Hi,
I have looked through the previous postings, because I was sure this question was asked before, but I have not as yet been successful at finding a relevant post. I have a mixed experimental design, with blocks and events, and I would like to model the events with time and dispersion derivatives, but not my blocks. However, as far as I can tell, I have to choose to model the derivatives for the whole model, and not for individual regressors. Is there a way to do this within SPM without creating the time and dispersion derivatives manually and adding them to the design matrix? If so how? Or is there some statistical assumption that I am overlooking here as to why I should not model only SOME of my regressors additionally with time and dispersion derivatives?
Thanks for your help.
-V
|