[R-sig-ME] follow-up (posthoc) analyses from an LMER model

[Tobias Verbeke]

Hi Antoine,

> I'm wondering how to do follow-ups for an LMER model (e.g., Y ~ A * B
> + (1|Subject) + (1|Item) )
> 
> If you fit a model using the function aov(), naturally without the
> random effects terms, you can use the TukeyHDS() function to do
> multiple comparisons or the pairwise.t.test() function. Unfortunately,
> these two functions can't be used to do posthocs analyses from an LMER
> model (at least to the extent of my knowledge).

Have a look at the multcomp package. It knows how
to deal with lmer objects (and has lmer examples
in the package vignettes).

http://cran.r-project.org/web/packages/multcomp/index.html

HTH,
Tobias