Dr. Walter Leite demonstrates how to evaluate covariate balance of a continuous treatment dose using inverse probability weights using the R statistical software. The weights were estimated with different methods, such a generalized linear modeling, covariate balancing propensity score, and Bayesian additive regression trees (BART). This example is from Chapter 7 of his book Practical Propensity Score Methods Using R. The data and code are available at www.practicalpropensityscore.com

свернуть