Factor within factor

Gil

Hello,

 

I want to test the effect of 3 categorical factors on 5 lots of donor cells and 5 lots of patient cells. The question I'm trying to answer is what the effect on the donor cells as whole (not on the individual lots) and the same for the patient cells. In addition,if possible, I want to set blocking in order to split the experiment  to few days but at the same time refrain from using a lot in more then one day (cost perspective). Can you suggest a way to set up the DoE?

 

Thank you,

 

Gil 

Mark_Bailey

First of all, I recommend using Custom Design in JMP. It is the most flexible method of designing an experiment. It also reproduces most of the classical designs when they are optimal, but it often produces a better design.

 

It sounds like you have four categorical factors. There are three factors that you mentioned that are associated with the treatment and then one for type of cell (donor, patient). Does that make sense or you do not want to model/test this aspect? Do you see lot as a replicate or as a block? If you intend to use only a few days, then it might not be worth blocking on day because you will have very few samples of day to estimate the random effect.