[R-sig-ME] Help file inconsistency for VarCorr in lme4?

[Ben Bolker]

On 13-07-05 09:20 AM, S Ellison wrote:
>> This does look like a documentation typo.  Will be fixed in lme4.0
>> on R-forge; we probably won't release another version of lme4 to
>> CRAN before the 1.0 release.
> 
> No problem; the present format is reasonably easily deciphered,
> especially if you have an older version's help file about.
> 
> An aside, though; I found it because I was trying to get numbers out
> of VarCorr for further use (specifically, I needed the SDs rather
> than variances), and made sure I was getting the right data by
> comparing with the summary output. Which leads me to a suggestion.
> All the numbers (at least for my model) are packaged much more tidily
> in the formatted table of random effects in summary() for lmer
> objects. That table, though, is apparently in character format.
> Wouldn't it be more useful to create that as a numerical data frame
> and use the default methods for data frame output to display it, or
> to format at print time? That way the summary object would hold the
> numbers, and we could get the random effect variances or sd's easily
> from the summary object..

  Maybe.  I'd say we (or at I at least) am still thinking about the best
way to return var-corr information in the most useful/general way
possible.  I had some ideas for this (see vcconv.R in the source code if
you're interested), but I don't think they're going to make it in for
release 1.0 ... the main thing about exporting the var-corr info (the
reason it's done as character()) is that for more complex structures
there are often empty correlation cells ...  Probably the quickest
solution to this would be to write an as.data.frame.VarCorr.merMod
(yuck!) method based on converting lme4::formatVC() output to numeric ...

  If you want to post this as an issue on github that would be fine ...


> 
> Steve E
> 
> 
> ******************************************************************* 
> This email and any attachments are confidential. Any u...{{dropped:9}}