[R] package relimp

David Firth david.firth at nuffield.oxford.ac.uk
Wed Jul 10 11:37:48 CEST 2002

Christian, no, that's not correct.

The "sd ratio" is a ratio, so if the numerator exceeds the denominator the 
ratio exceeds one; and vice-versa.

The "log(sd.ratio)" is the natural logarithm of that ratio.  It is 
positive when the numerator exceeds the denominator.

In the output you sent, it is not possible to be conclusive: the estimated 
ratio is too close to 1.

I hope that's clearer.  The documentation for relimp does need improving, 
but the improvements I have in hand would not have answered this for you.  
Did you look at the article by Silber, Rosenbaum and Ross (cited in the 


On Tuesday, July 9, 2002, at 16:02 , Christian Schulz wrote:

> Hi,
> i'm newbie for this, but it's very interesting, but how i have to 
> interpret the results
> if i get i.e. this results ?
> Is it correct -
> if the "Ratio of effect sd"  is positiv than the Numerator effects are 
> bigger , and the negative case vice-versa ?
> Ratio of effect standard deviations: 0.954
> Log(sd ratio):                 -0.047   (se 0.828)
> Approximate 95% confidence interval for log(sd ratio): (-1.669,1.575)
> Approximate 95% confidence interval for sd ratio:      (0.188,4.833)
> Thanks for advance & perhaps a link/pdf   but i find nothing related with 
> google.
> regards,Christian
> -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
> .-.-.-
> r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-
> FAQ.html
> Send "info", "help", or "[un]subscribe"
> (in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
> _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
> ._._._

r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch

More information about the R-help mailing list