Re: goodness of fit test for conditional logistic regression for 1:1 matched pair data

["Im, Kelly" <im@edc.pitt.edu>]

Greetings,
 
I'm looking for a way to test goodness of fit for a matched (1:1) data
analysis.   One can look at deviances or model chi-squares, I guess, but
that seems to provide more of a "relative" goodness of fit (compared to
other models) than an "absolute" goodness of fit.

Is there a Splus function to calculate a Hosmer-Lemeshow "like" measure
that gives a single statistic  that is also an absolute, rather than
relative measure of goodness of fit?

I'd very much appreciate your advice,

With regards,

kelly


++++++++++++++++++++++++++++++++++++
 KyungAh (Kelly)  Im, 
 Statistician
 130 Desoto St. 127 Parran Hall. 
 Epidemiology Data Center 
 Graduate School of Public Health
 Department of Epidemiology
 University of Pittsburgh
 PA 15261
 PHONE:(412) 624- 4612 FAX: (412) 624- 3775
 Email: Im@edc.pitt.edu
 http://www.edc.gsph.pitt.edu
++++++++++++++++++++++++++++++++++++