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

st: =?utf-8?Q?RE=3A_help_margin-_beginner=C2=B4s_ask_?==?utf-8?Q?=28again=21=29?=

From   "Sanz De Galdeano, Anna" <>
To   <>, <>
Subject   st: =?utf-8?Q?RE=3A_help_margin-_beginner=C2=B4s_ask_?==?utf-8?Q?=28again=21=29?=
Date   Thu, 20 Nov 2003 16:49:50 +0100

If I am not wrong, xtlogit, fe estimates Chamberlain's conditional logit model. 
I think that in order to predict the success probabilities it is necessary to have an estimate of the fixed effects, which is not directly obtained using the Chamberlain technique (see Wooldridge for a discussion of this). Hence, after estimating the conditional logit model, you have to estimate the fixed effects and then use them to compute the marginal effects. I think that you can do that by using the conditional logit coefficient estimates (betas) to compute the value of the fixed effects from the first order conditions correspondingh to the fixed effects parameters in the joint maximum likelihood estimation. This is, I think, the best you can do, altough these fixed effects estimates are not consistent unless you have a big number of periods.
I would like to know how mfx or margin deal with these issues.

	-----Original Message----- 
	From: Nick Varian [] 
	Sent: Thu 20/11/2003 13:58 
	Subject: st: help margin- beginner´s ask (again!)

	Hi all, 
	Could you help me once more? 
	i´m trying to use margin with the dummies option after 
	xtlogit, fe , but I got the message "-margin- does not 
	work with clogit; use -mfx- instead" . But, in the 
	help file, the author says that it works with xtlogit 
	and clogit. Am I doing anything wrong? 
	As I said before, I am a still learnig stata, so 
	perhaps the next question may be a very simple one: In 
	the help file of margin, the author says to code the 
	dummy variable as 0/1. How can I do that? 

	Best Regards, 
	Nick Varian 


	Yahoo! Mail: 6MB, anti-spam e antivírus gratuito! Crie sua conta agora: 

	*   For searches and help try: 

*   For searches and help try:

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