https://www.jstage.jst.go.jp/article/jea/31/8/31_JE20210145/_article

PMC

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8275441/

Highlights

• A general introduction to causal inference and propensity score methods is provided.
• Relative advantages of propensity score methods over multivariable regression are discussed.
• Two popular propensity score methods — matching and inverse probability weighting — are compared.
• The alternative methods rely on similar assumptions for identification, with subtle differences.
• The alternative methods make different modeling assumptions and answer different questions. 

Selected Result