A study on the criticism and alternatives about statistical hypothesis testing of sport management in Korea
Purpose This study was aimed at laying out criticism regarding statistical hypothesis testing and presenting realistic alternatives focused on published studies of sports management in Korea. Methods Among 202 studies compiled by the Korean Society for Sport Management, vol. 19, no. 1 through vol. 2...
Main Author: | |
---|---|
Format: | Article |
Language: | Korean |
Published: |
Korea Institute of Sport Science
2020-06-01
|
Series: | Cheyuk gwahag yeon-gu |
Subjects: | |
Online Access: | http://kjss.sports.re.kr/upload/pdf/KISS_2020_v31n2_216.pdf |
Summary: | Purpose This study was aimed at laying out criticism regarding statistical hypothesis testing and presenting realistic alternatives focused on published studies of sports management in Korea. Methods Among 202 studies compiled by the Korean Society for Sport Management, vol. 19, no. 1 through vol. 23, no. 6, 115 studies which used the null hypothesis significant testing were finally selected. After data coding for selected studies, p-curve, the distribution of p-values which reported in individual studies, was schematized, and than adequacy about the sampling method and result descriptions were analyzed. Results The ratio of p-values close to zero was relatively very high in p-curve although there was no clear evidence of p-hacking. Also, approximately 82% of the studies used convenience sampling method, and incorrect descriptions in part of result and discussion due to lack of understanding of statistical hypothesis testing was found in some studies. In conclusion, the shortcomings of statistical hypothesis testing which commonly used in academic field were depended on not a defect in the method itself but researchers who misused statistical hypothesis testing. Conclusions Researchers in sport management field need to understand the advantages and disadvantages of statistical hypothesis testing and consider to use both confidence interval and effect sizes to compensate the disadvantages of p-value. |
---|---|
ISSN: | 1598-2920 2233-7938 |