A Comparison of the Power of Test of Test Statistics for 2×2 Factorial Designs

Abstract

The objective of this research is to compare the power of test of parametric and nonparametric test statistics, namely F-test, ranked transform test (RT-Test) and adjusted ranked transform test (ART-Test) for balanced 2×2 factorial designs including tests of main effects and interaction effect. The power of tests was calculated with 10,000 replications in each situation. The data was conducted from error distributed as chi-square. The results indicated that ART-Test performs well for interaction effect test in the most of situations. For main effect tests, the RTTest gives the most power of test.