# AI-assisted and Big-Unit teaching enhance speed-skating performance through psychological mechanisms in adolescents: evidence from a three-arm intervention study

## TL;DR

AI-assisted and Big-Unit teaching methods improve speed-skating performance in adolescents by boosting psychological factors like motivation and self-efficacy.

## Contribution

This study is the first to compare AI-assisted and Big-Unit teaching models and identify psychological mechanisms in adolescent winter sports.

## Key findings

- AI-assisted and Big-Unit teaching improved 500-m speed-skating performance significantly more than conventional methods.
- Innovative teaching enhanced psychological outcomes like motivation, self-efficacy, and resilience in adolescents.
- Learning motivation and self-efficacy mediated the relationship between teaching methods and performance gains.

## Abstract

Innovative instructional approaches are increasingly advocated in physical education to enhance both motor skill development and psychological adaptation. However, few studies have directly compared micro-level (AI-assisted) and macro-level (Big-Unit) teaching models, or examined the psychological mechanisms underlying performance improvements in adolescent winter-sport environments.

A three-arm, quasi-experimental longitudinal study was conducted with 129 first-year middle school students (AI-assisted: n = 42; Big-Unit: n = 43; Conventional: n = 44). Participants completed an 8-week speed-skating intervention consisting of 24 on-ice lessons. Learning motivation, self-efficacy, psychological resilience, and related psychological constructs were assessed at baseline (T1), mid-intervention (T2), and post-intervention (T3). Skating performance was evaluated using electronic 500-m timing. Linear mixed-effects models, ANCOVA, and structural equation modeling were applied to assess Group × Time interactions and mediation pathways.

Both AI-assisted and Big-Unit teaching produced significantly larger improvements in 500-m performance than conventional instruction (AI: −5.59 s; Big-Unit: −7.60 s; Conventional: −1.80 s; all p < 0.001). All 13 psychological outcomes showed strong Group × Time interactions favoring the innovative groups [all χ2(4) > 137.28, q < 0.001]. ANCOVA confirmed substantial adjusted Group effects for changes in learning motivation, self-efficacy, psychological resilience, and anxiety/stress (partial η2 = 0.650–0.927). Mediation analyses identified a statistical suppression pattern, in which increases in learning motivation and self-efficacy served as significant indirect pathways linking innovative instruction to performance gains. However, the direct technical impact remained the dominant driver.

AI-assisted and Big-Unit teaching substantially enhance both technical performance and psychological functioning in adolescent speed skating. Statistical mediation models support learning motivation as a plausible mechanism linking teaching mode to performance, with self-efficacy providing additional support. These findings highlight the complementary potential of technology-enhanced and mastery-oriented pedagogies to modernize physical education through both direct technical renovation and indirect psychological adaptation.