MCPAN (version 1.1-21)

PQln: Generalized pivotal quantity for log expected values of several lognormal samples.

Description

Computes samples of the generalized pivotal quantities for the log of expected values of lognormal samples according to Krishnamoorthy and Mathew (2003). For internal use in lndci, lnrci.

Usage

PQln(nx, mlx, varlx, B = 10000)

Arguments

nx

a numeric (integer) vector, the sample sizes for each of K groups

mlx

a numeric vector, the sample estimates of the log mean for each of K groups

varlx

a numeric vector, the (unbiased) sample estimates of the log variance for each of K groups

B

a single (integer) value, the number of samples to draw, defaults to 10000

Value

A BxK matrix

Details

Generates B samples from independent standard normal and chi^2 distributions for each of K independent groups in a one-way layout and computes the generalized pivotal quantities for the log expected value for each of the groups, according to Krishnamoorthy and Mathew (2003). The input vectors nx, mlx, varlx must be vectors of length K.

References

Krishnamoorthy K, Mathew T (2003): Inferences on the means of lognormal distributions using generalized p-values and generalized con dence intervals. Journal of Statistical Planning and Inference 115, 103-121.