On 15 Sep 2003 14:10:35 GMT, Hiu Chung Law <[EMAIL PROTECTED]> wrote:

> In sci.stat.math Bruce Weaver <[EMAIL PROTECTED]> wrote:
> 
> > Conover, W.J.  Practical nonparametric statistics.
> 
> Thank you for the suggestion. However, when I read about its
> description in amazon, it seems that this book contains sections on
> using ranks to analysis traditional data, and not about how to process
> ranking data, i.e., the raw data are already rankings. 
> I am mainly looking for references of the second category. 
> 
> Please correct me if I am wrong. Anyway, thank you for the suggestion!

Well, it seems to me that you are wrong....  Conover 
tells how to analyze ranks.

If your data start out as something better, then you do 
have a choice; maybe you can figure out how to preserve 
some information  by (say) using transformations.  If your
data start out as ranks, you are stuck with it.  But Conover
gives the analyses, and they are proper tests in either case.

Am I missing some logical point here?

I think it is true that someone else might tell more about
the problems of making fancier  *models*  out of rank-data.
(But, for the most part, they are going to enumerate the 
'problems'  and not, I  think, lay out many solutions.)


-- 
Rich Ulrich, [EMAIL PROTECTED]
http://www.pitt.edu/~wpilib/index.html
"Taxes are the price we pay for civilization." 
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to