KKSA.test: Kharrati-Kopaei and Sadooghi-Alvandi's test for interaction

In sdateam/combinIT: Combine Interaction Tests for testing interaction in unreplicated two-way tables

Description

computes Kharrati-Kopaei and Sadooghi-Alvandi's test statistics and corresponding p-value. They proposed a test procedure for testing non-additivity based on splitting the b-by-t data matrix into two non empty sub-matrixes by dividing the rows, each having at least two rows. Suppose that b<e2><89><a5>t and b<e2><89><a5>4.Divided mean square errors for any two sub-matrixes such that maximize its ratio and computed the corresponding p-value. Their test statistics are the minimum value of p-values over all possible configurations: 2^(b-1)-b-1.

Usage

KKSA.test(x, nsim = 1000, distr = "sim", ...)

Arguments

x	A b-by-t data matrix, which rows corresponding to b-block effects and columns are t-treatment effects
nsim	Number of simulation for compueting exact p-value
dist	If dist="sim", we used Monte Carlo simulation for estimating exact p-value, and if dist="asy", Bonferroni-adjusted p-value computed. The defaut value is "sim"

Details

If rows numer(b) of data matrix is less than it's columns number(t) at fiest transposing data matrix. In addition requires that data matrix has more than three rows or columns

Value

A p-value for input

Author(s)

Zahra. Shenavari, ...

References

Kharrati-Kopaei, M., Sadooghi-Alvandi, S. M. (2007). A New Method for Testing Interaction in Unreplicated Two-Way Analysis of Variance. Communications in Statistics-Theory and Methods 36:2787<e2><80><93>2803.

Examples

## Not run: this is an example
data(impurity)
KKSA.test(impurity,nsim=1000,dist = "sim")

sdateam/combinIT documentation built on May 6, 2019, 12:10 p.m.