A method of statistical analysis in the field of sports science when assumptions of parametric tests are not violated
Introduction: Application of statistical software typically does not require extensive statistical knowledge, allowing to easily perform even complex analyses. Consequently, test selection criteria and important assumptions may be easily overlooked or given insufficient consideration. In such cases,...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Kazimierz Wielki University
2016-12-01
|
| Series: | Journal of Education, Health and Sport |
| Subjects: | |
| Online Access: | http://www.ojs.ukw.edu.pl/index.php/johs/article/view/4278 |
| _version_ | 1818311835191869440 |
|---|---|
| author | Elżbieta Sandurska Aleksandra Szulc |
| author_facet | Elżbieta Sandurska Aleksandra Szulc |
| author_sort | Elżbieta Sandurska |
| collection | DOAJ |
| description | Introduction: Application of statistical software typically does not require extensive statistical knowledge, allowing to easily perform even complex analyses. Consequently, test selection criteria and important assumptions may be easily overlooked or given insufficient consideration. In such cases, the results may likely lead to wrong conclusions.
Aim: To discuss issues related to assumption violations in the case of Student's t-test and one-way ANOVA, two parametric tests frequently used in the field of sports science, and to recommend solutions.
Description of the state of knowledge: Student's t-test and ANOVA are parametric tests, and therefore some of the assumptions that need to be satisfied include normal distribution of the data and homogeneity of variances in groups. If the assumptions are violated, the original design of the test is impaired, and the test may then be compromised giving spurious results. A simple method to normalize the data and to stabilize the variance is to use transformations. If such approach fails, a good alternative to consider is a nonparametric test, such as Mann-Whitney, the Kruskal-Wallis or Wilcoxon signed-rank tests.
Summary: Thorough verification of the parametric tests assumptions allows for correct selection of statistical tools, which is the basis of well-grounded statistical analysis. With a few simple rules, testing patterns in the data characteristic for the study of sports science comes down to a straightforward procedure. |
| first_indexed | 2024-12-13T08:08:16Z |
| format | Article |
| id | doaj.art-06d7d50027f64bf1937ac688bcafb236 |
| institution | Directory Open Access Journal |
| issn | 2391-8306 |
| language | English |
| last_indexed | 2024-12-13T08:08:16Z |
| publishDate | 2016-12-01 |
| publisher | Kazimierz Wielki University |
| record_format | Article |
| series | Journal of Education, Health and Sport |
| spelling | doaj.art-06d7d50027f64bf1937ac688bcafb2362022-12-21T23:54:16ZengKazimierz Wielki UniversityJournal of Education, Health and Sport2391-83062016-12-0161327528710.5281/zenodo.2937614068A method of statistical analysis in the field of sports science when assumptions of parametric tests are not violatedElżbieta Sandurska0Aleksandra Szulc1Department of Genetics, Institute of Experimental Biology, Kazimierz Wielki UniversityFaculty of Management and Economics, Gdansk University of Technology, GdanskIntroduction: Application of statistical software typically does not require extensive statistical knowledge, allowing to easily perform even complex analyses. Consequently, test selection criteria and important assumptions may be easily overlooked or given insufficient consideration. In such cases, the results may likely lead to wrong conclusions. Aim: To discuss issues related to assumption violations in the case of Student's t-test and one-way ANOVA, two parametric tests frequently used in the field of sports science, and to recommend solutions. Description of the state of knowledge: Student's t-test and ANOVA are parametric tests, and therefore some of the assumptions that need to be satisfied include normal distribution of the data and homogeneity of variances in groups. If the assumptions are violated, the original design of the test is impaired, and the test may then be compromised giving spurious results. A simple method to normalize the data and to stabilize the variance is to use transformations. If such approach fails, a good alternative to consider is a nonparametric test, such as Mann-Whitney, the Kruskal-Wallis or Wilcoxon signed-rank tests. Summary: Thorough verification of the parametric tests assumptions allows for correct selection of statistical tools, which is the basis of well-grounded statistical analysis. With a few simple rules, testing patterns in the data characteristic for the study of sports science comes down to a straightforward procedure.http://www.ojs.ukw.edu.pl/index.php/johs/article/view/4278t-test, analysis of variance, mathematical transformations, statistics as topic, nonparamentric test |
| spellingShingle | Elżbieta Sandurska Aleksandra Szulc A method of statistical analysis in the field of sports science when assumptions of parametric tests are not violated Journal of Education, Health and Sport t-test, analysis of variance, mathematical transformations, statistics as topic, nonparamentric test |
| title | A method of statistical analysis in the field of sports science when assumptions of parametric tests are not violated |
| title_full | A method of statistical analysis in the field of sports science when assumptions of parametric tests are not violated |
| title_fullStr | A method of statistical analysis in the field of sports science when assumptions of parametric tests are not violated |
| title_full_unstemmed | A method of statistical analysis in the field of sports science when assumptions of parametric tests are not violated |
| title_short | A method of statistical analysis in the field of sports science when assumptions of parametric tests are not violated |
| title_sort | method of statistical analysis in the field of sports science when assumptions of parametric tests are not violated |
| topic | t-test, analysis of variance, mathematical transformations, statistics as topic, nonparamentric test |
| url | http://www.ojs.ukw.edu.pl/index.php/johs/article/view/4278 |
| work_keys_str_mv | AT elzbietasandurska amethodofstatisticalanalysisinthefieldofsportssciencewhenassumptionsofparametrictestsarenotviolated AT aleksandraszulc amethodofstatisticalanalysisinthefieldofsportssciencewhenassumptionsofparametrictestsarenotviolated AT elzbietasandurska methodofstatisticalanalysisinthefieldofsportssciencewhenassumptionsofparametrictestsarenotviolated AT aleksandraszulc methodofstatisticalanalysisinthefieldofsportssciencewhenassumptionsofparametrictestsarenotviolated |