Re: Likelihood

Chanda Nimbkar <cnimbka2_at_UNE.EDU.AU>
Sent by: ASReml users discussion group <ASREML-L_at_AGRIC.NSW.GOV.AU>
13/01/2005 02:54 PM
Please respond to ASReml users discussion group

        To: ASREML-L_at_AGRIC.NSW.GOV.AU
        cc:
        Subject: Likelihood

Hi ASRemlers,

Shouldn't log likelihood only increase (significantly or otherwise) when
an
effect (fixed or random) is added to a model?
However, I have found it to decrease with the addition of an effect, in
some cases. For example, if a non-significant effect is added or if an
effect confounded with another already present in the model, is added. Is
this happening because ASReml doesn't necessarily find the highest
likelihood every time? Or maybe this is why LogL can be used for a LRT to
test significance of fixed effects using the same random model (Lynch and
Walsh- "Genetics and analysis of quantitative traits", p.860).
****you are not correct i am afraid. the logl you refer to here is the
Residual log likelihood (REML) and it changes whenever the fixed effects
model is changed and cannot be used for testing the significance of a
fixed effect. There are many references about this which I can supply if
you like. If you add another variance parameter, then the REML logl will
increase that is all that can be said, but you cannot compare REML logl's
with different fixed effects models

Grateful for your comments.

Chanda Nimbkar

Ms. Chanda Nimbkar
Ph.D. student
Animal Science (Room 35, Postgrad Annexe)
School of Rural Science and Agriculture
University of New England
Armidale, NSW 2351
Australia

Phone:Work +61-2-6773-3008
          Home +61-2-6772-0809
Fax +61-2-6773-3922

This message is intended for the addressee named and may contain
confidential information. If you are not the intended recipient or
received it in error, please delete the message and notify sender. Views
expressed are those of the individual sender and are not necessarily the
views of their organisation.
Received on Wed Jan 13 2005 - 15:21:58 EST