Hello,
I'm using the gllamm program in Stata/SE 13.1 and working on complex survey data (i.e., the domestic violence module of DHS).
I have 1 outcome variable ('experience of intimate partner violence' Yes/No), 5 categorical predictor variables, 5 binary predictor variables, and one latent predictor variable. The variable 'ppts' is meant to restrict the sample to those who participated in the domestic violence module.
DHS recommends that their domestic violence data be weighted, and survey variables applied, as follows:
generate wgt = d005/1000000
svyset [pweight = wgt],psu(v021) strata(v022)
I ran my estimation command as follows:
meglm exipv epctrl adult sinc hsgov lit Wempow ib(first).zone ib(0).home2 ib(0).hwealth ib(first).rel ///
ib(0).agemarr4 if (ppts) || v021: R.v022, or family(bernoulli) link(logit) vce(cluster v021)
My questions: (a) Did I apply the cluster and strata variables correctly?
(b) How can I apply the pweights as well?
Thank you in advance.
Som
I'm using the gllamm program in Stata/SE 13.1 and working on complex survey data (i.e., the domestic violence module of DHS).
I have 1 outcome variable ('experience of intimate partner violence' Yes/No), 5 categorical predictor variables, 5 binary predictor variables, and one latent predictor variable. The variable 'ppts' is meant to restrict the sample to those who participated in the domestic violence module.
DHS recommends that their domestic violence data be weighted, and survey variables applied, as follows:
generate wgt = d005/1000000
svyset [pweight = wgt],psu(v021) strata(v022)
I ran my estimation command as follows:
meglm exipv epctrl adult sinc hsgov lit Wempow ib(first).zone ib(0).home2 ib(0).hwealth ib(first).rel ///
ib(0).agemarr4 if (ppts) || v021: R.v022, or family(bernoulli) link(logit) vce(cluster v021)
My questions: (a) Did I apply the cluster and strata variables correctly?
(b) How can I apply the pweights as well?
Thank you in advance.
Som
