[R] a ratio-variable predictor

[Greg Snow]

A good article to read before using ratios is:

Kronmal, RA. Spurious Correlation and the Fallacy of the Ratio Standard Revisited. Journal of the Royal Statistical Society. Series A, Vol 156, No. 3 (1993), 379-392.

It shows some of the traps that misusing ratios can lead to along with strategies for avoiding those traps.

Hope this helps,

--
Gregory (Greg) L. Snow Ph.D.
Statistical Data Center
Intermountain Healthcare
greg.snow at imail.org
801.408.8111


> -----Original Message-----
> From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-
> project.org] On Behalf Of Jeroen Ooms
> Sent: Thursday, December 18, 2008 3:03 AM
> To: r-help at r-project.org
> Subject: [R] a ratio-variable predictor
>
>
> Maybe this is more of a statistical question than an R question, but I
> am
> going to ask it anyway :) Cortisol and Testosteron are known to
> interact in
> the body, and some literature suggest that especially the ratio between
> the
> two is a good predictor. So I want to add the ratio predictor
> (y~cort/test)
> to a glm, and later a multilevel model, however I have no experience
> with
> ratio variables.
>
> Intuitively, I would think that a ratio is very similar to an
> interaction
> effect, only with an inverse scale for the second term (cort* 1/test).
> However, I found that even more than with normal interaction effects,
> the
> scaling of the variables becomes important. Furthermore there is the
> obvious
> problem (which is not too big of a problem in my case), of values that
> are
> close to zero on the second term.
>
> So what is an appropriate way to incorporate a ratio-predictor? Any
> tips on
> R procedure/package that I could use, or any other experiences with
> ratio-predictors are welcome.
> --
> View this message in context: http://www.nabble.com/a-ratio-variable-
> predictor-tp21069805p21069805.html
> Sent from the R help mailing list archive at Nabble.com.
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-
> guide.html
> and provide commented, minimal, self-contained, reproducible code.