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: Marginal effects plot for tobit with interaction term


From   Jan Duchoslav <j.duchoslav@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   st: Marginal effects plot for tobit with interaction term
Date   Sat, 27 Apr 2013 12:33:00 +0200

Dear all,

I use the marginal effect plots for a model with and interaction term
as suggested by Berry, Golder and Milton. 2012. "Improving Tests of
Theories Positing Interaction." The Journal of Politics 74(3):
653-671. Please find the code (largely based on that provided by
Golder at https://files.nyu.edu/mrg217/public/interaction.html) below:

tobit cash_pc treat scchiefbl  inter scvillbl pop, ll(0) level(95)

matrix b=e(b)
matrix V=e(V)

scalar b1=b[1,1]
scalar b2=b[1,2]
scalar b3=b[1,3]

scalar varb1=V[1,1]
scalar varb2=V[2,2]
scalar varb3=V[3,3]

scalar covb1b3=V[1,3]
scalar covb2b3=V[2,3]

scalar list b1 b2 b3 varb1 varb2 varb3 covb1b3 covb2b3

drop vill_id-scvillbl scvill-inter

set obs 10000

generate MVZ=((_n+680)/1000)
	replace MVZ = . if MVZ > 1
	replace MVZ = . if MVZ < 0.68
	drop if MVZ == .

gen conbx=b1+b3*MVZ
	label var conbx "ME estimate"

gen consx=sqrt(varb1+varb3*(MVZ^2)+2*covb1b3*MVZ)

gen ax95=2.03*consx
gen ax90=1.69*consx

gen upperx95=conbx+ax95
	label var upperx95 "95% conf. level"
gen upperx90=conbx+ax90
	label var upperx90 "90% conf. level"

twoway (line conbx MVZ, lc(black) yaxis(1)) ///
	(line upperx90 MVZ, lc(black) lpat("-") yaxis(1)) ///
	(line upperx95 MVZ, lc(black) lpat(".-") yaxis(1))

Since I use a tobit model, I would like to construct similar plots for
marginal effects conditional on being uncensored and for marginal
effects on the probability of being uncensored. I get the marginal
effects on each term using the -dtobit- command (from STB-56: sg144,
updated sg144_1), which runs a Moffitt-McDonald decomposition. This
has to be done in version 8 as starting with version 9, Stata uses a
different way to store estimation results (or so I understand):

version 8: tobit cash_pc treat scchiefbl  inter scvillbl pop, ll(0) level(95)
dtobit

Now my question is: Does anybody know if and how -dtobit- stores its
results so that they can be used to construct the plots in a similar
way as above? I would appreciate any leads.

Thanks for your consideration,

Jan
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


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