Leyhr · European journal of sport science 2025 · prospective cohort study · n=264

Multidimensional Performance Assessments in U15 Female Soccer: The Predictive Validity for Different Selection Levels in U17 and Success in Adulthood.

Level 3 - non-randomized controlled study

Prospective cohort study evaluating predictive validity of talent selection tests

PubMed 40569648 · doi:10.1002/ejsc.12335 · record verified 2026-08-26

What was done

A prospective cohort study tracked 264 female soccer players (under-15 [U15]) enrolled in the German Football Association's Talent Development program. Baseline testing included subjective evaluations by coaches (kicking, tactical, psychosocial skills, and endurance capacity) and five objective motor tests (linear sprint, change-of-direction speed agility, dribbling, ball control, and juggling). The primary outcomes were subsequent selection for the U17 Bundesliga (43.6% selection rate), U17 Youth National Team (6.4%), and adult Women's Bundesliga (professional level; 6.1%). Researchers assessed predictive validity using univariate comparisons and multivariate regression models combining objective and subjective measures.

What was found

Adding subjective coach ratings to objective motor assessments increased the explained variance in selection models for the U17 Bundesliga (Nagelkerke's R² = 16%) and the adult Women's Bundesliga (Nagelkerke's R² = 31%), but did not increase explained variance for the U17 Youth National Team (Nagelkerke's R² = 18%). Sprint performance was a statistically significant predictor across multivariate models. Tactical skill was the only individual subjective assessment that retained statistical significance for predicting selection into the adult Women's Bundesliga.

Why it matters

Female youth talent development has historically been understudied relative to male pathways. These results show that combining objective physical profiling (especially sprint speed) with subjective tactical ratings meaningfully improves long-term prediction of professional adult soccer participation over either assessment alone.

Limits

The sample is restricted to a single national development system (Germany), which may limit generalizability to other coaching cultures or player pools. The absolute number of players reaching the highest outcomes was small (6.1% adult professional selection, corresponding to approximately 16 players). The abstract does not provide specific odds ratios, confidence intervals, or granular test cutoffs.