Democracy's Data

 Dan Bouk. Democracy's Data the Hidden Stories in the U.S. Census and How to Read Them.  Ferrar, Straus and Giroux, 2022.

I had read about this conference and seen it mentioned in many places as I began studying the 1940 Census.  In the context of later controversies, particularly over a question about income, this conference was frequently invoked to justify the census's decisions, to show that they had been made by expert, responsible men.  Yet for the first couple of years of my research, I found no detailed records from the conference.  I dug through file after file, box after box in the National Archives, to no avail.  I did not think I would ever discover what actually happened in that room.  Then, in what would be my final trip to that archives, in the summer of 2019, I was working my way through Director Austin's papers, and there it was, the summary minutes of the meeting, one that (although it was not a complete transcript) would allow me to reconstruct, more or less, how the Question Men built the data's frame.   pp. 33-39

One day in the National Archives, I stumbled upon  a silence in Price's paper itself.  Working through a box of bureau records, I found a folder titled "Negro Enumeration of 1920" containing a typed reply to a scholar who had, in 1922, made a reasoned case for a significant undercount of African Americans.  This surprised me.  I had never heard of serious scientific estimates of such undercounting before Price, and Price's paper hadn't said anything about Black researchers who had previously uncovered this.  Price's paper hadn't prepared me for Kelly Miller. 131

 

COMMENT

Imagine the astonishment and joy at finding the missing document after years of searching!  One thing that stands out in this story is, the meeting was well known, and known to be important.  It just wasn't well documented. 

The second story illustrates the way that information errors propagate through time.  Price's paper is frequently cited as the first research about undercounting Black people in the census.  It turned out that Miller, a Black mathematician, had already detected Census undercounts and explored the underlying causes, but since Price hadn't cited Miller's research that scholarship had been neglected.