classical.test: PIM implementations of distributionfree tests

Description Usage Arguments Value Note

Description

PIM implementations of distributionfree tests

Usage

1
2
3
4
classical.test(test = c("WilcoxonMannWhitney", "KruskalWallis",
  "MackSkillings", "BrownHettmansperger", "JonckheereTerpstra", "MackWolfe"),
  data, out, group, block, varianceestimator = varianceestimator.H0(),
  alternative = c("two.sided", "greater", "less"), levelP, verbosity = 0)

Arguments

test

Type of "classical" distribution free test to perform.

data

Context where out, group and block are to be interpreted

out

Column of data that holds the outcomes (responses)

group

Column of data that holds the predicting variable

block

Column of data that holds the blocking variable. This can be left out if no blocking is present

varianceestimator

Function like the result of varianceestimator.sandwich() (the default). The default (varianceestimator.H0) will provide the classical test statistic. With the Sandwich estimator, you can achieve a Wald-type test.

alternative

As for other tests

levelP

The level of the grouping variable that is the "top" of the umbrella (only relevant for test="MackWolfe").

verbosity

The higher this value, the more levels of progress and debug information is displayed (note: in R for Windows, turn off buffered output)

Value

a list holding the following items:

statistic

The test statistic.

p.value

p-value for the test.

df

Degress of freedom (if relevant, otherwise NA).

conversion

Function to convert the pim estimates to the test statistic.

Note

These functions are merely provided for comparison. The pim function supports generic ways of estimating these values. For generosity, the parameters to all of them are the same, though some are not relevant (e.g. block for WMW).


pimold documentation built on May 2, 2019, 5:50 p.m.