Predictive Validity of Physical, Technical, and Psychological Indicators in Talent Scouting: A Systematic Review and Meta-Analysis

Abstract

Talent scouting plays a critical role in athlete development systems, yet uncertainty persists regarding which early indicators most accurately predict long-term performance or elite attainment. Although physical, technical, and psychological assessments are widely used, existing evidence is fragmented and often domain-specific, limiting evidence-based decision-making in talent identification. This study aimed to systematically synthesise and quantitatively compare the predictive validity of physical, technical, and psychological indicators used in talent scouting. Specifically, it sought to (1) estimate pooled effect sizes for each indicator domain, (2) examine contextual moderators such as age, sport type, sex, and follow-up duration, and (3) evaluate the methodological quality of contemporary longitudinal evidence. A systematic review and meta-analysis were conducted in accordance with PRISMA 2020 guidelines. Longitudinal studies published between 2020 and 2025 were identified through searches of Scopus, Web of Science, PubMed/MEDLINE, SPORTDiscus, PsycINFO, and Embase. Eligible studies examined baseline physical, technical, or psychological indicators and subsequent performance or selection outcomes. Effect sizes were harmonised using Fisher’s z transformation and synthesised using random-effects models. Risk of bias was assessed using an adapted Newcastle–Ottawa Scale. Physical indicators demonstrated small-to-moderate predictive validity, particularly for short-term outcomes, but the effects were highly context-dependent and attenuated when biological maturation was controlled for. Technical indicators, especially match-derived and decision-executed measures, showed moderate and more stable predictive validity. Psychological and perceptual-cognitive indicators yielded smaller but significant effects and frequently provided incremental predictive value in multidomain models. Talent prediction is inherently multidimensional. Integrated, longitudinal scouting frameworks combining physical, technical, and psychological indicators offer superior predictive utility compared to single-domain approaches and support fairer, developmentally informed talent identification practices.