Hello statalisters:
Non Stata question here--
Does anyone have advice or guidance about how one should determine the
number of groups for the Hosmer Lemeshow goodness of fit test? I have
over 21,000 records with 830 covariate patterns so it seems like it
would be sensible to use more than 10 groups (the default), but I am
unsure of how many I should choose. Is there some kind of rule of
thumb for this determination? From reveiwing the literature, it seems
most people just use 10. I also know that one flaw of this test is
that the number of groupings can make large differences in the result.
Thanks very much for any ideas, Tim
*
* 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/
