R/SSR.r.r

In GiulioCostantini/safeguardpower: Safeguard Power Analysis

SSR.r <- function(r, SSR = 1, conf = .80, sig.level = .05, power = .8, nmin = 4, nmax = 1e+6)
{
  tol = 1e-7
  init_nmax <- nmax
  while(abs(nmin-nmax) > tol)
  {
    n <- mean(c(nmin,nmax))
    sg <- suppressWarnings(safeguard.r(r = r, n = n, conf = conf, sig.level = sig.level, power = power))
    SSR2 <- sg["SSR"]
    if (SSR2 < SSR) nmax <- n
    if (SSR2 > SSR) nmin <- n
  }
  n <- round(n)
  sg <- safeguard.r(r = r, n = n, conf = conf, sig.level = sig.level, power = power)
  SSR2 <- sg["SSR"]
  if(SSR2 > SSR)
  {
    n <- n+1
    sg <- safeguard.r(r = r, n = n, conf = conf, sig.level = sig.level, power = power)
    SSR2 <- sg["SSR"]
  }
  if(n >= init_nmax) warning("the estimation might be inaccurate. Try to increase the parameter nmax.")
  c("N" = n, "SSR" = as.numeric(SSR2))
}