GNU R fast estimators for design-based inference

 Fast procedures for small set of commonly-used, design-appropriate
 estimators with robust standard errors and confidence intervals.
 Includes estimators for linear regression, instrumental variables
 regression, difference-in-means, Horvitz-Thompson estimation, and
 regression improving precision of experimental estimates by interacting
 treatment with centered pre-treatment covariates introduced by Lin
 (2013) <doi:10.1214/12-AOAS583>.