Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

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

st: Standar error of discrete change probit model

From   NOAM MANO <>
Subject   st: Standar error of discrete change probit model
Date   Tue, 31 Jan 2012 15:30:14 -0500

I want to write my own do file to compute the marginal effect of
discrete change in probit model.

David M. Drukker(2010) ; 2010 Italian Stata Users Group meeting
Bologna November 2010

Using margins to estimate partial effects
probit crash tickets traffic i.male, nolog

Or using hand

 generate double xb0 = _b[tickets]*tickets + _b[traffic]*traffic + _b[_cons]
 generate double de = normal(xb0 + _b[1.male]) - normal(xb0)
 summarize de

The mean of de give the marginal effect. We use stata function margin
( margins , dydx(male), we will have the same value of the marginal
effect. However the standard error is different from the one obtained
by summarized de.
My question is: how to compute the standar error by hand.

I have tried this:

(normalden(xb0 + _b[1.male]) +normalden(xb0) )^2 *var(_b[1.male])

But I am not sure if it's good.
Your help will be very appreciated

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index