From: Achim Zeileis <Achim.Zeileis_at_wu-wien.ac.at>

Date: Tue 12 Jul 2005 - 18:51:45 EST

R-help@stat.math.ethz.ch mailing list

https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html Received on Tue Jul 12 18:54:39 2005

Date: Tue 12 Jul 2005 - 18:51:45 EST

On Tue, 12 Jul 2005, S.O. Nyangoma wrote:

> > If they have the same degrees of freedom, use the test statistic

*> > and not
**> > the p value for comparing them.
**> > Z
**>
**> I appretiate your input to this discussion. Do you know of a reference
**> to your statement above?
*

?? Any basic statistics book? Distribution functions tend to be monotonous.

> I had actually used the test-statistic which in my case is r-squared

*> to compare them. This is in my view was adequate but I also think it
**> is more convincing to say something about the p-values
*

Not really `more' convincing, it's all pretty equivalent when the number
of estimated parameters is the same. You can also compare the fitted
models via their associated residual sum of squares which I would find
most intuitive because that is the objective function you are trying to
minimize via OLS.

Z

> (difficulties

*> in computing them, and hence the rationale of solely using the test-
**> stat).
**>
**> regards. Stephen.
**>
**>
**>
*

R-help@stat.math.ethz.ch mailing list

https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html Received on Tue Jul 12 18:54:39 2005

*
This archive was generated by hypermail 2.1.8
: Fri 03 Mar 2006 - 03:33:31 EST
*