Addressing Nonignorable Attrition with Double Sampling and Bounds R package attrition version 1.0.0 Implements the double-sampling bounds estimator of Coppock, Gerber, Green, and Kern (2017) <doi:10.1017/pan.2016.6> for randomized experiments with nonignorable missing outcomes. Provides worst-case (Manski) bounds, double-sampling bounds with analytic variance and Imbens-Manski confidence intervals, Lee (2009) <doi:10.1111/j.1467-937X.2009.00536.x> trimming bounds with analytic and bootstrap standard errors, covariate adjustment via poststratification, and a sensitivity analysis for violations of the outcome stability assumption. Software Depends: R (>= 4.1.0) Imports: generics, ggplot2, tibble Suggests: dplyr, estimatr, knitr, purrr, rmarkdown, testthat (>= 3.0.0), vayr (>= 1.1.0) Alexander Coppock <acoppock@gmail.com> Comprehensive R Archive Network (CRAN) Alexander Coppock [aut, cre], Alan S. Gerber [aut], Donald P. Green [aut], Holger L. Kern [aut] GPL-3 2026-09-18 application/tgz https://CRAN.R-project.org/package=attrition doi:10.32614/CRAN.package.attrition en-US