Hi all, I am starting to work on a new analysis of nationally representative longitudinal cohort data which uses a complex sampling strategy. There are replicate weights included in the dataset constructed from ~ 100 variables. I was planning on using a subset of the data (a subset for which specific weights have already been computed) of individuals completing all waves. The broad analytic plan was to regress a binary outcome onto a time (indicated by *wave*) x moderator interaction to examine how a longitudinal trend *x *in the prevalence of a behavior *y* are moderated by individual characteristic *m. *I am happy to provide more details if it would be helpful. My assumption was that I would still have to account for the clustering of observations (weights are provided at the person-level) within an individual to account for the covariance between these observations, but I could not find a package that let me use BRR sampling weights with a GLMM. So any help would be appreciated in terms of either: a) pointing me in the right direction in terms of tools to undertake this analysis, or b) correcting a potential misperception in the need to account for the clustering given the other design characteristics of the survey. Any advice would be helpful! All the best, -- Alex
*Alexander W. Sokolovsky, PhD* Assistant Professor Center for Alcohol and Addiction Studies Department of Behavioral and Social Sciences E: Alexander_Sokolovsky at Brown.edu P: (401) 863-6629(401) 863-6697 (Fax) A: Box G-S121-5, Providence, RI 02912 <https://maps.google.com/?q=Box%20G-S121-5%2C%20Providence%2C%20RI%2002912> W: https://vivo.brown.edu/display/asokolo1 ? @alexsokophd.bsky.social <https://bsky.app/profile/alexsokophd.bsky.social> ?The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge.? - Stephen Hawking [[alternative HTML version deleted]]