estimatePrb: Compute the marginal rejection level

In PoolBal: Balancing Central and Marginal Rejection of Pooled p-Values

Compute the marginal rejection level

Description

Estimates the marginal rejection level for an arbitrary pooled p-value function.

Usage

estimatePrb(
  poolFun,
  alpha = 0.05,
  b = 1,
  M = 2,
  interval = c(0, b),
  poolArgs = list(),
  ...
)

Arguments

poolFun	function accepting a vector of p-values
alpha	numeric between 0 and 1
b	numeric, the value of the M - 1 repeated p-values
M	integer, how many p-values are there?
interval	two numerics giving the bounds of root-searching
poolArgs	(optional) additional named arguments for poolFun
...	additional arguments to uniroot

Details

The marginal rejection level is the maximum p-value in a single test less than b which still results in rejection of the null when all other tests have a p-value of b.

This function is essentially a wrapper for uniroot, and accepts a pooling function which takes a numeric vector as its first argument and potentially other arguments given in poolArgs and returns a single value. Using this pooling function, a specified dimension M and a rejection level alpha, uniroot searches for the root to poolFun - alpha along one margin when all other p-values are equal to b.

Value

The uniroot output.

Author(s)

Chris Salahub

Examples

stopool <- function(p) pnorm(sum(qnorm(p, lower.tail = FALSE))/ sqrt(length(p)), lower.tail = FALSE)
estimatePrb(stopool, 0.05, M = 10, interval = c(.Machine$double.eps, 1))
estimatePrb(stopool, 0.05, M = 10, b = 0.5, interval = c(.Machine$double.eps, 1))

PoolBal documentation built on Nov. 22, 2023, 5:07 p.m.