Robustness of trimmed F statistic when handling non-normal data
When the assumptions of normality and homoscedasticity are met, researchers should have no doubt in using classical test such as t-test, to test for the equality of central tendency measures for two groups.However, in real life this perfect situation is rarely encountered.When the problem of non-nor...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
University of Malaya
2013
|
| Subjects: | |
| Online Access: | http://repo.uum.edu.my/10742/1/z.pdf http://repo.uum.edu.my/10742/ http://e-journal.um.edu.my/publish/MJS/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| id |
my.uum.repo.10742 |
|---|---|
| record_format |
eprints |
| spelling |
my.uum.repo.107422016-12-04T07:55:56Z http://repo.uum.edu.my/10742/ Robustness of trimmed F statistic when handling non-normal data Md Yusof, Zahayu Othman, Abdul Rahman Syed Yahaya, Sharipah Soaad QA Mathematics When the assumptions of normality and homoscedasticity are met, researchers should have no doubt in using classical test such as t-test, to test for the equality of central tendency measures for two groups.However, in real life this perfect situation is rarely encountered.When the problem of non-normality and variance heterogeneity simultaneously arise, rates of Type I error are usually inflated resulting in spurious rejection of null hypotheses.In addition, the classical least squares estimators can be highly inefficient when assumptions of normality are not fulfilled.The effect of non-normality on the trimmed F statistic was demonstrated in this study.We propose the modifications of the trimmed F statistic mentioned by using (1) a priori determined 15% symmetric trimming and (2) empirically determined trimming using robust scale estimators such as MAD n , T n and LMS n .The later trimming method will trim extreme values without prior trimming percentage. Based on the rates of Type I error, the procedures were then compared. Data from g-and h-distributions were considered in this study.We found the trimmed F statistic using robust scale estimator LMS n as trimming criterion provided good control of Type I error compared to the other methods. University of Malaya 2013 Article PeerReviewed application/pdf en http://repo.uum.edu.my/10742/1/z.pdf Md Yusof, Zahayu and Othman, Abdul Rahman and Syed Yahaya, Sharipah Soaad (2013) Robustness of trimmed F statistic when handling non-normal data. Malaysian Journal of Science, 32 (1). pp. 73-77. ISSN 1394-3065 http://e-journal.um.edu.my/publish/MJS/ |
| institution |
Universiti Utara Malaysia |
| building |
UUM Library |
| collection |
Institutional Repository |
| continent |
Asia |
| country |
Malaysia |
| content_provider |
Universiti Utara Malaysia |
| content_source |
UUM Institutionali Repository |
| url_provider |
http://repo.uum.edu.my/ |
| language |
English |
| topic |
QA Mathematics |
| spellingShingle |
QA Mathematics Md Yusof, Zahayu Othman, Abdul Rahman Syed Yahaya, Sharipah Soaad Robustness of trimmed F statistic when handling non-normal data |
| description |
When the assumptions of normality and homoscedasticity are met, researchers should have no doubt in using classical test such as t-test, to test for the equality of central tendency measures for two groups.However, in real life this perfect situation is rarely encountered.When the problem of non-normality and variance heterogeneity simultaneously arise, rates of Type I error are usually inflated resulting in spurious rejection of null hypotheses.In addition, the classical least squares estimators can be highly inefficient when assumptions of normality are not fulfilled.The effect of non-normality on the trimmed F statistic was demonstrated in this study.We propose the modifications of the trimmed F statistic mentioned by using (1) a priori determined 15% symmetric trimming and (2) empirically determined trimming using robust scale estimators such as MAD n , T n and LMS n .The later trimming method will trim extreme values without prior trimming percentage. Based on the rates of Type I error, the procedures were then compared. Data from g-and h-distributions were considered in this study.We found the trimmed F statistic using robust scale estimator LMS n as trimming criterion provided good control of Type I error compared to the other methods. |
| format |
Article |
| author |
Md Yusof, Zahayu Othman, Abdul Rahman Syed Yahaya, Sharipah Soaad |
| author_facet |
Md Yusof, Zahayu Othman, Abdul Rahman Syed Yahaya, Sharipah Soaad |
| author_sort |
Md Yusof, Zahayu |
| title |
Robustness of trimmed F statistic when handling non-normal data |
| title_short |
Robustness of trimmed F statistic when handling non-normal data |
| title_full |
Robustness of trimmed F statistic when handling non-normal data |
| title_fullStr |
Robustness of trimmed F statistic when handling non-normal data |
| title_full_unstemmed |
Robustness of trimmed F statistic when handling non-normal data |
| title_sort |
robustness of trimmed f statistic when handling non-normal data |
| publisher |
University of Malaya |
| publishDate |
2013 |
| url |
http://repo.uum.edu.my/10742/1/z.pdf http://repo.uum.edu.my/10742/ http://e-journal.um.edu.my/publish/MJS/ |
| _version_ |
1644280447900844032 |
| score |
13.211869 |