Dan Bouk researches the history of bureaucracies, quantification, and other modern things shrouded in cloaks of boringness. He studied computational mathematics as an undergraduate at Michigan State, before earning a Ph.D. in history from Princeton University. His work investigates the ways that corporations, states, and the experts they employ have used, abused, made, and re-made the categories that structure our daily experiences of being human. His first book, How Our Days Became Numbered: Risk and the Rise of the Statistical Individual (Chicago, 2015), explored the spread into ordinary Americans' lives of the United States life insurance industry's methods for quantifying people, for discriminating by race, for justifying inequality, and for thinking statistically. His new book, Democracy's Data: The Hidden Stories in the US Census and How to Read Them will be published by MCD/Farrar, Straus and Giroux in August 2022. In an age when we often hear that good governance requires that we depend on good data, it is crucial that everyone (and not just those in quantitative fields) understand and can work to improve the processes that make data from people. Democracy's Data is a history of the 1940 census that will prepare its readers to examine and critique the data-driven systems that surround us. Dan blogs about his on-going research at shroudedincloaksofboringness.com.
When did you first become interested in census data as a means of thinking about politics and public policy?
Historians use census data as a way to get a peek into past societies and how they worked. It’s used to document the lives of people who otherwise wouldn’t have records kept about them. It becomes available in the US 72 years after it takes place, in manuscript form, and it’s remarkable because it preserves folks and their stories. So, as a historian, I knew and used the census for that purpose.
But, in the case of my current project, I was in the National Archives, researching a woman named Elbertie Foudray and her role in the Census Bureau’s work in the 1920s and 1930. In the midst of that research, I came across these letters, written by the enumerators of the 1940 census, to their supervisors, complaining about not being paid properly.
What clicked for me in that moment was that what I’m actually looking at when I look at census data is a ridiculously large and complex labor system that generates knowledge about a population. And that means that if I try to open up that infrastructure, there’s going to be a lot of interesting stories and much more to this data than I had thought about before.
Could you talk a little bit about Elbertie Foudray? I think her story is so interesting.
I first started digging into her story because she produced some really interesting statistical tables that Alfred Lotka, a preeminent demographer, used to build simulations of populations in a time before computers. I wanted to know more about this person who provided these tools that he relied on.
Then, when I was researching those letters in the Census Bureau, her name would be mentioned very frequently — people would talk about getting her to check their math. It became clear that she was by far the most accomplished mathematician in the Census Bureau for a number of decades.
And yet, when the Bureau decides it’s time to hire an actuary, they hire a man. It’s a classic story of how gender discrimination gets in the way of a remarkable career. Through some digging, I know a little more about her life. But ultimately, there’s still so much more about her I don’t know.
How has the original vision for Census data changed from its inception to the form that Americans fill out today?
Let’s talk about two big changes. The original reason that there is a Census is because the Constitution of the United States required there to be an actual enumeration of individuals as a basis for them assigning to each state its share of representation in the House of Representatives. At a fundamental level, the purpose of the Census was to ensure that as state populations changed, the representation would adjust accordingly.
At its inception, the Constitution counted enslaved people as three-fifths of an individual and counted no Native Americans. One of the major changes is that Census data has become more inclusive and the Census now seeks to count all Americans and count them equally.
Another big change is the number of questions asked. Most people in the 2020 census only answered seven questions. That’s pretty similar to back in the earliest days, though then the head of the household (usually a man) was the only one who answered the questions.
During the time in between, in the 1940 and 1950 Censuses for example, there were twenty-something questions asked of everyone. Then, however, sampling was introduced to ask extra questions of individuals and get quick and reliable information about a lot of people. But the general population is asked fewer questions, in order to get accurate information from them, or so it is hoped. Sampling was brought in as a means of trying to get better coverage of everyone, but at the cost of minimizing what future generations will know about each individual.
It seems like a lot is lost in that minimization. Speaking for the Irish census, we have a section called the “Time Capsule” where you can write whatever you like. It’s great — people got really creative with it, even putting in memes and things like that.
