Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


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

st: Easy estimation methods for discrete-time duration models


From   "Stephen P. Jenkins" <stephenj@essex.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: Easy estimation methods for discrete-time duration models
Date   Thu, 2 Dec 2010 10:33:58 -0000

------------------------------
Date: Wed, 1 Dec 2010 13:50:40 +0100
From: "Modena, Francesca" <francesca.modena@unitn.it>
Subject: st: Easy estimation methods for discrete-time duration models

Dear all,
I estimated a discrete time duration model using the easy eastimation
method (Jenkins, 1995).
I reorganized the data set (converting the unit of analysis) and then
I estimated the logit model.
In order to compute the marginal effects, may I use the mfx (or
prchange) command as in the standard logit model?
Thanks
Francesca Modena
University of Trento, Department of Economics
--------------------------------

The poster could have said the article is: "Easy estimation methods
for discrete-time duration models", Oxford Bulletin of Economics and
Statistics, 57(1) February 1995, 129?138.

My main reaction to the Question posed is: which marginal effect? On
the hazard? [If so, you can use standard tools of the sort you cite --
but also see -margins-.]  Arguably, it's more interesting to look at
impacts on statistics such as the median duration -- which can be
derived from the hazard. There is no canned routine out there for
this, but calculations are relatively straightforward using -predict-.

Historical note. The OBES article's main claim to fame, if any, is
showing that the trick of "data expansion step followed by fitting
using a binary regression program" also maximizes the right likelihood
for the case when there are left-truncated (and possibly
right-censored) survival times. Although the logit model was cited in
my old paper, you can use others. The -cloglog- has particular
attractions, given its link with proportional hazards models (see the
URL below signature). Of course, the 'easy estimation' method assumes
no frailty (a.k.a. unobserved heterogeneity) -- if that is present,
life isn't so easy.


Stephen
-------------------------------------------------------------
Professor Stephen P. Jenkins <stephenj@essex.ac.uk>
Institute for Social and Economic Research
University of Essex, Colchester CO4 3SQ, U.K.
Tel: +44 1206 873374.  Fax: +44 1206 873151.
http://www.iser.essex.ac.uk ; 
Survival Analysis using Stata:
http://www.iser.essex.ac.uk/survival-analysis
Downloadable papers and software: http://ideas.repec.org/e/pje7.html 


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   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   |   Site index