[R-sig-eco] AIC / BIC in lmer and glmer

[Chris Howden]

Better to be safe than sorry.

And I for one am glad u mentioned it. It's they type of knowledge one
might not pick up when using R to fit a mixed model for the first time.

It's certainly something I'll be keeping in mind should I ever use R for
mixed models!!!

Chris Howden
Founding Partner
Tricky Solutions
Tricky Solutions 4 Tricky Problems
Evidence Based Strategic Development, IP development, Data Analysis,
Modelling, and Training
(mobile) 0410 689 945
(fax / office) (+618) 8952 7878
chris at trickysolutions.com.au

-----Original Message-----
From: r-sig-ecology-bounces at r-project.org
[mailto:r-sig-ecology-bounces at r-project.org] On Behalf Of Philip Dixon
Sent: Friday, 6 August 2010 7:32 AM
To: r-sig-ecology at r-project.org
Subject: [R-sig-eco] AIC / BIC in lmer and glmer

My apologies to the list.  My e-mail this morning about interpreting AIC
in lmer() was off-target.  I did not notice that Chris's call to lmer()
included family = binomial, which means the actual estimation is done by
glmer().  glmer() ignores the REML= argument.  glmer() only does
approximate ML.  In a generalized linear model, AIC is reasonable to
compare models with different fixed effects.  My caution is appropriate
only for models with gaussian errors.

I get the daily burst from the list.  If others have already pointed out
the difference between glmer() and lmer(), thank you!

Philip Dixon

_______________________________________________
R-sig-ecology mailing list
R-sig-ecology at r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-ecology