Recent expertise development studies have used retrospective recall methods to explore developmental biographies and/or practice histories of current or past athletes. This methodological approach limits the generalizability and trustworthiness of findings. As such, a gap exists for research exploring key multidisciplinary features in athlete development using prospective longitudinal research designs. The present research aimed to holistically model the development of talent in Olympic Weightlifting using such a design. We observed the holistic profiles of 29 junior weightlifting athletes longitudinally over a 10-month period, and subsequently classified six of the 23 athletes as high performing based on their performances in competitions up to 12 months following the study. This holistic profile was based on a framework of expertise development themes: (1) demographics and family sport participation, (2) anthropometrics and physiological factors, (3) psychosocial profiling, (4) sport participation history, and (5) weightlifting specific practice activities. A summary model was produced which selected a critical set of nine features that classified group membership with 91% average accuracy. Odds ratio calculations uncovered discriminating features in the holistic profiles of performance groups, from which empirically derived logical statements could inform the description of high-performance attainment.
Keywords: coaching; expertise; machine learning; performance; talent development.
Copyright © 2022 Anderson, Gottwald and Lawrence.