Displacing Judgment: Statistical Graphic Methods and Corporate Executive Management in the United States, 1910s–20s

This paper examines the adoption of graphic methods—techniques for analyzing and displaying data using statistical maps and graphs—as a means of automating decision-making at the American Telephone & Telegraph Company (AT&T) and the chemical manufacturer E.I. du Pont de Nemours & Company (Du Pont) in the 1910s and 20s. In the early twentieth-century United States, leading business experts rallied behind calls for a more ‘scientific’ approach to business. Spurred by a series of financial panics, they sought to replace corporate executives’ experience and intuition with ‘facts’—defined as statistical data, properly analyzed—as the basis for executive management. Much of this data analysis was accomplished through graphic methods, which were thought to bring order to overwhelming masses of statistical data. That is, graphic methods used visual form—such as curves, dots, and colors—to transform so-called 'raw' data into comprehensible patterns and trends that could be used to guide business decisions. Thus, we can understand AT&T and Du Pont's adoption of graphic methods as an early effort in managerial automation, as well as a forerunner of today’s regime of algorithmic- and AI-based decision-making. And yet, by attending to graphic methods’ visual, material, and conceptual dimensions, I demonstrate that these visual-analytical practices did not eliminate judgment from corporate decision-making so much as displace it, from executives to statisticians. This finding complicates conventional understandings of statistical analysis and data science as rational and objective enterprises, showing that intuition was integral to modern statistical practices.