07-18-2014 01:29 PM
I went through 2 different books and they are not too clear because they are testing other things (slopes,etc) at the same time. How different is the Proc Mixed code from a complete block design vs an incomplete block design?
07-18-2014 01:36 PM
I suppose the answer depends on how you code the two different analyses ... this is a very vague and undefined question
The answer to "how different" also depends on whatever you choose as a scale of differences, and I am not familiar with such a scale
07-18-2014 01:57 PM
Thank you Paige;
Here is how I would code as a complete block design for this specific experiment.
PROC mixed data=growth ;
CLASS Trt block;
MODEL response = Trt;
How would one code if it was an incomplete block design?
07-21-2014 09:40 AM
The code is exactly the same. That is the beauty of mixed model procedures like MIXED. Anyone interested in mixed models (i.e., models with multiple random effects) should read SAS for Mixed Models, 2nd edition (2006).