The new military standard for inspection by variables

The Standard described here was prepared to meet a growing need for the use of standard sampling plans for inspecting products procured by the Government. It may serve as an alternate to MIL‐STD‐105A, which provides sampling plans for lot‐acceptance inspection by attributes. The variables sampling plans apply to a single quality characteristic and maybe used whenever the quality characteristic for test can be measured on a continuous scale. Like the pattern of MIL‐STD‐105A, risks of rejecting acceptable material were planned to decrease with increase in size and lot. In the new Standard, however, samples sizes are considerably smaller than those in MIL‐STD‐105A for corresponding sampling plans The Standard outlines procedures to be followed when the variability is unknown—either a standard‐deviation or an average‐range method—as well as those to be used when the variability is known. A new concept in the Standard is a lot‐acceptance criterion obtained by comparing an unbiased minimum‐variance estimate, r, of the true percent defective with a suitably chosen constant, M, which represents the maximum allowable percent defective for sample estimates. This procedure, when used for the two‐sided specification‐limit case, provides operating characteristics, for given percent defective, which are essentially the same as those for the one‐sided specification‐limit case, regardless of the total percent defective in the lot, which consists of the sum of the percent defective above the upper and below the lower specification limits. For the one‐sided specification‐limit case, the acceptance criterion provides the identical probability of acceptance as obtained with the familiar ks statistic for suitably chosen k. Values of k also are included in the standard, thus, permitting use of either form of the acceptance criterion.