unitest: Univariate F or t testing In ffmanova: Fifty-Fifty MANOVA

 unitests R Documentation

Univariate F or t testing

Description

The functions perform `F` or `t` testing for several responses based on a matrix of hypothesis observations and a matrix of error observations.

Usage

``````unitests(xyObj)

unitest(modelData, errorData, dfError = dim(errorData)[1])
``````

Arguments

 `xyObj` a design-with-responses object created by `xy_Obj` `modelData` matrix of hypothesis observations `errorData` matrix of error observations `dfError` Degrees of freedom for error needs to be specified if `errorData` is incomplete

Details

`modelData` and `errorObs` correspond to `hypObs` and `errorObs` calculated by `xy_Obj`. These matrices are efficient representations of sums of squares and cross-products (see `xy_Obj` for details). This means the univariate `F`-statistics can be calculated straightforwardly from these input matrices. Furthermore, in the single-degree-of-freedom case, `t`-statistics with correct sign can be obtained.

`unitests` is a wrapper function that calls `unitest` for each term in the `xyObj` (see `xy_Obj` for details) and collects the results.

Value

`unitest` returns a list with components

 `pValues` `p`-values `stat` The test statistics as `t`-statistics (when single degree of freedom) or `F`-statistics

`unitests` returns a list with components

 `pRaw` Matrix of `p`-values from `unitest`, one row for each term. `stat` Matrix of test statistics from `unitest`, one row for each term.

Note

The function calculates the `p`-values by making a call to `pf`.

Author(s)

Øyvind Langsrud and Bjørn-Helge Mevik

`rotationtest`, `rotationtests`