A SURVEY OF METHODS FOR ANALYZING CLUSTERED BINARY RESPONSE DATA

A comprehensive survey of regression-type models for clusters of corre lated binary outcomes, Including longitudinal data, is presented, In p articular, we focus on models which can accommodate both between- and within-cluster categorical and continuous covariates, Emphasis is give n to motivation of the model specification, interrelationships among m odels, parameter testing and interpretation, estimation methods (inclu ding both likelihood and non-likelihood approaches), computational iss ues, availability of software and other implementation issues, and to the advantages and disadvantages of the various approaches, Models dis cussed include naive and response feature models, conditionally specif ied models, marginal models, and cluster-specific models, Extensions t o ordinal data and relationships to graphical representations of model s are also discussed.