[Case Study] Introduction to [Causal Identification and Estimation Approaches]

[PhilippBach]

1. Description of the Case Study

Illustration and demonstration of causal identification and estimation approaches

2. Idea for the data product

A report, blog post or app illustrating the main idea and implementation (if available) of quasi-experimental approaches, for example

• regression discontinuity
• instrumental variable estimation
• estimation of local average treatment effects under imperfect compliance
• synthetic control
• matching and propensity score matching
• difference-in-differences
• heterogeneous treatment effects (conditional/group average treatment effects)

3. Available resources and references

There are plenty of resources available, examples include

• Chapter 6 of Cunningham (2020) on regression disontinuity
• Chapter 7 of Cunningham (2020) on instrumental variable estimation
• Chapter 7.6 of Cunningham (2020) on heterogenous treatment effects and IV (LATE)
• Chapter 5 of Cunningham (2020) on matching etc.
• This blogpost by M. Courthoud on synthetic control methods

4. Comments

The difficulty here is to first become familiar with the identification and estimation approach and develop a nice illustrating example. It is probably also necessary to dive into specific R packages that implement these appraoches.

This issue can be customized to the respective estimation approach - open a new, more specific issue then.