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

st: re: marginal effects after ivtobit

From   Kit Baum <>
Subject   st: re: marginal effects after ivtobit
Date   Sun, 23 Nov 2008 08:07:40 -0500

Martin said

Kit, maybe I misunderstand the word "deprecated", but are you implying that users should not use the official -dprobit-? I cannot find anything in its
help file that would discourage its use. In the case of the
user-written -dtobit-, the author has not yet withdrawn the command, which
is a different story...

Indeed, I was suggesting that. There was a discussion at last week's FNASUG in which someone complained that -dprobit- could not do this or that. It is essentially an orphaned command; when StataCorp thought about creating one such command for every limited dep var estimation command, they must have figured out that it would be easier to write a single -mfx- and set it up to deal with every estimation command. - dprobit- predates that and is still with us because Stata does not usually drop existing commands. But if you want the full set of capabilities available to you, use -probit- and -mfx-. Limitations of - dprobit- include vce options, allowed prefix commands, allowed weight types, etc.

Personally, in my research I use -probit- followed by -mfx-. I throw away the -probit- results and put the -mfx- results in a table (which cannot be done with -est table-, btw, but is handled nicely with - estout- from Ben Jann using the -marg- option).

Kit Baum, Boston College Economics and DIW Berlin
An Introduction to Modern Econometrics Using Stata:

*   For searches and help try:

© Copyright 1996–2015 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index