Re: st: Marginal Effects using subsetted data

[Richard Williams <richardwilliams.ndu@gmail.com>]

At 02:54 PM 10/31/2010, MOSTARDO, NICK wrote:
> Hello,
>
> I am running an ordered probit model using data that I wish to subset by
> gender and race (both dichotomous measures). I am running the following
> commands:
>
> 'by gender race, sort: ologit depvar indvars'
>
> At that point, I get the correct output of 4 tables for the proper
> combinations of the sorted variables (0,0)(0,1)(1,0)(1,1); however, when
> I try to run marginal effects as:
>
> 'mfx, predict(p outcome(1))' (my depvar is 3 ordered categories)
>
> I get the error:
>
> 'no observations
> r(2000);'

As a sidelight, you might want to use mfx2 or margeff instead of mfx 
when running commands like oprobit.  Both will give you the marginal 
effects for all the outcomes at once, instead of having to give a 
command for each outcome.  Both also give you the option of storing 
the results, which can be handy for creating tables.  Both are 
available from SSC. Slightly tweaking Kit's example,

clear all
set more off
sysuse auto
drop if rep78==1
su price,mean
g hip = price>r(mean)
forv p=0/1 {
        forv f=0/1 {
                di in r _n "High price==`p', Foreign==`f'"
                oprobit rep78 turn mpg if hip==`p' & foreign==`f', nolog
                mfx, predict(p outcome(#1))
                mfx2, nolog
        }
}

If I ever get ambitious, I may try to update mfx2 so it works with 
factor variables and the margins command.  Having to give separate 
commands for each outcome is kind of a pain and it makes the output 
harder to read.

-------------------------------------------
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
HOME:   (574)289-5227
EMAIL:  Richard.A.Williams.5@ND.Edu
WWW:    http://www.nd.edu/~rwilliam

*
*   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/