Use Sequential Sampling 2018 - Use Sequential Sampling to Reduce Attribute Sampling Costs

Description :
ANSI/ASQ Z1.9 (formerly MIL-STD 105) is an internationally recognized set of sampling plans for attribute (pass/fail, good/bad, or defect count) data. The standard specifies, on the basis of (1) the lot size N, (2) the inspection level, and (3) the acceptable quality level (AQL), a sample size n and acceptance number c. If the number of nonconformances exceeds the acceptance number, the lot is rejected. This does not mean the entire lot is thrown away, but rather that 100% inspection is required to remove all nonconforming items and possibly replacing them with known good ones.
Inspection is, while a necessary or required (value-assisting activity), a non-value-adding one that does not actually create saleable product. This is why it is desirable to perform as little inspection as possible while continuing to meet customer requirements for quality assurance. ANSI/ASQ Z1.9 therefore offers double and multiple sampling plans that reject extremely bad lots very quickly and accept extremely good ones quickly. This reduces, on average, the total inspection performed.
A sequential sampling plan, in which the successive sample size is 1, is the ultimate extension of a multiple sampling plan. It is relatively easy to define one on the basis of the corresponding ANSI/ASQ Z1.9 plan's AQL and producer's risk ?, and de facto rejectable quality level (RQL) and consumer's risk. (ANSI/ASQ Z1.9 does not have formal RQLs but the procedure pretends that the nonconforming fraction p for which the acceptance probability, or consumer's risk, is 10%, is the RQL.) The operating characteristic (OC) curve of the sequential plan will then match that of the ANSI/ASQ Z1.9 plan at (AQL, 1-?) and (RQL, ?)—that is, the chance of accepting a lot at the AQL will be 1-? (generally close to 95%) and accepting it at the RQL will be ? (10%).
Sequential plans minimize the average sample number (ASN), but their standard form—in which each increment has its own acceptance and rejection number, and which can therefore consist of 100 or more tabular rows—is not particularly convenient for everyday use. Neither is the graphical version that requires the inspector to plot a point for each item inspected. This webinar offers a greatly reduced tabular format, often of 10 or fewer rows, in which the acceptance or rejection decision is based on the number of nonconformances found versus the number of items inspected.