R/boots_g.R

In mattblackwell/DirectEffects: Estimating Controlled Direct Effects for Explaining Causal Findings

#' Coefficient Estimates across Bootstrapped Samples
#' 
#' Performs a simple bootstrap of a fitted DirectEffects model by re-estimating
#' the model with bootstrap samples. 
#'
#' @param seqg A fitted sequential_g estimate, computed by \code{\link{sequential_g}}.
#' @param boots The number of bootstrap replicates. Defaults to 1000.
#'
#' @return An object of type \code{seqgboots} which is a matrix with \code{boots}
#' rows and columns for each coefficient in the \code{seqg} model. Use \code{summary} 
#' to provide summary statistics, such as mean and quantiles.
#'
#' @examples
#' data(ploughs)
#' s1 <- sequential_g(women_politics ~ plow +
#'  agricultural_suitability + tropical_climate + large_animals + rugged |
#'  years_civil_conflict + years_interstate_conflict  + oil_pc +
#'  european_descent + communist_dummy + polity2_2000 |
#'  centered_ln_inc + centered_ln_incsq, ploughs)
#'
#' out.boots <- boots_g(s1, boots = 100)
#'
#' summary(out.boots)
#' 
#' @export
#' @importFrom stats quantile
boots_g <- function(seqg, boots = 1000) {
  stopifnot(boots %% 1 == 0 & boots > 0)

  B <- matrix(NA,
              nrow = boots,
              ncol = length(coef(seqg)),
              dimnames = list(NULL, names(coef(seqg))))
  nobs <- nrow(seqg$model)

  # init
  for (i in 1:boots) {
    # bootstrap model frame
    draw <- sample(1:nobs, replace = TRUE)

    # run seqg
    seqg.draw <- sequential_g(formula = seqg$formula,
                              data = seqg$model[draw, ])

    # store coef
    B[i, ] <- coef(seqg.draw)
  }

  class(B) <- "seqgboots"
  return(B)
}