I am conducting an analysis from a survey that used the following sampling frame:
It is a 2 stage sample where the first phase was done with 3 strata and 1 cluster with a varying inclusion probability (which I think would be a sampling rate) and the second phase was done with 1 strata (different from the other 3) and a different inclusion probability. How would I code this to create correct standard errors when creating means? I am really trying to figure out how to account for the 2 different probabilities for each stage in the rate=sampling_rate statement. How would I set up that dataset? Or if I am completely wrong, please lead in a better direction!
I have this so far:
proc surveymeans data=rhs_food rate=sampling_rate;
title1 'Strata and Cluster specified';
title2 'Complex Sample';
strata region_num sub_region_num comm_size_num age_gender_group;
cluster residence_merge;
weight weight_final;
var food_secure;
run;
Thank-you so much for any help!
Carole