[R-sig-ME] [R] understanding I() in lmer formula

Don Cohen don-r-help at isis.cs3-inc.com
Sat Jun 17 20:59:15 CEST 2017

 > >  vignette("lmer",package="lme4")
 > > gives me
 > >  Warning message:
 > >  vignette 'lmer' not found
 > That's surprising ... what's packageVersion("lme4") ?
> packageVersion("lme4")
[1] '1.1.14'
[actually I'm replacing left and right non-ascii quotes above]

 > Pinheiro and Bates (2000),
I'll look, but it seems that for something that old, even if it does 
contain these details, they are likely to be out of date.
I noticed http://lme4.r-forge.r-project.org/lMMwR/lrgprt.pdf
which is from 2010 says on p.15
 The vector u is available in fm01ML at re.  The vector Beta and the model matrix 
 X are in fm01ML at fe. 
and this seems no longer true - I get
  Error: no slot of name "fe" for this object of class "lmerMod" 

 >   Yes.    (So in particular this model are 3*(3+1)/2=6 parameters
 > (equivalent to s^2{1}, s^2(x),s^2(y), cov(1,x), cov(1,y), cov(x,y))

That might explain why changing (x+y+z+u+v+w||g) to single | seems to
run a lot longer (I interrupted it so I don't know how much longer).
Do you have an estimate of how number of terms affects run time?

If a paper describes various things being used as random effects
would you assume | or || ?
Perhaps it depends on the number of effects?  The example I sent
contained quite a few, and the original from which that was extracted
contained over a dozen.

More information about the R-sig-mixed-models mailing list