UCLA Academic Technology Services HomeServicesClassesContactJobs
Search

Stata Paper Examples
Using SAS Proc Mixed to Fit Multilevel Models, Hierarchical Models, and Individual Growth Models
by Judith Singer

This paper can be downloaded from Professor Singer's web site at http://gseweb.harvard.edu/%7Efaculty/singer/Papers/sasprocmixed.pdf . The Stata data files used are hsb12.dta and willett.dta.  

Note:  The xtmixed procedure is new to Stata 9.

Note on comparing the Stata and SAS results.  Stata reports results for fit statistics (e.g. AIC) in "smaller is better" form.  Since the printing of the article, SAS has started showing its fit statistics in "smaller is better" form as well.


This first set of examples use the hsb12.dta data file which you can obtain from within Stata like this

use http://www.ats.ucla.edu/stat/paperexamples/singer/hsb12, clear

Example on page 329.

xtmixed mathach || school: , variance

Performing EM optimization: 

Performing gradient-based optimization: 

Iteration 0:   log restricted-likelihood = -23558.397  
Iteration 1:   log restricted-likelihood = -23558.397  

Computing standard errors:

Mixed-effects REML regression                   Number of obs      =      7185
Group variable: school                          Number of groups   =       160

                                                Obs per group: min =        14
                                                               avg =      44.9
                                                               max =        67


                                                Wald chi2(0)       =         .
Log restricted-likelihood = -23558.397          Prob > chi2        =         .

------------------------------------------------------------------------------
     mathach |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
       _cons |   12.63697   .2443937    51.71   0.000     12.15797    13.11598
------------------------------------------------------------------------------

------------------------------------------------------------------------------
  Random-effects Parameters  |   Estimate   Std. Err.     [95% Conf. Interval]
-----------------------------+------------------------------------------------
school: Identity             |
                  var(_cons) |   8.614034   1.078805      6.739129    11.01056
-----------------------------+------------------------------------------------
               var(Residual) |   39.14832   .6606446      37.87466    40.46481
------------------------------------------------------------------------------
LR test vs. linear regression: chibar2(01) =   986.12 Prob >= chibar2 = 0.0000

estat ic

------------------------------------------------------------------------------
       Model |    Obs    ll(null)   ll(model)     df          AIC         BIC
-------------+----------------------------------------------------------------
           . |   7185           .    -23558.4      3     47122.79    47143.43
------------------------------------------------------------------------------

Example on page page 331.

xtmixed mathach meanses || school: , variance

Performing EM optimization: 

Performing gradient-based optimization: 

Iteration 0:   log restricted-likelihood = -23480.642  
Iteration 1:   log restricted-likelihood = -23480.642  

Computing standard errors:

Mixed-effects REML regression                   Number of obs      =      7185
Group variable: school                          Number of groups   =       160

                                                Obs per group: min =        14
                                                               avg =      44.9
                                                               max =        67


                                                Wald chi2(1)       =    263.15
Log restricted-likelihood = -23480.642          Prob > chi2        =    0.0000

------------------------------------------------------------------------------
     mathach |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     meanses |   5.863538    .361458    16.22   0.000     5.155094    6.571983
       _cons |   12.64944   .1492801    84.74   0.000     12.35685    12.94202
------------------------------------------------------------------------------

------------------------------------------------------------------------------
  Random-effects Parameters  |   Estimate   Std. Err.     [95% Conf. Interval]
-----------------------------+------------------------------------------------
school: Identity             |
                  var(_cons) |   2.638707   .4043388      1.954153    3.563067
-----------------------------+------------------------------------------------
               var(Residual) |   39.15708   .6608016      37.88312    40.47389
------------------------------------------------------------------------------
LR test vs. linear regression: chibar2(01) =   239.95 Prob >= chibar2 = 0.0000

estat ic

------------------------------------------------------------------------------
       Model |    Obs    ll(null)   ll(model)     df          AIC         BIC
-------------+----------------------------------------------------------------
           . |   7185           .   -23480.64      4     46969.28     46996.8
------------------------------------------------------------------------------

Example on page 335.

xtmixed mathach cses || school: cses, variance cov(un)

Performing EM optimization: 

Performing gradient-based optimization: 

Iteration 0:   log restricted-likelihood =  -23357.18  
Iteration 1:   log restricted-likelihood = -23357.118  
Iteration 2:   log restricted-likelihood = -23357.118  

Computing standard errors:

Mixed-effects REML regression                   Number of obs      =      7185
Group variable: school                          Number of groups   =       160

                                                Obs per group: min =        14
                                                               avg =      44.9
                                                               max =        67


                                                Wald chi2(1)       =    292.40
Log restricted-likelihood = -23357.118          Prob > chi2        =    0.0000

------------------------------------------------------------------------------
     mathach |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
        cses |   2.193192   .1282582    17.10   0.000      1.94181    2.444574
       _cons |   12.64934   .2445134    51.73   0.000      12.1701    13.12858
------------------------------------------------------------------------------

