Regression modelQuasi-experimental / causal inference
Panel Data Inverse Probability Weighting
Panel Data Inverse Probability Weighting (panel IPW) estimates the causal effect of a time-varying treatment by reweighting observed units to create a pseudo-population in which treatment is independent of measured confounders at each time point. It extends the cross-sectional IPW framework to longitudinal settings where treatment status and confounders both evolve across multiple periods.
Open in MethodMindSoonVideoSoon
Read the full method
Members only
Sign inSign in with a free account to read this section.
Sources
- Robins, J. M., Hernan, M. A., & Brumback, B. (2000). Marginal structural models and causal inference in epidemiology. Epidemiology, 11(5), 550-560. DOI: 10.1097/00001648-200009000-00011 ↗
- Hernan, M. A., & Robins, J. M. (2020). Causal Inference: What If. Boca Raton: Chapman & Hall/CRC. link ↗