Is anyone aware of a way to calculate log-likelihood for GEE models?
No. GEE fitting is based on quasi-likelihood.
However, it is possible to derive AIC-like measures based on the quasi-likelihood. Lebreton et al (1992) suggested a simple adjustment in the GLM case, i.e. when using family=quasibinomal or quasipoisson. For GEE models, Pan (2001) has introduced QIC. None of these measures are implemented in R or in any add-on package as far as I know.
Henric
______________________________________________ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
