testing contrasts in ASREML

Dear ASREML-users,

I have a question about the best way to fit contrasts and test for their
significance.

The data set contains varieties with contrasting effects, for example tall vs
short.
I have fitted the following model where
tvs = dummy variable for contrasting tall versus short
gen = genotypes 1 to 96
and the usual design variables rep, column and row

yield ~ mu mv tvs !r gen rep column,
!f lin(row)
1 2 0
12 column AR 0.2
16 row AR 0.4
predict tvs

From the predict statement I can determine the effects of each level in the
contrast and the SED. Is this what I should use to test the significance of the
effects or can I use a Wald test?
I expected the fixed effects (with accompanying Wald statistics) to appear
under Analysis of Variance in the *.asr file but they don't. Is there are
reason for that?

Thanks
Ky
Received on Tue Nov 02 2004 - 12:17:32 EST