Statistical Analysis with Missing Data

Fall 2018



It is very common to come across missing-data in various scientific research. Such missing-data could lead to a loss of statistical power and may introduce bias into the statistical analysis. We will begin the course by introducing various missing-data patterns and mechanisms. We will review different simple ad-hoc fixes for handling missing-data and bring out their shortcomings. We will then discuss different modern approaches for statistical analysis with missing data. We will cover single imputation, resampling methods for estimating imputation uncertainty, EM algorithm and multiple imputation. Derivation of formulas will be presented wherever necessary to explain some of the advanced topics. The course emphasizes data analysis via modern computer methods and R freeware packages that are introduced and used throughout the course.