[R] predicted means on mixed models
Francisco J. Zagmutt Vergara
fzagmutt at hotmail.com
Wed Apr 23 21:27:54 CEST 2003
I am trying to move from SAS to R and there is a couple of things that I
have not been able to obtain from the nlme3 module (mixed models)
I want to compare the differences between different levels of a categorical
variable in presence of a non-ignorable interaction compromising this
variable. To do this I SAS gives you the possibility to extract predicted
population margins (they call them least squares means or LS means) for each
level of a categorical fixed effect in the model, so these values can then
be used to perform multiple comparisons and look for meaningful differences
between levels of a categorical fixed effect. I guess that in R I could do
this manually by using the augPred() function for each possible combination
and then performing multiple comparisons over these values but I just want
to know if there is and already built-in function to do this. Also In SAS
exists an option to "slice" the LSMeans output so you can obtain
separated outputs for each level of the variable (test of simple effects).
This is very handy when it comes to evaluate levels of nested
and/or interaction terms. Is there sometihing similar available in R?
Thanks for your help!
More information about the R-help