R/esc_2x2.R

In esc: Effect Size Computation for Meta Analysis

#' @title Compute effect size from 2 by 2 Contingency Table
#' @name esc_2x2
#'
#' @description Compute effect size from a 2 by 2 frequency table.
#'
#' @param grp1yes Size of treatment group with successes (outcome = yes).
#' @param grp1no Size of treatment group with non-successes (outcome = no).
#' @param grp2yes Size of control group with successes (outcome = yes).
#' @param grp2no Size of control group with non-successes (outcome = no).
#'
#' @inheritParams esc_beta
#' @inheritParams esc_t
#'
#' @return The effect size \code{es}, the standard error \code{se}, the variance
#'         of the effect size \code{var}, the lower and upper confidence limits
#'         \code{ci.lo} and \code{ci.hi}, the weight factor \code{w} and the
#'         total sample size \code{totaln}.
#'
#' @note If \code{es.type = "r"}, Fisher's transformation for the effect size
#'       \code{r} and their confidence intervals are also returned.
#'
#' @references Lipsey MW, Wilson DB. 2001. Practical meta-analysis. Thousand Oaks, Calif: Sage Publications
#'             \cr \cr
#'             Wilson DB. 2016. Formulas Used by the "Practical Meta-Analysis Effect Size Calculator". Unpublished manuscript: George Mason University
#'
#' @examples
#' # effect size log odds
#' esc_2x2(grp1yes = 30, grp1no = 50, grp2yes = 40, grp2no = 45)
#'
#' # effect size odds ratio
#' esc_2x2(grp1yes = 30, grp1no = 50, grp2yes = 40, grp2no = 45, es.type = "or")
#'
#' @export
esc_2x2 <- function(grp1yes,
                    grp1no,
                    grp2yes,
                    grp2no,
                    es.type = c("logit", "d", "g", "or", "r", "f", "eta", "cox.d"),
                    study = NULL,
                    ...) {
  # match  arguments
  es.type <- match.arg(es.type)


  # evaluate ellipses
  ell <- match.call(expand.dots = FALSE)$`...`

  # do we have info string?
  if (!is.null(ell$info))
    info <- ell$info
  else
    # if not, set default
    info <- "2x2 table (OR)"


  es <- (grp1yes * grp2no) / (grp1no * grp2yes)
  v <- 1 / grp1yes + 1 / grp1no + 1 / grp2yes + 1 / grp2no

  totaln <- grp1yes + grp1no + grp2yes + grp2no

  # which es type to be returned?
  if (es.type == "or") {
    return(structure(
      class = c("esc", "esc_2x2"),
      list(es = es, se = sqrt(v), var = v, ci.lo = exp(lower_d(log(es), v)),
           ci.hi = exp(upper_d(log(es), v)), w = 1 / v, totaln = totaln,
           measure = "or", info = paste0(info, " coefficient to effect size odds ratio"),
           study = study)
    ))
  }

  # which es type to be returned?
  if (es.type == "d") return(convert_or2d(or = es, v = v, totaln = totaln, es.type = "d", info = paste0(info, " to effect size d"), study = study))

  # which es type to be returned?
  if (es.type == "cox.d") return(convert_or2d(or = es, v = v, totaln = totaln, es.type = "cox.d", info = paste0(info, " to effect size Cox d"), study = study))

  # which es type to be returned?
  if (es.type == "g") return(convert_or2d(or = es, v = v, totaln = totaln, es.type = "g", info = paste0(info, " to effect size Hedges' g"), study = study))


  # convert to plain d
  es <- log(es) * sqrt(3) / pi
  v <- v * 3 / pi^2

  # which es type to be returned?
  if (es.type == "logit") return(convert_d2logit(d = es, v = v, totaln = totaln, info = paste0(info, " to effect size logits"), study = study))

  # which es type to be returned?
  if (es.type == "r") return(convert_d2r(d = es, v = v, grp1n = (grp1yes + grp1no), grp2n = (grp2yes + grp2no), info = paste0(info, " coefficient to effect size correlation"), study = study))
}