Beyond the p-value: Pitfalls in Reporting and Applying Statistical Tests in Sports Science

Abstract

Statistical inference underpins the credibility of sports science research, yet concerns remain regarding the rigor of statistical reporting and analysis. This study assessed the adequacy of statistical practices in sports science research according to established methodological guidelines. A total of 167 original studies published in 2020 across 15 sports science journals indexed in the Journal Citation Reports (Q1–Q3) were systematically reviewed. Up to 50 studies per journal were randomly selected, and data regarding study characteristics and statistical methods were extracted. Statistical adequacy was classified according to predefined criteria involving reporting omissions and/or statistical misapplications. Overall, 87.4% of studies presented at least one inadequacy. The most frequent reporting omissions involved failure to report assumptions such as normality, homogeneity of variances, and sphericity when required. Common statistical misapplications included the use of inappropriate dependent variables and incorrect treatment of independent observations as dependent measures. Deficiencies were identified across all journal quartiles, with no association between adequacy and journal ranking (p=0.586). These findings raise concerns regarding validity, reproducibility, and practical interpretation in sports science research, reinforcing the need for greater statistical literacy, stricter reporting standards, and improved peer review practices.