Re: st: Predicted probabilities

[Jeph Herrin <junk@spandrel.net>]

Claude Francoeur wrote:
> Hello Richard,
>
> I am trying to generate a prediction-classification table.  It shows how the 
> model predicts the binary dependent correctly. The cut-off value is usually 
> 50%. For example, EViews produces the following table :
I think Richard is correct, you will need to make your own. But rather than
assume a cutpoint of 0.5, a better way to assess classification is with
a C-statistic (area under the ROC)

 xtprobit pw boardcomp compensation shrights disclosure tech mv endet cot
       volat march conc, pa i( i)
 predict predPW, mu
 gen PW=round(predPW,0.01)         /* create 100 categories */
 roctab pw PW

This will indicate how well your model classifies the outcome independent
of any assumption about the cutpoint.

hth,
Jeph

*
*   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/