This online course covers the key concepts and methods for handling missing data in statistical analyses using R, including:
- Rubin's missing data mechanism assumptions, how to investigate them using the observed data, using directed acyclic graphs (DAGs) to think about them
- problems with ad-hoc methods like adding a missing category or last observation carried forward - when is complete case analysis unbiased (more often than many think!)
- multiple imputation, via the chained equations method in the MICE package, and practicalities such as how many imputations to use and how to choose which variables to include
- advanced imputation topics, including imputation of covariates with survival and competing risks outcomes, non-linear effects and interactions, and incorporating survey design
Concepts and methods are introduced through mini-lecture videos. How to apply the methods in R is demonstrated through videos, with datasets and R scripts downloadable. Multiple choice quizzes are used throughout to test and develop understanding. Click the Free Preview button at https://thestatsgeek.thinkific... to sign up and view some of the lessons for free.