Chow test statistic

[Yatharth Garg]

Hi everyone! I tried doing a chow test using -test- and the following command:

. suest Brahmin UC OBC Dalit Adivasi, cluster(STATE_DISTID)

Simultaneous results for Brahmin, UC, OBC, Dalit, Adivasi

                                                Number of obs     =     79,325

                            (Std. Err. adjusted for 370 clusters in STATE_DISTID)
---------------------------------------------------------------------------------
                |               Robust
                |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
----------------+----------------------------------------------------------------
Brahmin_mean    |
logurbanisation |   .0962082   .0682512     1.41   0.159    -.0375617     .229978
      EDUCATION |   .0710476   .0059827    11.88   0.000     .0593217    .0827736
     OCCUPATION |  -.0056143   .0007989    -7.03   0.000    -.0071801   -.0040486
          URBAN |   .3803489   .0731118     5.20   0.000     .2370523    .5236455
           RO_3 |  -.1307126   .0551441    -2.37   0.018    -.2387931   -.0226322
            RO5 |    .006796   .0020625     3.30   0.001     .0027536    .0108385
            mar |  -.0032889   .0545881    -0.06   0.952    -.1102796    .1037017
           time |   .1482965   .0494396     3.00   0.003     .0513967    .2451962
          _cons |   10.42243   .1790616    58.21   0.000     10.07147    10.77338
----------------+----------------------------------------------------------------
Brahmin_lnvar   |
          _cons |  -.4670615   .0481963    -9.69   0.000    -.5615245   -.3725985
----------------+----------------------------------------------------------------
UC_mean         |
logurbanisation |   .0672352   .0710858     0.95   0.344    -.0720905    .2065609
      EDUCATION |   .0609063   .0048479    12.56   0.000     .0514046     .070408
     OCCUPATION |  -.0056759   .0006935    -8.18   0.000    -.0070352   -.0043166
          URBAN |   .3897185   .0488907     7.97   0.000     .2938945    .4855424
           RO_3 |  -.0240296   .0453255    -0.53   0.596    -.1128659    .0648067
            RO5 |   .0056959   .0017892     3.18   0.001     .0021891    .0092027
            mar |  -.1139842    .032705    -3.49   0.000    -.1780849   -.0498835
           time |   .1698457   .0468221     3.63   0.000      .078076    .2616153
          _cons |   10.47544    .146294    71.61   0.000     10.18871    10.76217
----------------+----------------------------------------------------------------
UC_lnvar        |
          _cons |  -.4558201   .0598579    -7.62   0.000    -.5731395   -.3385007
----------------+----------------------------------------------------------------
OBC_mean        |
logurbanisation |   .1213767   .0505292     2.40   0.016     .0223414    .2204121
      EDUCATION |   .0415183   .0033229    12.49   0.000     .0350056    .0480311
     OCCUPATION |  -.0044012   .0004984    -8.83   0.000     -.005378   -.0034244
          URBAN |   .4143306   .0356688    11.62   0.000     .3444209    .4842402
           RO_3 |   .0059099   .0204806     0.29   0.773    -.0342315    .0460512
            RO5 |   .0026357   .0009078     2.90   0.004     .0008564     .004415
            mar |  -.1232065   .0257449    -4.79   0.000    -.1736656   -.0727475
           time |   .2058166    .029633     6.95   0.000      .147737    .2638962
          _cons |   10.49678   .0966064   108.66   0.000     10.30743    10.68612
----------------+----------------------------------------------------------------
OBC_lnvar       |
          _cons |  -.5106735   .0337939   -15.11   0.000    -.5769082   -.4444387
----------------+----------------------------------------------------------------
Dalit_mean      |
logurbanisation |   .0599791   .0636714     0.94   0.346    -.0648145    .1847728
      EDUCATION |   .0323772   .0043063     7.52   0.000      .023937    .0408174
     OCCUPATION |  -.0029499   .0006317    -4.67   0.000     -.004188   -.0017119
          URBAN |   .4792115    .057459     8.34   0.000     .3665939    .5918292
           RO_3 |   .0139581   .0248035     0.56   0.574    -.0346558     .062572
            RO5 |   .0027359   .0010757     2.54   0.011     .0006276    .0048442
            mar |  -.1224813   .0279643    -4.38   0.000    -.1772903   -.0676724
           time |   .2930478   .0382048     7.67   0.000     .2181678    .3679279
          _cons |   10.25986     .13977    73.41   0.000     9.985914     10.5338
----------------+----------------------------------------------------------------
Dalit_lnvar     |
          _cons |  -.6977499   .0406441   -17.17   0.000    -.7774108    -.618089
----------------+----------------------------------------------------------------
Adivasi_mean    |
logurbanisation |   .1301019   .1127704     1.15   0.249    -.0909239    .3511278
      EDUCATION |   .0377194   .0071847     5.25   0.000     .0236376    .0518013
     OCCUPATION |  -.0040962   .0014458    -2.83   0.005    -.0069299   -.0012625
          URBAN |   .6741165   .0834127     8.08   0.000     .5106306    .8376024
           RO_3 |  -.0177608   .0261091    -0.68   0.496    -.0689338    .0334121
            RO5 |   .0046253   .0019444     2.38   0.017     .0008142    .0084363
            mar |  -.1385599    .057765    -2.40   0.016    -.2517772   -.0253426
           time |   .1967442   .0699005     2.81   0.005     .0597418    .3337467
          _cons |   10.27268   .2418091    42.48   0.000     9.798742    10.74662
----------------+----------------------------------------------------------------
Adivasi_lnvar   |
          _cons |  -.5216917   .1057542    -4.93   0.000    -.7289661   -.3144173
---------------------------------------------------------------------------------

. test [Dalit_mean]logurbanisation=[OBC_mean]logurbanisation

 ( 1)  - [OBC_mean]logurbanisation + [Dalit_mean]logurbanisation = 0

           chi2(  1) =    0.83
         Prob > chi2 =    0.3631

The test statistic is chi2 however, generally the chow test statistic is F. I am concerned that I am doing or interpreting something incorrectly here.

[Joro Kolev]

You are testing whether the effect pf logurbanisation is the same across the two groups. Whether this is what you want to test only you know.

You can see this FAQ about the relationship between the F and Chi^2 test: https://www.stata.com/support/faqs/s...distributions/

[Yatharth Garg]

Hi, that is indeed what I want to test, but I want to do it through a chow test that is used for a subset of coefficients and not for all the coefficients. My question is, would it be appropriate to report chi2 when I am writing the results of a chow test?

[Joro Kolev]

Yes, of course.

When you rely on large sample results, as we do most of the time, the F and Chi^2 tests are interchangeable.

Hi, that is indeed what I want to test, but I want to do it through a chow test that is used for a subset of coefficients and not for all the coefficients. My question is, would it be appropriate to report chi2 when I am writing the results of a chow test?

[Yatharth Garg]

Okay, I understand. Thank you very much!