R help - Dec 2003 - R^2 analogue in polr() and prerequisites for polr()

Hi

(1)In polr(), is there any way to calculate a pseudo analogue to the
R^2. Just for use as a purely descriptive statistic of the goodness of
fit?

(2) And: what are the assumptions which must be fulfilled, so that the
results of polr() (t-values, etc.) are valid? How can I test these
prerequisites most easily: I have a three-level (ordered factor)
response and four metric variables.

many thanks

Christoph

-- 
Christoph Lehmann <christoph.lehmann at gmx.ch>

On 8 Dec 2003, Christoph Lehmann wrote:
> (1)In polr(), is there any way to calculate a pseudo analogue to the
> R^2. Just for use as a purely descriptive statistic of the goodness of
> fit?
First define the statistic you are interested in!  There is an absolute 
measure of fit, the residual deviance.
> (2) And: what are the assumptions which must be fulfilled, so that the
> results of polr() (t-values, etc.) are valid? How can I test these
> prerequisites most easily: I have a three-level (ordered factor)
> response and four metric variables.
This is discussed with worked examples in the book that MASS supports, so 
please consult your copy.

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

many thanks! I was just asking for a r-square analogue, since the
students I will present the results to, might like to know, how the
measure of fit in an ordinal regression (e.g. the residual deviance)
compare to measures they know (from introductory courses to linear
regression) (such as the r-square), means: how much of the variance of
the dependent variable can be explained by the variance of the
independent variables.

thanks and best regards

christoph
On Tue, 2003-12-09 at 07:26, Prof Brian Ripley wrote:
> On 8 Dec 2003, Christoph Lehmann wrote:
> 
> > (1)In polr(), is there any way to calculate a pseudo analogue to the
> > R^2. Just for use as a purely descriptive statistic of the goodness of
> > fit?
> 
> First define the statistic you are interested in!  There is an absolute 
> measure of fit, the residual deviance.
> 
> > (2) And: what are the assumptions which must be fulfilled, so that the
> > results of polr() (t-values, etc.) are valid? How can I test these
> > prerequisites most easily: I have a three-level (ordered factor)
> > response and four metric variables.
> 
> This is discussed with worked examples in the book that MASS supports, so 
> please consult your copy.
-- 
Christoph Lehmann <christoph.lehmann at gmx.ch>