Abstract
The paper opens with a description of the applications of “qualitative” inspection. Briefly, it is used where “quantitative” inspection, or inspection by direct measurement on a continuous scale, is impossible, as in the case of a destructive test or a visual inspection for blemishes, or as an economical alternative through limit gauges to “quantitative” inspection.
The industrial and theoretical factors affecting the choice of a sampling scheme for qualitative inspection are discussed, particularly producer's risk (the risk that good material may be rejected), and consumer's risk (the risk that bad material may be passed). It is pointed out that a sampling scheme can be devised to give an agreed maximum producer's and consumer's risk. It is shown that the maximum allowable percentage of defectives cannot be set in an arbitrary manner, but only after consideration of the process average percentage of defectives. If the ratio of the maximum allowable to the process average is too low (e.g. if the requirement is too severe), the minimum sample size for reasonable producer's and consumer's risk is impossibly high.
A simple chart is provided, from which, with the industrial factors limiting the size of lots and the size of samples, the desired percentage of inspection, and the process average percentage of defectives as the “known” quantities, it is possible to read off the best single or double sampling scheme to give agreed maximum producer's and consumer's risks. The chart gives the best size of sample and the corresponding “pass up to” number of defectives per sample.
The paper concludes with instructions on when and how to change to more severe sampling conditions, when results from successive samples show a deterioration in quality from the process average; and discusses the need to keep the new conditions in force until there is satisfactory evidence that the quality has improved again to the process average.
Get full access to this article
View all access options for this article.