cancel
Turn on suggestions
Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.
Showing results for 
Show  only  | Search instead for 
Did you mean: 
New to using JMP? Hit the ground running with the Early User Edition of Discovery Summit. Register now, free of charge.
Register for our Discovery Summit 2024 conference, Oct. 21-24, where you’ll learn, connect, and be inspired.
Choose Language Hide Translation Bar
DoEDorkily123
DoEDorkily123
Level I

Suitability of %RSD as a response?

A week ago (138 views)

Hello,

 

I'm performing DoE analysis where I have %RSD as a potential response. Is %RSD a suitable response factor? It's my understanding that derived values of two or more individual values may not be suitable since they are not in direct response.

 

Does anybody have any experience or thoughts on this?

 

Thank you

0 Kudos
1 ACCEPTED SOLUTION

Accepted Solutions
MRB3855
Super User MRB3855
Super User
Solution

Re: Suitability of %RSD as a response?

Created: Sep 4, 2024 05:39 AM  |  Last Modified: Sep 4, 2024 2:55 AM (135 views)  |  Posted in reply to message from DoEDorkily123 09-04-2024

Hi @DoEDorkily123 : I assume you mean Relative Standard Deviation (100*[Sample SD/Sample Mean]). There is no reason it can't be a response. Generally speaking, any statistic (function of your data) could be a response. That said, RSD will not have a normal distribution; it tends to be skewed right so often a log (natural log) transformation is often used to make it approximately normal so that you can use the most common types of analyses (which assume normality); there is no guarantee that the log transformation will be appropriate (or even necessary), but in my experience it often is.  

View solution in original post

1 REPLY 1
MRB3855
Super User MRB3855
Super User
Solution

Re: Suitability of %RSD as a response?

Created: Sep 4, 2024 05:39 AM  |  Last Modified: Sep 4, 2024 2:55 AM (136 views)  |  Posted in reply to message from DoEDorkily123 09-04-2024

Hi @DoEDorkily123 : I assume you mean Relative Standard Deviation (100*[Sample SD/Sample Mean]). There is no reason it can't be a response. Generally speaking, any statistic (function of your data) could be a response. That said, RSD will not have a normal distribution; it tends to be skewed right so often a log (natural log) transformation is often used to make it approximately normal so that you can use the most common types of analyses (which assume normality); there is no guarantee that the log transformation will be appropriate (or even necessary), but in my experience it often is.