npc: Non-parametric combination of tests

In statlab/permuter: permuter: Permutation Tests

Description

Combines partial p-values from individual hypothesis tests H_{0i} against H_{1i} to test the global null hypothesis

\cap_{i} H_{0i}

against the alternative

\cup_{i} H_{1i}

using a combining function.

Usage

npc(statistics, distr, combine = "fisher", alternatives = "greater")

Arguments

statistics	Vector of observed statistics for each partial test
distr	Matrix or dataframe, columns are approimate null distribution for each partial test
combine	Combining function (default is 'fisher'). May either be the name of a function ('fisher', 'liptak', 'tippett') or a function itself
alternatives	Optional, vector of alternatives for each test (default is all 'greater')

Details

For details on the combining functions, see fisher, liptak, and tippett.

Alternative options are 'greater', 'less', or 'two-sided'. If specified, length of alternatives must either be 1 or match the number of statistics.

Value

A single p-value for the global test

statlab/permuter documentation built on May 30, 2019, 9:45 a.m.