[R] Robust SE for lrm object

David Winsemius dwinsemius at comcast.net
Sat Mar 6 06:10:37 CET 2010


On Mar 5, 2010, at 11:54 PM, Patrick Shea wrote:

>
> I'm trying to obtain the robust standard errors for a multinomial  
> ordered logit model:
>
> mod6 <- lrm(wdlshea ~   initdesch + concap + capasst + qualrat +  
> terrain,data=full2)
>
> The model is fine but when I try to get the RSE I get an error.
>
> coeftest(mod6, vcov = vcovHAC(mod6))
>
> Error in match.arg(type) :
>  'arg' should be one of “ordinary”, “score”, “score.binary”,   
> “pearson”, “deviance”, “pseudo.dep”, “partial”, ........etc.
>
> I'm a novice R user and am not sure how to address this problem. I  
> have also tried to use alternatives   (zelig, polr) but have had no  
> luck. Any assistance on generating RSE for a multinomial order logit  
> model would be appreciated

Have you loaded the library that contains the vcovHAC function?

(And do you know whether coeftest works with Design/rms objects?)

-- 

David Winsemius, MD
West Hartford, CT



More information about the R-help mailing list