Skip to content

DeclareDesign/estimatr

Repository files navigation

estimatr: Fast Estimators for Design-Based Inference

CRAN status CRAN RStudio mirror downloads Build status Code coverage

estimatr is an R package providing a range of commonly-used linear estimators, designed for speed and for ease-of-use. Users can easily recover robust, cluster-robust, and other design appropriate estimates. We include two functions that implement means estimators, difference_in_means() and horvitz_thompson(), and three linear regression estimators, lm_robust(), lm_lin(), and iv_robust(). In each case, users can choose an estimator to reflect cluster-randomized, block-randomized, and block-and-cluster-randomized designs. The Getting Started Guide describes each estimator provided by estimatr and how it can be used in your analysis.

You can also see the multiple ways you can get regression tables out of estimatr using commonly used R packages such as texreg and stargazer. Fast estimators also enable fast simulation of research designs to learn about their properties (see DeclareDesign).

Installing estimatr

To install the latest stable release of estimatr, please ensure that you are running version 3.4 or later of R and run the following code:

install.packages("estimatr")

If you would like to use the latest development release of estimatr, please ensure that you are running version 3.4 or later of R and run the following code:

install.packages("estimatr", dependencies = TRUE,
                 repos = c("http://r.declaredesign.org", "https://cloud.r-project.org"))

Easy to use

Once the package is installed, getting appropriate estimates and standard errors is now both fast and easy.

library(estimatr)

# sample data from cluster-randomized experiment
library(fabricatr)
library(randomizr)
dat <- fabricate(
  N = 100,
  y = rnorm(N),
  clusterID = sample(letters[1:10], size = N, replace = TRUE),
  z = cluster_ra(clusterID)
)

# robust standard errors
res_rob <- lm_robust(y ~ z, data = dat)
# tidy dataframes on command!
tidy(res_rob)
#>          term estimate std.error statistic p.value conf.low conf.high df
#> 1 (Intercept)    0.065      0.14      0.46    0.64    -0.21      0.34 98
#> 2           z   -0.067      0.21     -0.32    0.75    -0.48      0.35 98
#>   outcome
#> 1       y
#> 2       y

# cluster robust standard errors
res_cl <- lm_robust(y ~ z, data = dat, clusters = clusterID)
# standard summary view also available
summary(res_cl)
#> 
#> Call:
#> lm_robust(formula = y ~ z, data = dat, clusters = clusterID)
#> 
#> Standard error type:  CR2 
#> 
#> Coefficients:
#>             Estimate Std. Error t value Pr(>|t|) CI Lower CI Upper   DF
#> (Intercept)   0.0653      0.145   0.452    0.678   -0.358    0.489 3.53
#> z            -0.0670      0.202  -0.331    0.750   -0.544    0.410 7.05
#> 
#> Multiple R-squared:  0.00105 ,   Adjusted R-squared:  -0.00915 
#> F-statistic: 0.11 on 1 and 9 DF,  p-value: 0.748

# matched-pair design learned from blocks argument
data(sleep)
res_dim <- difference_in_means(extra ~ group, data = sleep, blocks = ID)

The Getting Started Guide has more examples and uses, as do the reference pages. The Mathematical Notes provide more information about what each estimator is doing under the hood.

Fast to use

Getting estimates and robust standard errors is also faster than it used to be. Compare our package to using lm() and the sandwich package to get HC2 standard errors. More speed comparisons are available here. Furthermore, with many blocks (or fixed effects), users can use the fixed_effects argument of lm_robust with HC1 standard errors to greatly improve estimation speed. More on fixed effects here.

dat <- data.frame(X = matrix(rnorm(2000*50), 2000), y = rnorm(2000))

library(microbenchmark)
library(lmtest)
library(sandwich)
mb <- microbenchmark(
  `estimatr` = lm_robust(y ~ ., data = dat),
  `lm + sandwich` = {
    lo <- lm(y ~ ., data = dat)
    coeftest(lo, vcov = vcovHC(lo, type = 'HC2'))
  }
)
estimatr median run-time (ms)
estimatr 21
lm + sandwich 41

This project is generously supported by a grant from the Laura and John Arnold Foundation and seed funding from Evidence in Governance and Politics (EGAP).