The Impact of Residual and Unmeasured Confounding in Epidemiologic Studies: A Simulation Study
The Impact of Residual and Unmeasured Confounding in Epidemiologic Studies: A Simulation Study
Measurement error in explanatory variables and unmeasured confounders can cause considerable problems in epidemiologic studies. It is well recognized that under certain conditions, nondifferential measurement error in the exposure variable produces bias towards the null. Measurement error in confounders will lead to residual confounding, but this is not a straightforward …