Sharp RDD treatment effect at the cutoff
Usage
morie_rdd_sharp(
data,
outcome,
running,
cutoff = 0,
bandwidth = NULL,
p = 1,
kernel = "triangular",
cluster = NULL,
covariates = NULL,
alpha = 0.05
)Arguments
- data
Data frame.
- outcome
Outcome column.
- running
Running variable column.
- cutoff
Threshold (default 0).
- bandwidth
Optional bandwidth; if
NULL, CCT MSE-optimal.- p
Local-polynomial order.
- kernel
Kernel name.
- cluster
Optional cluster column.
- covariates
Optional character vector of covariate names.
- alpha
Significance level.
Value
A named list with elements estimate, std_error, t_stat, p_value, ci_lower, ci_upper, n_obs, method, details.
Examples
set.seed(33)
n <- 1500
x <- runif(n, -1, 1)
y <- 0.5 * x + 0.3 * x^2 + 1.5 * (x >= 0) + rnorm(n, sd = 0.5)
d <- data.frame(y, x)
res <- morie_rdd_sharp(d, outcome = "y", running = "x", cutoff = 0)
res$estimate
#> [1] 1.506227