Discerning excellence from mediocrity in swimming: new insights using Bayesian quantile regression
MetadataShow full item record
AbstractPurpose: Previous research has captured point estimates for population means of somatic variables associated with swimming speed across strokes, but have not determined if predictors of swimming speed operate the same at the upper tails of the distribution (τ =0.9) as they do at the median levels (τ =0.5) and lower levels (τ =0.1). Method: Three hundred sixty-three competitive-level swimmers (male [n=202]; female [n=161]) participated in the study. To identify key somatic variables associated with 100-m swimming across and between strokes controlling for age, we used a Bayesian allometric quantile regression model, refined using Bayes Factors and Leave-one-out cross validation. Results: High probabilities (>99%) were found for arm-span, seated-height and shoulder-breadth being the strongest somatic predictors across strokes. For individual strokes, Bayesian quantile regression demonstrated that the relative importance of predictors differs across quantiles. For swimmers in the 0.9 quartile, shoulder-breadth is a more important than height for front-crawl, wide shoulders are important for breaststroke swimmers but can be detrimental when combined with narrow hips, seated-height and hip-width are important for backstroke swimming speed, and calf girth for butterfly. Conclusion: These results highlight the importance of considering key somatic variables for talent identification in swimming and ensure young swimmers focus on strokes compatible with their somatic structure. The most important new insight is that predictors differ for the best swimmers compared to average or poorer swimmers. This has implications beyond swimming, pointing to the importance of considering the upper tails of distributions in performance and talent identification contexts.
CitationMyers, T.D., Negra, Y., Sammoud, S., Chaabene, H. and Nevill, A.M. (2020) Discerning excellence from mediocrity in swimming: new insights using Bayesian quantile regression, European Journal of Sport Science. https://doi.org/10.1080/17461391.2020.1808080
PublisherTaylor & Francis
JournalEuropean Journal of Sport Science
DescriptionThis is an accepted manuscript of an article published by Taylor & Francis in European Journal of Sport Science on 10/08/2020, available online: https://doi.org/10.1080/17461391.2020.1808080 The accepted version of the publication may differ from the final published version
Except where otherwise noted, this item's license is described as https://creativecommons.org/licenses/by-nc-nd/4.0/