Skip to content

MLE for pretest-posttest design

1 message · Shoeayb Qasemi

#
Hi All,

I have data from pre-post design. The necessary assumption of ANCOVA,
homogeniety of slopes, is not satisfied. I tried the following model
as an alternative.
#variables: group(0=control, 1=experimental); time(0=pretest, 1=posttest);
mod1<-  lmer(score ~ group*time + (1+time|id), data=dat1)

My questions are:
 1. Is the model correctly specified?

2. Which effect tests the effectiveness of treatment?

Thanks in advance,

Shoeayb