Applying weights to mixed-effects logistic regression model

[Rosa Blau]

Hello Statalist,

I am trying to add weights to my mixed-effects logistic regression model. I could use regression commands melogit or meqrlogit, but I don't know how to add weights to either.

So far, I have been working with simple logistic regression, and weighting with svyset and svy: . The difference is very significant depending on whether or not I weight the models. I expect the weighting to be similarly important with a mixed-effects logistic regression.

However, svyset and svy: do not seem to work in combination with melogit or meqrlogit (although I am not sure, as I find the help entry to be a bit confusing here).

So my question is whether somebody can tell me how to combine my weighting with either melogit or meqrlogit.

Here is the data and command information you may require:

My weighting variable

* Example generated by -dataex-. To install: ssc install dataex
clear
input double Compwt
2.09944
 .41773
 .36181
 .87289
 .41773
 .87289
 .41773
 .87289
 .36181
 .87289
 .36181
 .87289
 .87289
2.09944
 .36181
 .52707
 .87289
 .52707
 .87289
 .52707
 .87289
2.09944
 .87289
 .87289
2.09944
 .87289
2.09944
 .87289
 .87289
 .52707
 .36181
 .52707
 .52707
2.09944
2.09944
2.09944
 .52707
 .87289
 .87289
 .87289
2.09944
 .87289
 .87289
 .87289
 .36181
 .36181
 .87289
 .54305
 .52707
 .52707
 .87289
 .87289
2.09944
 .87289
 .87289
 .87289
 .87289
 .87289
 .87289
 .41773
 .36181
2.09944
2.09944
 .52707
 .87289
 .52707
 .41773
 .87289
 .52707
 .36181
2.09944
2.09944
 .36181
 .87289
 .52707
 .52707
2.09944
 .52707
2.09944
 .52707
 .52707
 .87289
 .87289
 .87289
 .87289
 .87289
 .41773
 .87289
 .54305
 .87289
2.09944
 .87289
2.09944
 .36181
 .36181
 .52707
 .87289
 .87289
 .52707
 .36181
end
label values Compwt Compwt

My dependent variable

* Example generated by -dataex-. To install: ssc install dataex
clear
input float childcare
0
0
1
1
1
0
1
1
1
1
1
0
0
0
1
0
0
0
1
1
1
0
0
1
0
1
0
1
1
0
0
1
0
1
1
1
0
0
0
1
0
0
0
1
1
1
1
1
0
1
0
1
.
1
0
1
0
0
0
.
1
0
0
1
0
1
1
1
1
1
1
0
1
0
1
0
0
1
1
1
0
1
0
0
0
1
1
1
1
0
1
0
1
1
1
1
1
1
0
1
end
label values childcare childcare
label def childcare 0 "Not provided", modify
label def childcare 1 "Provided", modify

My group variable

* Example generated by -dataex-. To install: ssc install dataex
clear
input float p_regime
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
end
label values p_regime p_regime
label def p_regime 1 "Pre-K only", modify
label def p_regime 2 "Both", modify

My weighting command

 svyset [pweight = Compwt]

My weighted logit command (which works perfectly)

 svy: logit childcare i.p_regime i.skillprofile gdp size fem_execs

Many thanks for your time and consideration.

[Rich Goldstein]

if you look at the manual for -melogit- (e.g.), you will see examples of how to use various types of weights, including pweights; you get to that entry by typing "help melogit" and clicking on the blue link near the top of the help file; you can also use the svy prefix (but do not specify weights if you do so)