I am currently writing a paper comparing various packages in
different programs, and how they handle GEE. I was wondering if you
could help me out a bit.
Beyond the options included in each package, I've noticed a variance
in the various estimates and standard errors for most working
correlation options between packages. The most is observed in
Unstructured, and none in independence obviously. From what I've been
able to gather the difference is coming from how the working
correlations are calculated. I'm betting it's coming from the
residuals being used. SAS for instance uses the Pearson Residuals.
I was wondering if you could tell me how geepack or gee calculates
correlation.
Could you also include full citation for yourself so I can properly
cite you,
Thank you Very Much,
Cameron Redsell-Montgomerie
MSc Statistics
University of Guelph
MACN rm:501
[EMAIL PROTECTED]
[[alternative HTML version deleted]]
_______________________________________________
R-SIG-Mac mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-mac
