Bias study of the naive estimator in a longitudinal binary mixed-effects model with measurement error and misclassification in covariates.

摘要

When covariates in Longitudinal data are subject to errors, the naive estimatesudof the model parameters are often biased. In this research, we exploit a DynamicudBinary Mixed-effects Model using a Generalized Quasi-likelihood approach. Throughudsimulations, we shall study the patterns in the bias of the naive estimator of theudparameters that ignores the errors in the covariates.