Bayesian and Non-Bayesian Sensitivity Analysis for Correcting Measurement Error

Measurements of the exposure are often contaminated by the systematic error. The systematic error then causes inconsistent exposure effect estimates and leads to erroneous conclusions to various degrees in statistical analysis. While some methods develop to correct for measurement error, techniques for correcting this still are controversial. In this paper, we compare Bayesian and Non-Bayesian methods to correct measurement error. This is performed both through theoretically and simulations. We investigate the efficiency of these methods by estimating the effect of hypothetical exposure which is subject to measurement error by sensitivity analysis.