From: Stan Hilliard
Date: 16 Oct 2003
Time: 00:09:04
Eddie,
The sampling plan does not give you a basis for making a statement about a specific lot. For that, you need to calculate the confidence limit for that specific sample. Software program TP105 calculates confidence limits for any user-specified confidence. (95%, 90%, etc.)
A special case is for lots that pass sampling plans having Ac=0. In that case, there is only one possible number of defectives in the sample that will pass: zero. For lots that pass, you can make a statement like this example for the plan having a consumer's risk of beta=0.05 and RQL=0.10. (both decimal fractions.)
You can be 100x(1-0.05)=95 percent confident that the lot contains less than 100x0.10=10 percent defective items.
For lots with Ac>0 there is not a single probability statement for lots that pass. It depends on the number of defectives in the sample. Once that is known, the confidence limits can be calculated. The software program TP105 makes that calculation based on the binomial distribution. www.samplingplans.com/programtp105.htm