mass: Value of the definite (two-dimensional) integral of L,...

In AldoCP/MetaBiasR: Assessment of Bias in a Meta-Analysis in a Bayesian framework

Description

Value of the definite (two-dimensional) integral of L, evaluated around the distribution's mean (for mu), and between 0 and 1 (for eta)

Usage

mass(y, sig, sig_prior = 2, Z = 1.96)

Arguments

y	a vector of length n (n=number of studies) representing the effect (standardized mean difference)
sig	a vector of length n (n=number of studies) representing the corresponding sampling standard deviation sqrt(variance)
sig_prior	a numeric entry representing the standard deviation's prior; default: 2
Z	a numeric entry representing the Z value; default: 1.96

Details

This is an auxiliary function to obtain the definite integral of L, evaluated around the distribution's mean (for mu) and between 0 and 1 for eta. It calls the logL function, and solves the integral using integrate from the cubature package. The output is later called by BFbias to compute the Bayes Factor.

Value

the value of the definite (two-dimensional) integral of the L distribution, evaluated around the distribution's mean (+-10 standard deviations, for mu), and between 0 and 1 (for eta)

Author(s)

Aldo Cordova-Palomera

See Also

logL, BFbias

Examples

library(metafor)
dat <- get(data(dat.bangertdrowns2004))
mass(dat$yi, sqrt(dat$vi))

AldoCP/MetaBiasR documentation built on May 5, 2019, 1:36 p.m.