mcmc.pval {MCMC.qpcr}

calculates p-value based on Bayesian z-score or MCMC sampling

Description

Accessory function to HPD... function series

Usage

mcmc.pval(dat, testlim = 0, sided = 2, ptype="z")

Arguments

dat

a table of MCMC samples

testlim

positive value to test whether the posterior crosses it (for variance components)

sided

sidedness of the test

ptype

'z' for z-score based p-values, 'mcmc' for standard sampling based p-values

Value

A vector of calculated Bayesian p-values. For standard 'mcmc' p=values, the minimal possible value is 1/(MCMC sample size). z-score based p-values are useful to approximate very low p-values with limited MCMC sample sizes.

Author(s)

Mikhail Matz, UT Austin <[e-mail]>