[R-sig-ME] Teaching to use lme4

Martin Schmettow schmettow at web.de
Thu Oct 20 23:56:43 CEST 2011


Dear Philipp,

to me that sounds very ambitious. Psychology students are often spoiled by
the prevalent experimental paradigm, where hardly someone is ever analyzing
individual differences and most people only do simple null hypothesis
testing. Especially with the Gaussian model, variance is just seen as a
nuisance factor. 
Recently, I gave a class on generalized linear models to psychology
students. On a few applied research examples I explained to them that other
distribution types have a strict variance-mean relations. In that case,
overdispersion directly interprets as individual differences (more or less).
Maybe this is a viable approach to motivate GLMMs, especially when you are
teaching psychometrics. I didn't go that route but introduced GEE instead. 
Are you aware that the newest version of SPSS comes with GLMM support
(finally)? Wouldn't that be just more convenient? You can teach modern
statistics to psychology students and you can learn them programming. But,
better do it one at a time.

--Martin

> -----Original Message-----
> From: r-sig-mixed-models-bounces at r-project.org [mailto:r-sig-mixed-models-
> bounces at r-project.org] On Behalf Of Philipp Doebler
> Sent: Donnerstag, 20. Oktober 2011 12:01
> To: r-sig-mixed-models at r-project.org
> Subject: [R-sig-ME] Teaching to use lme4
> 
> Hello everyone,
> 
> in around 2 months time I intend to explain to psychology students
(master's
> degree) the use of lme4. My experience shows, that viewing mixed models as
> multilevel models is a good way to motivate the use of random effects
(pupils
> within classes within schools...). My intention is though not only to
teach
> "recipes" that are to be copied later, but to enable students to really
apply mixed
> models, beyond the mutlilevel perspective.
> 
> On the other hand translating any mixed model to a formula requires some
> effort. Many students, at least initially when we start R with simple
things like
> lm(), struggle a lot with the syntax. I do not doubt that my students are
clever,
> but as psychologists they have not been exposed a lot to formal languages,
so I
> would like to ask for advice on how to teach, say, arriving at a formula.
> 
> 
> Best regards, Philipp
> 
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models




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