[Case Study] Matching, Subclassification and D-Separation

[PhilippBach]

Thanks for contributing to the Digital Causality Lab!

1. Description of the Case Study

Illustrate matching and subclassification approaches based on d-separation with simulated and/or real data sets

2. Idea for the data product

• Implement one or several basic estimation approach(es), like subclassification or exact/approximate matching
• Include DAG with confounding (d-separation)
• Illustrate/compare estimation outcome based on a simulated/real data set

3. Available resources and references

• This could be mainly based on Chapter 5 of Cunningham (2021) which contains many intuitive data examples with nice explanation
• The idea of the data product would include an implementation and maybe a comparison of one or several estimation approaches and should intuitively illustrate the idea (combining DAGs, data, estimation output)

4. Comments

This issue can be used for several case study - each focusing on a separate estimation approach and/or data example.