UCLA Academic Technology Services HomeServicesClassesContactJobs
Search

Stata Textbook Examples
Multilevel Analysis Techniques and Applications by Joop Hox
Chapter 7: Cross-Classified Multilevel Models


Table 7.1 on page 126 using data set  pupcross.

Part 1: Intercept only.

use http://www.ats.ucla.edu/stat/stata/examples/mlm_ma_hox/pupcross, clear

xtmixed achiev  || _all: R.sschool||_all: R.pschool, var ml
Performing EM optimization: 
Performing gradient-based optimization: 
Iteration 0:   log likelihood = -1158.9243  
Iteration 1:   log likelihood = -1158.9243  
Computing standard errors:
Mixed-effects ML regression                     Number of obs      =      1000
Group variable: _all                            Number of groups   =         1
                                                Obs per group: min =      1000
                                                               avg =    1000.0
                                                               max =      1000
                                                Wald chi2(0)       =         .
Log likelihood = -1158.9243                     Prob > chi2        =         .
------------------------------------------------------------------------------
      achiev |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
       _cons |   6.348653   .0783121    81.07   0.000     6.195164    6.502142
------------------------------------------------------------------------------
------------------------------------------------------------------------------
  Random-effects Parameters  |   Estimate   Std. Err.     [95% Conf. Interval]
-----------------------------+------------------------------------------------
_all: Identity               |
              var(R.sschool) |   .0654012   .0213101      .0345329    .1238622
-----------------------------+------------------------------------------------
_all: Identity               |
              var(R.pschool) |   .1693465   .0393136      .1074411    .2669205
-----------------------------+------------------------------------------------
               var(Residual) |   .5131688   .0239004      .4683993    .5622174
------------------------------------------------------------------------------
LR test vs. linear regression:       chi2(2) =   235.17   Prob > chi2 = 0.0000
Note: LR test is conservative and provided only for reference

Part 2: intercept plus pupil level variables

xtmixed achiev  pupsex pupses || _all: R.sschool||_all: R.pschool, var ml
Performing EM optimization: 
Performing gradient-based optimization: 
Iteration 0:   log likelihood = -1121.7414  
Iteration 1:   log likelihood = -1121.7414  
Computing standard errors:
Mixed-effects ML regression                     Number of obs      =      1000
Group variable: _all                            Number of groups   =         1
                                                Obs per group: min =      1000
                                                               avg =    1000.0
                                                               max =      1000
                                                Wald chi2(2)       =     77.33
Log likelihood = -1121.7414                     Prob > chi2        =    0.0000
------------------------------------------------------------------------------
      achiev |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
      pupsex |   .2613131   .0456365     5.73   0.000     .1718672    .3507591
      pupses |   .1140858   .0161001     7.09   0.000     .0825302    .1456414
       _cons |   5.755482   .1052697    54.67   0.000     5.549157    5.961807
------------------------------------------------------------------------------
------------------------------------------------------------------------------
  Random-effects Parameters  |   Estimate   Std. Err.     [95% Conf. Interval]
-----------------------------+------------------------------------------------
_all: Identity               |
              var(R.sschool) |   .0636083   .0205936      .0337235    .1199763
-----------------------------+------------------------------------------------
_all: Identity               |
              var(R.pschool) |    .168996   .0387797       .107783    .2649734
-----------------------------+------------------------------------------------
               var(Residual) |   .4742561   .0220886      .4328804    .5195866
------------------------------------------------------------------------------
LR test vs. linear regression:       chi2(2) =   255.60   Prob > chi2 = 0.0000
Note: LR test is conservative and provided only for reference

