1

I have searched extensively for solutions but still unsure about what to do. Specifically, I have two groups (one A treatment, one B treatment) that were assessed twice (pre-treatment, post-treatment). And I have at least 10 dependent variables (with missing values, that's why i don't want to use MANOVA since it throws out the entire participant). My goal is to run an omnibus test on these dependent variables and I was instructed by a colleague that lme4 in R is the way to go. But I'm not sure how would I set the model or deal with multiple dependent variables in a single model.

Thanks!

Catherine Tang
  • 11
  • 2
  • I asked a similar question in the past. Have a look on [this](http://stackoverflow.com/q/7831243/707145). – MYaseen208 Mar 25 '17 at 17:13
  • Another [useful link](http://rpubs.com/bbolker/3336). – MYaseen208 Mar 25 '17 at 17:15
  • With only two observations per subject, then I would not use a mixed-effects model. Is your dependent continuous? If so, one approach is to create a new variable , post - pre, and use this as the dependent, also adding pre as an independent. – user20650 Mar 25 '17 at 21:16
  • I am voting to move this to http://stats.stackexchange.com/questions as it is needing stats advice rather than programming. But you should add more details to your question, describing the variable types, number of observations, your hypothesis that you are interested in etc. – user20650 Mar 25 '17 at 21:17

0 Answers0