|
I am trying to estimate simple contrasts for interaction terms using LMATRIX
in a MANOVA setting
and I keep getting the following error:
"This L matrix is not estimable. Hypothesis tests cannot be computed."
This does not change even if I try a simple UNIANOVA with two factors and
request a contrast for the main effect.
I have been able to obtain similar contrasts with different datasets and it
has worked. This made me think that there was a problem with the dataset.
However, when I ran the same model and contrast (see syntax below) in SAS I
had no problem getting results.
Is there any problem with the syntax below? If not, can someone give me some
hints re how to address this issue?
Thanks,
GLM
q13 q14 q15 q16 q17 q18 q19 q20 q21 BY q2 q3 q4
/METHOD = SSTYPE(3)
/INTERCEPT = INCLUDE
/EMMEANS = TABLES(q3*q4)
/PRINT = TEST(LMATRIX)
/LMATRIX q3*q4 1 -1 -1 1 .
/CRITERIA = ALPHA(.05)
/DESIGN = q2 q3 q4 q2*q3 q2*q4 q3*q4 q2*q3*q4 .
|