Zero Acceptance Number Sampling Plans for Isolated Lots
- Thread starter mdurivage
- Start date
is there a reference that says it's not?
The statistics of acceptance sampling is NOT based on a contiguous stream of lots. I would be hard pressed to find a citation that states this explicitly (WAY too much reading), although any good stats book or paper that explains how to determine an appropriate sample size given your 'acceptable quality level' and alpha risk (or confidence) - and in the case of 'small' lots the lot size so that the appropriate sample size correction factor can be applied - will describe conditions of both isolated lots and contiguous process streams. THere will alsways be the caveat of RANDOM sampling but this is a concern with isolated lots as well as a contiguous stream...The FDA is loaded with talented highly educated statisticians who will know this. (It's stats 101 stuff) If you have someone who claims otherwise I would challenge THEM to site a reputable reference and reasoning.
The classic tome on acceptance sampling is "Acceptance Sampling in Quality Control" by Edward G. Schilling
The statistics of acceptance sampling is NOT based on a contiguous stream of lots. I would be hard pressed to find a citation that states this explicitly (WAY too much reading), although any good stats book or paper that explains how to determine an appropriate sample size given your 'acceptable quality level' and alpha risk (or confidence) - and in the case of 'small' lots the lot size so that the appropriate sample size correction factor can be applied - will describe conditions of both isolated lots and contiguous process streams. THere will alsways be the caveat of RANDOM sampling but this is a concern with isolated lots as well as a contiguous stream...The FDA is loaded with talented highly educated statisticians who will know this. (It's stats 101 stuff) If you have someone who claims otherwise I would challenge THEM to site a reputable reference and reasoning.
The classic tome on acceptance sampling is "Acceptance Sampling in Quality Control" by Edward G. Schilling
ah - that does change things but only slightly.
isolated lots are typically inspected with concern that the defect rate of the lot not exceed some level. (refered to as the limiting level, LTPD, RQL, yada yada)
An AQL plan is the Acceptable Quality level
A Limiting plan gives the Rejectable level.
c=0 plans are not limited to an AQL plan or a Limiting plan. It is simply the sample size that restricts acceptance of the lot to zero defects in the sample. Many limiting plans are designed to allow for more discrimination in determining the true nature of the lot. But they do not preclude a sample size size that would reject on 1 defect and accept on 0.
The critical descriptive terminology is NOT c=0, but Limiting Quality plan.
The confusion comes from the fact that most c=0 tables are for AQLs which is a more common sampling technique used for contiguous lots...c=0 plans are common for AQL sampling but are not restricted to AQL sampling...
my answer is still yes.
isolated lots are typically inspected with concern that the defect rate of the lot not exceed some level. (refered to as the limiting level, LTPD, RQL, yada yada)
An AQL plan is the Acceptable Quality level
A Limiting plan gives the Rejectable level.
c=0 plans are not limited to an AQL plan or a Limiting plan. It is simply the sample size that restricts acceptance of the lot to zero defects in the sample. Many limiting plans are designed to allow for more discrimination in determining the true nature of the lot. But they do not preclude a sample size size that would reject on 1 defect and accept on 0.
The critical descriptive terminology is NOT c=0, but Limiting Quality plan.
The confusion comes from the fact that most c=0 tables are for AQLs which is a more common sampling technique used for contiguous lots...c=0 plans are common for AQL sampling but are not restricted to AQL sampling...
my answer is still yes.
Similar threads
