R/Naive-SE.R

In MatchItSE: Calculates SE for Matched Samples from 'MatchIt'

#' Parametric bootstrap Standard Error for the ATT
#'
#' Calculates the Standard Error for the  Average Treatment Effect with a
#' a parametric bootstrap method.
#' This procedure mimics the output of the zelig packages as shown
#' on pages 12 and 13 in the MatchIt vignette.
#'
#' @param obj MatchIt Object
#' @param Y Response Vector
#' @return SE for the ATT of \code{Y}
#' @references
#' \href{http://imai.princeton.edu/research/files/matchit.pdf}{MatchIt vignette}
#' @examples
#' \dontrun{
#'   library(MatchIt)
#'   data("lalonde")
#'   m.out  <- matchit(treat ~ educ + black, data = lalonde)
#'   att(obj = m.out, Y = lalonde$re78)
#'   zelig_se(obj = m.out, Y = lalonde$re78)
#' }
#' @export
zelig_se <- function(obj, Y){
  stopifnot(methods::is(obj, "matchit"))
  ww <- obj$weights
  tt <- obj$treat
  XX <- cbind(1,obj$X)
  ic <- ww > 0 & tt == 0  # Index used control units
  it <- ww > 0 & tt == 1  # Index used treatment units
  fitc <- fastReg(Y[ic], XX[ic,], ww[ic])
  beta.hat <- matrix(stats::rnorm(1000 * ncol(fitc$vcov)), nrow = 1000, byrow = TRUE) %*% chol(fitc$vcov)
  beta.hat <- sweep(beta.hat, 2, fitc$coef, "+")  # Add mean
  pool.expected <- beta.hat %*% t(XX[it,])
  att.expected  <- colMeans(Y[it] - t(pool.expected))
  return(c(est = mean(pool.expected), sd = stats::sd(pool.expected),
           att.est = mean(att.expected), att.se = stats::sd(att.expected)))
}