Causal Inference with Survey Data

Explore the concepts of causal inference in survey data, learn some of the underlying theory of causality, and focus on empirical methods to identify causality in data.

Introduction

• Causality unlocked: A primer for data analysts
• What you can learn
• What you should know

1. Cause and Effect

• Why causal inference matters
• The gold standard: Experimental data
• What is different about survey data?
• Observables vs. unobservables causes
• What are treatment effects?
• An applied example: The LaLonde debate

2. Experimental Survey Designs

• Setting up a randomized controlled trial
• Analyzing a randomized controlled trial

3. Cross-Sectional Survey Designs

• Surveys with cross-sectional data
• Regression analysis
• Propensity score matching
• Regression discontinuity designs
• Instrumental variable models

4. Longitudinal Survey Designs

• Surveys with longitudinal data
• Regression models with time effects
• Fixed effects regression models
• Difference-in-difference estimation
• Synthetic control methods

5. Other Models

• How to evaluate causal robustness
• How to present causal statistics

Conclusion

• Next steps and additional resources