Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: logit with iweight


From   "Bontempo, Daniel E" <[email protected]>
To   "'[email protected]'" <[email protected]>
Subject   st: logit with iweight
Date   Thu, 15 Mar 2012 14:15:08 +0000

Hello -

I found iewights does not change the reported N for logit regression, although the desired weighting is reflected in point estimates. As a result, when adding CoxSnell R2 to stored results, the wrong N is used for calculation.

The only related post I found in the archives was Bradley Fedy's message (below) indicating that before v11, the weighted N was stored/displayed. I did not find any replies to Bradley's question.

Is there some reason CoxSnell (weighted N) is invalid with iweights, or is this a problem that should be reported?

Thanks


> From   Bradley Fedy <[email protected]> 
> To   [email protected] 
> Subject   st: iweight (iw) behaves differently between ver. 10.1 and 11 
> Date   Thu, 19 Aug 2010 10:30:38 -0600 
>
> --------------------------------------------------------------------------------
>
> I am using iweights to adjust the "importance" of observations in STATA
> 10.1 using the logit function.  I sent the data and code to a colleague who
> is working in STATA ver. 11.  The coefficient estimates from our models are
> the same.  However, any metric based on likelihood (e.g. AIC, BIC) is
> different.  It appears that ver 10.1 is "down weighting" observations as
> specified and thus calculating BIC based on a smaller sample size, but ver.
> 11 is not.  Can anyone explain change between the two versions?




*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index