[R] GEE for three-level hierarchical data?

Bram Van Moorter bram.van.moorter at gmail.com
Tue Dec 7 10:53:20 CET 2010

Are there R packages for Generalized Estimating Equations that provide
the possibility to implement a three-level structure?
My data consist of repeated measures of subjects (AutoRegressive
correlation is expected) within different groups (I expect the
behavior of individuals from one group to be correlated), with
individuals not balanced over groups.  So, I have observations at time
t (level 1) nested within individuals (level 2) and individuals nested
within groups (level 3).  The implementation of a two-level structure
is rather straightforward in for instance "geepack".  At first, I
thought that by providing a user-defined correlation structure in
“geepack”, I could create such a 3-level structure.  However, after
further investigation (i.e. the single hierarchy id-vector) it seems
to me that this is not the way to go.
Are there suggestions for how to implement such a three-level structure?
Best regards,

Bram Van Moorter
Centre for Conservation Biology (NTNU),
Norwegian Institute for Nature Research (NINA)
Trondheim (Norway)
email:  Bram.Van.Moorter at gmail.com
website: http://ase-research.org/moorter
phone: +47 73596060

More information about the R-help mailing list