RDMono: Inference in RD under Monotonicity

In koohyun-kwon/rdadapt:

Inference in RD under Monotonicity

Description

Description to be supplied.

Usage

RDMono(
  X,
  Y,
  t.ind,
  C,
  mon.ind,
  C.l,
  C.u,
  method = c("mm", "adpt.L", "adpt.U"),
  se.method,
  se.init,
  alpha,
  N
)

Arguments

X	a vector or a matrix of the running variables, with ncol(X) being the dimension of the turnning variable when X is a matrix.
Y	a vector of the outcome variable.
t.ind	a vector of treatment indicators, with t.ind = 1 indicating treated observations and t.ind = 0 control observations.
C	bound on the first derivative of the regression function; if C is missing, C = Inf is used when method corresponds to one of c("adpt.L", "adpt.U") (see below).
mon.ind	index for monotone variables, a subset of 1:ncol(X) if X is a matrix, either 1 or 0 if X is a vector.
C.l	lower bound for the Lipschitz constants to which the adaptive one-sided CI adapt to.
C.u	upper bound for the Lipschitz constants to which the adaptive one-sided CI adapt to.
method	type of confidence interval to be used. The options are: "mm" Minimax two-sided confidence interval "adpt.L" Adaptive one-sided lower confidence interval "adpt.U" Adaptive one-sided upper confidence interval
se.method	method for estimating the standard error of the estimate. (to be edited later)
se.init	method for estimating the initial variance for computing the optimal bandwidth. (to be edited later)
alpha	determines confidence level, 1 - alpha.
N	number of nearest neighbors to be used when "nn" is specified in se.method.

Details

Under development.

Value

a list of following components

ci

Confidence interval

ht, hc

Bandwidths used for treated and control observations

Examples

print("under development")

koohyun-kwon/rdadapt documentation built on May 8, 2022, 8:49 p.m.