------------------------------------------------------------------------------
  Random-effects Parameters  |   Estimate   Std. Err.     [95% Conf. Interval]
-----------------------------+------------------------------------------------
school: Unstructured         |
                   var(cses) |   .6939724   .2807827      .3140105    1.533699
                  var(_cons) |   8.681651   1.079627      6.803763    11.07785
             cov(cses,_cons) |   .0507474   .4063922     -.7457667    .8472615
-----------------------------+------------------------------------------------
               var(Residual) |    36.7002   .6257441      35.49403    37.94736
------------------------------------------------------------------------------
LR test vs. linear regression:       chi2(3) =  1065.70   Prob > chi2 = 0.0000

Note: LR test is conservative and provided only for reference

estat ic

------------------------------------------------------------------------------
       Model |    Obs    ll(null)   ll(model)     df          AIC         BIC
-------------+----------------------------------------------------------------
           . |   7185           .   -23357.12      6     46726.24    46767.51
------------------------------------------------------------------------------

Example on page 337.

generate msesXcses = meanses*cses
generate secXcses = sector*cses

xtmixed mathach meanses sector cses msesXcses secXcses || school: cses, variance cov(un)

Performing EM optimization: 

Performing gradient-based optimization: 

Iteration 0:   log restricted-likelihood = -23252.888  
Iteration 1:   log restricted-likelihood = -23251.835  
Iteration 2:   log restricted-likelihood = -23251.834  

Computing standard errors:

Mixed-effects REML regression                   Number of obs      =      7185
Group variable: school                          Number of groups   =       160

                                                Obs per group: min =        14
                                                               avg =      44.9
                                                               max =        67


                                                Wald chi2(5)       =    746.22
Log restricted-likelihood = -23251.834          Prob > chi2        =    0.0000

------------------------------------------------------------------------------
     mathach |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     meanses |   5.339122   .3693012    14.46   0.000     4.615305    6.062939
      sector |   1.216671   .3063874     3.97   0.000     .6161622    1.817179
        cses |   2.938756   .1551034    18.95   0.000     2.634759    3.242753
   msesXcses |   1.038844   .2989198     3.48   0.001     .4529717    1.624716
    secXcses |  -1.642572   .2398074    -6.85   0.000    -2.112586   -1.172558
       _cons |   12.11359   .1988085    60.93   0.000     11.72393    12.50324
------------------------------------------------------------------------------

------------------------------------------------------------------------------
  Random-effects Parameters  |   Estimate   Std. Err.     [95% Conf. Interval]
-----------------------------+------------------------------------------------
school: Unstructured         |
                   var(cses) |   .1016223   .2134717      .0016555    6.238148
                  var(_cons) |   2.381904   .3717541      1.754178    3.234261
             cov(cses,_cons) |   .1924953    .204535     -.2083859    .5933765
-----------------------------+------------------------------------------------
               var(Residual) |   36.72101   .6261091      35.51414     37.9689
------------------------------------------------------------------------------
LR test vs. linear regression:       chi2(3) =   220.57   Prob > chi2 = 0.0000

Note: LR test is conservative and provided only for reference

estimates store M1

estat ic


------------------------------------------------------------------------------
       Model |    Obs    ll(null)   ll(model)     df          AIC         BIC
-------------+----------------------------------------------------------------
          M1 |   7185           .   -23251.83     10     46523.67    46592.47
------------------------------------------------------------------------------

Example on page 339. 

xtmixed mathach meanses sector cses meanses*cses sector*cses || school:  , variance

Performing EM optimization: 

Performing gradient-based optimization: 

Iteration 0:   log restricted-likelihood = -23252.397  
Iteration 1:   log restricted-likelihood = -23252.397  

Computing standard errors:

Mixed-effects REML regression                   Number of obs      =      7185
Group variable: school                          Number of groups   =       160

                                                Obs per group: min =        14
                                                               avg =      44.9
                                                               max =        67


                                                Wald chi2(5)       =    765.44
Log restricted-likelihood = -23252.397          Prob > chi2        =    0.0000

------------------------------------------------------------------------------
     mathach |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
     meanses |   5.342945   .3689877    14.48   0.000     4.619742    6.066148
      sector |   1.214627   .3061252     3.97   0.000     .6146327    1.814621
        cses |   2.935841   .1507053    19.48   0.000     2.640464    3.231218
meansesBYc~s |   1.044086   .2910422     3.59   0.000     .4736542    1.614519
sectorBYcses |   -1.64207   .2330966    -7.04   0.000     -2.09893   -1.185209
       _cons |   12.11382   .1986485    60.98   0.000     11.72448    12.50317
------------------------------------------------------------------------------

