Stata’s mixed for fitting linear multilevel models supports survey data. Sampling weights and robust/cluster standard errors are available.

 

Sampling weights are handled differently by mixed:

Weights can (and should be) specified at every model level unless you wish to assume equiprobability sampling at that level.

Weights at lower model levels need to indicate selection conditional on selection of the higher-level cluster and not merely indicate overall selection.

The scaling of weights at lower levels needs to be considered. Unlike a standard analysis where the scale of the sampling weights is not an issue (only their relative sizes matter), in multilevel models weight scales need to be made “consistent” across lower-level clusters.

 

We demonstrate using mixed to fit a two-level model for data from a two-stage sampling design with sampling weights at both stages. Schools were sampled at the first stage, students at the second.

 

. webuse pisa2000
(Programme for International Student Assessment (PISA) 2000 data)

. mixed isei female high_school college one_for both_for test_lang > [pw=w_fstuwt] || id_school:, pweight(wnrschbw) pwscale(size) nolog

 

Mixed-effects regression                                                                      Number of obs = 2,069
Group variable: id_school                                                                   Number of groups = 148
Obs per group:
                                                                                                                   min = 1
                                                                                                                   avg = 14.0
                                                                                                                   max = 28
                                                                                                                   Wald chi2(6) = 187.23
Log pseudolikelihood = -1443093.9                                                 Prob > chi2 = 0.0000
                                                                                                  (Std. Err. adjusted for 148 clusters in id_school)

 

isei

Robust

Coef.               Std. Err.           z               P>|z|           [95% Conf. Interval]    [95% Conf. Interval]

female .59379             .8732886       0.68         0.497                   -1.117824                       2.305404
high_school 6.410618         1.500337        4.27         0.000                   3.470011                       9.351224
college 19.39494        2.121145          9.14          0.000                  15.23757                        23.55231
one_for -.9584613       1.789947        -0.54        0.592                   -4.466692                     2.54977
both_for -.2021101       2.32633          -0.09        0.931                   -4.761633                       4.357413
test_lang 2.519539         2.393165        1.05          0.292                  -2.170978                       7.210056
_cons 28.10788        2.435712        11.54        0.000                  23.33397                        32.88179

 

Random-effects Parameters

Robust

Estimate                     Std. Err.                                   [95% Conf. Interval]

id_school: Identity

var(_cons)

34.69374                   8.574865                               21.37318                 56.31617
var(Residual) 218.7382                   11.22111                                 197.8147                 241.8748

 

In the above, we specified the student-level weights using standard Stata weight syntax [pw=w_fstuwt] and the school-level weights with the pweight(wnrschbw) option as part of the school random-effects equation. We also specified pwscale(size) to rescale the student-level weights using one of three available methods.

 

Sampling weights imply robust standard errors, and in the case of mixed, standard errors are clustered at the highest level (schools in this example) unless you specify otherwise.