I am working with complex survey data, multi-stage sampling. My current code includes a cluster variable for the primary sampling unit (PSU) and a strata variable: PROC SURVEYLOGISTIC DATA= data; cluster psu; strata strata; In addition to PSU clustering, the data has clustering at the household level where multiple individuals within a household are surveyed. I want to account for the correlation between observations in a household. I am unclear if adding the household variable to the cluster statement will resolve this. My new revised code is: PROC SURVEYLOGISTIC DATA= data; cluster psu household; strata strata; Is the above correct? Please advise.
... View more
