Advice on extracting data from regression and exporting into a matrix

Micha

Join Date: Apr 2014
Posts: 12

Advice on extracting data from regression and exporting into a matrix

27 Mar 2019, 14:36

Dear all,

I am trying to extract data from a pvar and mold it into a matrix. The output of the pvar is as follows

Code:

Final GMM Criterion Q(b) =      .156
Initial weight matrix: Identity
GMM weight matrix:     Robust
                                                   No. of obs      =     44475
                                                   No. of panels   =        38
                                                   Ave. no. of T   =  1170.395


------------------------------------------------------------------------------
             |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
value_dif    |
   value_dif |
         L1. |  -.0113901   .0330503    -0.34   0.730    -.0761674    .0533873
             |
    IV_c_dif |
         L1. |   1918.093   476.1653     4.03   0.000     984.8264     2851.36
             |
     val_dif |
         L1. |    .000953   .0005652     1.69   0.092    -.0001548    .0020608
             |
     sat_dif |
         L1. |   .0000536   .0005408     0.10   0.921    -.0010063    .0011136
             |
    qual_dif |
         L1. |  -.0008847   .0007642    -1.16   0.247    -.0023824     .000613
             |
    sent_pos |
         L1. |  -.0014811   .0003022    -4.90   0.000    -.0020735   -.0008887
             |
    sent_neg |
         L1. |   .0016537   .0004451     3.72   0.000     .0007813     .002526
             |
         fgc |
         L1. |  -.0273508   .0277591    -0.99   0.324    -.0817577     .027056
             |
         MVE |
         L1. |   1.32e-06   6.93e-07     1.91   0.057    -3.69e-08    2.68e-06
             |
          at |
         L1. |  -1.48e-06   7.05e-07    -2.10   0.036    -2.86e-06   -1.00e-07
             |
       adv_s |
         L1. |   4.14e-06   9.14e-06     0.45   0.651    -.0000138     .000022
-------------+----------------------------------------------------------------
IV_c_dif     |
   value_dif |
         L1. |   4.10e-07   4.44e-07     0.92   0.356    -4.61e-07    1.28e-06
             |
    IV_c_dif |
         L1. |  -.0403185   .0202629    -1.99   0.047    -.0800332   -.0006039
             |
     val_dif |
         L1. |   4.59e-09   3.98e-08     0.12   0.908    -7.34e-08    8.26e-08
             |
     sat_dif |
         L1. |  -4.87e-09   4.36e-08    -0.11   0.911    -9.03e-08    8.06e-08
             |
    qual_dif |
         L1. |  -6.98e-08   6.18e-08    -1.13   0.259    -1.91e-07    5.13e-08
             |
    sent_pos |
         L1. |   2.26e-08   2.48e-08     0.91   0.361    -2.59e-08    7.12e-08
             |
    sent_neg |
         L1. |  -2.50e-08   3.81e-08    -0.66   0.512    -9.97e-08    4.97e-08
             |
         fgc |
         L1. |  -.0000328   3.36e-06    -9.77   0.000    -.0000394   -.0000262
             |
         MVE |
         L1. |  -6.84e-11   8.14e-11    -0.84   0.400    -2.28e-10    9.11e-11
             |
          at |
         L1. |  -9.97e-11   3.92e-11    -2.54   0.011    -1.77e-10   -2.28e-11
             |
       adv_s |
         L1. |  -1.19e-09   9.44e-10    -1.26   0.209    -3.04e-09    6.65e-10

---------------------------------------------------------------------------

<from here repeating for all variables>

My matrix should look like the following

	value_dif		IV_c_dif		val_dif
value_dif			coefficient	***	coefficient	***
			t-statistic		t-statistic
IV_c_dif	coefficient	***			coefficient	***
	t-statistic				t-statistic
val_dif	coefficient	***	coefficient	***
	t-statistic		t-statistic

for all variables listed above.

I tried to get the coefficient and t-statistics via

Code:

estout, cells (b(star) t(par))

which works fine in extracting the data, but I am not sure on how to put it into a matrix.

Can anyone provide advice on how to achieve the matrix?

Any help would be most appreciated.

Best
Micha

Tags: None

Andrew Musau

Join Date: Oct 2014

Posts: 10284
#2

27 Mar 2019, 15:36

After running the gmm command, start with

Code:

mat Results= r(table) mat list Results

and then you can create other matrices based on the elements of this matrix.
1 like
Comment
Micha

Join Date: Apr 2014

Posts: 12
#3

30 Mar 2019, 09:07

Thanks for your reply Andrew! The command you mentioned you gives me an empty 1x1 matrix. Do i need to change the initial mat command?

Thanks again

Best
Micha
Comment

Andrew Musau

Join Date: Oct 2014
Posts: 10284

30 Mar 2019, 10:26

You should get a results matrix, similar to below

Code:

webuse docvisits, clear
gmm (docvis - exp({xb:private chronic female income} + {b0})),instruments(private chronic female age black hispanic) deriv(/xb= -1*exp({xb:} + {b0})) deriv(/b0 = -1*exp({xb:} + {b0})) twostep

Result:

Code:

GMM estimation

Number of parameters =   5
Number of moments    =   7
Initial weight matrix: Unadjusted                 Number of obs   =      4,412
GMM weight matrix:     Robust

------------------------------------------------------------------------------
             |               Robust
             |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     private |    .535335    .159904     3.35   0.001      .221929     .848741
     chronic |   1.090126   .0617659    17.65   0.000     .9690668    1.211185
      female |   .6636579   .0959885     6.91   0.000      .475524    .8517918
      income |   .0142855   .0027162     5.26   0.000     .0089618    .0196092
-------------+----------------------------------------------------------------
         /b0 |  -.5983477    .138433    -4.32   0.000    -.8696714    -.327024
------------------------------------------------------------------------------
Instruments for equation 1: private chronic female age black hispanic _cons

. mat Results= r(table)

.
. mat list Results

Results[9,5]
                xb:         xb:         xb:         xb:         b0:
           private     chronic      female      income       _cons
     b   .53533501   1.0901257   .66365793   .01428549  -.59834766
    se   .15990395   .06176589   .09598845   .00271624     .138433
     z   3.3478535   17.649315   6.9139351   5.2592952  -4.3222906
pvalue    .0008144   1.030e-69   4.714e-12   1.446e-07   .00001544
    ll   .22192902   .96906681   .47552402   .00896177  -.86967135
    ul   .84874101   1.2111847   .85179184   .01960922  -.32702396
    df           .           .           .           .           .
  crit    1.959964    1.959964    1.959964    1.959964    1.959964
 eform           0           0           0           0           0

.

Here the matrix is 9\(\times\)5.

Comment

Xavier Babu

Join Date: Jun 2023

Posts: 1
#5

11 Jun 2023, 17:29

I would like to create PVAR results in a matrix form using esttab. Please help me to find a code. My output is not using GMM, it is using the command PVAR, that means many depvars. Please help.
Comment
Andrew Musau

Join Date: Oct 2014

Posts: 10284
#6

14 Jun 2023, 04:28

Originally posted by Xavier Babu View Post

I would like to create PVAR results in a matrix form using esttab.

I do not use pvar from the Stata Journal (FAQ Advice #12). However, you can check whether the command stores results in matrices from either

Code:

return list

or

Code:

ereturn list

Then look at

Code:

help matrix list

to show us any such matrices. From here, it should be possible to illustrate how to use estout to output the contents of the matrices.
Comment

Announcement

Advice on extracting data from regression and exporting into a matrix

Comment

Comment

Comment

Comment

Comment