Abstract
Goodness-of-fit tests are primarily designed to verify the null hypothesis, which asserts that the observations of a given sample conform to a specific probability distribution—a scenario frequently encountered in practical applications across various fields, particularly in genetics, medical research, and biological studies. In 1965, Shapiro and Wilk proposed an intuitive goodness-of-fit test that accounts for nuisance location and scale parameters, a method that has since garnered significant attention within statistical literature. This work was further advanced by Sen in 2002, who extended the Shapiro-Wilk test to accommodate cases involving nuisance regression and scale parameters. Sen’s approach developed a test based on a pair of probability estimators, utilizing Maximum Likelihood Estimation and L-estimators for standard deviation within a linear regression model. This development facilitated a comparison between the Shapiro-Wilk test and the test introduced by Jureckova in 2003, involving evaluations of the Shapiro-Wilk test in both its original and extended forms across various sample sizes.
DOI
10.33095/jeas.v15i54.1176
Subject Area
Statistical
First Page
269
Last Page
282
Recommended Citation
Mahdi, D. I., & Mohammed, W. J. (2009). Comparison of the Shapiro-Wilk Test Jureckova Test Using Simulation and Multiple Distributions. Journal of Economics and Administrative Sciences, 15(54), 269-282. https://doi.org/10.33095/jeas.v15i54.1176
