Could you please advise whether studentized residuals are meaningful when computed on a robust linear regression model using an M-estimator?

I'd like to use it to detect outliers by doing something like this:

rfit = rlm(y~x, data=d)
pt(rstudent(rfit), df=nrow(d)-3) 

Is this reasonable? I'd be quite happy with a rather crude measure and I'd rather err on the conservative side.

I'm also wondering whether I should run some kind of diagnostic of the general goodness of fit on this robust model before doing this.

  • 1
    $\begingroup$ Have a look at the fit.models function in the fit.models package. This provides a comparison of regression diagnostics of model fits using lm and rlm. $\endgroup$ – Tony Ladson Feb 8 '16 at 2:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.