This function computes the p-value for an association score using
interpolation from a precomputed table
(fast - only available for the bivariate case) or using a permutation test.
Usage
ma.test(d, maStruct, permPartition, lookup, reps)
Arguments
d
the data set
maStruct
the structure returned by a call to ma.
Must be provided. No default.
permPartition
a permutation partition for the data set depending on
the statistical question being asked.
Defaults to the partition returned by the call to ma.
lookup
logical, if TRUE then a p-value is interpolated from a table of
precomputed p-values.
If FALSE then a monte carlo simulation is carried out.
Table lookup is only accurate for associations on bivariate data
with number of samples
reps
no of repeats for the monte carlo simulation.
defaults to 1000.
Value
Returns a real number in the range [0,1].
Details
Computes a p value for an association from the likelihood ratio statistic
computed by ma. Uses interpolation by default for bivariate associations,
and permutation testing for multivariate associations.
d <- shpd(n=100,m=2,Rsq=0.2)
aScore <- ma(d)
ma.test(d, aScore)
# or set lookup to FALSE if you have computing time available# ma.test(d, aScore, lookup=FALSE, reps=10000)