Overview
This 1953 paper is one of Deming's most intellectually significant contributions, distinguishing between enumerative studies (describing what exists in a finite population) and analytic studies (predicting what will happen under changed conditions). While it appears to be a narrow methodological point about statistical inference, the distinction is foundational to Deming's entire epistemology and connects directly to his later system-of-profound-knowledge. The paper was published during the period between the juse-lectures-1950 and the later technical works like sample-design-in-business-research.
The Core Distinction
An enumerative study aims to describe a population that exists — how many defective items are in this shipment, what percentage of households have television sets, what is the average income in this county. The statistical framework is well understood: draw a sample, estimate parameters, calculate confidence intervals. An analytic study aims to predict what will happen — will changing this process reduce defects, will this policy intervention improve outcomes, will this management practice lead to better results. Analytic studies require a fundamentally different approach because the population of interest does not yet exist; it is the future output of a process.
Why This Matters for Management
The distinction is devastating for most management practice. Managers who rely on last quarter's numbers to make decisions about next quarter are treating an analytic problem as if it were enumerative. They are describing what was, not predicting what will be. Deming later argued in out-of-the-crisis and the-new-economics-for-industry-government-education that management is inherently about prediction — and therefore inherently about analytic studies. This means management requires theory, not just data. You cannot manage by the numbers alone because the numbers describe the past, and management must deal with the future.
Connection to Theory of Knowledge
The enumerative/analytic distinction is one of the intellectual roots of Deming's theory of knowledge — one of the four pillars of the system-of-profound-knowledge. Deming drew on C.I. Lewis's pragmatic epistemology to argue that knowledge requires prediction, prediction requires theory, and theory must be tested through the PDSA cycle. The 1953 paper plants the seed: if analytic studies (prediction) require theory beyond what the data alone can provide, then all management decisions require a theoretical framework. This is why Deming insisted that "there is no knowledge without theory" — a claim that sounds philosophical but is grounded in this statistical distinction.
Significance
This paper is underappreciated because it sits at the intersection of statistics and philosophy — too philosophical for most statisticians, too statistical for most philosophers. But it is the bridge between Deming the sampling theorist (the author of some-theory-of-sampling) and Deming the management philosopher. Without the enumerative/analytic distinction, the System of Profound Knowledge lacks its epistemological foundation. The paper was produced during a time when Deming's statistical career was at its technical peak but his management philosophy had not yet reached American audiences. It represents the intellectual pivot point in Deming's career — the moment when statistical methodology began to generate philosophical conclusions about the nature of knowledge and management.
Most industrial statistics is analytic in nature — managers want to know whether changing a process will improve future output — yet most statistical theory assumes enumerative conditions, where one draws inferences about a fixed, existing population. Deming's insight is that the standard tools of statistical inference (confidence intervals, hypothesis tests) are designed for enumerative problems and do not directly apply to the analytic problems that dominate management and process improvement. This gap between the questions managers need answered and the questions statistical theory is designed to answer is, in Deming's view, a fundamental source of management error.
Source
Free full PDF available from the Deming Institute at the URL above. This is one of the most accessible of Deming's technical papers and rewards careful reading by anyone interested in the philosophical foundations of his management thinking.