st: xtprobit postestimation-predict

 From gmc107@york.ac.uk To Subject st: xtprobit postestimation-predict Date Mon, 24 Oct 2005 12:55:19 +0100

```I want to extract the error terms following the estimation of a random
effects probit model (xt probit command).
If we had a linear model then
Y=xb+e and hence e=y-xb is the error term.
What does xb (linear prediction) actually calculate for xtprobit?
I have read in Econometrics Stewart and Gill (2nd edition) that it is better
to calculate the estimated probability that the individual has the attribute
measured by the dummy dependent variable eg the pu0 option.
Using
predict xb, xb
predict pu0, pu0
and just calculating phixb=norm(xb)
I end up with norm(xb)=pu0
Is xb the "probit index"???
To calculate the errors is it sufficient to use
predict xb, xb
gen errors=y-xb where y is the dependent variable which is a dummy taking a
value of one or zero?
Or do I have to use y-puo=y-norm(xb).
Would be most grateful if someone could help.
Georgios

*
*   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/
```