[R-sig-ME] Order of terms for random slopes
thierry@onkelinx @ending from inbo@be
Thu Aug 30 09:14:36 CEST 2018
IMHO you shouldn't use an overfitted model for didatic purposes. Teach
students that you need a sufficiently large data set depending on the
complexity of the model.
ir. Thierry Onkelinx
Statisticus / Statistician
Vlaamse Overheid / Government of Flanders
INSTITUUT VOOR NATUUR- EN BOSONDERZOEK / RESEARCH INSTITUTE FOR NATURE AND
Team Biometrie & Kwaliteitszorg / Team Biometrics & Quality Assurance
thierry.onkelinx using inbo.be
Havenlaan 88 bus 73, 1000 Brussel
To call in the statistician after the experiment is done may be no more
than asking him to perform a post-mortem examination: he may be able to say
what the experiment died of. ~ Sir Ronald Aylmer Fisher
The plural of anecdote is not data. ~ Roger Brinner
The combination of some data and an aching desire for an answer does not
ensure that a reasonable answer can be extracted from a given body of data.
~ John Tukey
2018-08-29 20:51 GMT+02:00 Stefan Th. Gries <stgries using gmail.com>:
> > Thanks. This is a known issue: https://github.com/lme4/lme4/issues/449
> Ohh, ok, I had googled a bit on 'order of terms', 'random effects'
> etc. but hadn't come across this, sorry.
> > - it's not terribly surprising that a model with 11 parameters fitted to
> 48 observations is numerically unstable ...
> Absolutely, the example is from a workshop and was used only for
> didactic purposes, and ...
> > there don't seem to be any _substantive_ differences in the estimate ...
> ... yes, we only wanted to make sure there wasn't something
> superobvious but important we had missed.
> Thanks for the quick feedback!
> R-sig-mixed-models using r-project.org mailing list
[[alternative HTML version deleted]]
More information about the R-sig-mixed-models