Incorporate Time lag in DoE

BivariateKoala1

How should I go ahead with my experiments (interpretation) when I know before hand that my response (continuous) will have a time delay.

How do I incorporate time delay in interpreting the results of my experiments?

 

Thanks,

Ashish

Victor_G

Hi @BivariateKoala1,

 

Welcome in the Community !

 

I'm not sure to understand the situation about the time delay :

• Is the time delay for the response the same, no matter the experimental combinations of factors ? Or is it dependent on the combination of factors ?
• Can you record the response from initial time T0 to the time where you have a change in the response ?

Depending on your answers, here are some suggestions :

• You could create a time-robust DoE, by taking into account the possible effect of time/order in the design. Here is a similar topic where this type of DoE has been used : Covariates in defined order in custom design 
  With this kind of design, you balance the levels of factors accross time, to avoid confounding/aliasing between time or experiments order and any of the studied factors.
• If you can record the response across time, you could maybe transform the response into a time-independent responses, or use the functional response and analyze it through different platforms, like Functional Data Explorer in JMP Pro to directly create a functional model for your response, or using Fit Curve/Fit Y by X platforms to extract the interesting coefficients of your curves and analyze how the change of your factors impact these coefficients.
  You might be interested in closely related topics : DoE design for a chemical reaction with time point data 
  How to obtain the optimum formulation 

 

A combination of these two techniques (if possible) could help you mitigate the effects of time in your design, as well as improving the understanding of the link between your functional response and your factors.

 

Hope this first discussion starter might help you,