Quantile foliation for modelling performance across body mass and age in Olympic weightlifting. (December 2021)

• Record Type: Journal Article
• Authors: Perperoglou, Aris; Huebner, Marianne
• Is Part Of: Statistical modelling. Volume 21:Number 6(2021)
• Page Start: 546
• Page End: 563
• Publication Date: 2021-12
• Subjects: quantile regression -- p-splines -- 3D smoothing -- Tensor products -- performance development -- sex differences; Linear models (Statistics) -- Periodicals; Mathematical models -- Periodicals; Modèles linéaires (Statistique) -- Périodiques; Modèles mathématiques -- Périodiques; Modèle statistique; Modèle linéaire; Modélisation statistique; Périodique électronique (Descripteur de forme); Ressource Internet (Descripteur de forme); 519.5011
• DOI: 10.1177/1471082X20940156
• Languages: English
• ISSNs: 1471-082X

Abstract

In this work, we develop 'quantile foliation' to predict outcomes for one explanatory variable based on two covariates and varying quantiles. This is an extension of quantile sheets. Data from World Championships in Olympic weightlifting with athletes aged 13 to 90 are used to study performances across the life span. Weightlifters of all ages compete in body weight classes, and we study performance development for adolescents, age at peak performance and decline for Masters athletes who are 35 years or older. In prior studies, weightlifting performances were compared with a body mass adjustment formula developed using world records. Although intended for elite athletes with highest performances, this formula was applied to weightlifters of all ages, and age factors for Masters were estimated based on these body mass adjustments. A comparison of youth athletes' performances for different body mass has not been done. With quantile foliation, it is possible to examine age-associated patterns of performance increase for youth and to study the decline after reaching the peak performance. This can be done for athletes with different body mass and different performance levels as measured by quantiles. R code and example data are available as supplementary materials.