Hi Oliver,
These marginal effects should not add up. See for instance Breen (1996) or Long (1997), for very accessible accounts of these different marginal effects.
Richard Breen (1996) "Regression Models; Censored, sample selected, or Truncated Data", Thousand Oaks: Sage. (one of those books in the series Quantitative Applications in the social Sciences, a.k.a. "little green sage books")
J. Scott Long (1997) "Regression Models for Categorical and Limited Dependent Variables", Thousand Oaks: Sage.
HTH,
Maarten
-----------------------------------------
Maarten L. Buis
Department of Social Research Methodology
Vrije Universiteit Amsterdam
Boelelaan 1081
1081 HV Amsterdam
The Netherlands
visiting adress:
Buitenveldertselaan 3 (Metropolitan), room Z214
+31 20 5986715
http://home.fsw.vu.nl/m.buis/
-----------------------------------------
-----Original Message-----
From: [email protected] [mailto:[email protected]]On Behalf Of olivier lohest
Sent: vrijdag 9 december 2005 12:02
To: [email protected]
Subject: st: Dtobit mfx and Mcdonal and Moffit decomposition - tobit
Hi
I am estimating a tobit model for migration (cross section)
I am using the dtobit "procedure"* to compute the unconditionnal expected
value of the marginal effect but also to decompose this effect into two
components : the marginal effect of the conditional to being unconsered and
the marginal effect of the probability of being unconsored....The dtobit
command is useful to compute quieckly this result.
My question is :
when I compare the marginal effect of the unconditionnal expected value and
the sum of the marginal effect of the conditional to being unconred and of
the probability of being unconsored, I don't have the same result.... the
sum of these two components is not equal of the marginal effect of the
unconditionnal expected value. I am using stata 8....
When I am looking to the empirical literature, I found also the same
difference if the people have used stata
Do you any explications for these differences
Thank your for your help
best
Olivier
* I have also trying to compute these marginal effects via the command mfx
(I obtain the same result)
*
* 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/
