ERIC Number: ED092583
Record Type: Non-Journal
Publication Date: 1974-Apr
Reference Count: N/A
Monte Carlo Study of the Power of H-Test Compared to F-Test When Population Distributions Are Different in Form.
Srisukho, Dirake; Marascuilo, Leonard A.
Based on a Monte Carlo simulation, this study is designed to investigate the power of the Kruskal-Wallis's H-test compared to the power of the F-test for three equal moderate sample sizes drawn at random from distributions of common or different shapes but for which the population distributions have equal variances. The distributions are the Normal, Uniform, and Double Exponential. It was found that the F-test is robust to violating assumptions of non-normality for sample of size 10, 15, and 20 but the power of H-test is affected by the shape of the population distributions. The power of the H-test increase faster when all samples are drawn from Double Exponential distribution than the power of H-test drawn from all Normal or all Uniform distribution. It is also found that the power of H-test is greater than the power of the F-test when all samples are drawn from double exponential distribution and the combinations of double exponential and normal distributions. The power of H-test is almost identical to the power of F-test when 2 samples are drawn from the double exponential and one sample is from the uniform distribution and when 3 samples are drawn from 3 different shaped distributions. The power of H-test is less than the power of the F-test when all samples are from normal or from uniform distributions. (Author/BB)
Publication Type: Speeches/Meeting Papers
Education Level: N/A
Authoring Institution: N/A