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

Re: st: marginal effect in random effect logit model


From   May Boggess <mboggess@stata.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: marginal effect in random effect logit model
Date   22 Mar 2004 15:39:36 -0600

On Monday,  Pallab Mozumder wrote:

> I am running a simple random effect logit model with 3 independent variable 
> [2 of them are dummy variables (with 0 and 1) and the other one is 
> continuous variable]. When I try to compute the marginal effets at the mean 
> level (with "mfx compute"), i get the same value of (dy/dx), st. error and 
> significance level as I got for the coefficients of the original random 
> effect model. It looks a little odd to me.

This is happening because when you do not use the -predict- option with
-mfx- it uses the default prediction of the predceding estimation
command. In the case of -xtlogit, re- this is xb, the linear predictor.
The marginal effects of that prediction function are, as Pallab pointed
out, the coefficients, because the derivative of xb with respect
to xb is b: 

clear
webuse union
drop if _n>5000
xtlogit union age grade ,i(id) nolog re
mfx compute, predict(xb)

You may want to try the prediction option pu0, as follows:

clear
webuse union
drop if _n>5000
xtlogit union age grade ,i(id) nolog re
mfx compute, predict(pu0)


-- May
mmb@stata.com


*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



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