Software‎ > ‎

ui - inference with non-ignorable missing data, and with unobserved confounders

ui is an R-package which provide confidence intervals (sampling variation) and uncertainty intervals (sampling variation+uncertainty due to non-ignorability assumptions) for (i) for regression (linear and probit) parameters when outcome is missing not at random (non-ingorable missingness); (ii) for double robust and outcome regression estimators of average causal effects (on the treated) with possibly unobserved confounding.

Keywords: Double Robust estimator, Regression Imputation estimator, Average Causal effect, Estimation with missing outcome data, Estimation with non-ignorable dropout. 

Package: ui (download zip file below)

Genbäck, M. & de Luna, X. (2017). Causal inference taking into account unobserved confounding. Working paper soon available. arXiv:1712.00292 To appear in Biometrics

Genbäck, M. & de Luna, X. (2018). Predictors of decline in self-reported health: addressing non-ignorable dropout in longitudinal studies of ageing. European Journal of Ageing, 15(2), 211-220.

Genbäck, M., Stanghellini, E., & de Luna, X. (2015). Uncertainty intervals for regression parameters with non-ignorable missingness in the outcome. Statistical Papers56(3), 829-847.
Minna Genbäck,
Dec 7, 2018, 4:12 AM