A simulation study to compare the power of normality tests

A common assumption of many statistical procedures during data analysis is that the data is normally distributed. Several statistical tests have been developed for the determination of the validity of this assumption. Now, the question is: which is the most powerful? We performed an extensive Monte...

Full description

Saved in:
Bibliographic Details
Main Authors: Erdélyi Péter
Rajkó Róbert
Corporate Author: International Symposium on Analytical and Environmental Problems (23.) (2017) (Szeged)
Format: Book part
Published: 2017
Series:Proceedings of the International Symposium on Analytical and Environmental Problems 23
Kulcsszavak:Statisztikai hipotézisvizsgálat
Online Access:http://acta.bibl.u-szeged.hu/56237
Description
Summary:A common assumption of many statistical procedures during data analysis is that the data is normally distributed. Several statistical tests have been developed for the determination of the validity of this assumption. Now, the question is: which is the most powerful? We performed an extensive Monte Carlo simulation study to answer this question. We found that compared to six other tests, the Shapiro–Wilk test performs the best under most conditions.
Physical Description:397-400
ISBN:978-963-306-563-1