Abstract
Acceptance inspection and compliance testing often require levels of protection for both the consumer and the producer that make for large sample size relative to lot size. A given sample size can, however, be made to apply to several lots jointly if the lots can be shown to be homogeneous. This reduces the economic impact of a necessarily large sample size. Grand lot schemes, as introduced by L. E. Simon, can be used to effect such a reduction. This paper greatly simplifies application of the grand lot scheme by incorporating graphical analysis of means procedures in verifying the homogeneity of a grand lot. The resulting approach can be applied to attributes or variables data, is easy to use, provides high levels of protection economically, and can reduce sample size by as much as 80 percent. It may be applied to unique “one-off” lots, isolated lots from a continuing series, an isolated sequence of lots, or to a continuing series of lots.
Additional information
Notes on contributors
E. G. Schilling
Dr. Schilling is Manager of the Statistics and Quality Systems Operation at the General Electric Company, F. J. Borch Lighting Research Center. He is a Fellow of ASQC.