Downloaded from https://www.everyspec.com AMCP 706-198 TABLE 3.4 COMPARISON OF SINGLE, MULTIPLE, AND SEQUENTIAL SAMPLE PLANS (Adapted and modified from Ralf. 1) clw telsle1-uample rn-sample seq-sample $samote Size Known lcain be a random Average can be computed for Avurage can be computed variable) various Incoming quality for various Incoming levels, Often less than quality levels. Often less I ~.50nipia. then m-sarnplo.
Decision Accept or reject Accept, e-ject, or tike Accept, reject, or test1 Choices another Aarnple until final another Item wtinple Is selected Predetermtnes Two of the three Same as 1 -sarmple Fix awsid 0; N Is a random Characteristics quantities A, ax, or Variable Statistical Must know dlstribu- Same as 1 -tample Samo as 1sample, C4 A 3 are JConsiderations tion of temple rarely known exactly, statistic Usual formulas are very approximate. Personnel Requires least Better trained people re.
Requires most training Twialning training quircid than for single Ease of Easiest, Scheduling can More difficult tihan 1-sample Mast difficult In terms of Administration be fairly precise and qince th~i exact number of testing, scheduling, and precise toot-cost osti. tests is unknown, Only overail administration, metes can ho mode average test costs can be Most time consuming. estimated of administration Is 'seucond chance" by taking Can bc truncated and still . most Important and L jher samples If first maintain good ax, i3, Cu,,- cost of tasting Is rela. sample results Indicate a suit statistician knowledge. tively unimportant, marginal lot. able In this area.
Miscellaneous Best used for tastimj Has psychological advantage !',on be meat "effilnent" un- situations whero vaso In that supplier Is given a der many circumstances, 3 - ~~~~ ~3-16- . tMv4 . . 5 IlI