Statistical Effects of Imperfect Inspection Sampling III. Screening (Group Testing)

This paper, the third in a series, describes the effects of imperfect inspection on screening (group testing) acceptance sampling procedures. These procedures are based on first grouping items and testing the group as a whole for presence or absence of nonconforming (NC) items and proceeding to individual testing only if such presence is indicated. If sampling is perfect, the procedure can reduce the expected amount of testing, especially if the proportion of NC items is low. Modifications of the basic procedure—curtailed and hierarchal group testing—aimed at further reduction in the expected amount of testing are also discussed. Some new formulas are presented.