s-news
[Top] [All Lists]

is anova for glmmPQL valid?

To: "S-News (E-mail)" <s-news@lists.biostat.wustl.edu>
Subject: is anova for glmmPQL valid?
From: Rob Campbell <rob@robertcampbell.co.uk>
Date: Tue, 21 Mar 2006 16:00:46 +0000
User-agent: Thunderbird 1.5 (X11/20051201)
Hi,

I know it's not meaningful to compare two glmmPQL objects with anova, but what about passing a single glmmPQL fitted model object to it? Is that ok? It seems to give sensible results.

If it's not then how am I supposed to asses the significance of factors or their interactions with a glmmPQL object. The output of summary showing t-tests of the coefficients is not always what I'm most interested in.

Thanks,
Rob



--
  Rob Campbell   -   Research Student
 Autistic Bacteriophage Research Group
 www.autisticBacteriophage.notlong.com

               Oxford

      www.robertcampbell.co.uk

<Prev in Thread] Current Thread [Next in Thread>
  • is anova for glmmPQL valid?, Rob Campbell <=