06-04-2012 07:48 PM
Split plot design with a ordinal whole-plot factor.
I have an experimental design where the whole plot factor (call it A) has five unreplicated ordinal levels. Imagine this as five water tanks each maintained at one of five different temperatures. Within each is a replicated treatment (call it B) with two levels and continuous dependent variable. So instead of the typical split plot with each level of factor A [whole-plot] minimally replicated, I have five unreplicated ordinal levels (temperature) of Factor A each represented by a single tank. So I would like to use a split plot regression approach using factors A and B (similar to that recommended by Jones and Nachstheim) but how do I specify AND interpret the output in Proc Mixed? So this is split plot design with a ordinal whole-plot factor. Any experience with this?
06-05-2012 08:16 AM
No experience, so I'm just spitballing here. If the temperatures differ from tank to tank but are a known continuous value, could you just include temperature as a covariate in the model?
Maybe something like:
model response=B temperature B*temperature;
Comparisons of means for the levels of B can be controlled by using the AT= option in an LSMEANS statement. If you suspect heterogeneity of response due to temperature, then this may get more complicated, but I would think this would be appropriate as a start.