Define factors and responses

[#1]

Hello! 

I am a bit confused about how to define the responses for my experiments appropriately. I want to see how the processing factors influence the polymer crystallinity, but then the crystallinity will affect the material's mechanical strength. So in this case, should I consider both the polymer crystallinity and mechanical strength as the responses? Or the crystallinity is a factor/response?  I am confused...Hope someone could help me resolve my issue. Thanks a lot!

[#2]

In DOE work any characteristic of the output of the experiment is eligible to be a response. Try thinking of the output as something of value to someone or something. Now, what are the critical to function characteristics of that output from the 'someone' or 'something's' points of view? Those characteristics should be on your list of responses.

[#3 · Accepted solution]

Hi @DendrogramSteer,

From what I understand from your experimental setup, it seems you're trying to find the influence of process factors, and maybe raw materials/monomers choice/mixtures on polymer crystallinity.
Can you freely choose polymer crystallinity or is it a consequence of the process factors levels/choice ? If it's the first option, you might use it as a factor, if it's the second option, it sounds more like a response.

Since polymer crystallinity seems to be correlated with the mechanical strength (and is influenced by the choice/levels of your factors), it seems that both polymer crystallinity and mechanical strength are your responses.

Depending on your objectives, you may choose one of them or both.
Depending on the level of informations both responses may provide, the way to measure/collect information and their measurement accuracy/capability, you may choose one as your primary response, and analyze later the correlation between the two responses, and link it back to your factors thanks to your DoE. For example, I'm not sure how "detailed" polymer crystallinity can be measured (how can you characterize it ? Microscopy, chromatography, mass spectroscopy, ... ?), but mechanical strength seems to be a quite straightforward and numerical response, so it may be easier to use it in your design, and then analyze the link between mechanical strength, polymer crystallinity and your factors thanks to your DoE's model.

I hope this first answer will help you,

[#4 · Accepted solution]

I would treat both as responses.  Use Multivariate Methods to assess the correlation between the 2 responses.  My advice for experimentation is once your have created the experimental units (and spent the resources to create them), you should measure every possible response.