Dear S users,
I'm exploring robust versions of logistic regression; the output I get using
glm( ... family= robust(binomial) ) looks sensible but I haven't found much
documentation on what's actually being done.
Examining the robust(binomial) object there seems to be something in there akin
to M-estimation with a Huber influence function - could someone please clarify
this and (even better) provide a reference that I can cite?
Many thanks
Ken Rice
_____________________________________________________
Kenneth Rice
MRC Biostatistics Unit
Inst. of Public Health
Forvie Site, Robinson Way
Cambridge, CB2 2SR
Tel: (+44) 01223-330-369
|