Chapter 12: Missing Data
NoteComing soon
This chapter is planned but not yet written. The outline below shows what it is intended to cover.
[12a] Missing Data Theory
Topics
- Missing completely at random, at random, and not at random
- Why the mechanism is an assumption about unobserved data
- Describing and visualizing missingness patterns
- Why listwise deletion, pairwise deletion, and mean imputation all fail
- What “unbiased” and “efficient” mean for a missing-data method
[12b] Imputation & FIML
Topics
- Multiple imputation with {mice}: impute, analyze, pool
- Choosing an imputation model, and why it should be at least as rich as the analysis model
- How many imputations, and how to tell if it was enough
- Full information maximum likelihood with {lavaan}
- Choosing between multiple imputation and FIML, and reporting either one
This chapter is deliberately two lectures rather than one. Mechanisms alone give students a vocabulary for a problem they cannot yet do anything about; pairing the theory with a lecture where they actually fit and pool a model is what makes it worth teaching.