What could have I possibly done wrong?

[Hanna Hi]

Hello,
I just did a regression analysis and displayed a scatterplot and the points are I can't see no pattern for nearly all the variables.
Maybe its just like that but I have the strong feeling that I did something wrong..
Do you have any idea what could possibly cause that?

[David Benson]

Hi Hanna Hi ,

Sorry I am jumping into this late. I created a video on how to use dataex and code delimiters (obviously, you don't need to watch since you figured it out and successfully posted it here). I'm listing it here so future readers can find it. On YouTube it's at https://youtu.be/bXfaRCAOPbI

[Carlo Lazzaro]

Hanna:
- the number of predictors to be included in the right-hand side of your regression equation should be consistent with the aim of giving a fair and true view of the data generating process;
- even in your original regression model the R-sq was low (and significance was due to the large sample size);
- as expected, re-running the post estimation tests on a larger sample size gives back different results (see and compare the degrees of freedom of the F-tests).

[Hanna Hi]

I just compared your output with mine and I have different results.
Also for the first test (using the same command) stata displays this:
F(3, 5446) = 1.88
Prob > F = 0.1312

[Hanna Hi]

Carlo Lazzaro Thanks a lot for your help!
Can I just repeat if I understood it correct?
So by too many categorical variables, you mean that I should probably take less control variables?
R-squared is too low.. but that is not really something I can do about because it refers to the whole model, is that correct?
About the two tests you did, that means that their aren't omitted variables and there is also no heteroskedasticity, right?
Is there any possibility to improve my results?

[Carlo Lazzaro]

Hanna:
at a first glance, your regression suffers an excess of categorical predictors:

. regress V130 Womansright_index V240 V248 V239 V115 V147 V24 V85 V23

      Source |       SS           df       MS      Number of obs   =        93
-------------+----------------------------------   F(9, 83)        =      0.53
       Model |  2.11822605         9   .23535845   Prob > F        =    0.8524
    Residual |  37.1936019        83  .448115686   R-squared       =    0.0539
-------------+----------------------------------   Adj R-squared   =   -0.0487
       Total |   39.311828        92  .427302478   Root MSE        =    .66941

-----------------------------------------------------------------------------------
             V130 |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------+----------------------------------------------------------------
Womansright_index |  -.1009144   .3927768    -0.26   0.798    -.8821316    .6803028
             V240 |   .1016473   .1495953     0.68   0.499    -.1958917    .3991862
             V248 |   .0837466    .054947     1.52   0.131    -.0255407    .1930339
             V239 |  -.0358435   .0633795    -0.57   0.573    -.1619028    .0902157
             V115 |   .0249463   .0918601     0.27   0.787    -.1577596    .2076523
             V147 |   .1121565   .1581337     0.71   0.480     -.202365    .4266781
              V24 |  -.1023259   .2083322    -0.49   0.625    -.5166903    .3120385
              V85 |   .0955919    .133333     0.72   0.475    -.1696022    .3607859
              V23 |  -.0062662   .0361575    -0.17   0.863    -.0781821    .0656496
            _cons |   2.078786   .7702723     2.70   0.008     .5467454    3.610827
-----------------------------------------------------------------------------------

. estat ovtest

Ramsey RESET test using powers of the fitted values of V130
       Ho:  model has no omitted variables
                  F(3, 80) =      0.95
                  Prob > F =      0.4205

. estat hettest

Breusch-Pagan / Cook-Weisberg test for heteroskedasticity
         Ho: Constant variance
         Variables: fitted values of V130

         chi2(1)      =     0.05
         Prob > chi2  =   0.8153

. estat vif

    Variable |       VIF       1/VIF 
-------------+----------------------
        V239 |      1.50    0.667012
        V115 |      1.37    0.728434
        V248 |      1.26    0.792934
         V24 |      1.15    0.868171
Womansrigh~x |      1.14    0.875247
        V147 |      1.13    0.881785
        V240 |      1.13    0.884258
         V85 |      1.12    0.892015
         V23 |      1.12    0.893010
-------------+----------------------
    Mean VIF |      1.21

Actually, despite -regress postestimate- tests do not report any problem, the R-sq is simply too low (and non-significant) to make your model outperforming a mean estimate of the dependent variable:

. mean V130

Mean estimation                   Number of obs   =         97

--------------------------------------------------------------
             |       Mean   Std. Err.     [95% Conf. Interval]
-------------+------------------------------------------------
        V130 |   2.927835   .0659445      2.796936    3.058734
--------------------------------------------------------------

[Carlo Lazzaro]

Hanna:
exactly!
Very well done!
Thanks for your efforts.

