How do matching and inverse weighting use the same propensity score differently?
StatisticsLevel: IntermediateCausal identification and targets
How do matching and inverse weighting use the same propensity score differently?
Hints
- Selection versus contribution.
- Both rely on identification assumptions.
Answer
Matching selects comparable units; weighting changes their contribution to a target population. Neither automatically fixes unmeasured confounding or absent overlap.
Source and licence: DoWhy — propensity-based effect estimation · MIT · answer: Quantoros editorial