with a lot of manual entries for forms which are filled out in a sloppy fashion), how are the number crunchers going to cross-reference that many records against multiple other databases? I somehow doubt we have that capability built into the system already, so it’s going to require some serious coding work.

And what happens if they do find instances of people answering the question differently than they did on previous government forms? It’s being suggested that the government can simply change the answer on the form, but any such changes are open to clerical errors and subsequent court challenges. Technically it’s a violation of federal law to lie on your census form under 18 U.S.C. § 1001 as well as Title 13, U.S. Code, Sections 141 and 193. But the reality is that the U.S. hasn’t prosecuted anyone over this in almost fifty years. Each instance of incorrect citizenship status recorded on the 2020 census could, in theory, result in prosecution, but do we really have the resources to go after all those cases? (Or even an interest in doing so?)

This is yet another case where I can fully agree with the intent of a proposal but immediately become skeptical about the underlying realities. Yes, we would like to have an accurate count of everyone in the country as required by the Constitution as well as getting a grip on the number of illegals we have to deal with. And the census is certainly the largest repository of data imaginable to achieve those goals. But this sort of journey down the Big Data wormhole sounds like it would be a daunting challenge even for someone with the resources of Google. Is it really practical? I think the Commerce Department should provide us with more specifics as to how they plan to pull it off, how long it will take and what it will cost before we sign on.