multivariance.pvalue {multivariance}

transform multivariance to p-value

Description

Computes a conservative p-value for the hypothesis of independence for a given multivariance / m-multivariance / total multivariance.

Usage

multivariance.pvalue(x)

Arguments

x

value of a normalized multivariance scaled by the sample size (i.e., computed with Nscale = TRUE)

Details

This is based on a distribution-free approach. The p-value is conservative, i.e. it might be much smaller. This is the counterpart to rejection.level. For a less conservative approach see resample.pvalue or pearson.pvalue.

p-values larger than 0.215 might be incorrect, since the distribution-free estimate on which the computation is based only holds up to 0.215.

References

For the theoretic background see the references given on the main help page of this package: multivariance-package.