Inquire about the DOE in JMP

Question

Hi :D

I am a student at university and would like to use DOE in my thesis work. My purpose is to study the of the main effect and interactive influence of three/four parameters I used in my experiment.

But I am a little bit confused about the difference between 2 levels and 3 levels factor in the full factorial design, also the use of center point. In my mind, it looks like the center point in the 2-level design is one combination of medium level factors in the 3-level design and repeat that combination many times. Does that mean I will lose many combinations of medium level factors in 2-level design? Or the 3-level design will be more helpful for further analysis?

Thank you for your help.

P_Bartell · Answer

My suggestion is if you don't have access to a DOE course, then get a hold of a copy of "Statistics for Experimenters" by Box, Hunter and Hunter. The answers you seek are there. SAS also offers some great educational experiences in the DOE realm. The beauty of these classes is you learn DOE as a method (which is at the heart of your original question...) AND 'how to' in JMP. Here's a link to the SAS educational resources:

https://support.sas.com/training/us/paths/jmp.html?_ga=2.136921972.590949608.1558359341-144957384.1553798939

Mark_Bailey · Answer

The factorial designs are the oldest. They were first published in 1923. They still work today. Many other methods have been developed since and the latest method is embodied in the JMP custom design.

You are thinking about factor levels and combinations. Instead, think about the nature of each factor and the potential effects that changing these factors will have on your response. Define the factors individually and then define the model in terms that will represent those effects. The custom design method will decide the best combinations for estimating the model.

I don't think that you will be satisfied with a few quick replies in this discussion area. DOE is powerful but with power comes the requirement to acquire the knowledge and skill to use it properly for your purpose. Your thesis is important to you. Be sure to learn about the methods that will help you complete it. I suggest:

* Select Help > Books > Design of Experiments. This guide has a lot of information about DOE, in general, and specific methods, in particular.
* See "Optimal Design of Experiments: A Case Study Approach" by Peter Goos and Bradley Jones.
* Attend a JMP training course for DOE.