Differential Learning Enhances Motor Performance and Retention in Sport Science Students

A two-arm parallel-group randomized controlled trial. One hundred undergraduate sport science students (age 18-25 years) were randomly assigned to differential learning (DL, n=50) or blocked practice (BP, n=50). The DL group performed a table tennis topspin serve with continuous pseudo-random variations across five movement dimensions (stance, backswing amplitude, execution speed, trunk rotation, wrist action) without corrective feedback. The BP group repeated a standardized movement pattern with trial-by-trial corrective feedback. Both groups completed six training sessions over three weeks (two sessions per week, 45 min each). Motor performance (accuracy measured as Euclidean distance from target center) was assessed at baseline, post-intervention (within 48h), and retention (7 days later). Movement variability was quantified using approximate entropy (ApEn) from kinematic data. Theoretical knowledge of motor learning was assessed as a potential moderator. Linear mixed-effects models, moderation analysis, and mediation analysis (bootstrapping) were performed. The trial was retrospectively registered at ClinicalTrials.gov.