st: AW: Re: AW: treatment effect estimation with an ordinal 1st step and a continuous 2nd step

["Martin Weiss" <martin.weiss1@gmx.de>]

<> 


The official example seems to run into the same problem:


*************
cmp setup
webuse womenwk, clear
gen selectvar = wage<.
gen wage3 = (wage > 10)+(wage > 30) if wage < .
cmp (wage3 = education age) (selectvar = married children education age), ind(selectvar*$cmp_oprobit $cmp_probit) qui
*************



HTH
Martin

-----Ursprüngliche Nachricht-----
Von: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] Im Auftrag von ???
Gesendet: Samstag, 17. Juli 2010 22:17
An: "Martin Weiss"
Cc: statalist@hsphsun2.harvard.edu
Betreff: st: Re: AW: treatment effect estimation with an ordinal 1st step and a continuous 2nd step

Dear Martin,

Thanks for your comment.
It was helpful.

However, the following syntax

xi: cmp (y1  = x1) (y2 = x2 y1), ind(5 1)

where y1 = ordinal numbers like 1, 2, 3, 4
          y2 = continuous

gives an error sign  like "matrix___00000B not found"

And the program shows (y1 = x1) result only, which is exactly identical to 
the outcome from an (oprobit y1 x1) regression.

Thanks in advance.

Jaemin





----- Original Message ----- 
From: "Martin Weiss" <martin.weiss1@gmx.de>
To: <statalist@hsphsun2.harvard.edu>
Sent: Saturday, July 17, 2010 10:45 PM
Subject: st: AW: treatment effect estimation with an ordinal 1st step and a 
continuous 2nd step


>
> <>
>
> Try
>
> *************
> ssc d cmp
> *************
>
>
>
> HTH
> Martin
>
> -----Ursprüngliche Nachricht-----
> Von: owner-statalist@hsphsun2.harvard.edu
> [mailto:owner-statalist@hsphsun2.harvard.edu] Im Auftrag von ???
> Gesendet: Samstag, 17. Juli 2010 15:29
> An: statalist@hsphsun2.harvard.edu
> Betreff: st: treatment effect estimation with an ordinal 1st step and a
> continuous 2nd step
>
> Hi all!
>
> I want to run a treatment effects model.
>
> In my case, the 1st step dependent var. is ordinal (for exmaple, 
> "perfectly
> not matched", "not matched", "matched", and "perfectly matched"), and the
> 2nd step dependent var. is continuous (for example, wage in log).
> I have run the similar model using "treatreg" command if the 1st step
> depednat is binary.
>
> Is there any command working in this case.
>
> I found "mtreatreg" is available if
> (1) 1st step dependent var. is multivariate and
> (2) 2nd step DV is continuous
>
> Of course, 1st step ordinal treatment variable should be shown in the 2nd
> step equation as an independent variable explicitly.
>
> P.S.: I don't think running "oprobit" as the 1st step, and insertting IMR
> from it as independent variable into the 2nd step OLS regression is valid.
>
> Thanks,
> jaemin
>
> *
> *   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/