setting variance components to be equal

Dear all,

I would like to constrain two of my variance components in the model to optimise LogL assuming they are equal (doesn't matter what value they take). How does one code this in ASreml?

Example: Two populations and a bivariate model of the same trait.
Question: do the two pop differ signf from each other in Va?

example code:
T1 T2 ~ Fixed !r at(Trait,1).animal at(Trait,2).animal

1 1 2

at(Trait,1).animal 2
1 2 US
5
animal

at(Trait,2).animal 2
1 2 US
5
animal

How could I specify that the two populations should have equal Va without actually giving them a number? I.e is it possible to say that the two var comp should be the same, get the LogL and compare to a unconstrained model. If no better fit then that suggests that Va are not signf different in the two pop.

Thanks,
Arild

-------------------- m2f --------------------

Sent using Mail2Forum (http://www.mail2forum.com).

Read this topic online here:
http://www.vsni.co.uk/forum/viewtopic.php?p=373#373

-------------------- m2f --------------------
Received on Fri Feb 05 2009 - 19:26:15 EST