I don't see why you couldn't use a standard method for goodness of fit.
I don't know whether it's implmented.
The weights should be used in the test.
I dont' think this will affect the chi2 distribution (as long as
individual weights don't dominate which would reduce the effective
sample size too much ) - but I can't give you a formal proof.
matt
Julien Junod wrote:
Hello Stata list,
I would like to know, if there exist routines for computing goodness of
fit or other diagnostics (like lfit, etc.) for a logistic regression on
complex surveys using stratification.
If it is not the case, as Hosmer and Lemeshow write in the 2nd ed. of
their book, can I use the standard method ? What is the effect of the
weights on the distribution ? Is it still (approximately) chi2 ?
Thanks,
Julien Junod
*
* 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/
--
Matt Schonlau
310-393-0411 x7682
http://www.rand.org/methodology/stat/members/matt/
*
* 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/
