Hi
I have a response variable (response) measured hourly 48 hours (hour) and repeated this again for 5 weeks (week). I would appreciate if you suggest how would you tell this to lme in R with autoregressive first order covariance structure?
The following discusses using correlation structures in nlme. You'd still need to figure out how to specify the nested effects for the model.
http://rcompanion.org/handbook/I_09.html
(search for Define model and conduct analysis of deviance)
Let me see if I understand. You have only two time points and one response variable at each. Details of cov structure are in several places, including the linked pub below, but can you say more about your data so that it is clear what you are trying to model.
Article Multilevel modelling: Beyond the basic applications
Hi Daniel,
Thanks for your response. The same subject (lets say rat) was measured hourly for 24 hours (24 data points). Then this was repeated for 5 weeks (lets say from week 1 to 5). Here both hour and week are repeated measures at the the same experimental unit (rat). Then how would tell model that two repeated measures nested with each other (hour is nested within week) in both random statement (for random slope model) and covariance structure (autoregressive first order)? I would appreciate
Okay, that makes sense. The UCLA web page for the ALDA book (good to have is you are doing lots of longitudinal work) shows this and other R codings. It is at http://www.ats.ucla.edu/stat/examples/alda/
Hi Daniel, Thanks once again for your response. I had a quick look at this book. Mostly, such references have focused on nested random variables rather than nested repeated measures and way to specify this into a model. If you come across into those materials, I would very much appreciate if you let me know about it! Thanks!!
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics