An investigation of the effects of two types of inspector error on sampling inspection plans
Résumé
The field of statistical quality control plays an important part in the manufacture of consumer, industrial, and military products. Sampling inspection constitutes one of the most important areas of statistical quality control.
In order to provide an alternative to one hundred percent inspection, especially in situations where testing is of a destructive or exceedingly expensive nature, statistical sampling procedures have been developed. Generally, inspection procedures are performed disregarding any possible effects of human involvement, even though thesuccessful operation of inspection plans may depend to a large extent on a human inspector. The involvement of the inspector may vary from recording inspection machine data to making a subjective evaluation of product quality. The classical formulations of sampling inspection plans assume that the human inspector is always perfect, when in reality the human inspector is capable of making two types of errors--he may classify a good item as bad or he may classify a bad item as good.