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

st: RE: Modelling with probit


From   "Naji Nassar \(MIReS\)" <naji.nassar@mires.fr>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: Modelling with probit
Date   Mon, 10 Nov 2003 21:18:44 +0100

Joan

I hope I've understood your 'problem' & my remarks will help :
- The group weight : if the group represents, say 80% of the people
concerned with the estimated model, it's 'normal' that the model fits the
group better than the last 20%
- heterogeneity : when you calibrate a standard probit model, you do suppose
that the explanatory has a homogenous impact over all the population. If
it's not the case (latent class approch, mixed modelling..) the standard
model will look like a median between several unobserved models. Your group
can be close to one of those unobserved models
- processing the explanatory variable : suppose that one is modelling buying
a luxuary car (illustrative case), and the group is related to for medium
income. Given their financial & status implication, this group will perhaps
pays more attention to the car & service characteristics. The model will
exhibit higher probability even though the explanatory variable impact is
low..


Regards
Naji
-----Message d'origine-----
De : owner-statalist@hsphsun2.harvard.edu
[mailto:owner-statalist@hsphsun2.harvard.edu]De la part de Joan A.
Williams
Envoye : lundi 10 novembre 2003 19:32
A : statalist@hsphsun2.harvard.edu
Cc : jawillia@princeton.edu
Objet : st: Modelling with probit


Hello to all:

I am familiar with the standard probit model for evaluating to what effect
an explanatory variable may have on a categorical dependent variable
assuming (in my case) a binary value of 1 as well as the Stata commands
probit, dprobit, mfx, etc.

In my model, I am trying to assess why a given group of people (dummy
variable)exhibits a higher probability (in regards to the dependent
variable)than people not in the group within my probit model.
Is the usual xi interaction expansion kosher (in terms of technique and
finesse)? Or should I develop another model apart from the original probit.
Critiques, suggestions are welcomed. Thanks

*
*   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/

*
*   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