------------------------------------------------------------------------------
  Random-effects Parameters  |   Estimate   Std. Err.     [95% Conf. Interval]
-----------------------------+------------------------------------------------
school: Identity             |
                  var(_cons) |   2.375333   .3709728      1.748986    3.225988
-----------------------------+------------------------------------------------
               var(Residual) |   36.76611   .6206516      35.56956    38.00291
------------------------------------------------------------------------------
LR test vs. linear regression: chibar2(01) =   219.44 Prob >= chibar2 = 0.0000

estimates store M2

estat ic

------------------------------------------------------------------------------
       Model |    Obs    ll(null)   ll(model)     df          AIC         BIC
-------------+----------------------------------------------------------------
          M2 |   7185           .    -23252.4      8     46520.79    46575.83
------------------------------------------------------------------------------

estimates stats M1 M2

------------------------------------------------------------------------------
       Model |    Obs    ll(null)   ll(model)     df          AIC         BIC
-------------+----------------------------------------------------------------
          M1 |   7185           .   -23251.83     10     46523.67    46592.47
          M2 |   7185           .    -23252.4      8     46520.79    46575.83
------------------------------------------------------------------------------
 

The next set of examples use the willett.dta data which you can obtain from within Stata like this.

use http://www.ats.ucla.edu/stat/paperexamples/singer/willett, clear

Example on pages 341/342.

xtmixed y time || id: time, variance cov(un)

Performing EM optimization: 

Performing gradient-based optimization: 

Iteration 0:   log restricted-likelihood = -633.41137  
Iteration 1:   log restricted-likelihood = -633.41137  

Computing standard errors:

Mixed-effects REML regression                   Number of obs      =       140
Group variable: id                              Number of groups   =        35

                                                Obs per group: min =         4
                                                               avg =       4.0
                                                               max =         4


                                                Wald chi2(1)       =    154.84
Log restricted-likelihood = -633.41137          Prob > chi2        =    0.0000

------------------------------------------------------------------------------
           y |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
        time |      26.96   2.166604    12.44   0.000     22.71353    31.20647
       _cons |   164.3743   6.118849    26.86   0.000     152.3816     176.367
------------------------------------------------------------------------------

------------------------------------------------------------------------------
  Random-effects Parameters  |   Estimate   Std. Err.     [95% Conf. Interval]
-----------------------------+------------------------------------------------
id: Unstructured             |
                   var(time) |   132.4006    40.2107       73.0089    240.1065
                  var(_cons) |   1198.777   318.3812      712.3102    2017.472
             cov(time,_cons) |  -179.2556   88.96347     -353.6208   -4.890378
-----------------------------+------------------------------------------------
               var(Residual) |   159.4771   26.95655      114.5036    222.1149
------------------------------------------------------------------------------
LR test vs. linear regression:       chi2(3) =   120.90   Prob > chi2 = 0.0000

Note: LR test is conservative and provided only for reference

estat ic

------------------------------------------------------------------------------
       Model |    Obs    ll(null)   ll(model)     df          AIC         BIC
-------------+----------------------------------------------------------------
           . |    140           .   -633.4114      6     1278.823    1296.473
------------------------------------------------------------------------------

Example on page 344.

xtmixed y time ccovar timeBYccovar || id: time, variance cov(un)

Performing EM optimization: 

Performing gradient-based optimization: 

Iteration 0:   log restricted-likelihood = -630.14238  
Iteration 1:   log restricted-likelihood = -630.14238  

Computing standard errors:

Mixed-effects REML regression                   Number of obs      =       140
Group variable: id                              Number of groups   =        35

                                                Obs per group: min =         4
                                                               avg =       4.0
                                                               max =         4


                                                Wald chi2(3)       =    191.86
Log restricted-likelihood = -630.14238          Prob > chi2        =    0.0000

------------------------------------------------------------------------------
           y |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
        time |      26.96   1.993881    13.52   0.000     23.05207    30.86794
      ccovar |  -.1135527   .5040119    -0.23   0.822    -1.101398    .8742925
timeBYccovar |   .4328577   .1619278     2.67   0.008      .115485    .7502305
       _cons |   164.3743   6.206096    26.49   0.000     152.2106     176.538
------------------------------------------------------------------------------

------------------------------------------------------------------------------
  Random-effects Parameters  |   Estimate   Std. Err.     [95% Conf. Interval]
-----------------------------+------------------------------------------------
id: Unstructured             |
                   var(time) |   107.2492    34.6767      56.90841    202.1211
                  var(_cons) |   1236.413   332.4024      730.0001    2094.132
             cov(time,_cons) |  -178.2333   85.42982     -345.6726   -10.79389
-----------------------------+------------------------------------------------
               var(Residual) |   159.4771   26.95656      114.5036    222.1149
------------------------------------------------------------------------------
LR test vs. linear regression:       chi2(3) =   120.72   Prob > chi2 = 0.0000

Note: LR test is conservative and provided only for reference

estat ic

------------------------------------------------------------------------------
       Model |    Obs    ll(null)   ll(model)     df          AIC         BIC
-------------+----------------------------------------------------------------
           . |    140           .   -630.1424      8     1276.285    1299.818
------------------------------------------------------------------------------

Examples on pages 346 & 348, cannot be done in Stata at this time. Stata does not have a repeated option for the fixed effects.


How to cite this page

Report an error on this page

UCLA Researchers are invited to our Statistical Consulting Services
We recommend others to our list of Other Resources for Statistical Computing Help
These pages are Copyrighted (c) by UCLA Academic Technology Services


The content of this web site should not be construed as an endorsement of any particular web site, book, or software product by the University of California