Relative age effects in track-and-field: Identification and performance rebalancing

Audrey Difernand; Quentin De Larochelambert; Sébastien Homo; Florian Rousseau; Juliana Antero; Jean-François Toussaint; Adrien Sedeaud

doi:10.3389/fphys.2022.1082174

Relative age effects in track-and-field: Identification and performance rebalancing

Front Physiol. 2023 Jan 12:13:1082174. doi: 10.3389/fphys.2022.1082174. eCollection 2022.

Authors

Audrey Difernand^{1

2}, Quentin De Larochelambert^{1

2}, Sébastien Homo³, Florian Rousseau⁴, Juliana Antero^{1

2}, Jean-François Toussaint^{1

2

5}, Adrien Sedeaud^{1

2}

Affiliations

¹ IRMES-URP 7329, Institut de Recherche Médicale et d'Epidémiologie du Sport, Université Paris Cité, Paris, France.
² INSEP, Institut National du Sport, de l'Expertise et de la Performance, Paris, France.
³ Fédération Française d'Athlétisme, Paris, France.
⁴ Fédération Française de Cyclisme, Montigny-le-Bretonneux, France.
⁵ Centre d'Investigation en Médecine du Sport, Assistance Publique-Hôpitaux de Paris, Paris, France.

Abstract

Introduction: Relative Age Effect (RAE) consists of a biased distribution of the dates of birth in a same-age group. Objectives: This study aimed to investigate Relative Age Effect among French athletes in different track-and-field events, and propose a corrective adjustment method to highlight the true potential of an athlete with respect to his/her relative age. Methods: 358,610 performances from 2009 to 2019 of female and male athletes between 12 and 21 years old were collected. Relative age distributions of performances were analyzed by level of competitiveness ("All," "Top50%," "Top10%" where "all" represents all athletes, top50% and top10% represent the best 50% and 10% of athletes per age category respectively) and age category, with chi-square and odd-ratio statistics. A linear relationship between distribution of performances and age leads to a calibration coefficient allowing to rebalance the performance by considering the effect of Relative Age Effect. Validation is obtained by Wilcoxon statistical test on actual athlete data. Results: Relative Age Effect is present in all types of events. It is larger when the level of competitiveness increases. In male 100 m sprint, 1 year difference between two athletes birth date represents an average gain of 931.01 ms (6.5%) in the U13 (Under 13 years old) and 229.65 ms (1.9%) in the U17 (Under 17 years old) categories. Our validated rebalancing methods allows to compensate for the biases induced by the relative age effect. By comparing the rebalanced performance and the realised performance of each athlete, we cannot say that they are significantly different. On average, there is no significant difference between these two performances. Conclusion: This study showed that there is a relative age effect among young French athletes, with an even greater effect as the level of competition increases. Thanks to the rebalancing method that has been validated, performances can now be better appreciated according to category and event.

Keywords: athlete development; corrective adjustment procedure; rebalancing method; talent identification; youth performance.