incremental sample size due to outcome
- March 23, 2011 at 3:04 pm #53765
ÖrtlundParticipant@Exiguus Include @Exiguus in your post and this person will
be notified via email.
Im trying to figure out a practical way for a problem at hand.
We will evaluate a test equipment that sorts parts in GO/NO GO (visual inspection).
The thing I want to do is to say if it does correct decisions to a certain level.
Say for example It shall be able to a correct decision in 99,8%. (eg 0,2% bad decisions)
And I would like to be 95% certain that it is so.
So the numbers of test to run (for each criteria), with zero faulty outcome will be approximately 1500 in order to have a CI 95% that is less 0,2% (using the rule of 3/n, n=1496 more accurate calculated).
Fine so, but if there is an faulty judgment? Then I need to go to n=2781 (with no more faulty judgment) in order to get the CI level to sink under 0,2%. And so on
It seams not that practical and a little bit rough. Anyone that have a better solution that is more practical?
You must be logged in to reply to this topic.