[R-sig-ME] Pointer in right direction

[Sokolovsky, Alexander]

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 using 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]]