How to apply a weight variable for logistic regression?

Question

Here comes a probably very easy question about how to apply a weight variable in R.

I use the ANES 2013 Internet Recontact Survey dataset. Here I have the variable weight, which must be used in order to generalize to population.

My question is very simple: How do I apply the weight for logistic regression?

My best guess will be:

fit.gov.trust <-glm(gov.trust ~ party.contact:weight, family=binomial(link="logit"))

But I am not sure at all.

Answer 1

This can be done using the R survey package by Thomas Lumley, specifically svyglm {survey}

The home page below provides a nice introduction:

http://r-survey.r-forge.r-project.org/survey/