[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: ststa

From   Maarten buis <>
Subject   Re: st: ststa
Date   Mon, 15 Jun 2009 13:08:50 +0000 (GMT)

--- On Mon, 15/6/09, wrote:
> dear sir
> i have to analyse the marginal effect and estimates the
> probability by using STATA.please sir if i do that
> estimation i could not get the probability measures and
> i got the same answers from tobit estimation and marginal
> effects.i have the date zero,one and between zero and one
> as the dependent variables,plsease sir replyme

It is hard to answer this question as you are not very clear
about what model you use and what you have already tried to 
do. Apperently you have at least two model: one with a 
binary depedent variable and one with a continuous dependent
variable (restricted between 0 and 1)  on which you use a 
-tobit- model. In case of -tobit-, -mfx- returns by default 
the marginal effects with respect to the latent variable, 
which happens to be the same as the coefficients you get 
directly from -tobit-. If you want other types of marginal
effects you have to specify that with the -predict- option 
within -mfx-. As to your binary model, you have just given
too little information to be able to give you more than 
some very general advise, like: "You can often get predicted 
probabilities with the -predict-, and the -adjust- command
can sometimes also be useful". 

Hope this helps,

PS. If this does not answer you question, than it may help
to read the advise on how to formulate answerable questions 
on: ,
and than sent a refrased version of this question to the list.

PPS. It is often safer to avoid any reference to the gender
of the potential readers of your message.

Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen


*   For searches and help try:

© Copyright 1996–2022 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index