What is the confidence level for a given AQL in zero sampling plan

Hi folks,

How to answer a such question like "What is the confidence levell on the lots when we are given a AQL in the zero sampling plan"?

As I know we can only studying the level of the PPM in the zero sampling plan.

Thanks and regards
YJ

AQLs are typically set at an alpha level of .05 (the consumer's risk). If you look at the original source of the sampling plan this will probably be defined. You can also create the Operating Characteristic (OC) curve that will allow you to see the AQL and LTPD (producer's risk) points.

Then do the confidence which is 1 - alpha will always be same when we use the zero acceptance sampling plan?

Of there is any related to the AQL itself?

What parameter do you want to be confident about?

In Z.14 plans (aka MIL-STD-105E), it is approximately true that lots with a defect rate equal to the AQL will be accepted approximately 95% of the time. Thus the producer is 95% confident that a "good" lot will be accepted.

However, that doesn't not really seem like the sort of confidence level you would be interested in. A more likely parameter to be confident about would be "I'm 95% confident that I would reject a lot with x% defective". For this sort of confidence, c=0 plans are the worst!

For example, suppose you have a lot of 50,000 and AQL = 0.1. For c=0, the only choice (for any size lot) would be a sample of 200. Looking at the table, you find that you would need 2.4% bad for c=0, N=200 before you would be 95% confident of rejecting the lot. But if you go up to the sample size of 500 suggested by the table for Normal Level II, then a lot with 0.95% bad for c=1, N=500 will get 95% rejection rate.



Tim F

Thanks Tim, I think we suppose not to use the confidence as this is dealing with human beings which is not deterministic at times.

In the sampling plan, we are interesting on how often the submitted lot will be accepted if the process running at defective level p.