Course Content

The course will cover:

• The effects of missing data on statistical inferences
• Missingness mechanism assumptions include missing completely at random, missing at random, and missing not at random.
• Multiple imputation for missing data, based on joint models and fully conditional specification approaches, and Rubin's pooling rules.
• Multiple imputation accommodating non-linearities and interactions.
• Multiple imputation for sensitivity analysis.
• Multiple imputation in the context of propensity score analysis.
• Multiple imputation in the context of prognostic model development and deployment.

Through computer practicals using R, participants will learn how to apply the statistical methods introduced in the course to realistic datasets.

Course Certificate and Assessment

There will be no formal assessment, but participants will receive a Certificate of Attendance.