why one of the level of a factor is not included in the prediction profiler?

[#1]

Hi, I am doing analysis of my bioprocess data in which I am trying to screen the effects of factors such as seeding, feeding etc on antibody productivity. For seeding I have three level 1.2, 2 and 4, however, I found that in the prediction profiler, the level 4 is missing. I am wondering why this lease square model don't take level 4 into consideration?  thanks

[#2 · Accepted solution]

I think it is because you used a "Disallow Combinations" constraint when you designed the experiment.  Note that if you move the level of the 2nd factor then you see the available factor levels for the first facts switch

There is a Disallowed Combinations script saved to the data table that has an expression that evaluates as a boolean and that is what is used by the Profiler to determine the factor constraints.  

If JMP detects that in the table, then it will apply those disallowed combinations to the profiler.  

If you delete the Disallowed Combinations script, then the profiler will show all the levels and not use that constraint.  However, the constraint was put in for a reason, so it may not make sense to remove it.

[#3]

Thank you so much for your nice answer. Yes, you are right. I did set disallowed combination. It doesn't make sense in reality that combine the lowest level of seeding density with the highest level of feeding.