Journals Information
Mathematics and Statistics Vol. 10(1), pp. 215 - 221
DOI: 10.13189/ms.2022.100120
Reprint (PDF) (453Kb)
The Power of Test of Jennrich Statistic with Robust Methods in Testing the Equality of Correlation Matrices
Bahtiar Jamili Zaini *, Shamshuritawati Md Sharif
School of Quantitative Sciences, Universiti Utara Malaysia, 06010 UUM Sintok, Kedah, Malaysia
ABSTRACT
Jennrich statistic is a method that can be used to test the equality of 2 or more independent correlation matrices. However, Jennrich statistic begins to be problematic when there is presence of outliers that could lead to invalid results. When exiting outliers in data, Jennrich statistic implications will affect Type I errors and will reduce the power of test. To overcome the presence of outliers, this study suggests the use of robust methods as an alternative method and therefore, it will integrate the estimator into Jennrich statistic. Thus, it can improve the testing performance of correlation matrix hypotheses in relation to outlier problems. Therefore, this study proposes 3 statistical tests, namely Js-statistic, Jm-statistic, and Jmad-statistic that can be used to test the equation of 2 or more correlation matrices. The performance of the proposed method is assessed using the power of test. The results show that Jm-statistic and Jmad-statistic can overcome outlier problems into Jennrich statistic in testing the correlation matrix hypothesis. Jmad-statistic is also superior in testing the correlation matrix hypothesis for different sample sizes, especially those involving 10% outliers.
KEYWORDS
Correlation, Jennrich Statistic, Robust Correlation Methods, Outlier
Cite This Paper in IEEE or APA Citation Styles
(a). IEEE Format:
[1] Bahtiar Jamili Zaini , Shamshuritawati Md Sharif , "The Power of Test of Jennrich Statistic with Robust Methods in Testing the Equality of Correlation Matrices," Mathematics and Statistics, Vol. 10, No. 1, pp. 215 - 221, 2022. DOI: 10.13189/ms.2022.100120.
(b). APA Format:
Bahtiar Jamili Zaini , Shamshuritawati Md Sharif (2022). The Power of Test of Jennrich Statistic with Robust Methods in Testing the Equality of Correlation Matrices. Mathematics and Statistics, 10(1), 215 - 221. DOI: 10.13189/ms.2022.100120.