From: Stan Hilliard
Date: 23 Sep 1999
Time: 01:12:23
Greetings David,
The amount to sample and the decision rule you should use will depend of the context. Please explain a little more about the situation - what is the population, what is the quality characteristic, and what is the objective.?
Also, assuming that you are dealing with attribute data, do you have the current process capability in terms of the lot fractions defective of what you consider OK lots? It would be useful to have this data on a number of lots to see how they differ. Also, think about how bad the lot should be before the sampling plan should have a slim chance of accepting it. This will make it easier to decide which levels you want the sampling plan to accept and which to reject. View this issue in terms of an operating characteristic curve:
www.samplingplans.com/usingoccurves.htm
Also, do you have a planned course of action if a lot is rejected?
Sincerely, Stan Hilliard