Zero Acceptance Sampling - Is anyone is actually using this?

[datigerdad]

If you have to sample - this is the way to go - I would think. But do you know if anyone is actually using this? There is a Plan devleped by Nicholas Squegila (ISBN 0-87389-305-0) that has been in print since 1996. Does anyone use it? What is the statistacal basis for it?

Dad

 

[Marc]

I can't answer your question - I've never read it, but also see this brief discussion thread: Zero Acceptance Sampling By N.L. Squeglia

Also see these discussion threads: Zero Acceptance Sampling

 

[e006823]

If you have to sample - this is the way to go - I would think. But do you know if anyone is actually using this? There is a Plan devleped by Nicholas Squegila (ISBN 0-87389-305-0) that has been in print since 1996. Does anyone use it? What is the statistacal basis for it?

Dad

We’re using the C = 0 plans in Squeglia’s book. The intent of the plans is to give equal or greater consumer protection with less inspection then the corresponding Mil-Std-105 plans. I actually prefer the plans in Mil-Std-1916, mostly due to the switching rules. I’m intend to compare the Mil-Std-1916 plans and Squeglia’s in the near future in order to convince program management to (hopefully) switch to Mil-Std-1916. I’ve attached some OC curves for our C=0 and the equivalent Z1.4 AQL’s that we typically use so that you can compare the plans performance. As you can see the C=0 plans consumer protection (LTPD, P(a) .10) is better for a smaller sample size. In theory for the plan to be effective you either sort defective lots or send them back to the supplier to sort. Due to production needs the reality is different, we either sort or disposition “Use As Is

 

[fuzzy]

Really do and have

Used them for a number of years in the automotive industry at both QS and ISO certified businesses. Using Squeligia's chart right now for Receiving Inspection at my current ISO certified location. Never had much exposure to Mil-spec stuff so that scheme has always seemed foreign to me. I'm no statistician, just a pragmatic user...

 

[Howard Atkins]

We have been using it for years.
The big thing about it is the sample number. I have never believed that companies have really been sampling as per Mil/Z and this standard gives a realistic number to check.
The idea that you can discover bad parts and accept the lot is an impossible concept to explain to staff

 

[Jim Wynne]

We have been using it for years.
The big thing about it is the sample number. I have never believed that companies have really been sampling as per Mil/Z and this standard gives a realistic number to check.
The idea that you can discover bad parts and accept the lot is an impossible concept to explain to staff

The funny thing about sampling is that everyone does it, but samples are almost never randomly selected (i.e., with every member of the population having an equal chance of being selected). This means that the sampling plan that so many people fret over is out the window right from the git.