[Hanna Hi]

That's what you meant, isn't it?

[Hanna Hi]

dataex V130 Womansright_index V240 V248 V239 V115 V147 V24 V85 V23 in 1/100

copy starting from the next line -----------------------

* Example generated by -dataex-. To install: ssc install dataex
clear
input int V130 float Womansright_index int(V240 V248 V239 V115 V147 V24 V85 V23)
.  .3333333 0 7 6 2 2 0 2  8
3  .4166667 0 9 6 3 2 0 2  6
3  .4166667 1 7 6 3 3 0 2  9
3  .4166667 1 7 6 3 3 0 1  3
3  .6666667 1 7 6 3 2 0 1  9
3 .16666667 1 7 6 2 3 0 2  5
3        .5 1 9 5 2 3 0 2  8
3       .25 0 7 5 3 3 0 2  8
3       .25 0 7 5 3 2 0 1  6
3  .4166667 1 6 4 3 3 0 1  4
4  .5833333 1 9 6 3 2 0 1  8
3  .3333333 0 7 4 2 2 0 1  4
3  .5833333 1 9 8 2 2 0 3  8
2 .16666667 0 7 7 4 2 0 1  7
4  .3333333 1 9 7 4 2 0 1 10
3  .4166667 1 9 6 2 2 0 1  9
4  .5833333 1 9 8 3 3 . 1  7
1  .4166667 0 9 5 3 3 0 2 10
3  .6666667 1 7 5 4 2 1 1  8
2  .5833333 0 7 3 4 2 0 1  4
4  .5833333 0 7 5 4 2 0 1  7
3  .5833333 1 4 4 4 2 0 1  8
4 .16666667 0 7 4 4 2 0 1  9
4  .5833333 1 7 5 3 2 1 3  5
4        .5 0 5 3 4 2 0 1 10
3       .75 1 7 3 4 3 0 1  8
3  .3333333 1 7 4 4 2 . 1  7
3  .3333333 0 7 4 4 2 0 1  8
3        .5 1 9 5 4 2 0 1  9
3        .5 0 9 4 3 2 0 3  3
3        .5 1 7 3 3 3 0 1 10
2        .5 0 9 3 3 2 0 1  6
3        .5 0 7 6 4 2 0 1  8
3        .5 0 7 6 4 3 . 1  9
2        .5 1 7 5 3 2 0 1  9
2  .5833333 1 5 5 4 2 0 1  6
3  .6666667 1 9 5 4 2 0 1 10
3  .6666667 1 7 5 4 2 1 1 10
3  .5833333 1 7 5 4 2 0 1  8
2  .5833333 1 7 5 4 2 0 1  5
3        .5 0 7 4 4 2 0 1  9
2        .5 0 7 4 4 3 1 1  9
3        .5 0 5 6 4 2 0 1 10
3        .5 0 8 5 4 2 0 1  8
2        .5 1 9 5 4 2 0 1  9
3        .5 1 7 5 4 2 0 1  5
2        .5 1 5 5 3 2 0 1  7
3        .5 1 7 5 4 2 0 1  6
4  .4166667 0 8 5 2 2 0 2  5
.       .25 0 9 8 3 2 0 3  5
3        .5 1 7 5 3 2 0 1  9
3        .5 1 5 5 4 2 0 1  9
3        .5 0 7 6 3 2 0 1  9
2        .5 0 7 5 3 2 0 1  6
3        .5 1 7 3 3 2 . 3  5
3  .5833333 1 5 4 4 3 0 3  9
3        .5 0 5 5 4 2 0 1  5
3        .5 1 7 4 3 2 0 1 10
3  .5833333 1 7 4 4 3 0 1  5
3        .5 0 7 4 3 2 0 1  7
3  .4166667 1 9 6 1 2 0 3  7
3        .5 0 5 5 2 3 0 1  7
3  .6666667 0 9 8 3 3 0 1  7
4  .9166667 1 9 5 3 2 0 2  5
3  .6666667 0 9 8 1 2 0 1 10
3       .75 0 9 8 1 2 0 1  9
3  .6666667 0 9 4 4 3 0 1  7
3  .4166667 1 7 4 4 2 0 1  1
3  .6666667 1 9 8 3 3 0 1  8
4  .3333333 1 7 6 2 3 0 1  4
1  .3333333 1 5 5 2 2 0 1  6
3 .08333334 0 7 6 3 3 1 1  9
4  .6666667 1 9 6 2 3 0 1  6
3  .3333333 1 9 5 3 2 0 1  5
3  .6666667 1 6 3 2 2 0 1  7
3         0 0 9 7 2 2 0 1  8
3  .9166667 1 7 6 3 3 0 1  5
3         0 0 7 7 3 3 1 1  5
3        .5 1 9 8 3 3 1 1  6
4  .3333333 1 7 6 2 3 0 1  5
3  .5833333 0 9 5 3 2 1 1 10
4  .5833333 1 9 6 3 2 0 1 10
3  .6666667 1 9 7 1 2 0 1 10
3       .25 1 9 7 4 3 0 1 10
3  .6666667 0 9 7 3 2 0 1  8
3 .16666667 1 9 7 3 3 1 2  7
3       .25 1 9 7 1 3 0 1  6
3       .75 1 9 6 1 2 0 1  4
4  .3333333 1 9 8 3 3 0 1 10
3  .9166667 0 9 5 4 2 1 1 10
3 .16666667 0 9 7 2 3 1 3  6
