sglr (version 0.8)

sglr-package: A package for computing the boundaries for sequential generalized likelihood ratio test for pre-licensure vaccine studies

Description

This package is an implementation of the methodology of Shih, Lai, Heyse, and Chen (to appear in Statistics in Medicine) for computing Generalized Likelihood Ratio test boundaries for pre-licensure vaccine studies

Details

Package:	sglr
Type:	Package
Version:	0.05
Date:	2010-04-20
License:	GPL (version 2 or later)
LazyLoad:	yes

The package provides several functions. The function glrSearch computes boundaries for testing a given \(p_0\) versus \(p_1\) (specified as a two-dimensional vector) given a significance level \(\alpha\) and a type II error \(\beta\). The function computeBoundary computes the boundary in terms of a more understandable and usable quantity, such as the number of adverse events in a pre-licensure vaccine study for example. It takes as input a set of given boundaries for the GLR statistic. The third function is plotBoundary which also takes the same arguments as \(computeBoundary\) and produces a plot. The last two functions can make use of statistics computed previously for the problem, which can be specified as an argument; otherwise, the statistics are computed from scratch.

References

``Sequential Generalized Likelihood Ratio Tests for Vaccine Safety Evaluation'' 10.1002/sim.4036.

Examples

# NOT RUN {
library(sglr)
result <- glrSearch(p=c(.5, .75), alpha=0.05, beta=0.10)
## print(result)  ## large amounts of output possible!
result[1:3]
# }