[R] Robust estimators for ordered logistic regression

L.Stirton at uea.ac.uk L.Stirton at uea.ac.uk
Mon Aug 13 14:49:16 CEST 2007


I am trying to estimate an ordered logistic regression in R, and I wish to
use robust estimators. Can this be (simply) done? More specifically, can
this be done using polr (in MASS) together with sandwich. I am using R
2.5.1 on Mac OS 10.4.10. If not, is there any other way?

I can fit a model using polr, and then try:

>coeftest(fit8.polr, df=NULL, vcovHC(fit8.polr, type="HC3"))

When I do this R returns the error:

Error in estfun(x) : no applicable error for "estfun"

This is the first time I have tried to use sandwich, so I wouldn't be
surprised if I was doing something dumb--but I can't see what. I have been
working through Achim Zeileis's "Econometric Computing with HC and HAC
Covariance Matrix Estimators".

Any help greatly appreciated.

Lindsay Stirton

More information about the R-help mailing list