Doubly robust estimation of local average treatment effects with a binary instrument.
drlate estimates the local average treatment effect (LATE) and the local
average treatment effect on the treated (LATT) from observational data with a
binary instrument, a continuous/binary/count treatment, and a
continuous/binary/count outcome. It implements the IPWRA, AIPW, IPW, and RA
estimators of Słoczyński, Uysal, and Wooldridge (2022), with robust,
cluster-robust, and bootstrap standard errors. It is an R port of the Stata
drlate command and is validated against it to roughly 1e-8 on the SIPP
example data.
# install.packages("remotes")
remotes::install_github("xiangao/drlate")library(drlate)
set.seed(1)
n <- 200
x <- rnorm(n)
z <- rbinom(n, 1, plogis(0.3 * x)) # binary instrument
d <- rbinom(n, 1, plogis(-0.2 + 0.8 * z + 0.3 * x)) # treatment
y <- 1 + 0.5 * d + 0.4 * x + rnorm(n) # outcome
dat <- data.frame(y, d, z, x)
# Default: doubly robust IPWRA LATE
drlate(y ~ x, d ~ x, z ~ x, data = dat)
# AIPW with inverse-probability-tilting PS; LATT; bootstrap SEs
drlate(y ~ x, d ~ x, z ~ x, data = dat, method = "aipw", ps = "ipt")
drlate(y ~ x, d ~ x, z ~ x, data = dat, estimand = "latt")
drlate(y ~ x, d ~ x, z ~ x, data = dat, vce = "bootstrap", boot_reps = 200)The three formulas give, in order, the outcome model, the treatment model, and
the instrument propensity-score model. Methods are print/summary/coef/
vcov/nobs/broom::tidy.
Full documentation: https://xiangao.github.io/drlate/
- Doubly robust LATE and LATT with drlate — worked vignette reproducing the Stata help-file examples on the SIPP data, with the matching Stata reference numbers.
drlate()— the estimation function.- Reference index — all functions on one page.
Słoczyński, T., Uysal, S. D., and Wooldridge, J. M. (2022). Doubly Robust Estimation of Local Average Treatment Effects Using Inverse Probability Weighted Regression Adjustment. arXiv:2208.01300. https://doi.org/10.48550/arXiv.2208.01300