Digital learning tools drive higher teaching quality in PE courses

A new study has identified a close link between the use of digital technologies and improved teaching quality in physical education (PE) at the university level. The research team leveraged advanced machine learning techniques and structural equation modeling to analyze the effects of digital integration across several educational platforms and tools.

Published in Applied Sciences, the study, titled “The Impact of Digital Technology Use on Teaching Quality in University Physical Education: An Interpretable Machine Learning Approach,” combines large-scale empirical analysis with AI-driven interpretability frameworks. With a sample of 1,158 university students across China, the study investigates how different forms of digital technology contribute to educational outcomes, and what psychological mechanisms mediate these effects.

Which digital tools make the biggest impact on teaching quality?

The study evaluated seven types of digital technologies frequently used in PE instruction: multimedia courseware, e-textbooks, question banks, management software, educational websites, multimedia materials, and online communication tools. To determine their effectiveness, the research team employed nine machine learning models alongside SHAP (Shapley Additive Explanations) and feature permutation importance techniques.

Among the digital tools assessed, multimedia courseware, e-textbooks, and question banks emerged as the most influential in enhancing perceived teaching quality. These tools were associated with improved clarity, structure, and interactivity in classroom delivery. Multimedia courseware, in particular, helped instructors break down complex physical techniques and rules using visual demonstrations, thereby increasing student understanding.

Management software and educational websites were also found to play key roles. These tools supported administrative efficiency and provided students with access to additional learning materials beyond class time. Importantly, the study showed that certain technologies worked best in combination. For example, the simultaneous use of multimedia materials and educational websites generated a synergistic effect that contributed more substantially to perceived instructional quality than either tool alone.

The ML models confirmed the consistency of these findings across different statistical configurations, lending strong credibility to the conclusions. In particular, the SHAP analysis clarified how individual features within each technology influenced learning outcomes, offering interpretable results often absent in black-box AI models.

What psychological mechanisms link digital use to learning quality?

Beyond the identification of effective digital tools, the study explored the psychological mechanisms that connect technology use to teaching outcomes. Two mediating variables, perceived teacher support and academic self-efficacy, were found to play critical roles in this process.

The structural equation modeling showed that digital technology use had a significant direct effect on teaching quality. More importantly, perceived teacher support and self-efficacy served both as independent mediators and as part of a chain mediating mechanism. In other words, the mere presence of digital tools was not enough; their effectiveness hinged on how these tools influenced students’ perceptions of their instructors and their own academic capabilities.

Perceived teacher support, which refers to students’ sense that their instructor is invested in their progress and available for help, was strongly enhanced by technologies that increased teacher-student interaction. For example, online communication platforms and learning management systems allowed more flexible, real-time dialogue, which contributed to a greater sense of instructional presence and responsiveness.

Academic self-efficacy, defined as a student's belief in their ability to perform tasks successfully, was boosted by technologies that allowed for iterative learning and instant feedback. E-textbooks and question banks enabled students to test their understanding in low-stakes environments, increasing their confidence before formal assessments.

Notably, the study found a chain mediating effect in which digital tools improved teacher support, which in turn strengthened students’ academic self-efficacy, ultimately leading to better perceptions of teaching quality. The statistical confidence intervals for these mediating effects were robust, reinforcing the credibility of these psychological pathways.

How can these findings be applied in higher education?

While physical education may appear resistant to digitization due to its emphasis on movement and real-time demonstration, the findings indicate that properly selected and implemented technologies can meaningfully enhance the instructional experience.

The authors recommend a more intentional deployment of digital tools, emphasizing those with proven pedagogical value such as multimedia courseware and adaptive learning platforms. Universities can also prioritize investments in teacher training to ensure that instructors understand how to use these tools not only for content delivery but also for reinforcing psychological dimensions such as support and efficacy.

To sum up, the study asserts that technology should be viewed not as a replacement for traditional instruction but as a complementary asset that can reinforce student-teacher relationships and support self-regulated learning. By framing digital tools within a pedagogical and psychological framework, institutions can more effectively align technological adoption with educational outcomes.