[R-sig-ME] Questions on migrating from nlme to lme4

Dieter Menne dieter.menne at menne-biomed.de
Fri Jun 22 19:03:34 CEST 2007

Martin Maechler <maechler at ...> writes:

> Hmm, we are getting off-topic here, but I think these referees
> are not quite fit for the 21st century.

Only partially off-topic, because I think it's important that some accepted
solution in the df discussion of lmer should be found. I rarely had problems
with professional statistical referees in Journals like Lancet or B.Journal of
Dermatology; or if so, they had damned goods reasons to complain.

The problem is that you rarely publish in such high-ranked journals with
statistical referees. Normally you get trouble with medical referees who argue
in a complex cross-over design with several nested repeats "why don't you give
simple standard error of mean instead of these incomprehensible stuff?". For
example, at the University Hospital around your corner.

> MANY modern statistical procedures depend on random numbers to
> some extent:
... True, but
> If referees ask researchers to refrain from all such methods, 
> a good journal editor should switch referees,
> or a good author should switch to a better journal  

If you publish in statistical journals, no problem. If you publish statistics in
a medical journal: good luck. They want Michelin-guide style *** tables.


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