[Sugestion] MMRM (GLS) for Afex: Kenward-Roger and Satterthwaite for free :)

[Generalized]

Dear Authors,

Recently, a new package for MMRM models (used widely in clinical trials) was born. It employs the GLS estimation to obtain marginal model with lots of residual covariance structures and both KR and S. degrees of freedom.

I thought you might find it useful, just to explore a new calculation engine. It's always good to know alternatives! It's quite fast and powerful. In case you were interested, please find: https://github.com/openpharma/mmrm

[singmann]

This looks pretty interesting. However, I noticed that you cannot use it to estimate models in case you have replicates at the level of the grouping factor and time point which is pretty much the type of data you need for lme4. So it seems to be targetting a different type of data then what mixed models are usually used for. Do you have any relevant references?