From: Stan Hilliard shilliard@samplingplans.com
Date: 06 May 2003
Time: 22:54:29
Greetings Vandu,
For attribute sampling plans, AQL describes the percentage of defective items that a lot can contain and still have a high probability of acceptance (of the lot) by the sampling plan.
To explain the AQL of a sampling plan, one has to also consider the probability of acceptance associated with it.
Example: AQL=2.5%, probability of acceptance =Pa=95%
In terms of this example, AQL=2.5% means that if the sampling plan is applied to a lot that truly contains exactly 2.5% defective items, then that lot has a good chance (95%) of being accepted.
I have found that most of the people that specify sampling plans by picking AQL don't understand that if the plan is applied to a lot containing a percentage of defective items much higher than AQL, it (the lot) can also be accepted. For example, at sampling plan with AQL=2.5%, when applied to a lot that ismuch worse, say 5.0% defective, could have a 50% probability of accepting the lot.
This common misconception is explained in my article at:
www.samplingplans.com/aqlprimer.htm