Yesterday, I had the pleasure of finishing Desrosières’ The Politics of Large Numbers: A History of Statistical Reasoning. I say the pleasure of finishing, because the book is a bit of a slog but it picks up nicely in the last couple chapters. The first part of the book is an often jumpy history of certain ideas and debates in statistics (both the administrative data collection procedures known as statistics in the 18th and 19th century, and the mathematical statistics and probability we now associate with that word), jump-y because the examples come from four or five countries and 2 or 3 centuries. But the payoff is in excellent quotes like this one, describing the purpose of statistics and, I will argue, enchantment growing with distance*:

“The aim of statistical work is to make a priori separate things hold together, thus lending reality and consistency to larger, more complex objects. Purged of the unlimited abundance of the tangible manifestations of individual cases, these objects can then find a place in other constructs, be they cognitive or political.” (236)

Desrosières is arguing here that the job of statistics is to make certain kinds of black boxes by purging cases of their particularities and making portable objects (usually numbers) that hold together previously disparate information. This process cannot work if it is constantly interrogated, if enchantment does not grow with distance. The national income can’t be an input into the mathematical Keynesianism of the post-WWII era if national income statistics aren’t portable and black boxed. If they are continually interrogated, if we must always ask how they were made and what they mean, they cannot serve as input to other processes. In other words, enchantment is a feature, not a bug. That doesn’t imply we can’t fight it when the results are ones we don’t like, but rather that we should not expect to make every user of statistical data a completely reflexive sociologist – nothing could hold together if we spent all of our time critiquing the underlying concepts and methodologies. So, pick your battles (unless you really think all of these constructions are less than worthless).

* Enchantment grows with distance is a phrase that often gets invoked to describe how the users of statistical data and datasets who were not involved in their creation often have inflated views of the reliability and meaningfulness of the data as compared to those who helped create the data.