Ok Nick, this is a bit of details.
I am mainly looking at the impact of binge drinking on violence
related outcomes.
I run probit but then biprobit. And I have controls such as age, sex, race,
ethnicity, religious affiliation, year in school, parents education,
parents drinking patterns, living arrangements,type of college,
region, and year.
So as I said below, I have to force -mfx- to give me s.e but then
these s.e would not even be reported in estout. And what are the
costs of this -force-?
(Results copied from Stata Results window0
. mfx compute;
warning: predict() expression unsuitable for standard-error calculation;
option nose imposed
Marginal effects after biprobit
y = Pr(Bdrinking=1,Damage=1) (predict)
= .06381258
-------------------------------------------------------------------------------
variable | dy/dx X
---------------------------------+---------------------------------------------
age | -.0104554 39.3722
age2 | .0001959 1675.15
x1 *| .0438169 .567348
x2*| -.030192 .945393
x3 | .0451067 1.10193
race1*| .038111 .469056
race2*| -.0596293 .202464
jewi*| .020694 .201064
noreligion*| .011279 .627275
catholic*| .1514319 .16718
Bdrinking*| .0064071 .288435
-------------------------------------------------------------------------------
(*) dy/dx is for discrete change of dummy variable from 0 to 1
.
end of do-file
mfx compute, force
Marginal effects after biprobit
y = (predict)
= .06381258
------------------------------------------------------------------------------
variable | dy/dx Std. Err. z P>|z| [ 95% C.I. ] X
---------+--------------------------------------------------------------------
age | -.0104554 .00174 -6.00 0.000 -.013868 -.007043 39.3722
age2 | .0001959 .00002 9.30 0.000 .000155 .000237 1675.15
x1*| .0438169 .00569 7.70 0.000 .032663 .054971 .567348
x2*| -.030192 .01535 -1.97 0.049 -.060277 -.000107 .945393
x3 | .0451067 .00877 5.15 0.000 .027924 .06229 1.10193
race1*| .038111 .00625 6.10 0.000 .02586 .050362 .469056
race2*| -.0596293 .00519 -11.50 0.000 -.069795 -.049464 .202464
jewi*| .020694 .0478 0.43 0.665 -.072991 .114379 .201064
inorel~n*| .011279 .04088 0.28 0.783 -.06884 .091398 .627275
icatho~c*| .1514319 .0804 1.88 0.060 -.006142 .309006 .16718
Bdrinking*| .0064071 .00484 1.32 0.186 -.003085 .015899 .288435
------------------------------------------------------------------------------
(*) dy/dx is for discrete change of dummy variable from 0 to 1
On 10/17/07, Nick Cox <[email protected]> wrote:
> The econometric costs of force would seem to depend on y, x1, x2
> and x3. Otherwise put: is this an answerable question?
>
> Nick
> [email protected]
>
> Nirina F
>
> > I would appreciate your help on this:
> > I run
> > probit y x1 x2 x3
> > mfx, force
> > estimates store mfx
> > estout mfx using feuille1, margin cells (b(fmt(%9.4f)) se(par
> > fmt(%9.2f)))
> >
> > where I have to use the option force as mfx alone wouldn't
> > report the s.e.
> > In the stata window, I can see results of se after I use force but
> > then in estout the se 's are not reported.
> > Also, what are the cost of using force in my econometric results?
>
> *
> * 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/
>
*
* 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/
