meglm updated question

[Hillel Alpert]

I have overcome the second-level weight issue with meglm for multilevel two-stage sampled survey data having only the final weight.
Individual students were randomly selected from classes (psu) in schools (strata), which were randomly selected in each state.
The first-level weight in these data is set to 1.

The remaining question is how to specify the svyset command.

Does svyset state, weight(weight2) || _n, weight(weight) look correct,
where state is the first level (cluster) variable; weight2 is its weight, which is constant; _n is to indicate random sampling of individuals at the second level, and weight is the first level weight?

The svy: melogit command is now running with the above survey specifications and is taking a long time.

Also, strata are not accounted for explicitly in the model? Is this okay? Can it be incorporated in the model outside of svyset?

[Dirk Enzmann]

You state that your question is a follow up of a previous question: I would like to know the context of your question (this does not mean that I can answer your question if I would know it).

You should have answered to your own previous question, thus continue the previous thread with a new post, or (if you think a new topic would be more appropriate) insert a link to the previous post by copying the link to it with a right click on the respective number of the post in the first thread (#.. top right of the question window) and inserting the resulting link into your new question using the symbol of the "link" function in the top line of the advanced editor, e.g. "I am refering to a previous question").

[Hillel Alpert]

Thanks.