Title: Estimation in GEE models for clustered/longitudinal data with covariate measurement error
Authors: Jeff Buzas - The University of Vermont (United States) [presenting]
Abstract: Estimation and inference in cluster/longitudinal population averaged models is considered in the presence of covariate measurement error, with emphasis on logistic regression. A general approach to bias correction, resulting in approximately consistent estimators, is considered when auxiliary data in the form of replicates or instrumental variables are available. For the logistic model, the approach yields unbiased estimating equations resulting in fully consistent estimators.