st: Compute intraclass correlation coefficient after xtmelogit

[Raquel Rangel de Meireles Guimarães <raquelrguimadem@gmail.com>]

Hi all,

Could you please give me and advice on how to compute the intraclass 
correlation coefficient after xtmelogit varying intercept model?

It seems that sd_resid is not reported...

Below you may find the output:

. xtmelogit excluido_leitura masculino branco pardo atrasado nse_transf 
c_nse_escola c_atraso_escola capitalcultural_
> transf ///
> envolvimento_transf motivacao_transf || escola: , or laplace

Refining starting values:

Iteration 0:   log likelihood = -1123816,5
Iteration 1:   log likelihood = -1119658,1
Iteration 2:   log likelihood = -1119658,1  (backed up)

Performing gradient-based optimization:

Iteration 0:   log likelihood = -1119658,1
Iteration 1:   log likelihood =   -1119378
Iteration 2:   log likelihood = -1119371,6
Iteration 3:   log likelihood = -1119371,6

Mixed-effects logistic regression               Number of obs      =   
2102433
Group variable: escola                          Number of groups   =     
37300

                                                Obs per group: min 
=         1
                                                               avg 
=      56,4
                                                               max 
=       518

Integration points =   1                        Wald chi2(10)      =  
98039,45
Log likelihood = -1119371,6                     Prob > chi2        =    
0,0000

------------------------------------------------------------------------------
excluido_l~a | Odds Ratio   Std. Err.      z    P>|z|     [95% Conf. 
Interval]
-------------+----------------------------------------------------------------
   masculino |   1,486187   ,0050551   116,49   0,000     1,476312    
1,496128
      branco |   ,7596705   ,0040705   -51,30   0,000     ,7517342    
,7676907
       pardo |   ,6731994   ,0034304   -77,66   0,000     ,6665094    
,6799565
    atrasado |    2,03682   ,0077668   186,56   0,000     2,021654      
2,0521
  nse_transf |   1,053706   ,0015861    34,75   0,000     1,050602     
1,05682
c_nse_esco~f |   ,5189814    ,003619   -94,06   0,000     ,5119365    
,5261232
c_atraso_e~a |   ,9864226   ,0229888    -0,59   0,557     ,9423789    
1,032525
capitalcul~f |   ,9296524   ,0014114   -48,05   0,000     ,9268902    
,9324228
envolvimen~f |   ,8468481   ,0013563  -103,80   0,000      ,844194    
,8495105
motivacao_~f |   1,008081   ,0013549     5,99   0,000     1,005428     
1,01074
------------------------------------------------------------------------------

------------------------------------------------------------------------------
  Random-effects Parameters  |   Estimate   Std. Err.     [95% Conf. 
Interval]
-----------------------------+------------------------------------------------
escola: Identity             |
                   sd(_cons) |   ,6075827   ,0033232      ,6011041    
,6141311
------------------------------------------------------------------------------
LR test vs. logistic regression: chibar2(01) = 70351,01 Prob>=chibar2 = 
0,0000

Note: log-likelihood calculations are based on the Laplacian approximation.

Thank you very much.

Best,

Raquel

--
Raquel Rangel de Meireles Guimarães
MA Student International&  Comparative Education
School of Education, Stanford University
http://stanford.academia.edu/RaquelGuimaraes

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/