Measurement Error and Misclassification for Epidemiologists

Harvard HSPH

This course focuses on methods for the analysis of data when covariates are misclassified or measured with error. It will cover theory for valid estimation and inference in this setting, as well as application of the theory to current epidemiologic studies using computer software developed for this purpose. Methods for contingency tables, generalized linear models, and survival data analysis will be addressed. Topics include likelihood-based methods and regression calibration. Examples from…