[R-sig-ME] Predictions from lme and lmer models

[Patrick Connolly]

So just to check that I've understood this correctly, population-level
predictions exclude random coefficients, but the coefficients it does
use were themselves calculated concurrently with those random term/s?
(As distinct from a model with only fixed terms.)

On Sat, 05-Jan-2019 at 07:00PM -0500, Ben Bolker wrote:

|> 
|>   See the "re.form" argument ... re.form=~0 or re.form=NA gives
|> population-level predictions (level-0). The interface has changed a bit
|> because lme4 can handle more complex models than nlme (specifically, it
|> can fit non-nested models, so "which random effects are included" is
|> potentially a more complex question than "which levels are included").
|> 
|> On 2019-01-05 4:29 p.m., Patrick Connolly wrote:
|> > 
|> > 
|> > The call to predict with lme models has an argument 'level' which, I
|> > understand, produces population predictions for level = 0 and group
|> > specific predictions for level = 1.
|> > 
|> > There doesn't seem to be a similar argument for lmer (lme4 package)
|> > models.  Is it just a matter of specifying the appropriate newdata
|> > data frame to achieve the same end?  Or is there a more direct way?
|> > 
|> > TIA
|> >
|> 
|> _______________________________________________
|> R-sig-mixed-models using r-project.org mailing list
|> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models

-- 
~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.   
   ___    Patrick Connolly   
 {~._.~}                   Great minds discuss ideas    
 _( Y )_  	         Average minds discuss events 
(:_~*~_:)                  Small minds discuss people  
 (_)-(_)  	                      ..... Eleanor Roosevelt
	  
~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.