error obtaining scores for robust variance

[Fred Lee]

mixed y cl.x##c.w age edu i.ind || id:cl.x, vce(robust) cov(exc)

here is the data:
https://www.dropbox.com/s/oef3hk8ui8rquu1/test.dta?dl=0

how to solve this?

[William Lisowski]

Different symptoms but the same essential problem as your previous one discussed at

https://www.statalist.org/forums/for...ed-not-concave

Run the following experiment.

regress y i.industry i.id
regress y i.id i.industry

The first regression tells us

. regress y i.industry i.id
note: 21.id omitted because of collinearity
note: 55.id omitted because of collinearity
note: 61.id omitted because of collinearity
note: 66.id omitted because of collinearity
note: 71.id omitted because of collinearity
note: 75.id omitted because of collinearity
note: 76.id omitted because of collinearity

The second regression tells us

. regress y i.id i.industry
note: 4.industry omitted because of collinearity
note: 5.industry omitted because of collinearity
note: 6.industry omitted because of collinearity
note: 7.industry omitted because of collinearity
note: 8.industry omitted because of collinearity
note: 9.industry omitted because of collinearity
note: 11.industry omitted because of collinearity

which - along with 3.industry which is omitted as the base category - means all categories of industry are dropped. That is, the variable industry is perfectly predicted by id and adds no information to that which is already given by id.

The complicated formulation of the mixed model hid this fact from you.

Your model is flawed by including both industry and id. You must omit one or the other.

If you omit industry the model completes as expected.

. mixed y cl.x##c.w age edu || id:cl.x, vce(robust) cov(exc)

Performing EM optimization: 

Performing gradient-based optimization: 

Iteration 0:   log pseudolikelihood =  -233.6067  
Iteration 1:   log pseudolikelihood = -232.37291  
Iteration 2:   log pseudolikelihood = -232.37141  
Iteration 3:   log pseudolikelihood = -232.37103  
Iteration 4:   log pseudolikelihood = -232.37093  
Iteration 5:   log pseudolikelihood = -232.37091  
Iteration 6:   log pseudolikelihood = -232.37091  

Computing standard errors:

Mixed-effects regression                        Number of obs     =        474
Group variable: id                              Number of groups  =         76

                                                Obs per group:
                                                              min =          3
                                                              avg =        6.2
                                                              max =          7

                                                Wald chi2(5)      =      18.04
Log pseudolikelihood = -232.37091               Prob > chi2       =     0.0029

                                    (Std. Err. adjusted for 76 clusters in id)
------------------------------------------------------------------------------
             |               Robust
           y |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
           x |
         L1. |  -.2645623   .2294345    -1.15   0.249    -.7142456     .185121
             |
           w |   .0846382   .1984724     0.43   0.670    -.3043605    .4736369
             |
    cL.x#c.w |   .0442101   .0628712     0.70   0.482    -.0790151    .1674353
             |
         age |    .002277   .0077217     0.29   0.768    -.0128572    .0174112
         edu |   .0516818   .0635339     0.81   0.416    -.0728423     .176206
       _cons |   3.840026   .8525146     4.50   0.000     2.169128    5.510924
------------------------------------------------------------------------------

------------------------------------------------------------------------------
                             |               Robust           
  Random-effects Parameters  |   Estimate   Std. Err.     [95% Conf. Interval]
-----------------------------+------------------------------------------------
id: Exchangeable             |
              var(L.x _cons) |   .0102613   .0020181      .0069791    .0150872
              cov(L.x,_cons) |   .0102613   .0020181       .006306    .0142167
-----------------------------+------------------------------------------------
               var(Residual) |   .1124081   .0269964      .0702052    .1799806
------------------------------------------------------------------------------