ESHS/HSS Annual Meeting

Statistics under Scarcity: Jerzy Neyman and the Design of Hypothesis Testing, 1923-1933

• In Event: New Perspective on the Histories of Mathematics

English Abstract

As articulated in papers from 1927 to 1933 by Polish statistician Jerzy Neyman, hypothesis testing—a method for deciding whether evidence supports or rejects a statistical claim—has since become fundamental to statistical inference. In 1925, traveling to London—the epicenter of early 20th century statistics—under a newly independent Poland’s imperative to build its own statistical infrastructure, Neyman initially employed techniques that bore great resemblance to those of his mentors, the eugenicists Karl Pearson and Ronald Fisher. Yet upon his return to Poland in 1929, the techniques he deployed moved decisively away from British influences. I argue that this divergence emerged as a strategic response to conditions of severe resource scarcity that plagued interwar Poland. Further overwhelmed by the Great Depression, Neyman found that materials routine to British statistical life—made possible by ample royal and state backing—such as mathematical tables, calculating machines, and trained labor had become scarce or unattainable in Poland. Amid these difficulties, the 1933 framework championed a logic of reduction, mobilizing techniques novel to contemporary statistics—drawn from linear algebra and set theory—to recast complex problems into simpler ones. As Neyman himself cheerfully noted, instruments and tables designed for simpler problems could now be stretched to solve complex ones otherwise intractable or computationally costly, making his mathematical techniques literal resources deployed against material constraints. Conditions of material scarcity revealed how mathematical instruments such as tables and data infrastructures imposed boundaries on valid and calculable statistical knowledge. Viewing the field from Neyman’s vantage point at the periphery—attempting to build statistical infrastructure while emulating established traditions—makes visible underexamined disciplinary assumptions when the history of statistics is told from a center such as Britain.

Author

• Iris Wu, UC Berkeley