Right. So, is there a way to either a) ask mfx to treat a scalar (or a matrix) like a variable or b) get mfx to treat the lincom return as a variable? If not, what's the appropriate syntax to use to get an interpretable value from lincom after using dprobit using mfx? I'm assuming such syntax exists and that others have used it successfully.....
M
________________________________________
From: [email protected] [[email protected]] On Behalf Of Martin Weiss [[email protected]]
Sent: January 21, 2010 3:41 PM
To: [email protected]
Subject: RE: st: dprobit and lincom
<>
" I know there's more than one error there"
Not quite, the only error in there is the -varlist()- option right at the
end, which expects, you guessed it, a variable, not a -scalar-...
HTH
Martin
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Melanee Thomas
Sent: Donnerstag, 21. Januar 2010 21:25
To: [email protected]
Subject: RE: st: dprobit and lincom
Well, what I have is as follows:
lincom highed+whighed
return list
estadd scalar lincom=r(estimate)
mfx, predict(xb) varlist(lincom)
I know there's more than one error there, but it's unclear to me how to get
mfx to grab the coefficient and se generated by lincom. That the above
doesn't work is intuitive to me, but what's not it how to grab the correct
info and run with it.
Thanks!
Melanee
________________________________________
From: [email protected]
[[email protected]] On Behalf Of Nick Cox
[[email protected]]
Sent: January 21, 2010 3:08 PM
To: [email protected]
Subject: RE: st: dprobit and lincom
Please show exactly what you typed to maximise your chance of helpful
comment on what did you wrong (if anything) and on what you should do.
This is longstanding advice in the FAQ, and for good reason.
Nick
[email protected]
Melanee Thomas
Thanks! I must admit that I had tried using mfx (I'm using Stata 10).
The problem is that I'm having difficulties getting Stata to recognize
the lincom result as the coefficient I'd like the marginal effect for.
I've tried adding the lincom result as a scalar, but I get an error
message indicating that too few variables are specified. If there's no
option outside mfx, what's the best way to get Stata to recognize what
it needs to generate the marginal effect for?
[email protected]] On Behalf Of Maarten buis
--- On Thu, 21/1/10, Melanee Thomas wrote:
> I have a question about dprobit and lincom. Specifically,
> when I run lincom after dprobit, the returned lincom result
> is a probit coefficient, rather than the dF/dx reported by
> dprobit. Is there an option I can use to transform the
> probit coefficient returned by lincom to the dF/dx reported
> by dprobit?
Depending on your version of Stata -mfx- (< 11) or -margins-
(11).
> Stated differently, is there a way to get lincom
> to deliver an unexponentiated result?
This is a common misunderstanding, the exponentiated coeficients
are only meaningful after -logit- but not after -probit-. So,
-dprobit- doesn't exponentiate coefficients.
*
* 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/
*
* 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/
*
* 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/
*
* 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/
