At 09:29 PM 12/7/2005, Sunhwa Lee wrote:
Hello,
I am writing an ml program that replicates "oprobit" code in stata. With
all many trials, I have not succeeded in getting the exactly identical
estimates as with "oprobit" command. Below is the basic program I used to
compare with oprobit estimates.
**************************************
capture program drop myoprobit
program define myoprobit
args lnf xb t1 t2 t3
tempvar p1 p2 p3 p4
qui gen double `p1'=ln(norm(`t1'-`xb'))
qui gen double `p2'=ln(norm(`t2'-`xb')-norm(`t1'-`xb'))
qui gen double `p3'=ln(norm(`t3'-`xb')-norm(`t2'-`xb'))
qui gen double `p4'=ln(norm(`-t3'+`xb'))
This line is wrong. It should be
qui gen double `p4'=ln(norm(-`t3'+`xb'))
The two results look similar at the first glance, but if you take a closer
look, _Iturn_32 and _Iturn_46 are different. The differences may be
negligible with auto.dta, but they are amplified with my dataset, again on
dummies.
i assume the oprobit command was
xi: oprobit rep mpg i.turn
Under either your program or oprobit, the standard errors for
_Iturn_32 and _Iturn_46 are enormous. oprobit gives the error
Note: 4 observations completely determined. Standard errors questionable.
The actual estimates reported by oprobit for those two parameters are
------------------------------------------------------------------------------
rep78 | Coef. Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
_Iturn_32 | 10.03019 9014619 0.00 1.000 -1.77e+07 1.77e+07
_Iturn_46 | -8.896943 5353941 -0.00 1.000 -1.05e+07 1.05e+07
There may or may not be other problems in your program, but I don't
think a small dataset using lots of dummy variables, where the counts
for some categories are as low as 1, is a good way of testing it.
-------------------------------------------
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
FAX: (574)288-4373
HOME: (574)289-5227
EMAIL: [email protected]
WWW (personal): http://www.nd.edu/~rwilliam
WWW (department): http://www.nd.edu/~soc
*
* 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/