[JMP SEA] JMP on Tuesday Session #11 : DOE Tutorial

[#1]

Thank you for your questions during the session. We're sharing the recorded session and answers to the Q&A here for everyone to improve our JMP skills together.

You can also check JMP DOE Introductory Toolkit : https://www.jmp.com/en_us/events/getting-started-with-jmp/doe-intro-kit.html 

Q1: How does JMP simulate the response?  Random numbers from a normal distribution?

Q2: How can I see that simulation algorithm (formula)?

To find out more about the Simulate Responses options, please check out this link: JMP Help: Simulate Responses

Q3: How can blocking be used together in custom design?

To find out how to add blocking factors in Custom Design, please check out this example at: Response Surface Design with Flexible Blocking

Q4: How about replicates and duplicates? When to use?

Replicates are multiple experimental runs with the same factor settings. They are used to increase the accuracy of the estimate and detect smaller effects. Between unit replicates require additional units run at the same settings. Within unit duplicates are multiple measurements taken on the same unit. Within unit duplicates are sometimes referred to as pseudo replicates.

Q5: When we add in more replicates in the No. of runs, which is the condition that the model will choose to have more replicate? Which are the set that will have duplicates, because not all set will be replicates?

Replicates are multiple experimental runs with the same factor settings. You can specify the number of experimental runs that matches your budget in Custom Design. A 10-run design with 1 replicate will have a total of 20 runs and the model will be based on the 20 runs.

Q6: What are the covariate factors? Can you give some examples?

Material properties and humidity are good examples of covariate factors. They cannot be controlled; however, may influence the response.

Q7: What are the other responses we can expect from an experiment apart from yield?

Output, test time and cost are some other common measures for responses.

Q8: What if add 3 factors with different levels? Then how can we design it?

Custom Designer allows combinations of any number of levels for categorical factors and 2 levels for continuous factors. To add the third factor, go to the Factors outline in Custom Design window, click ‘Add Factor’ and pick a factor type and level.

Q9: Why only RSM interactions are there and quadratic terms are not included? Is there a reason for it?

When RSM is selected only the interaction and quadratic terms up to the 2^nd order for all continuous factors are added.

Q10: If we have the data table, why the values are not automatically populated in the custom design window?

If you are referring to a table containing factor information, you can load them automatically using the saved table by clicking the Custom Design red triangle and select ‘Load Factors’.

Q11: Which DoE suitable for screening and development of Dissolution method?

My recommendation is Custom Designs because they are used more often in practice. Custom Designs allow you to create a design based on your goal, budget and problem. Also, they can be used for factor screening, response optimization and mixture problems all within the same experiment.