A relatively simple type of acceptance procedure currently in use has the dual acceptance requirements that the average value of all items of a sample must equal or exceed a limit and that each individual value must equal or exceed some lower limiting value. Provided the sampling is from a normal population, the probability of passing either of these criteria separately can be calculated quite readily. However, determining the compound probability of passing both requirements of the acceptance procedure is complicated by the fact that the individual probabilitis are correlated. A solution is presented that involves the calculation of upper and lower bounds for the desired probabilitry. This solution is then generalized to apply to any number of multiple acceptance criteria. An example is presented to demonstrate that the bounds are usually sufficiently close together to make this a practical approach. Because the calculations are quite tedious, a computer program has been written to simplify this part of the procedure. The speed and convenience of the computerized approach will permit specification writers to experiment with different acceptance limits to determine those that produce suitable producer's and consumer's risks. For the producer, it will provide a means to determine the target value necessary to ensure that the probability of acceptance will be at least the amount desired. (Author)

Media Info

  • Media Type: Print
  • Features: Figures; References;
  • Pagination: pp 20-23
  • Monograph Title: Computer Aids and Statistics in Quality Assurance
  • Serial:

Subject/Index Terms

Filing Info

  • Accession Number: 00319329
  • Record Type: Publication
  • ISBN: 0309030544
  • Files: TRIS, TRB
  • Created Date: Oct 27 1980 12:00AM