R/effectSizeReplicationSuccess.R

In ReplicationSuccess: Design and Analysis of Replication Studies

effectSizeReplicationSuccess <- function(zo, 
                                         c = 1, 
                                         level = 0.025, 
                                         alternative = "one.sided",
                                         type = "golden"){
  
  mRV <- mapply(FUN = function(zo, c, level){
    if (!is.numeric(level) || (level <= 0 || level >= 1)) 
      stop("level must be numeric and in (0,1)!")
    
    alphas <- levelSceptical(level = level, 
                             alternative = alternative, 
                             type = type)
    zalphas <- p2z(alphas, alternative = alternative)
    K <- zo^2/zalphas^2
    
    if (zalphas > zo) {
      warning(paste("Replication success is not achievable at this level as",
                    zo, " < ", round(p2z(levelSceptical(level = level,
                                                            alternative = alternative,
                                                            type = type)),
                                     3)))
      d <- NA
    } else {
    d <- ifelse(c < Inf, sqrt(1 + c/(K - 1))/(sqrt(K * c)), 1/sqrt(K * (K - 1)))
    }
    return(d)
  }, zo, c, level)
  
  return(mRV)
}