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

Re: st: margfx?

From   "Dimitriy V. Masterov" <[email protected]>
To   [email protected]
Subject   Re: st: margfx?
Date   Tue, 10 Aug 2004 20:24:03 -0500 (CDT)

I downloaded margfx from

after someone on this list recommended it to me. Jonah B. Gelbach is the
author, but I thought someone might be familiar with it on this list.

I am familiar with mfx, but in my understanding it calculates dphi/dx
evaluated at a given x (with the average x as the default). I would call
the latter derivative the marginal average effect (MAE). I want the AME,
which is the average of all the dphi/dxs evaluated for each x in
e(sample). Are there any user-written programs that accomplish this?


On Tue, 10 Aug 2004, Christopher F Baum wrote:

> Dmitri wrote
> Could someone please explain why these two versions of margfx do not
> work?
> According to the help, they are equivalent, but I get a cryptic syntax
> error after the second.
> Margfx also sometimes gives the AME for a subset of the explanatory
> variables without any indication of why it dropped the others. The
> probit
> runs without dropping anything.
> If someone could explain this, I would be most grateful. This is my
> code:
> use auto, clear
> gen sex=round(uniform(),1) /* I wanted to add a binary var on RHS */
> /* Method 1 */
> margfx for*, contin(price mpg headroom) binary(sex) model(probit)
> /* Method 2 */
> probit for* price mpg headroom sex
> margfx, q
> According to findit margfx, margfx is neither a Stata command nor a
> user-contributed routine. Do you mean mfx?  mfx runs just fine after
> the probit you specify. But I have no idea how you might do the "method
> 1" estimation (or where you got margfx).
> Kit

Dimitriy V. Masterov

Center for Social Program Evaluation
1155 East 60th St. Room 038
Chicago, IL 60637
Work: (773)256-6005
Fax: (773)256-6313

1312 East 53rd St., Apt.309
Chicago, IL 60615
Mobile: (773)220-2760

*   For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index