Re: st: probit mean

[Marcello Pagano <pagano@hsph.harvard.edu>]

I don't know about what was meant by the first part, but the second part 
"My colleague estimates a probit model and then puts mean projections 
into the explanatory variables..." is a silly way to proceed.  Think 
about it.  You would be finding the probability for the "average" 
person, i.e. one with a gender of  0.46, for example.  It is far better 
to average out the probabilities.

m.p.

Dirk Nachbar wrote:

> All
>
> A colleague received a reply from an academic saying " Remember that 
> the probit does not predict at the mean and an adjustment needs to be 
> made for this." Unfortunately I couldn't tell him what she meant by 
> that. My colleague estimates a probit model and then puts mean 
> projections into the explanatory variables to predict a dependent 
> probability.
>
> Does any of you know what the academic meant by that?
>
> Dirk Nachbar
> Assistant Economist
> WWEG FPWD CPSA
> Department for Work and Pensions
> Level 2, The Adelphi
> 1-11 John Adam St
> WC2N 6HT London
> 020 796 28531 uk
*
*   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/