Shewhart, Walter Andrew

Copyright notice

This article Walter Andrew Shewhart was adapted from an original article by Denis Bayart, which appeared in StatProb: The Encyclopedia Sponsored by Statistics and Probability Societies. The original article ([http://statprob.com/encyclopedia/WalterAndrewSHEWHART.html StatProb Source], Local Files: pdf | tex) is copyrighted by the author(s), the article has been donated to Encyclopedia of Mathematics, and its further issues are under Creative Commons Attribution Share-Alike License'. All pages from StatProb are contained in the Category StatProb.

Walter Andrew SHEWHART

b. 18 March 1891 - d. 11 March 1967

Summary. Shewhart was the father of statistical methods of quality control. His methods were closely related to an epistemology of physics and application to economics and industry in his time.

Born in Illinois, Shewhart first pursued his tertiary education at the University of Illinois, and was awarded his PhD in Physics from the University of California in 1917. He was an Assistant in Physics at the University of Illinois in 1916-17, and then became Head of the Physics Department at the Wisconsin Normal School at Lacrosse in 1917-18. In 1918, he was offered a job with the Western Electric Company, and was later transferred to the Bell Telephone Laboratories in 1925; he remained in their employ until his retirement in 1956.

Shewhart's name is associated with methods of statistical quality control in industrial mass production (Shewhart control charts), but his contribution to the field of statistical techniques must be placed within the background of a theory of knowledge which he developed in the specific context of industrial production. Shewhart was profoundly influenced by the philosophy of pragmatism of C.S. Peirce, W. James and C.I. Lewis; this he put to work operationally in the field of engineering. At the same time, he tried to conceptualize the collective aspect of work in an enterprise; this caused him to be equally well known in the field of management, particularly thanks to W.E. Deming (q.v.) who promoted his views. Through Deming, Shewhart became well known in Japan where he consequently played an important role in the development of methods of management and quality which are associated with the name of K. Ichikawa.

Work on control charts originated in a political decision made by the Engineering Department management at Western Electric, which wished to define scientifically a method for ensuring "a Company-wide view" of what constituted quality assurance. Under the direction of R.L. Jones and G.D. Edwards, an entire team began to attack these problems: in addition to Shewhart, one should also mention H.F. Dodge, E.C. Molina, T.C. Fry, and H.G. Romig. When the Bell Telephone Laboratories were set up officially in 1925, engineering inspection appeared as part of their organizational chart. In 1931, Shewhart published his entire work on control charts, in a masterly treatise which still remains relevant today. The application of these statistical quality assurance methods is an achievement of which Bell Labs is still proud (Fagen, 1975).

Control chart methods are particularly well suited for use in workshops, as a result of their elegant usage of graphical representations, which limit the operator's need to resort to calculations. Since product quality is defined by a set of measurable physical quantities, which are assumed to follow statistical distributions, the control charts provide a relatively simple procedure for the visual representation of the evolution of the population mean and standard deviation. So far as statistical background was concerned, Shewhart first made use of Karl Pearson's (q.v.) theory of distributions, which he complemented with the small sample theory of "Student" (Gossett, q.v.) and Fisher (q.v.). He introduced the concept of ``rational subgroup" to define the method and frequency of sampling.

Control charts were, however, only the tip of his personal iceberg. Although he worked in industry, Shewhart remained a physicist, interested in problems of measurement in experimental physics, and in the nature of the acquisition of knowledge. He believed profoundly in the statistical nature of all physical laws, and attacked any deterministic concepts:

We ordinarily think of the physical and engineering sciences as being exact. In a majority of physical measurements this is practically true. ... With the introduction of the molecular theory and the theory of quanta, it has been necessary to modify some of our older conceptions. Thus, more and more we are led to consider the problem of measuring any physical quantity as that of establishing its most probable value. We are led to conceive of the physico-chemical laws as a statistical determinism to which the "law of large numbers" imparts the appearance of infinite precision. (Shewhart (1913))

To provide a theoretical foundation for the use of control charts, Shewhart, developed the concept of a ``constant system of chance causes". Intuitively, the idea is simple: such a system is a random sytem governed by a statistical distribution independent of time. The control chart allows us to detect the occurrence of significant causes of perturbation (called assignable causes), which can then be tracked down and eliminated. One can thus progressively (as experience proves, according to Shewhart) eliminate all assignable causes, and the system can then be said to be ``under statistical control".

