Work

Selection and Validation of Comparative Study of Normality Test

Year: 2018

Type: article

Abstract: The goal of this study is to investigate the best goodness-of-fit test among five selected normality tests under various continuous non-normal distributions using power as criteria. The tests were compared when the normal parameters are unknown and sample sizes are 10, 30, 50, 100, 300, 500 and 1000 were iterated 1000 times each with 0.01, 0.05, and 0.10 level of significance, using the Monte Carlo technique. We study the procedures based on five well-known normality tests: the Anderson–Darling, Cramer–von Mises, Shapiro–Wilk, Jarque–Bera and Chi-Square. Evidence from the simulation study reveals that the performance of the five normality test statistics varies with the level of significance, sample size and alternative distributions.

Source: American journal of mathematics and statistics

Authors: Babalola A. Mayowa, Obubu Maxwell, Oluwaseun A. Otekunrin, Obiora-Ilono Happiness

Language: English

Topic: Statistical Distribution Estimation and Applications

Subfield: Statistics and Probability

Field: Mathematics

Domain: Physical Sciences

Open Access status: diamond

OpenAlex ID: W2911695285

FWCI: 0.575

Cites:

Cited by: 5

Related to: 10