Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

RE: st: weighted logit - predict residuals?


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: weighted logit - predict residuals?
Date   Sat, 6 Nov 2004 15:58:51 -0000

Regardless of the views expressed here, whereas 

. logit foreign mpg [aw=weight] 

. predict res, res

doesn't work, 

. glm foreign mpg [aw=weight], link(logit) 

. predict res, res

does. 

Nick 
n.j.cox@durham.ac.uk 

Stas Kolenikov
> 
> From what I see in the help file, Stata does not provide for [aweight]
> estimation of a logit model. At any rate, if you can come up with a
> clear explanation as to how you want to generalize the residual /
> influence diagnostics to weighted data, that would be quite an
> achievement -- I don't think I've seen it done before in literature.
> It is one of those cases when something obvious is not implemented in
> Stata means it is not statistically justified. If you estimate the
> model without the weights, get the diagnostic variables, and multiply
> them by the weights, you still won't have exactly what you need, as
> the model in which they were estimated is a different (and probably
> incorrect) one.
> 
lisa@alumni.princeton.edu
 
> > I understand that Stata won't let me "predict r,
> > residuals" or "predict h, hat" after an "aweight"
> > weighted logit, but I would still like to find a way to
> > look at the residuals and leverage to find observations
> > that have an undue effect on the fitted model. Any
> > thoughts? Thanks very much,

*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index