AI Insight
This three-wave longitudinal study of 511 Chinese physical education students examined how physical activity self-efficacy relates to continued use of AI learning tools over a seven-month period. The research found that students with higher physical activity self-efficacy were more likely to continue using AI learning tools, with this relationship mediated through a chain of factors including perceived ease of use, perceived usefulness, positive attitudes, and behavioral intentions. All five tested mediation pathways were statistically significant, supporting the Technology Acceptance Model framework in this educational context.
Why it matters
The findings suggest that building students' confidence in physical activity may indirectly support their adoption of educational technology in physical education settings. This has practical implications for educators seeking to increase student engagement with AI-based learning tools by first addressing domain-specific self-efficacy.
Understand the Science
ObjectiveGrounded in the context of physical education and based on the Technology Acceptance Model (TAM), this study examined the time-separated associations between physical activity self-efficacy and the continuance use of AI learning tools among physical education students through multiple chained mediation pathways involving perceived ease of use, perceived usefulness, attitude toward use, and behavioral intention to use.MethodsA three-wave longitudinal design was employed. Data were collected from physical education students in April 2025 (T1), July 2025 (T2), and November 2025 (T3), resulting in 511 successfully matched samples. Physical activity self-efficacy was measured at T1; perceived ease of use, perceived usefulness, attitude toward use, and behavioral intention to use were measured at T2; and continuance use behavior was measured at T3. Structural equation modeling (SEM) and the Bootstrap method (5,000 resamples) were used to test multiple chained mediation effects within the TAM framework.ResultsAll five chained mediation pathways (H1–H5) were statistically significant, and the overall chained mediation effect was 0.113 (95% CI [0.083, 0.143]). The chained mediation effect for path aegh was significant (β = 0.011, 95% CI [0.006, 0.018], p < 0.01). The serial mediation effect for path afh was significant (β = 0.047, 95% CI [0.029, 0.067], p < 0.01). The chained mediation effect for path bdgh was significant (β = 0.022, 95% CI [0.013, 0.034], p < 0.01). The chained mediation effect for path bcegh was significant (β = 0.006, 95% CI [0.004, 0.011], p < 0.001). The chained mediation effect for path bcfh was also significant (β = 0.027, 95% CI [0.018, 0.040], p < 0.01).ConclusionAmong Chinese physical education students, the continuance use of AI learning tools appears to be associated with physical activity self-efficacy through a time-separated process involving perceived ease of use, perceived usefulness, attitude toward use, and behavioral intention to use. Based on three-wave longitudinal data, this study provides evidence that physical activity self-efficacy may function as a domain-specific psychological antecedent associated with self-reported continuance use behavior through TAM-based cognitive, attitudinal, and intentional processes in the context of physical education.