Hartung: Hartung's combination test for dependent p-values

In punitroots: Unit Roots in Panels of Time Series

Description

This function implements the procedure for combining dependent tests of significance proposed by Hartung (1999).

Usage

  Hartung(p, lambda=rep(1,length(p)), kappa=0.2, alpha=0.10)

Arguments

p	the vector of p-values.
lambda	a vector of weights. It must be of the same length as p.
kappa	adjustment parameter. It can be either a positive scalar (0.2 is the default value) or it can take the character value "formula". When k = "formula" is used, then it is computed as in Hartung (1999), p. 853.
alpha	level for the 1-alpha confidence interval for rho (0.10 is the default).

Value

The function returns a list of class "htest" containing:

statistic	the Ht test statistic.
parameter	the number of combined tests (p-values).
p.value	the p-value of the combination test.
conf.int	the confidence interval for the estimated correlation.
estimate	the estimated correlation.
null.value	the specified hypothesized value under the null.
alternative	a character string describing the alternative hypothesis.
method	a character string indicating the type of combination test.
data.name	a character string giving the name of the vector of p-values.

Author(s)

Claudio Lupi

References

Hartung, J (1999). A Note on Combining Dependent Tests of Significance, Biometrical Journal, 41 (7), 849–855.

Examples

  fake.pvalues <- runif(20)
  Hartung(fake.pvalues)

punitroots documentation built on May 2, 2019, 5:16 p.m.