Exporting Marginal effects from stata to excel/word.

[Palana Bhatia]

Hello All,

I am beginner to stata, trying to learn. For my dataset I am using Oprobit model, to export marginal effects I have used esttab, and outreg2 however for both I am unable to export margins everytime it exports the coefficients from oprobit model.
I have also tried using asdoc but asdoc is giving me the entire table where in I only want coefficient with pvalues sidewaves


Thanks
Palana

[Andrew Musau]

You need the -post- option of margins to overwrite the previous estimation results.

margins, dydx(*) post

[Maarten Buis]

Here is a way of doing so with the new etable command.

. // open example data
. sysuse auto, clear
(1978 automobile data)

.
. // use some sensible units
. replace weight = weight*0.00045359237
variable weight was int now float
(74 real changes made)

. label var weight "weight (1000kg)"

. replace price = price/1000
variable price was int now float
(74 real changes made)

. label var price "price (1000$)"

.
. // estimate the model
. oprobit rep78 price weight

Iteration 0:   log likelihood = -93.692061  
Iteration 1:   log likelihood = -84.869304  
Iteration 2:   log likelihood = -84.824693  
Iteration 3:   log likelihood = -84.824643  
Iteration 4:   log likelihood = -84.824643  

Ordered probit regression                               Number of obs =     69
                                                        LR chi2(2)    =  17.73
                                                        Prob > chi2   = 0.0001
Log likelihood = -84.824643                             Pseudo R2     = 0.0946

------------------------------------------------------------------------------
       rep78 | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
       price |   .1317976   .0548115     2.40   0.016     .0243691    .2392262
      weight |  -1.927944     .46756    -4.12   0.000    -2.844345   -1.011543
-------------+----------------------------------------------------------------
       /cut1 |  -3.945445   .6535409                     -5.226361   -2.664528
       /cut2 |  -3.063049   .6031377                     -4.245177    -1.88092
       /cut3 |  -1.641716   .5561489                     -2.731748   -.5516842
       /cut4 |  -.6576951   .5204988                     -1.677854    .3624639
------------------------------------------------------------------------------

.
. // compute marginal effects
. margins, dydx(*) post

Average marginal effects                                    Number of obs = 69
Model VCE: OIM

dy/dx wrt: price weight

1._predict: Pr(rep78==1), predict(pr outcome(1))
2._predict: Pr(rep78==2), predict(pr outcome(2))
3._predict: Pr(rep78==3), predict(pr outcome(3))
4._predict: Pr(rep78==4), predict(pr outcome(4))
5._predict: Pr(rep78==5), predict(pr outcome(5))

------------------------------------------------------------------------------
             |            Delta-method
             |      dy/dx   std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
price        |
    _predict |
          1  |  -.0087722   .0059431    -1.48   0.140    -.0204204     .002876
          2  |  -.0178039   .0086162    -2.07   0.039    -.0346914   -.0009164
          3  |   -.017579   .0078997    -2.23   0.026    -.0330621   -.0020959
          4  |   .0171142   .0073932     2.31   0.021     .0026238    .0316045
          5  |   .0270409   .0116786     2.32   0.021     .0041513    .0499305
-------------+----------------------------------------------------------------
weight       |
    _predict |
          1  |     .12832    .076828     1.67   0.095      -.02226    .2789001
          2  |   .2604368   .0867301     3.00   0.003      .090449    .4304246
          3  |   .2571468   .0774315     3.32   0.001     .1053839    .4089097
          4  |  -.2503475   .0698052    -3.59   0.000    -.3871632   -.1135317
          5  |  -.3955562   .1041235    -3.80   0.000    -.5996344   -.1914779
------------------------------------------------------------------------------

.
. // create a table from the marginal effects
. etable, showeq

------------------------------
                          1   
------------------------------
price (1000$)                 
  _predict                    
    1                   -0.009
                       (0.006)
    2                   -0.018
                       (0.009)
    3                   -0.018
                       (0.008)
    4                    0.017
                       (0.007)
    5                    0.027
                       (0.012)
weight (1000kg)               
  _predict                    
    1                    0.128
                       (0.077)
    2                    0.260
                       (0.087)
    3                    0.257
                       (0.077)
    4                   -0.250
                       (0.070)
    5                   -0.396
                       (0.104)
Number of observations      69
------------------------------

.
. // put the outcome categories in the columns
. collect layout (coleq#result[_r_b _r_se]) ///
>                (colname[]#stars[value])

Collection: ETable
      Rows: coleq#result[_r_b _r_se]
   Columns: colname[]#stars[value]
   Table 1: 4 x 5

-------------------------------------------------------
                                _predict               
                   1       2       3       4       5   
-------------------------------------------------------
price (1000$)    -0.009  -0.018  -0.018   0.017   0.027
                (0.006) (0.009) (0.008) (0.007) (0.012)
weight (1000kg)   0.128   0.260   0.257  -0.250  -0.396
                (0.077) (0.087) (0.077) (0.070) (0.104)
-------------------------------------------------------

.                            
. // give those columns names                        
. collect label levels colname ///
>     1._predict "Poor"        ///
>     2._predict "Fair"        ///
>     3._predict "Average"     ///
>     4._predict "Good"        ///
>     5._predict "Excellent"

. collect preview

---------------------------------------------------------
                  Poor    Fair  Average   Good  Excellent
                                                         
---------------------------------------------------------
price (1000$)    -0.009  -0.018  -0.018   0.017     0.027
                (0.006) (0.009) (0.008) (0.007)   (0.012)
weight (1000kg)   0.128   0.260   0.257  -0.250    -0.396
                (0.077) (0.087) (0.077) (0.070)   (0.104)
---------------------------------------------------------

.
. // add the number of observations
. collect addtags colname[1._predict], fortags(result[N])
(1 items changed in collection ETable)

. collect layout (coleq#result[_r_b _r_se] result[N]) ///
>                (colname[]#stars[value])

Collection: ETable
      Rows: coleq#result[_r_b _r_se] result[N]
   Columns: colname[]#stars[value]
   Table 1: 5 x 5

----------------------------------------------------------------
                         Poor    Fair  Average   Good  Excellent
                                                                
----------------------------------------------------------------
price (1000$)           -0.009  -0.018  -0.018   0.017     0.027
                       (0.006) (0.009) (0.008) (0.007)   (0.012)
weight (1000kg)          0.128   0.260   0.257  -0.250    -0.396
                       (0.077) (0.087) (0.077) (0.070)   (0.104)
Number of observations      69                                  
----------------------------------------------------------------

You can now export the table to your prefered system using collect export

[Palana Bhatia]

Andrew Musau I have tried margins, dydx(*) post. Still when I visit the file it shows the coefficient from the probit model.
Thank you so much I will try this Maarten Buis.

[Andrew Musau]

You don't show the actual commands and Stata output, but you will need to reference the active results

margins, dydx(*) post
esttab . using myfile.rtf, replace

or store the estimates and refer to them

margins, dydx(*) post
est sto m1
esttab m1 using myfile.rtf, replace