Measurement error correction techniques in case-control studies.

Measurement error affecting covariates is a common problem in many scientific areas. In this paper, we focus on the application of likelihood-based techniques to correct for measurement errors in case-control studies. The likelihood approach has received less attention in literature with respect to alternative correction techniques, mainly because of the computational complexity and the difficulties in specifying the relationships among variables which are required. The performance of the likelihood approach is evaluated here through simulation studies, under a multiplicative error structure. Moreover, an extension of the likelihood-based method is proposed, which has the aim of robustifying the analysis with respect to misspecifications of the unobservable and mismeasured phenomenon