Enumerative and Analytic Studies

An excerpt from The Handbook for Quality Management (2013, McGraw-Hill) by Paul Keller and Thomas Pyzdek

Deming (1975) defines enumerative and analytic studies as follows:

Enumerative study.A study in which action will be taken on the universe.

Analytic study.A study in which action will be taken on a process to improve performance in the future.

The term universe is somewhat synonymous with population: the entire group of interest. For example, the expected voters in a specific election might constitute a population of interest.

In an analytic study the focus is on a process and how to improve it. The focus is the future. Thus, unlike enumerative studies, which make inferences about the universe actually studied, analytic studies are interested in a universe that has yet to be produced. Table 9.2 compares analytic studies with enumerative studies (Provost, 1988).

Consideration	Enumerative Study	Analytic Study
Aim	Parameter estimation	Prediction
Focus	Universe	Process
Method of access	Counts, statistics	Models of the process (e.g., flow charts,causes and effects, mathematical models)
Major source of uncertainty	Sampling variation	Extrapolation into the future
Uncertainty quantifiable?	Yes	No
Environment for the study	Static	Dynamic

Table 9.2 - Important Aspects of Analytic Studies

With regard to the analysis of processes, Deming (1986) comments:

Analysis of variance, t-tests, confidence intervals, and other statistical techniques taught in the books, however interesting, are inappropriate because they provide no basis for prediction and because they bury the information contained in the order of production.

In organizations, processes are carried out as repeatable activities, carried out time and time again. The element of time is lost in the enumerative tools of confidence intervals and hypothesis testing. While useful for analyzing short-term data from a planned experiment, for example, these enumerative tools pool the variation that occurs over time into a single estimate of sample variation.

This is a persistent and unfortunate problem with the use of histograms. Apparently,most practitioners learn that histograms are useful to graphically show the shape of the data, which is fundamentally true. Unfortunately, the shape of the data and the expected shape of the process are completely different if the process is not stable. An example of this will be shown shortly.