RDLPreg: Local Polynomial Regression in Sharp RD

Description Usage Arguments Value

View source: R/RDfunctions.R

Description

Calculate sharp RD estimate and its variance given a bandwidth using local polynomial regression of order order.

Usage

1
2
RDLPreg(d, hp, kern = "triangular", order = 1, hm = hp,
  se.method = "nn", no.warning = FALSE, J = 3)

Arguments

d

object of class "RDData"

hp, hm

bandwidth for treated (units with positive running variable), and control (units with negative running variable) units. If hm is not supplied, it is assumed to equal to hp. If neither bandwidth is supplied, optimal bandwidth is computed according to criterion given by opt.criterion.

kern

specifies kernel function used in the local regression. It can either be a string equal to "triangular" (k(u)=(1-|u|)_{+}), "epanechnikov" (k(u)=(3/4)(1-u^2)_{+}), or "uniform" (k(u)= (|u|<1)/2), or else a kernel function.

order

Order of local regression 1 for linear, 2 for quadratic.

se.method

Vector with methods for estimating standard error of estimate. If NULL, standard errors are not computed. The elements of the vector can consist of the following methods:

"nn"

Nearest neighbor method

"EHW"

Eicker-Huber-White, with residuals from local regression (local polynomial estimators only).

"demeaned"

Use EHW, but instead of using residuals, estimate sigma^2_i by subtracting the estimated intercept from the outcome (and not subtracting the estimated slope). Local polynomial estimators only.

"plugin"

Plug-in estimate based on asymptotic variance. Local polynomial estimators in RD only.

"supplied.var"

Use conditional variance supplied by sigma2 / d instead of computing residuals

no.warning

Don't warn about too few observations

J

Number of nearest neighbors, if "nn" is specified in se.method.

Value

list with elements:

estimate

point estimate

se

Named vector of standard error estimates, as specified by se.method.

wm,wp

Implicit weight functions used

sigma2m, sigma2p

Estimates of sigma^2(X) for values of running variable below (above) cutoff and receiving positive kernel weight. By default, estimates are based on squared regression residuals, as used in "EHW". If "demeaned" or "nn" is specifed, estimates are based on that method, with "nn" method used if both are specified.

eff.obs

Number of effective observations


kolesarm/RDHonest documentation built on April 3, 2018, 11:08 a.m.