Re: Any good way to deal with non-normal distributed data for Control Limit/Cpk analysis?

[chem engineer <chemical_engr@yahoo.com>]

you will need to do a data transform

---

From: Fred Zhang <ieaggie2009@gmail.com>
To: jmp-l@lists.biostat.wustl.edu
Sent: Tuesday, March 3, 2009 9:46:21 AM
Subject: [jmp-l] Any good way to deal with non-normal distributed data for Control Limit/Cpk analysis?

We have so many different types of non-normal distributed data from the practical scenarioes.

Now we try to determine the control limit, the Cpk, or the meaningful mean and deviation.
So any idea how we can do in JMP to propose good way to achieve the above analysis?

All data are continuous variable type.

Sometimes the data are always falling into limited categories,

or they have a very long rail to the right or the left,

or there are a very few points far away from the others although they are acceptale and reasonable,

etc.

Thanks a lot.