You should tell your readers that they should get the US Census Bureau to include a time capsule!
Seriously, a lot of people today who are interested in the histories of their communities or their genealogies use census records as a primary means of getting access to that kind of information. They’re going to be very surprised when the 1960s, 1970s, 1980s data is released and they realize how little information exists in those records compared to the materials we’ve gathered until now.
A time capsule would be a great gift to future generations! It’s a symbol from the state to its people. It’s saying that you can define yourself and you don’t have to exist purely within the narrowly constructed bureaucratic forms we’re giving you. So I think it’s just a great thing.
How has the representative power of the census changed as the American population has morphed and grown over time?
For the longest time, the census would produce its counts and Congress would figure out how to translate those totals into an apportionment of representation. That meant not just how many seats each state got, but how big the House would be. For the first half of American history, the most consistent response from Congress to increasing population size was to increase the size of the House.
The population was growing more substantially than the growth of the House however, and the number of people represented by each member had doubled by the 1920s. Since then, the House has been frozen, while the population has tripled.
What I’ve come to determine through my research is that over the course of the 1920s and 1930s, Congress essentially removed itself from the process of translating census data into apportionment outcomes. This began in 1920, when apportionment failed and Congress discovered that if they wanted to hold the size of the house steady, they couldn’t do it by legislative enactment. So, instead, they forced the process to be automatic.
I’m sure a lot of your readers are thinking “good, that gets politics out of the process”. But the major outcome here is not increased rationalization, but instead a decline in representative power for most individuals. The net result is that Americans are continuously less well represented.
How important is what I’ve heard you call “substantial accounting”? What is lost when it’s minimized down to seven questions?
It’s important to recognize that these are double edged swords. Every time I look at a census record from the past, I think of it as a moment in which I'm seeing a negotiation in which an individual is talking to an enumerator or looking at a form and engaging with this form. At the same time they’re trying to figure out how they're going to fit their own very complex life into whatever set of possible outcomes the form will allow, or the enumerators are willing to write down.
That can be a moment of creativity as people figure out how to make themselves fit. It can also be a distressing encounter as people are forced to accept identities that don’t fit them well. If we limit the number of ways people have to be put into boxes, then there’s a liberatory possibility of asking fewer questions—this is why I love the idea of putting in the time capsule!
But in terms of thinking about this as a historian, there are so many people who will never have a proper record of them kept. Even those of us with significant digital trails, do we trust Facebook or Google to maintain a record of our data? I think it’s quite likely that a lot of that data will evaporate over the next fifty years.
One of the remarkable things about the Census Bureau is that it is a public institution devoted to keeping a record of every single person in the US. That record is secured confidentially for 72 years and then held for posterity as long as the United States exists. That’s a beautiful model. It also means that the less we ask in each Census, the narrower the picture of those people will be in those records.
If there was an advocacy effort made to broaden out the 2030 census, when would that have to take place? What would it look like?
It has to start now. Three to four years out, the Census Bureau will be shopping around its questions and informing legislators what it's considering. But to make any significant change, advocacy work must start now.
One of my pet projects is trying to find people capable of organizing to lead an effort on apportionment. We need to think about how the apportionment system could be reconsidered and opened up again, which requires legislative change.
There are important changes to be considered in the Census itself. The gender and sex questions, for instance, are a place where we should see movement to open those up beyond just a binary. And we really need to get moving on that time capsule! It would be an incredible shame if we didn’t do that.
Can you tell us a little about your book coming out later this year?
We talk constantly about how we want to have data that will allow us to make better decisions about how to govern and organize. I agree with that, and I’m sure your readers feel the same way.
That means it’s really important that we understand how our data is made, where it comes from, what decisions and assumptions went into it. How are people made into data? And how can we use that data to better govern, represent, and improve their lives?
This book, I hope, will be a historical guide, one that can help readers think about the process by which people are turned into data and how that should help us think about the systems we construct.