2  .3333333 0 9 5 3 2 1 1  8
2       .75 1 9 6 2 2 0 1  8
2  .4166667 0 5 7 3 2 1 1  9
2        .5 0 5 6 3 2 0 1  9
. .16666667 1 9 4 2 2 0 1  6
1  .5833333 1 9 9 3 2 0 1  8
3  .3333333 0 9 7 3 3 1 1  7
3         0 1 7 6 2 2 0 1  8
2       .75 1 5 7 2 3 0 1  9
end
label values V130 V130
label def V130 1 "Very bad", modify
label def V130 2 "Fairly bad", modify
label def V130 3 "Fairly good", modify
label def V130 4 "Very good", modify
label values V240 V240
label def V240 0 "Male", modify
label def V240 1 "Female", modify
label values V248 V248
label def V248 4 "Incomplete secondary school: technical/ vocational type", modify
label def V248 5 "Complete secondary school: technical/ vocational type", modify
label def V248 6 "Incomplete secondary school: university-preparatory type", modify
label def V248 7 "Complete secondary school: university-preparatory type", modify
label values V239 V239
label def V239 3 "Third step", modify
label def V239 4 "Fourth step", modify
label def V239 5 "Fifth step", modify
label def V239 6 "Sixth step", modify
label def V239 7 "Seventh step", modify
label def V239 8 "Eigth step", modify
label values V115 V115
label def V115 1 "None at all", modify
label def V115 2 "Not very much", modify
label def V115 3 "Quite a lot", modify
label def V115 4 "A great deal", modify
label values V147 V147
label def V147 2 "Not a religious person", modify
label def V147 3 "A religious person", modify
label values V24 V24
label def V24 0 "no trust", modify
label def V24 1 "trust", modify
label values V85 V85
label def V85 1 "Would never", modify
label def V85 2 "Might do", modify
label def V85 3 "Have done", modify
label values V23 V23
label def V23 1 "Completely dissatisfied", modify
label def V23 3 "3", modify
label def V23 4 "4", modify
label def V23 5 "5", modify
label def V23 6 "6", modify
label def V23 7 "7", modify
label def V23 8 "8", modify

copy up to and including the previous line ------------------

Listed 100 out of 11400 observations

.

[Hanna Hi]

dataex V130 Womansright_index V240 V248 V239 V115V147 V24 V85 V23 in 1/100

[Carlo Lazzaro]

Hanna:
just type:

dataex V130 Womansright_index V240 V248 V239 V115V147 V24 V85 V23 in 1/100

then copy and paste on the General forum.

[Hanna Hi]

Carlo LazzaroI still don't get how to export it. Is there an other command? Or do I have to type in -dataex- somewhere here?

[Hanna Hi]

Did I understood it correct that you are basically saying that there might be a better way to display the scatterplot but there is not really an effect to see anyways?I also tried if a ordered logistic regression fits better but there wasn't much of a difference. Do you think that would make more sense?
And the other variables are things like, income, education social trust, religiosity etc. I'm really surprised why there seems to be no effect at all.

[Carlo Lazzaro]

Hanna:
yes, and -dataex- is the answer.

[Hanna Hi]

Carlo Lazzaro

is there an other way to export it than safe it as a picture?