Acceptance Sampling

So far, we explore the binomial distribution in the context of acceptance sampling.

Acceptance sampling is a method used heavily in quality control. Lots (or batches) of a product are sampled, and each item in the sample is determined to be “conforming” or “nonconforming” to a set of standards. If the number of nonconforming items is above a certain threshold (called the acceptable quality level), the lot is rejected.

This method was first utilized during World War II in order to test munitions.^[1] Sampling plans today are often more complex than this original model and may involve multi-stage sampling. There are quality control companies who work with organizations and businesses to apply statistical methods to a particular situation in order to determine what the proper sample size should be, as well as what the acceptable quality level of a sample should be.^[2]

Resources for learning about acceptance sampling:

• https://www.ncbi.nlm.nih.gov/books/NBK235721/ (discusses sampling plans for seafood safety determinations made by the FDA)
• https://hbr.org/1983/09/product-defects-and-productivity (general discussion of these methods in Harvard Business Review)
• https://www.itl.nist.gov/div898/handbook/pmc/section2/pmc22.htm (description of acceptance sampling methods given by the National Institute of Standards and Technology, part of the U.S. Department of Commerce)
• https://asq.org/quality-resources/history-of-quality, https://asq.org/quality-resources/sampling/attributes-variables-sampling (some information about the history of acceptance sampling and its basics as part of the larger quality control discussion by the American Society for Quality)