- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Dear everyone,
I seeking your advice on whether SAS will be able to fit my model, and if so, which procedure.
My data has a three-level structure (country, community, individual) with a binary outcome variable at level 1, and main independent variable at level 2. The data was collected using two-stage complex sampling, hence I hope to incorporate stratification and sample weight in the model. Sampling strata is not being analysed as a separate level.
Thank you for your help in advance!
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Without more details this sound like you may be looking at Proc SurveyLogisitic.
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
And if you have SAS/STAT14.1 installed, then PROC GLIMMIX has an example (45.18 Weighted Multilevel Model for Survey Data) that covers a two level model, and an example of a three level model using the quadrature method to fit the data (45.19 Quadrature Method for Multilevel Model), that has students, classes and schools. By combining the principles in these examples, I think you can get what you are after.
Steve Denham