Part 3:  primary by secondary School crossed with pupil and school variables

xtmixed achiev  pupsex pupses pdenom sdenom || _all: R.sschool||_all: R.pschool, var ml
Performing EM optimization: 
Performing gradient-based optimization: 
Iteration 0:   log likelihood = -1118.7376  
Iteration 1:   log likelihood = -1118.7376  
Computing standard errors:
Mixed-effects ML regression                     Number of obs      =      1000
Group variable: _all                            Number of groups   =         1
                                                Obs per group: min =      1000
                                                               avg =    1000.0
                                                               max =      1000
                                                Wald chi2(4)       =     83.77
Log likelihood = -1118.7376                     Prob > chi2        =    0.0000
------------------------------------------------------------------------------
      achiev |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
      pupsex |   .2630786   .0456102     5.77   0.000     .1736843    .3524729
      pupses |   .1135645   .0160943     7.06   0.000     .0820203    .1451087
      pdenom |    .204121    .124103     1.64   0.100    -.0391164    .4473585
      sdenom |   .1761507   .0946598     1.86   0.063    -.0093792    .3616806
       _cons |   5.518506   .1407722    39.20   0.000     5.242597    5.794414
------------------------------------------------------------------------------
------------------------------------------------------------------------------
  Random-effects Parameters  |   Estimate   Std. Err.     [95% Conf. Interval]
-----------------------------+------------------------------------------------
_all: Identity               |
              var(R.sschool) |   .0554237   .0185174      .0287941    .1066812
-----------------------------+------------------------------------------------
_all: Identity               |
              var(R.pschool) |   .1594093   .0368569      .1013229    .2507952
-----------------------------+------------------------------------------------
               var(Residual) |   .4741053   .0220801      .4327454    .5194182
------------------------------------------------------------------------------
LR test vs. linear regression:       chi2(2) =   236.89   Prob > chi2 = 0.0000
Note: LR test is conservative and provided only for reference

Part 4:  primary by secondary School crossed with pupil and school variables with variable pupses being modeled as a random effect.

xtmixed achiev  pupsex pupses pdenom sdenom || _all: R.sschool || pschool: pupses, var ml cov(un)

Performing EM optimization: 

Performing gradient-based optimization: 

Iteration 0:   log likelihood = -1112.2391  
Iteration 1:   log likelihood = -1112.2377  
Iteration 2:   log likelihood = -1112.2377  

Computing standard errors:

Mixed-effects ML regression                     Number of obs      =      1000

-----------------------------------------------------------
                |   No. of       Observations per Group
 Group Variable |   Groups    Minimum    Average    Maximum
----------------+------------------------------------------
           _all |        1       1000     1000.0       1000
        pschool |       50         10       20.0         31
-----------------------------------------------------------

                                                Wald chi2(4)       =     64.80
Log likelihood = -1112.2377                     Prob > chi2        =    0.0000

------------------------------------------------------------------------------
      achiev |      Coef.   Std. Err.      z    P>|z|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
      pupsex |   .2531604   .0453024     5.59   0.000     .1643692    .3419515
      pupses |   .1142274    .020469     5.58   0.000     .0741089    .1543459
      pdenom |   .1999041   .1176438     1.70   0.089    -.0306735    .4304817
      sdenom |   .1645517   .0934395     1.76   0.078    -.0185863    .3476897
       _cons |   5.532417   .1374673    40.25   0.000     5.262986    5.801848
------------------------------------------------------------------------------

------------------------------------------------------------------------------
  Random-effects Parameters  |   Estimate   Std. Err.     [95% Conf. Interval]
-----------------------------+------------------------------------------------
_all: Identity               |
              var(R.sschool) |   .0537343   .0180115      .0278569    .1036504
-----------------------------+------------------------------------------------
pschool: Unstructured        |
                 var(pupses) |   .0080183   .0038835      .0031033    .0207177
                  var(_cons) |   .1485845   .0752156      .0550913    .4007411
           cov(pupses,_cons) |  -.0156084   .0150028     -.0450135    .0137966
-----------------------------+------------------------------------------------
               var(Residual) |   .4583576   .0218485      .4174747     .503244
------------------------------------------------------------------------------
LR test vs. linear regression:       chi2(4) =   249.89   Prob > chi2 = 0.0000

Note: LR test is conservative and provided only for reference

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