Why do I need batch size in establishing control plan to certain AQL??

[Adam G]

Hi,

Why AQL table require batch size (N) to determine the required sample size and acceptance number (c) to certain level of acceptable defectiveness (AQL).
In fact, if my process is designed to be ~2% defective i can establish control plan for AQL = 2 independent on Lot size.
Moreover using Excel formula i can use binominal distribution function do get to know probability of acceptance any defectiveness having just sample size and acceptance no.
Can anyone explain when should i use General or Special inspection levels and what are they for?
Thanks

 

[Statistical Steven]

Hi,

Why AQL table require batch size (N) to determine the required sample size and acceptance number (c) to certain level of acceptable defectiveness (AQL).
In fact, if my process is designed to be ~2% defective i can establish control plan for AQL = 2 independent on Lot size.
Moreover using Excel formula i can use binominal distribution function do get to know probability of acceptance any defectiveness having just sample size and acceptance no.
Can anyone explain when should i use General or Special inspection levels and what are they for?
Thanks

You don't! That is the fallacy of the Z1.4 tables. The lot size does not have an impact on the calculations. If you want to use the hypergeometric to get the exact probability, then lot size matters. More over, the AQL values are "estimates" not the exact value for the different sampling plans. I highly recommend using the binomial formula to get the AQL and LQ values for a given sampling plan.

 

[Adam G]

Thanks Steven,

Thats exactly what i was hoping to get. I've been struggling for days to find in google some confirmation. Now i can rest in peace

 

[krolikk]

Batch size is clear now. What about selection of inspection plan type?

The difference in criteria is clear, number of samples taken, anything else?

What more interesting is the difference in consequence of using normal/tightened/reduced and regular and special?

I read sth about discrimination level and sampling risk, but what is behind in reality?

thanks in advance

Lukas

 

[Statistical Steven]

Batch size is clear now. What about selection of inspection plan type?

The difference in criteria is clear, number of samples taken, anything else?

What more interesting is the difference in consequence of using normal/tightened/reduced and regular and special?

I read sth about discrimination level and sampling risk, but what is behind in reality?

thanks in advance

Lukas

The normal/tightened/reduced are based on the switching rules (these vary between the ISO/Z1.4/Japan/BS). The regular versus special are explained in the standard as based on "special" situations where you need to reduce your sample size, typically because of cost or time. Typically you start at General Inspection Level II.