scientific article; zbMATH DE number 421721

From MaRDI portal
Publication:3135243





``In practice, if a lot is large, we can talk about being normally distributed without serious difficulty. This may not be true for small lots. We assume here that if a small lot has a bell-shaped distribution, with the same \(\gamma\) coefficients as the normal distributions, viz. \(\gamma_ 1'=0\), \(\gamma_ 2'=0\), it will be satisfactory to consider it as `normally distributed'. [\textit{A. J. Duncan}, Quality control and industrial statistics. (1965), p. 219-221.] Here it is shown that the last assumption is false, i.e. the discrete and finite distribution of the quality characteristic within a given lot cannot be approximated sufficiently well by a normal distribution in order to ``go from the mean and standard deviation to the fraction defective being the quantity of primary interest. For the entire collection see [Zbl 0771.00037].











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3135243)