Bayesian Correction for Misclassification in Multilevel Count Data Models

Covariate misclassifcation is well known to yield biased estimates in single level regression models. Te impact on hierarchical count models has been less studied. A fully Bayesian approach to modeling both the misclassifed covariate and the hierarchical response is proposed. Models with a single diagnostic test and with multiple diagnostic tests are considered. Simulation studies show the ability of the proposed model to appropriately account for the misclassifcation by reducing bias and improving performance of interval estimators. A real data example further demonstrated the consequences of ignoring the misclassifcation. Ignoring misclassifcation yielded a model that indicated there was a signifcant, positive impact on the number of children of females who observed spousal abuse between their parents. When the misclassifcation was accounted for, the relationship switched to negative, but not signifcant. Ignoring misclassifcation in standard linear and generalized linear models is well known to lead to biased results. We provide an approach to extend misclassifcation modeling to the important area of hierarchical generalized linear models.