Basic steps for matching
The following are the basic steps in performing a causal analysis using data preprocessing (Stuart, 2010)1:
- Decide on covariates for which balance must be achieved
- Estimate the distance measure (e.g., propensity score)
- Condition on the distance measure (e.g., using matching, weighting, or subclassification)
- Assess balance on the covariates of interest; if poor, repeat steps 2-4
- Estimate the treatment effect in the conditioned sample
一些有用的R包
cobalt R package
cobalt package 是一个集成了诸多匹配方法的R包,由Noah Greifer开发。使用说明
-
Stuart, E. A. (2010). Matching Methods for Causal Inference: A Review and a Look Forward. Statistical Science, 25(1), 1–21. http://doi.org/10.1214/09-STS313.↩
近期评论