Work

Reproducibility Package for "Beyond mean squad age: testing roster-age profiles at the 2026 FIFA World Cup"

Year: 2026
Type: dataset
Abstract: This repository contains the analysis-ready datasets, derived statistical outputs, sensitivity analyses, and R scripts supporting the study “Beyond mean squad age: testing roster-age profiles at the 2026 FIFA World Cup". The study analysed the complete census of 48 national teams and 1,248 registered players participating in the FIFA World Cup 2026. Player age was referenced to 11 June 2026 and classified into five operational age bands: ≤21, 22–25, 26–29, 30–33, and ≥34 years. Team-level age-band percentages were used as the reference descriptive multivariate representation of roster-age structure. Hierarchical agglomerative clustering was performed on z-score-standardised percentages using Euclidean distance and Ward.D2 linkage. The retained k = 4 solution was interpreted as an illustrative descriptive partition of the multivariate roster-age space rather than as evidence of four naturally discrete roster types or as an optimal age configuration. The four descriptive profiles were: C1 = 26–29 profile (n = 16), C2 = ≥30 profile (n = 15), C3 = ≤21 profile (n = 10), and C4 = 22–25 profile (n = 7). The repository includes the team-level roster-age dataset, conventional squad-age descriptors, hierarchical-clustering outputs, normalised Shannon entropy, descriptive principal component analysis, internal-validation results, compositional sensitivity analyses, age-band cut-off sensitivity analyses, exploratory tournament-outcome analyses, and the secondary birth-quarter analysis. Average silhouette width and gap-statistic profiles were used to characterise internal validation, while additional sensitivity analyses evaluated the dependence of the illustrative partition on compositional geometry and alternative operational age-band definitions. Tournament outcomes were analysed only as secondary exploratory outcomes. Because the dataset represents the complete census of the 2026 tournament, statistical tests are interpreted as exploratory assessments of compatibility with specific null models rather than as design-based inference or evidence of generalisability beyond the observed tournament. No statistically significant association was identified between the illustrative descriptive partition and tournament progression. The secondary birth-quarter analysis examined Q1–Q4 distributions overall and within the five operational age bands using chi-square goodness-of-fit tests, false discovery rate adjustment for age-band-specific analyses, and Cohen’s w as effect size. Birth-quarter variables were analysed independently and were not used in roster-age profiling. These analyses provide a complementary developmental perspective on age-related variation in the composition of senior national-team rosters. All underlying information was obtained from publicly available official FIFA sources. The deposited materials contain only variables required to reproduce the analyses. Player-level data have been minimised and do not include player names, clubs, full dates of birth, medical information, or individual-level performance data. This repository is intended to support transparency, reproducibility, and secondary methodological examination of the analytical framework used in the associated manuscript, including comparative analyses of generational roster composition across national teams and competitive cycles.
Language: English
Open Access status: green
OpenAlex ID: W7204668293
Cites:
Cited by:
Related to: 10