I will state that providing a specification that "I want to be x% confident that no more than y% are defective" is a common and effective quality control technique for go - no go testing. Nothing is 'wrong' in that approach. I will state if you can base the criteria on a Measurement - I want X% confidence that the measurement is Y +/- Z, then you can be more effective, and have a smaller sample size.
The EPA commonly uses 95/5 for cleanup standards. That implies 59 samples with no detected contamination. Now, is that sufficient risk control for say the residents of Palestine OH? Perhaps yes, perhaps no. This is where the "risk" and also perception of risk comes in.
The EPA commonly uses 95/5 for cleanup standards. That implies 59 samples with no detected contamination. Now, is that sufficient risk control for say the residents of Palestine OH? Perhaps yes, perhaps no. This is where the "risk" and also perception of risk comes in.
