Non-normal distributions in the Process Capability and Process Screening PlatformsHi all,
I am writing a script to aid my team in analyzing process capabilities for many properties across a variety of products. It seems like the Process Capability and Process Screening platforms are the ideal way to do that, but I'm struggling to get those platforms to handle non-normal distributions. When I assign Distribution or Process Capability Distribution properties to a column, the platforms either ignore them or try to use a bizzare (Johnson?) distribution. I'm using JMP 16 but could upgrade to 17 if it makes a difference.
- What distributions do the Process Capability and Process Screening platforms handle? The documentation mentions that they don't handle as many types as the Distribution platform but doesn't say which ones they do handle. I'm a simple person, I just use normal, log-normal, and Weibull.
- Should I assign distribution types to the "Distribution" property or the "Process Capability Distribution" property? And why are there two properties for the same thing?
- Will the Process Screening platform use either the distribution or process capability distribution?
Thanks!
Automation and ScriptingQuality and Process EngineeringRe: Non-normal distributions in the Process Capability and Process Screening PlatformsProcess capability seems to manage just fine with the property defined in Process Capability Distribution property

It will get added when you add the column to the platform selections so you might have to take this into account when scripting

Column which has no distributions defined

For Process Screening I'm not sure if you can define the distribution and it seems to have it's own column properties

Re: Non-normal distributions in the Process Capability and Process Screening PlatformsIn JMP 18 a warning is being given in Process Screening that now tells you if a distribution other than normal is specified in the column properties. Still doesn't calculate based on that column property, but at least tells you.