Shewhart's theories create a bridge between the epistemology \footnote {the theory of the methods of grounds of knowledge} of measurement and the notion of human organization which must make the process of measurement work. Statistical control according to him consists of three phases: specification, production and inspection, corresponding respectively to making a hypothesis, carrying out an experiment and testing the hypothesis. ``The three steps consitute a dynamic scientific process of acquiring scientific knowledge" (Shewhart, 1939, p. 45). These three steps were reformulated by Deming under the name ``Shewhart cycle", and they arise in recent developments in the theory of organizations, such as organizational apprenticeship, of which Shewhart appears to have been a forerunner. In the same book, he writes: ``mass production viewed in this way constitutes a continuing and self-corrective method for making the most efficient use of raw and fabricated material".

A paradox remains in the social persona of Shewhart. His works, full of references to physics and to very diverse technical areas, make difficult reading and his lectures were little understood by industrialists. On the other hand he is the inventor of quality control which has been well accepted by industry, and which is recognized as easy to use. Clearly the weight of AT&T in American industry played an essential role in the promotion of his work, from the moment when the firm decided to support it. But one must also stress, his great personal ability to form alliances and weave relationships among institutional teams, which was essential to the adoption of his ideas. He played masterfully the game of recognition by agencies, learned societies and engineering associations. Already in 1929, he obtained funds to create a Joint Committee of the American Society for Testing Materials and the powerful American Society of Mechanical Engineers. This resulted in the publication in 1935 of a standard which was later adopted during the Second World War for the requirements of the military. One should also mention the promotional effects of the lectures which he delivered in London in 1932, at the invitation of Egon Pearson (q.v.). The British, who did not wish to be left behind in this area, undertook several studies which later appeared as publications of the Royal Statistical Society.

During his lifetime, Shewhart was the recipient of many distinctions. Among these were his founding membership of the Institute of Mathematical Statistics, of which he was elected Fellow in 1937 and President in 1944. He was a founding member of the American Society for Quality Control in 1947, and was their first Shewhart Medallist in 1948. He was elected to the Presidency of the American Statistical Association in 1945, and was awarded the Holley Medal of the American Society of Mechanical Engineers in 1954. He was a member of the International Statistical Institute from 1950, and of the Royal Statistical Society which elected him to honorary Fellowship in 1954.

Shewhart was nevertheless a modest and magnanimous individual, who believed in the fertility of collective work, and the value of exchanges between disciplines and professions. He saw the role of the statistician in an organization as that of the coordinator of several experts involved in the process of the production of knowledge.

References

[1]	Deming, W.E. (1978). Shewhart, Walter, A. In International Encyclopedia of Statistics. Ed. W.H. Kruskal and J.M.Tanur, Free Press, New York. Vol. 2, 942-944.
[2]	Eisenhardt, C.(1990). Shewhart, Walter Andrew. In Dictionary of Scientific Biography, Vol. 18, Supplement II, 816-819. Scribner, New York. This contains a detailed bibliography.
[3]	Industrial Quality Control. (1967), 24. Special Issue in Memory of Shewhart, containing several testimonials.
[4]	Fagen, M.D. (Ed.) (1975). A History of Enginering and Science in the Bell System, The Early Years (1875-1925), Vol. 1. Bell Telephone Laboratories, N.J., Especially: Chapter 9 ``Quality Assurance".
[5]	Shewhart, W.A. (1931). Economic Control of Quality of Manufactured Products. Van Nostrand, New York and MacMillan, London (501 pages).
[6]	Shewhart, W.A. (1939). Statistical Methods from the Viewpoint of Quality Control. The Graduate School, U.S. Department of Agriculture, Washington D.C. (105 pages). Republished by Dover Publications with a new Foreword by W.E. Deming in 1986.

Reprinted with permission from Christopher Charles Heyde and Eugene William Seneta (Editors), Statisticians of the Centuries, Springer-Verlag Inc., New York, USA.