R/p.threshold.adjust.R

In drhutools: Political Science Academic Research Gears

#' Calculate Vector of p-value Thresholds for Multiple Inference
#'
#' Internal function. Calculates the adjusted thresholds for multiple significance
#' tests, assuming that the original threshold for a single test is p<0.05.
#' Used by [fsQCApermTest()] to calculate confidence intervals.
#' @param total.configs The total number of hypotheses tested, or the number of
#' configurations utilized by the Quine-McCluskey algorithm in fsQCA (including
#' logical remainders, if they are used in the analysis).
#' @param my.method The adjustment method used to calculate p-values (see
#' [stats::p.adjust()] for details).
#' @return Numeric vector giving adjusted p-value thresholds, from smallest to
#' largest.
#' @note Adopted from the archived CRAN package \pkg{QCAfalsePositive} by Bear
#'   Braumoeller. The original package has been removed from CRAN and is no
#'   longer maintained. It is included in \pkg{drhutools} for continued
#'   accessibility.
#' @concept p-value threshold fsQCA
#' @export
#' @examples
#' p.threshold.adjust(10, "holm")


p.threshold.adjust <- function(total.configs, my.method){
	pvec <- NULL
	for(m in 1:total.configs){
		pvec <- c(pvec, p.adjust(c(rep(1, m-1), 0.05, rep(0, total.configs-m)), method=my.method)[m])
	}
	out.vec <- rep(0.05, total.configs) / (pvec/0.05)
	rev(out.vec)
}