[R] Help with problem - multilevel model?

Crabb, David David.Crabb.1 at city.ac.uk
Thu Jan 18 15:15:45 CET 2007

I have what is probably a very simple problem but I would be very
grateful to the list for any help or pointers to implement a solution in

We have two types of measurements on the eye that we have collected in
62 patients at 5 fixed time points during a clinical visit over an
office day. We want to establish if there is an association between the
measurements. Obviously it would be wrong to generate a simple
scatterplot of all 62*5 observations and consider something like a
simple correlation coefficient. What I really need is to set up a
2-level model with measurements nested within patients and use, perhaps,
one variable as a response and the other as a predictor. I would also
like to add in time of day (probably as a categorical) to check this
isn't affecting things. I haven't used multi-level modelling in R and
wonder how I can simply go about fitting this model and looking at the
significance of the estimates in R. I also have a plot in my mind - 62
little regression lines (n=5) and think that this might be

Any help or pointers will be gratefully received - Many Thanks.

Dr. David Crabb
Department of Optometry and Visual Science,
City University, Northampton Square, London EC1V OHB
Tel: 44 207 040 0191   d.crabb at city.ac.uk

More information about the R-help mailing list