The Effect of Departures from Assumption when Manufacturing to a Specification

Abstract

When manufacturing to a specification articles of which some dimension varies by chance, the best conditions of manufacture can be found theoretically when the distribution of the dimension is known, often a normal distribution is assumed for this purpose. It is shown that even considerable departures from normality cause little change in the optimum conditions, and thus the theoretical calculations based on the normal distribution can be applied in practice over a wider class of situations.