Article contents
Educational Assessment and Statistical Robustness in University Athletes
Abstract
Background: Psychological and physical assessments commonly combine item responses and field tests into categorical summaries, yet their stability depends on scoring keys and decision thresholds. Objective: To describe an auditable workflow for tracing and assessing classification sensitivity using an existing, previously published university-athlete dataset, without re-presenting its original well-being/performance association as new evidence. Methods: We linked 40 athletes (20 per team) across the supplied consolidated, response and field-test workbooks. We checked 39 six-point well-being responses per participant, applied the documented 17-item reverse key, evaluated strict pairwise rank changes and quartile-class transitions, exhaustively varied two unverified source responses over their admissible values, and contrasted recorded with printed physical-index thresholds. Conditional fixed-margin Monte Carlo comparisons were interpreted as exploratory sensitivity analyses. Results: Of 1,560 expected original responses, 1,558 could be cross-checked. The corrected-key total mean was 167.825 versus 148.325 for the direct sum. Among 780 pairs, 314 reversed their strict score order. Twenty-three of 40 quartile categories changed in the consolidated scenario; across 36 admissible missing-response configurations this number was 23–25. Seven athletes with physical-index total 10 were assigned a stored class inconsistent with the earlier article’s printed thresholds. Conclusions: The study documents substantial classification sensitivity and provides an applied, traceable audit example. It neither establishes new psychological associations nor determines whether the original journal must issue a correction; that issue requires separate provenance review.
Aims & scope
Call for Papers
Article Processing Charges
Publications Ethics
Google Scholar Citations
Recruitment