Statistics and engineering practice

This paper was originally written merely as an elementary introduction to a discussion on the application to industrial problems of modern statistical methods. In view, however, of the alteration in the arrangements for meetings minor changes have been made, but the paper should still be read as an introduction to a general discussion on the subject. The authors have had in mind particularly the scrutiny of test data obtained in the course of the development of new manufacturing procedures and in the control of established ones. Having first dealt with the fundamental conception of variability and the basic effect of chance on the magnitudes of test values, an approach is made to the more elementary methods, illustrated by examples in order of increasing complexity. At the end of the paper it is indicated that extensions to the Analysis of Variance involving correlation and regression are useful aids in testing and research. The authors emphasize that the efficient use of the statistical tool can come only by practice, and that in no sense can it be substituted for sound technical knowledge, with which in all problems it must be combined.