Publication type: Article
Author(s): HW Lin, YH Chen
Publication: American Journal of Epidemiology
Year published: 2014
Why this Article: Missing or incomplete data on confounders is a common problem in epidemiology. This article discusses a two-stage calibration method using propensity scores to address missing data on confounders in observational studies. The main drawback is that this approach requires a small subset with complete data. However, in situations where gathering complete data on a subset is feasible this a method worth considering.