betaMeldTest {bpcp}

Melded Beta Test

Description

This function gives a two sample melded beta test together with the associated melded confidence intervals. It can be used when the confidence distributions (or upper and lower confidence distributions) for the one sample parameters are beta, and one is interested in either the difference, odds ratio, or ratio of those two one sample parameters. The betaMeldTest is usually called by bpcp2samp, and not called directly by the user.

Usage

betaMeldTest(betaParms1, betaParms2,
    nullparm = NULL, 
    parmtype = c("difference", "oddsratio", "ratio","cdfratio"), 
    conf.level = 0.95, conf.int = TRUE, 
    alternative = c("two.sided", "less", "greater"), 
    eps = 10^-8, dname = "", estimate1 = NA, estimate2 = NA)

Arguments

betaParms1

a list of the beta parameters for group 1

betaParms2

a list of the beta parameters for group 2

nullparm

null value of the parameter of interest, default of NULL gives 0 if parmtype='difference' and 1 otherwise

parmtype

parameter type for comparing the survival function of the two groups, either 'difference' 'ratio' 'oddsratio' or 'cdfratio'

conf.level

confidence level, e.g., 0.95

conf.int

logical, calculate confidence interval?

alternative

character, either 'two.sided','less', or 'greater'

eps

small value to make integration tractable

dname

name describing data

estimate1

estimate of mean for beta parameter of group 1 (statistic of htest object)

estimate2

estimate of mean for beta parameter for group 2 (parameter of htest object)

Details

If the upper and lower confidence distributions for both samples are described by beta distributions, then you can create a CD test using this function. For example, if you have sample 1 is binomial with x (with 0<x<n) out of n positive responses, then the 100(1-alpha) confidence interval is qbeta(alpha/2, x, n-x+1) and qbeta(1-alpha/2, x+1, n-x). So the lower confidence distribution is beta with parameters a=x and b=n-x+1, and the upper CD is beta with parameters a=x+1 and b=n-x.

See bpcp2samp for a full description of the parmtypes.

Value

an object of class 'htest'

Examples

fisher.test(matrix(c(4,5,2,22),2,2),alternative="greater")

betaMeldTest(
    betaParms1=list(alower=2,blower=22+1,aupper=2+1,bupper=22),
    betaParms2=list(alower=4,blower=5+1,aupper=4+1,bupper=5),
    alternative="greater",parmtype="oddsratio",
    estimate1=2/24,estimate2=4/9)