Include other distributions in the calculation of cpk (not just the normal distribution)

[guillaumebugnon]

What inspired this wish list request? Today, Process screening plateform assume only normal distribution.

What is the improvement you would like to see? Integrate all others calculation (percentiles) for non normal distribution.

Why is this idea important? To have a better plateform.

[hogi]

agree:
take the center part of a distribution to fit sigma - and then extrapolate to 6 sigma
-> dangerous.

on the other hand: Process Capability provides the option to use non-normal distributions.
Maybe, for Process Screening, there is an inherent assumption that nothing more than a simple sigma can be derived from the short term variations?
So: intentionally not possible?

[mia_stephens]

Thank you for the request @guillaumebugnon . Have added to a previous request for consideration.

[guillaumebugnon]

Thank you @mia_stephens 

[Valerie_Nedbal]

@guillaumebugnon Hello Guillaume, for your information, in JMP 19 the platform will supports Exponential, Gamma, Weibull, and Log-normal, though not the Capability measures that the Process Capability Platform supports. In addition, it will support count data.

[guillaumebugnon]

Hi @Valerie_Nedbal, nice to read that.