From: bart.debruyn@vlm.be
Date: 24 Sep 2001
Time: 09:50:56
greetings Stan,
I doubt that I have well understood your comment. In the given example there should be no conformance to the specifications: p'= 96% +/- 3% whereas f> 97% (95% confidence) has been specified and 93% is lower than 97%. The problem I see is that both upper and lower confidence limits depend on sample size. It is not possible that an acceptance decision could depend on sample size when controlling a lot.