RE: st: How do I test that two subsample have different coefficient of variation?

["Verkuilen, Jay" <JVerkuilen@gc.cuny.edu>]

One other point:

Many seemingly ratio scale variables are actually log-interval, e.g., gas mileage, which can be equally well represented as miles per gallon or gallons per mile. 

There is a very nice FAQ on this on the SAS web page: ftp://ftp.sas.com/pub/neural/measurement.html



-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Verkuilen, Jay
Sent: Thursday, July 10, 2008 5:42 PM
To: statalist@hsphsun2.harvard.edu
Subject: RE: st: How do I test that two subsample have different coefficient of variation?

Peter Lachenbruch wrote:

>Perhaps a na�ve approach would be to bootstrap the cv with a few thousand reps.  An obvious problem arises if the mean is near 0.  >I suspect one should restrict this to positive random variables, but I have no theory to go on.

Seems right to me. CV is meaningful for a ratio scale variable (e.g., mass, time, distance) but not an interval one (e.g,. Celsius temperature, utility). While it is, of course, possible to define a signed ratio scale, the vast majority of them are non-negative. 



*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/