Tomas,
It is also not a bad idea to look at the substantive interpretation of your
results and ask the question "Is this meaningful?" If you have an odds
ratio of 1.01 does that 1% increase mean anything to you? My point is that
in the social sciences 1% may not mean much to us but in the medical field
1% could be a big deal when life or death is a reality.
If you are going for parsimony in your model (which I assume you are) the
above logic also applies when adding variables.
Carter
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Blau Blau
Sent: Tuesday, December 20, 2005 5:03 AM
To: [email protected]
Subject: st: comparing logit models with large N
Dear all,
I would like to ask you if there is any statistical approach how to compare
logit models with large N in STATA?
It means to make difference in contribution to response variable among
substantive effects and nuisance effects of explanatory variables. The
problem is that if I add each new variable (or each new interaction between
two variables) in model, it always significantly contributes to response
variable and the fit of each complex model is always better than the
previous (more parsimonious) one. (BIC is always lower, LR is always higer
and D is alway lower).
I think that the problem is in large N. My data come from the whole
population. I did the random sample but the sample is still large to
differentiate substantive effects in explanatory variables. Does anybody
know what else I can do? Thanks.
Tomas
*
* 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/
*
* 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/
