Topic: The Balancing Property and Estimation

Why the Propensity Score Is the Coarsest Balancing Score

Why Treatment Prediction Is Not the Modeling Objective

Worked Comparison of Flexible Propensity Estimators

Worked Logistic Specification for Propensity Scores

Why Fitted Scores Do Not Guarantee Exact Sample Balance

Uncertainty in Estimated Propensity Scores

Sparse Categories and Instability in Score Estimation

Separation in Logistic Propensity Models

Ridge and Lasso Penalties for Propensity Estimation

Restricted Cubic Splines in Logistic Score Models