Cause-and-Effect-Diagrams are often used to visually relate Responses to possible Factors of a process in a brainstorming session (in JMP: Analyze -> Quality and Process -> Diagram).
I propose to extend the Diagram Platform by a Transfer-Tool in order to feed Responses and Factors from the Diagram directly into the DOE Custom Design Platform.
In the transfer protocol the user must define
- the Data/Modeling Type of Factors
- if they are controllable, uncontrolled, covariates or constants (with practical information what these mean, e.g. uncontrolled may be a logged room temperature, covariates might be specimen dimensions that are measured as is, and cannot be requested to meet a specific value. Typically, there will be many constants which are then carried along for documentation)
- if they can easily be changed (Split-Plot)
- if the relationship between a Factor and a Response is assumed linear, quadratic, polynomial, ... and what the -1/+1 levels are, respectively
- what the levels of nominal Factors are (and how many)
- which interactions shall be examined (cross-table with all 2-way interactions ticked by default, those unwanted can be unticked. Practical info what interactions are should be supplied)
- if certain level combinations don't work (constraints)
- additional free text info about Responses and Factors
- ...?
On send to DOE Custom Design the information will be used to fill the respective fields, upon request responses and factors will be saved, and the assumed model will be constructed using the info provided (so users don't necessarily need to concern themselves with the technique of constructing models in JMP - although it's of course a good idea to know how this works).
General information on Responses and Factors, such as Units, Limits, Goals should be supplied in the Transfer-Tool, as well.
In my experience, although we use Cause-and-Effect-Diagrams a lot, we never use them in JMP, because the information in there is not transferrable. From other tools I can't transfer this info into JMP-DOE either, but in JMP there would be a chance to give the Cause-and-Effect-Diagram additional value. Much more of process development and management could take place in JMP itself.