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

Re: st: Marginal Effects

From   Andreas Kuhn <>
Subject   Re: st: Marginal Effects
Date   Mon, 21 Jun 2004 14:40:34 +0200

peter harper wrote:

Dear Statalist
I am trying to do marginal effects after oprobit. I have run a oprobit regression on a dependent
variable, which is called price3. Its other
are price2 and price1, which obviously I have not
used. So after I have run
. oprobit price3 x1 x2 x3 I issued the command
. mfx compute, predict(outcome(3))
the message I got was put a value of price3 or
or worde to that effect
so I tried
. mfx compute, predict(outcome(#3))
the message I then got was there is no category 3,
there are only 2 categories.
I suspect, this is because this price3 itself is
categorised under 0 and 1. 0 is people who do not
think price is important and 1 is for people who
price is important.
so I ran a marginal effects using
.mfx compute, predict (outcome(#2)
All I should want is M.E. of price3 is it not? as I
have run the oprobit using that variable. So I ran
. mfx compute, predict(outcome(#2))
Is that the right thing to do, as I want the M.E.
the regression I ran with the dependent variable as
Please confirm, if I am doing it correctly.
Please let me know if you need further clarification.
Thanks in advance

___________________________________________________________ALL-NEW Yahoo! Messenger - sooooo many all-new ways to express yourself
* For searches and help try:

Dear Peter,

You should install the - spost - package (written by Jeremy Freese and Scott Long; by the way: they have written a very useful book concerning the estimation and the interpretation of the ordered probit model with Stata, among others. "Regression Models for Categorial Dependent Variables Using Stata", published by Stata Press 2001.), which is available for version 7 or 8.
Having installed this package, you can then use the command - prchange - for estimating marginal effects of the ordered probit model.

Best regards, Andreas

* For searches and help try:

© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index