WeightIt

Weighting for Covariate Balance in Observational Studies

Generates weights to form equivalent groups in observational studies with point or longitudinal treatments by easing and extending the functionality of the R packages 'twang' for generalized boosted modeling (McCaffrey, Ridgeway & Morral, 2004) <doi:10.1037/1082-989X.9.4.403>, 'CBPS' for covariate balancing propensity score weighting (Imai & Ratkovic, 2014) <doi:10.1111/rssb.12027>, 'ebal' for entropy balancing (Hainmueller, 2012) <doi:10.1093/pan/mpr025>, 'optweight' for optimization-based weights (Zubizarreta, 2015) <doi:10.1080/01621459.2015.1023805>, 'ATE' for empirical balancing calibration weighting (Chan, Yam, & Zhang, 2016) <doi:10.1111/rssb.12129>, and 'SuperLearner' for stacked machine learning-based propensity scores (Pirracchio, Petersen, & van der Laan, 2015) <doi:10.1093/aje/kwu253>. Also allows for assessment of weights and checking of covariate balance by interfacing directly with 'cobalt'.

Version History