Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: RE: Re: Question about Stata


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: Re: Question about Stata
Date   Fri, 20 Mar 2009 12:24:00 -0000

Neither the original author nor the reviser of -dtobit- now works for
StataCorp. That's just factual information; it's not for me to say what
it does or does not imply. 

What's key, however, is that -dtobit- was never part of official Stata. 
It remains a user-written program whose author and then reviser happened
also to work for StataCorp (as it now is). 

Thus, contrary to some statements, there is no question of suppressing
or discontinuing it. It will remain archived and accessible and
downloadable in STB/SJ archives indefinitely. 

Nick 
n.j.cox@durham.ac.uk 

Kit Baum

Mykhailo,

I cannot test the capabilities of -dtobit- because that routine  
(updated for sg144_1, from SJ 4:3) does not run the example in its  
help file:

.  tobit mpg trunk weight, ll(17)

Tobit regression                                  Number of obs    
=         74
                                                   LR chi2(2)       
=      73.86
                                                   Prob > chi2      
=     0.0000
Log likelihood = -163.74787                       Pseudo R2        
=     0.1840

------------------------------------------------------------------------
------
          mpg |      Coef.   Std. Err.      t    P>|t|     [95% Conf.  
Interval]
------------- 
+----------------------------------------------------------------
        trunk |  -.1487203   .1477163    -1.01   0.317    -. 
4431875    .1457468
       weight |  -.0063328   .0008709    -7.27   0.000     -.008069    
-.0045966
        _cons |   41.90603   2.094632    20.01   0.000      
37.73045     46.0816
------------- 
+----------------------------------------------------------------
       /sigma |   3.822488   .3637359                      3.097394     
4.547582
------------------------------------------------------------------------
------
   Obs. summary:         18  left-censored observations at mpg<=17
                         56     uncensored observations
                          0 right-censored observations

.  dtobit
[_se] not found
r(111);


Apparently some changes made to -tobit- have broken -dtobit-.

However I would wonder whether you can what you need with the various  
options available on -mfx-, see -help tobit postestimation-. The  
default output provides


. mfx

Marginal effects after tobit
       y  = Fitted values (predict)
          =  20.738523
------------------------------------------------------------------------
------
variable |      dy/dx    Std. Err.     z    P>|z|  [    95%  
C.I.   ]      X
--------- 
+--------------------------------------------------------------------
    trunk |  -.1487203      .14772   -1.01   0.314  -.438239  . 
140798   13.7568
   weight |  -.0063328      .00087   -7.27   0.000   -.00804 -. 
004626   3019.46
------------------------------------------------------------------------
------

and other options are available. -mfx- saves everything it produces,  
and those matrices can be added to -estimates- and output with - 
esttab- or other table-making programs.

Presumably someone from StataCorp will have a look at -dtobit-, as it  
was written and modified by StataCorp employees (but is not part of  
official Stata).

On Mar 20, 2009, at 24:42 , msyrotenko wrote:

> I am working with dtobit command to receive the marginal effects.
>
> . dtobit, brief
>
------------------------------------------------------------------------
------
>          | Marginal Effects at Means
>           
> |--------------------------------------------------------------------
>          |   Latent           Unconditional        Conditional  
> on          Probability
>   Name   |  Variable     Expected Value     being Uncensored      
> Uncensored
> --------- 
> +--------------------------------------------------------------------
>    l_gdp |  2.1510068       .03051076         . 
> 22277649              .00370306
>    l_dist | -11.960954      -.16965905         -1.2387778             
> -.02059134
>      dev*| -.95325248      -.01298252        -.09792999            -. 
> 00158439
> --------- 
> +--------------------------------------------------------------------
> (*) dF/dx is for discrete change of dummy variable from 0 to 1
>
> Now I would like to extract the columns of this table separately  
> (Latent variable, Unconditional expected value; Conditional expected  
> Value, Probability Uncensored) Than I would like receive the general  
> table for different regressions but sepparately for: Latent  
> variable, Unconditional expected value, Conditional expected Value,  
> Probability Uncensored.
>
> using command "estimates store"        a few times
> and then esttab I receive the table only for Latent variable(first  
> column).
> So "estimates store extract" only first column of the initial table
> Is it possible to extract other than first column of the initial  
> table?
> What command can help me for this purpose?
>

*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index