----- Original Message -----
From: "Constantine Daskalakis" <[email protected]>
To: <[email protected]>
Sent: Friday, December 05, 2003 11:06 AM
Subject: Re: st: PSEUDO R2
> At 09:28 AM 12/5/2003, you wrote:
> >I am estimating grouped data logistic models using glogit or simply
> >using regress with analytical weights
> >
> >generate double lcrp=log(ntprop/(pop-ntprop))
> >generate double wt1=ntprop*(pop-ntprop)/pop
> >
> >glogit ntprop pop clprop $xvars $zvars $tvars
> >regress lcrp clprop $xvars $zvars $tvars [w=wt1]
> >
> >If I want to get a measure of goodness of fit I might use the squared
> >correlation between y and yhat
> >
> >regress lcrp clprop $xvars $zvars $tvars
> >predict yhat1
> >corr yhat1 lcrp
> >
> >Is that correct?
> >
> >Antonio
>
> That is not ideal.
> See the book by Hosmer & Lemeshow (Applied Logistic Regression) for an
> excellent and extensive basic discussion on assessing fit.
> CD
>
Sealed Envelope has a Hosmer-Lemeshow goodness of fit test that can be used to
test whether observed binary responses, conditional on a vector of covariates,
are consistent with predictions.
Try:
net from http://www.sealedenvelope.com/
hope this helps,
Scott
*
* 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/
