A SIMEX approach for control risk regression with study-specific covariates

This thesis studies the problem of control risk regression while utilizing information given by study-specific covariates. The proposed approach is SIMEX, a simulation-based measurement error correction technique. The method is compared to the likelihood approach, which relies on additional hypothesis circa the underlying generating process of the available data, and the naïve approach of the usual least squares method. To carry out the comparison, simulated and real world data are utilized.