[Felipe Damasceno]

I'm sorry.

I'm founding the same thing. The empty variables are not being suppressed and are showing as omitted in my table.

[Rich Goldstein]

as the FAQ requests, please show us exactly what you typed and exactly what Stata gave you in a CODE block - see the FAQ for how to use code blocks

[Emad Shehata]

* Example generated by -dataex-. To install: ssc install dataex
clear
input byte y float(x1 x2)
1  96.7   101
1  98.1 100.1
0   100   100
0 104.9  90.6
1 104.9  86.5
1 109.5  89.7
1 110.8  90.6
0 112.3  82.8
1 109.3  70.1
0 105.3  65.4
0 101.7  61.3
1  95.4  62.5
0  96.4  63.6
0  97.6  52.6
1 102.4  59.7
1 101.6  59.5
1 103.8  61.3
end
gen x3=x1
probit y x1 x3 x2
probit y x1 x3 x2, noomitted
probit y x1 x3 x2, noomitted vsquish

. probit y x1 x3 x2

note: x3 omitted because of collinearity
Iteration 0:   log likelihood = -11.517405 
Iteration 1:   log likelihood = -11.359142 
Iteration 2:   log likelihood =  -11.35912 
Iteration 3:   log likelihood =  -11.35912 

Probit regression                                 Number of obs   =         17
                                                  LR chi2(2)      =       0.32
                                                  Prob > chi2     =     0.8536
Log likelihood =  -11.35912                       Pseudo R2       =     0.0137

------------------------------------------------------------------------------
           y |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
          x1 |    .009611   .0615281     0.16   0.876    -.1109818    .1302039
          x3 |  (omitted)
          x2 |   .0096197   .0193082     0.50   0.618    -.0282237    .0474631
       _cons |  -1.497935   6.229788    -0.24   0.810    -13.70809    10.71222
------------------------------------------------------------------------------

 probit y x1 x3 x2, noomitted

note: x3 omitted because of collinearity
Iteration 0:   log likelihood = -11.517405 
Iteration 1:   log likelihood = -11.359142 
Iteration 2:   log likelihood =  -11.35912 
Iteration 3:   log likelihood =  -11.35912 

Probit regression                                 Number of obs   =         17
                                                  LR chi2(2)      =       0.32
                                                  Prob > chi2     =     0.8536
Log likelihood =  -11.35912                       Pseudo R2       =     0.0137

------------------------------------------------------------------------------
           y |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
          x1 |    .009611   .0615281     0.16   0.876    -.1109818    .1302039
             |
          x2 |   .0096197   .0193082     0.50   0.618    -.0282237    .0474631
       _cons |  -1.497935   6.229788    -0.24   0.810    -13.70809    10.71222
------------------------------------------------------------------------------

 probit y x1 x3 x2, noomitted vsquish

note: x3 omitted because of collinearity
Iteration 0:   log likelihood = -11.517405 
Iteration 1:   log likelihood = -11.359142 
Iteration 2:   log likelihood =  -11.35912 
Iteration 3:   log likelihood =  -11.35912 

Probit regression                                 Number of obs   =         17
                                                  LR chi2(2)      =       0.32
                                                  Prob > chi2     =     0.8536
Log likelihood =  -11.35912                       Pseudo R2       =     0.0137

------------------------------------------------------------------------------
           y |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
          x1 |    .009611   .0615281     0.16   0.876    -.1109818    .1302039
          x2 |   .0096197   .0193082     0.50   0.618    -.0282237    .0474631
       _cons |  -1.497935   6.229788    -0.24   0.810    -13.70809    10.71222
------------------------------------------------------------------------------

[Nick Cox]

On outreg and outreg2:note that they are more different than the similarity of names might suggest.

My #3 in http://www.statalist.org/forums/foru...-after-capture still holds so far as I can see.

[Richard Williams]

I'm founding the same thing. The empty variables are not being suppressed and are showing as omitted in my table.

To be more precise, it is three categories of the categorical independent variable that are showing up as empty. I wonder if this is a bug in Stata, i.e. sometimes categories of a factor variable are immune to the noempty/ noomitted options. Why they are showing up as empty I don't know; maybe supersmall Ns in those categories? Or perfect collinearity because of other variables in the model? You might try tabbing the dependent variable by the categorical independent variable.

I have moral objections to suggesting this, but I might try preceding the probit command with xi: Doing so will actually create dummies out of your categorical variable. Then see if the empty dummies get correctly omitted because of the noomitted option. That won't solve the problem but it will at least indicate that it is something about factor variables that is causing the problem.

Anyway, I have a feeling that the empty categories reflect some sort of collinearity problem, and because they occur in a factor variable the usual fixes aren't working.

[Aihong Li]

Hi all, I came across the same issue of using "noomitted", and it didn't work well in producing output tables of my probit model results. But I searched "help esttab" and found out the option "nobaselevels" to omit the omitted/baseline obs. For example my output code is "esttab m1 m2 m3 m4, b(2) eform cons pr2 aic bic mtitles compress nogap wide lines nobaselevels", which has worked well.

Although the post has been kind of history now, I still want to share the solution in case someone comes across this issue again.