split-split plot with continuous subplot variable


I am trying to analyze data from a split-split-plot design. The sub-plot is a continuous factor and since we suspect a non-linear relationship, the quadratic form needs to be tested as well.

Factors: a-main plot-5 levels
c-sub-subplot-2 levels.

To test the quadratic form of b I am using: b1=b*b

I was thinking about using the following SAS code:

proc glimmix data=x;
class rep a c;
model y=a|b|b1|c;
random rep rep*a rep*a*b;

I am uncertain about the random effects. To test b1 (the quadratic form of b), should I include random effects for b1 as well (rep*a*b1)?
What I would like to know is the correct error to test the quadratic function. Do we need to specify individual error terms for quadratic, cubic etc functions or the linear error term would be sufficient?

I would appreciate any help from anyone with an idea about the correct SAS code.

Last edited: