R/dg612.R

In DasGuptR: Das Gupta Standardisation and Decomposition

#' Das Gupta equation 6.12 for a differences between 2 populations when standardised across N populations.
#' @param srates a dataframe output from dg2p
#' @param all_p character or numeric vector of all N populations
#' @param ps vector of length 2 specifying a possible pairwise comparison of populations
#' @param factor character string indicating name of factor
#' @return data.frame object including K-a standardised-rate-differences for each population for given factor a, across N populations
#' @export
dg612 <- function(srates, all_p, ps, factor) {
  n <- length(all_p)
  y1 <- ps[1]
  y2 <- ps[2]

  eq.612.a <- srates[srates$pop == y1 & srates$std.set == y2, "diff"]

  # eq.612.a = srates[srates$pop == y2 & srates$std.set == y1, 'rate'] -
  #   srates[srates$pop == y1 & srates$std.set == y2, 'rate'])

  eq.612.b <- sum(
    sapply(all_p[!all_p %in% c(y1, y2)], \(yy)
    # a_12
    eq.612.a +
      # a_2j
      srates[srates$pop == y2 & srates$std.set == yy, "diff"] -
      # -a_1j
      srates[srates$pop == y1 & srates$std.set == yy, "diff"])
  ) / n

  res <- data.frame(
    diff = eq.612.a - eq.612.b,
    pop = y1,
    diff.calc = paste0(y1, "-", y2),
    std.set = paste0(all_p[!all_p %in% c(y1, y2)], collapse = "."),
    factor = factor
  )

  return(res)
}