P
pangchiaboon
Can anybody please advise me how to determine producer risk? Is there any fix rule or derive from formula? How producer risk correlate with AQL for C=0 sampling plan? Thank you.
D
Duke Okes
It depends on what you mean by DETERMINE producer risk. If you're saying you have a sampling plan and you want to know what the AQL (consumer's risk) and producer's risk are, you can plot the OC curve by using a Poisson distribution.
The producer's risk is the probabilty of rejecting a lot that is equal to or less than the AQL (ACCEPTABLE quality level) for any given sample plan (sample size, n, and accept on number, c). Duke is correct that an OC curve for any given n,c plan will help you determine the 'risks' for that plan. An OC curve will return the probability of accepting a lot at various defect rates for a given n, c.
YOU must what risk level is acceptable to you. The general rule of thumb is a 5% probability (5% risk) of rejecting a lot at the AQL defect rate...or a 95% probability of accepting the lot...
YOU must what risk level is acceptable to you. The general rule of thumb is a 5% probability (5% risk) of rejecting a lot at the AQL defect rate...or a 95% probability of accepting the lot...
P
pangchiaboon
Thanks Duke, Bev D. I am in medical device industry and i have intention to apply c=0 sampling on packaging material incoming inspection. The sampling plan that we are using now has not statistically justified, means we simply pick a number for example 5/100.
I have tried to work backword from C=0 sampling plan. Base on specific range of lot size, i have calculated the probability of acceptance base on sample size through out a range of AQL i.e. 0.10, 0.15 ... 10 to understand the relationship between AQL and producer risk base on poisson distribution. Please refer attachment.
I do not know whether this is the right method to choose AQL. What i know is we need to use AQL figure in sampling plan to calculate the producer risk, and once we know the producer risk we can choose which AQL we prefer. Is this the right way? or my understanding is wrong?
After the study i have an idea to use historical rejection data on these packaging material to calculate ppm. I read one of the thread mentions that 0.065% AQL = 650ppm, means we can base on ppm to determine AQL also. Is this true? Is this a better way because it is base on actual performance?
Appreciate your guidance.
I have tried to work backword from C=0 sampling plan. Base on specific range of lot size, i have calculated the probability of acceptance base on sample size through out a range of AQL i.e. 0.10, 0.15 ... 10 to understand the relationship between AQL and producer risk base on poisson distribution. Please refer attachment.
I do not know whether this is the right method to choose AQL. What i know is we need to use AQL figure in sampling plan to calculate the producer risk, and once we know the producer risk we can choose which AQL we prefer. Is this the right way? or my understanding is wrong?
After the study i have an idea to use historical rejection data on these packaging material to calculate ppm. I read one of the thread mentions that 0.065% AQL = 650ppm, means we can base on ppm to determine AQL also. Is this true? Is this a better way because it is base on actual performance?
Appreciate your guidance.
P
pangchiaboon
Hi anybody can help in this? Thank you.
It's the Fourth of July week in the USA. Holiday and vacation time.
Someone will respond..just be patient with us..
Stijloor, Forum Moderator.
sushant_kulkarni
Involved In Discussions
why only poissions distribution?
why not binomial distribution or other distribution?????
Regards,
SHK
Similar threads
