Clustered data are found in many different types of studies, for examp
le, studies involving repeated measures, inter-rater agreement studies
, household surveys, crossover designs and community randomized trials
. Analyses based on population average and cluster specific models are
commonly used for estimating treatment (exposure) effects with cluste
red data. This paper discusses conditions when one or both types of mo
dels are appropriate for estimating causal effects and when there is a
greement between population average and cluster specific analyses. App
lications of survey sampling methods to the robust estimation of stand
ard errors of estimated treatment parameters are discussed.