The 2020 U.S. census has been mired in controversies, from the eventually struck-down citizenship question to staffing problems. But in the end, it was an unexpected hurdle — the Covid-19 pandemic — that posed the most serious threat to the census’ success.
Will more people fill out their census online because they’re stuck home with nothing better to do? Or will the inability of census workers to knock on doors drastically affect the population count for this decade? To find out, GEN called Andrew Whitby, author of a sweeping new book about the global history of censuses called The Sum of the People.
GEN: Bottom line up front: How does the Covid-19 pandemic affect the 2020 U.S. census?
Andrew Whitby: It delays it is the simple answer. There are really two parts of the census to think about: There’s the part that is conducted without any personal contact, which started in the middle of March and has been ongoing. That’s mostly focused on internet response — that’s the mailings people have been getting saying, “Go online and reply.” Any people who haven’t replied yet will be getting a mailing that has a paper form they can fill out, or they can respond by telephone. And all of that is going ahead close to what was planned, and that should capture about 60% of the households that the Census Bureau targets. That part’s not affected.
The rest of it is partly people who live in remote areas — certain parts of Maine, Alaska, Native American communities — and partly people who don’t respond on the internet, by telephone, or by mail. The Census Bureau [usually] sends people called enumerators out a little bit after Census Day, the first of April, to knock on doors and try to get those census responses. That part is on hold. It was delayed first for two weeks, and then for another two weeks, and now it’s on the order of months. That affects when they would initially report the actual figures to the president and then a later set of figures that would be used for redistricting.
I believe the bureau said this week that enumerators would get into the field “as quickly as possible” after June 1. It’s just hard to imagine that June is a realistic timeline for people to be opening their doors to strangers right now.
Yeah, it’s a really good point. My book is not just about the U.S. Census — it’s really a global perspective — and the U.S. is not the only country planning a census for 2020. There are probably a couple dozen countries, including big ones like China. Some of those countries, like Brazil, have already announced that they are postponing the census a full year, into 2021. It’s a bit more difficult for countries like the U.S. or Mexico, who were already doing their census when the pandemic really broke. It’s also a bit difficult in the U.S. because of those constitutional statutory timelines. Some countries have more flexibility to play with the dates than I think the U.S. does.
Has any country’s response to this problem particularly impressed you, either in style or substance?
There are countries that are going to have more robust censuses during this time, precisely because they have a totally different method of doing it. Singapore has a census planned for this year, but they do it from administrative records. A lot of the European countries have this as well: They’ll have a set of records of who lives in the country, and they keep that constantly up to date. That means taking a census is really just a case of pressing a button on a computer, and it spits out the population data as of that date.
There are pros and cons of having a system like that, and it’s not obvious that you could persuade Americans to participate in a system where the government has a complete list of everybody who lives here. But effectively the government has those records on nearly everybody anyway, it’s just that they’re distributed — the IRS has some information, and the Social Security Administration has some information, and maybe Medicare and Medicaid have some information.
It’s tricky in the U.S. because of the large undocumented population, and I think you would need to find the political solution to that before you could really propose something centralized. But countries that have that system set up, yeah, they’re not affected by the coronavirus at all.
You note in the book that international recommendations suggest avoiding censuses during “traditional festivals, pilgrimages, and fasting periods,” when people are away from home. Does this make sheltering in place a good time to take a census, at least for the self-reporting part?
In some ways, yes. As of Tuesday, the national self-response rate was 49.1%. They got a 66.5% mail response rate in 2010, and they cut their target for this year back to 60.5%, because it’s getting increasingly difficult to get people to respond to surveys. My guess is they’ll hit that.
The problem is that you have a lot of people who are not in their usual place of residence. Certainly, a lot of college students who would normally be in their student housing or college residence are now maybe back living with their parents, potentially in a different state. There have been news reports of New Yorkers who have second homes fleeing upstate or interstate. So there’s a risk that you’ll have people responding not in the place where they normally are.
Some of the predictions say perhaps 200,000 Americans will die from this virus. Is that statistically significant to the census?
You could argue that it makes the census immediately less relevant. If you collected the data on the first of April, some number of those people are no longer alive. We’re using the census for apportioning political representation, we’re using the census for distributing funds, and now you’re creating representation for people who are no longer alive.
In the scheme of things, I don’t think it’s really a noticeable change, because they’re not talking about 10% of the population dying. If you think about it, we set those apportionment numbers once every 10 years anyway, so they get quite far away from reality over that 10-year period, even in a totally normal decade. And we’re comfortable with that as a nation.
What were the main concerns for the U.S. census in 2020 before this pandemic hit?
It’s quite funny because they all seem fairly insignificant now compared with the problems they’re actually facing. But there were at least three that I can think of. One was that the Census Bureau had been underfunded. They did a major dress rehearsal test in 2018, which was supposed to occur in several parts of the country; instead, it was focused just on Providence County in Rhode Island. Potentially it’s still a good test; you still get a good sense of how you need to tweak things. But essentially it means that you’ve missed out on how people in a different part of the country with a different cultural mix respond. It increases the risk that something won’t go right when it comes to the actual census.
The second thing was the citizenship question debate. Last year, the Supreme Court cut the question on technical grounds. But still, the fact that it was in the news for a year or so meant there was certainly a concern that it would have an impact on response rates within the Latino community, the immigrant community in general, and among undocumented people.
Third, this was going to be the first census that offered an internet response option to the bulk of the community as the default way of responding. As we know, the internet today is an environment where you have to assume there are hostile actors trying to attack and create disinformation and misinformation. But as far as I can tell, it’s been completely smooth.
There was a strange question on the census this year — after I selected my race, it was mandatory to fill in a text field on my ethnic background. What was that about?
I don’t know the history of how that field came in. It’s the first time, certainly in recent years, that if you chose the White or Black race option, you got this additional text field to further specify that.
It’s interesting that you call it mandatory — it is mandatory, legally, to respond to every question on the census. And the internet response certainly doesn’t seem like you can avoid it. If you leave it blank and click “next,” it comes up with an error. But in fact, if you click “next” again, you can skip to the next page, so you actually can avoid it. But the Census Bureau hasn’t advertised that at all.
The same is true of the binary sex question — you can click “next” twice and go past it. The Census Bureau will just look at all your other answers and say, “Statistically, you probably would have replied male, or you probably would have replied female,” and they’ll allocate that to you.
If it makes anyone feel better, they can not answer that question. Legally, they have to, but probably no one’s going to prosecute you because you didn’t answer what kind of White you are or whether you’re male or female.
This interview has been edited and condensed for clarity.