[R-sig-ME] question about emmeans post-hoc analysis

Mengxia Fu m|u804 @end|ng |rom @uck|@ndun|@@c@nz
Tue Jul 16 05:15:54 CEST 2019

Hi list,

I have a question about post-hoc analyses using emmeans function. The model
I am using has four independent variables, pre-test scores, group (4
levels), time (4 levels), and the interaction between group and time. The
dependent variable is post-test scores. I would like to know the
differences between the four groups in terms of their post-test scores at
each time point while controlling for the effects of pre-test scores. My
model is:

*mod<-lmer(posttest scores~pre-test

For the post-hoc analyses, I used emmeans. The R code is:

*emmeans(mod, list(pairwise~group|time), adjust="bonferroni")*

My question is: Do the pair-wise comparisons carried out through emmeans
account for the influence of pre-test scores? Because the omnibus test
shows that the effect of pre-test scores is significant, I would like to
exclude the influence of pre-test scores on pair-wise comparisons.

Thank you in advance for any advice.



	[[alternative HTML version deleted]]

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