Published in 1943 by Wiley, "Statistical Adjustment of Data" is Deming's first solo book and his most rigorous technical contribution to mathematical statistics. The work covers least squares estimation, curve fitting, error propagation, and the statistical treatment of redundant observations — the mathematical machinery used when a physical system is measured multiple times and the measurements must be reconciled with each other and with a theoretical model.
Technical Content and Scope
The book addresses a specific class of problem: given a set of observations subject to error, and given constraints that the observations must satisfy (either physical laws or definitional relationships), how does one find the best-fitting values? The method of least squares, developed by Gauss and Legendre in the early nineteenth century, is the classical answer, and Deming's contribution is a rigorous and practically oriented treatment that includes error propagation through the adjustment process.
The audience is professional statisticians, geodesists, physicists, and engineers working with measurement systems — not managers, not quality practitioners, not organizational leaders. The technical level is demanding. The book reflects the training Deming received at Yale under E.H. Moore and his exposure to the methods of the western-electric-bell-laboratories tradition, and it belongs firmly within the academic statistics of its era.
Key topics include the method of least squares with constraints, weighted least squares, the propagation of measurement error through functional relationships, and the statistical testing of fitted models. These methods were essential in geodesy (where surveying measurements must satisfy geometric constraints), spectroscopy, and experimental physics. The application to quality and management would not have been obvious to a 1943 reader.
Evidence of the Technical Foundation
The existence of this book matters for understanding Deming's intellectual development. By the time Deming became famous as a management philosopher — with out-of-the-crisis in 1982 and the-new-economics-for-industry-government-education in 1993 — it was easy to read him as a management thinker who used statistics as supporting evidence. "Statistical Adjustment of Data" makes visible that the trajectory ran in the opposite direction: Deming was a rigorous mathematical statistician first, who built his management philosophy on a technical foundation that most management writers lack entirely.
This matters for evaluating the system-of-profound-knowledge. The four components — appreciation-for-a-system, common-cause-vs-special-cause-variation, theory-of-knowledge, and psychology-of-management — are not the intuitions of a generalist but the synthesis of someone who had spent decades thinking carefully about measurement, error, and inference. The common-cause-vs-special-cause-variation distinction, inherited from walter-a-shewhart, required exactly the statistical understanding of error that this book demonstrates.
Placement in the Publication Sequence
The book was published during the usda-and-census-bureau-1927-1946 period, when Deming was working as a mathematical advisor at the Census Bureau. This placement is telling: the same year the book appeared, Deming was applying sampling methods to the U.S. Census and preparing the statistical expertise that would take him to Japan for the 1947 census work (see first-trip-to-japan-1947).
The sequence of Deming's technical books traces the arc of his statistical career before the management philosophy emerged: "Statistical Adjustment of Data" (1943) established the mathematical foundations; some-theory-of-sampling (1950) addressed large-scale survey design; sample-design-in-business-research (1960) applied sampling to commercial problems. By the time of that last technical book, Deming was already deeply involved in the Japanese quality transformation that would eventually become his primary legacy.
The Distance to Out of the Crisis
The intellectual distance from "Statistical Adjustment of Data" to out-of-the-crisis is enormous — least squares curve fitting versus a comprehensive philosophy of management transformation — and the path between them is not obvious. What bridged them was Deming's encounter with Shewhart's work on statistical quality control (see walter-a-shewhart), which showed that the statistical concepts of variation and process control had direct application to manufacturing and management. The 1943 book is one end of an intellectual arc; system-of-profound-knowledge is the other. Seeing both ends makes the arc visible.