Re: [R] Logistic regression and robust standard errors

2016-07-01 Thread Achim Zeileis
On Fri, 1 Jul 2016, Faradj Koliev wrote: Dear Achim Zeileis,  Many thanks for your quick and informative answer.  I?m sure that the vcovCL should work, however, I experience some problems.  > coeftest(model, vcov=vcovCL(model, cluster=mydata$ID)) First I got this error:  Error in vcovCL(mod

Re: [R] Logistic regression and robust standard errors

2016-07-01 Thread Faradj Koliev
Dear Achim Zeileis, Many thanks for your quick and informative answer. I’m sure that the vcovCL should work, however, I experience some problems. > coeftest(model, vcov=vcovCL(model, cluster=mydata$ID)) First I got this error: Error in vcovCL(model, cluster = mydata$ID) : length of 'cl

Re: [R] Logistic regression and robust standard errors

2016-07-01 Thread Achim Zeileis
On Fri, 1 Jul 2016, Faradj Koliev wrote: Dear all, I use ?polr? command (library: MASS) to estimate an ordered logistic regression. My model: summary( model<- polr(y ~ x1+x2+x3+x4+x1*x2 ,data=mydata, Hess = TRUE)) But how do I get robust clustered standard errors? I??ve tried coeftest