On Some Statistical Aids Toward Economic Production

Abstract

This paper covers management's responsibility for (1) design of product; (2) specification of service offered; (3) measurement by simple statistical methods of the amount of trouble with product or with service that can be ascribed to causes that only management can act on; (4) action on the causes so indicated. It shows by principle and by example how management may observe week by week the effects of guided effort toward reduction of trouble. The paper upsets a number of commonly accepted principles of administration. For example, a job-description, for best economy, should require the production-worker to achieve statistical control of his work; to meet specifications without paying the high cost of inspection, rework, and replacement. Statistical evidence of performance replaces opinion of foreman and supervisor. As a second principle, it is demoralizing and costly to call the attention of a production-worker to a defective item when he is in a state of statistical control. The fault for the defective item is not chargeable to the worker, but to the system. Fewer defectives can come only from a change in the system, not from efforts of the production-worker. Third, it is better to shift to a totally different job a worker that has developed statistical control of bad habits in his present job. All variation in quality-characteristics (dimension, hardness, color) causes loss, whether the variation results in defective product or not. Economies in manufacture are a natural consequence of reduction in the variation of a quality-characteristic. The author divides causes of variation into two sources: (1) the system (common causes) , the responsibility of management; (2) special causes, which are under the governance of the individual employee. In the author's experience, losses from the system overshadow losses from special causes. The same principles apply to sales and to service.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 85 citations exist upstream.