Dear colleagues,
I'm analysing a cluster-randomized controlled trial with students nested within classes and within schools (three levels). Students have been measured three times within one year (continuous measurements).
I was wondering if I should model time as a fixed effect (as a categorical variable plus the interaction between intervention#time) or as a fourth level in the model.
Many thanks in advance,
Sebastián
Time is nested in individuals and schools. Experiment allows for modeling time wothout baseline-centering scores.
Hope this helps
Matt
Yes - you should include the time points as a fourth level in order to capture the dependency between repeated observations. This won't prevent you from adding fixed effects for time also.
Have a look at this straightforward example
http://www.bristol.ac.uk/media-library/sites/cmm/migrated/documents/repeated-measures.pdf
and Module 15 of this free online course
https://www.cmm.bris.ac.uk/lemma/
In my book on Multilevel synthesis (Springer, 2007) you have a whole chapter on Multilevel event-history analysis, which will show you how to introduce time in a multilevel analysis, with many examples, figures and maps. I hope that this will be helpful for you
- Badges
- Science topic
- Similar topics
- Indexes