I am very new to mixed/multilevel models. I have an experiment where we measured 2 scale variables (varA and varB) in 2 different groups of subjects (control and treatment) at 4 different time points.

I suspect that the relationship between varA and varB is different for the 2 groups (control vs treatment), however it should be roughly consistent across the 4 time points. Researching on the internet I've come to conclusion that I need to model this using a multilevel (mixed effects) model. So, I am interested in the difference between the regression line between varA and varB for the 2 groups, but I want to account for the repeated measures.

What is my experimental design in this case? It's not very clear to me what are the fixed and random effects. Additionally, is it also possible to test if the regression lines are significantly different across the time point? Would that call for a different model? Thanks very much, any help on how I should go about doing this in R/Matlab/SPSS would be greatly appreciated as well!

  • 1
    $\begingroup$ This is a bit open-ended... Broadly speaking "random effects" appear due to the use of a subset of a population, eg. your participant's IDs; "fixed effects" are (otherwise unknown) deterministic components, eg. if someone is in your control group or not. Stating the obviously probably: Go with R's lme4 package of starters, it is not a panacea but it approximates one. :) Check the "sleep deprivation data" (it comes in the sleepstudy dataset of lme4), I think it will help you build some intuitions with the worked examples it has. $\endgroup$ – usεr11852 Nov 3 '13 at 6:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.