[R-sig-ME] MixMod lsmeans question

[Alexandra Kuznetsova]

Thank you for the data and explanations! Now I understand the problem. This is a normal behaviour, and there is no error here - proc mixed of SAS program gives the same results. It could be that estimates of population means are significant and the effects of ANOVA table non-significant 
________________________________________
From: guillaume chaumet [guillaumechaumet at gmail.com]
Sent: Tuesday, September 11, 2012 8:43 AM
To: Alexandra Kuznetsova
Cc: r-sig-mixed-models at r-project.org
Subject: Re: [R-sig-ME] MixMod lsmeans question

Here's the data and the code :

aggTc=read.csv("aggTc.csv")

library(lme4)
library(MixMod)

CRT.lmer=lmer(log(CRT)~test*fatigue*wakefulness+ (1|subject),data=aggTc)

totalAnalysis(CRT.lmer,aggTc)

As you could observe, lmer part shows us no interaction effect except fatigue:wakefulness but second "lsmeans" part shows every significant effects...

Guillaume Chaumet




2012/9/10 Alexandra Kuznetsova <alku at imm.dtu.dk<mailto:alku at imm.dtu.dk>>
If for lsmeans table you use lsmeans function, then all the least squares means will be calculated for all effects (for non significant ones as well), if for lsmeans table you use totalAnalysis function, then by default only for significant effects lsmeans will be calculated. Does this answers your question?
_______________________________________________
R-sig-mixed-models at r-project.org<mailto:R-sig-mixed-models at r-project.org> mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models