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)
diagnostics
to iv_robust()
glance()
methods for all estimatorslh_robust()
for easy interface to car::linearHypothesis()
is.na(var)
in the covariates
formula in lm_lin()
(issue #283)broom
hack for tidy
method and instead relies on importing generics
lm_lin
sandwich
version 2.5-0 and off-diagonal blocks of multivariate regression vcov matriceslm_lin
preventing multivariate regressionmargins
difference_in_means
when using condition1
and condition2
to subset a treatment vector with more than two treatment conditions. Previous estimates and standard errors were incorrect.ci.lower
and ci.upper
to conf.low
and conf.high
to be in line with other tidy methodsfixed_effects
that are just one blockcondition_prs
in horvitz_thompson()
as a single numberhorvitz_thompson()
lm_robust
and iv_robust
commarobust
and starprep
for stargazer integrationtexreg
support for 2SLS IV modelsbroom::tidy
coefficient_names
-> term
se
-> std.error
p
-> p.values
ci_lower
-> ci.lower
ci_upper
-> ci.upper
tidy
; furthermore for tidy
objects one further name change from coefficients
-> estimate
has been madedifference_in_means
rlang
na.omit
handler in Rlm_lin()
from _bar
to _c
.Rbuildignore
, only available on website nowlm_robust_helper.cpp
algorithm to not catch own exception and to deal with valgrind
memory errors