The integration of virtual reality (VR) into education offers immersive learning experiences, particularly beneficial for practice-centered pre-service teacher training. This study addresses the need to understand the factors influencing pre-service teachers' adoption of VR training systems in China, where there's a strong emphasis on enhancing teachers' technological proficiency. Existing research has highlighted VR's potential in pre-service teacher training, focusing on application domains but lacking investigation into use behavior and behavioral intention. Most studies utilized the Technology Acceptance Model (TAM), neglecting other technology acceptance models. This study aims to fill this gap by developing a theoretical model, based on the UTAUT2 model, to analyze the factors influencing pre-service teachers' behavioral intention and use behavior toward adopting a VR training system. The research questions are: (1) What factors influence pre-service teachers' behavioral intention to adopt the VR training system, and what is their level of influence? (2) What factors influence pre-service teachers' use behavior in adopting the VR training system, and what is their level of influence?

Virtual Reality (VR) provides immersive experiences by stimulating multiple sensory channels (visual, auditory, tactile, etc.), enhancing learning comprehension, motivation, and effectiveness across various disciplines. VR's application in pre-service teacher training addresses challenges like limited training in classroom management and stress responses. Existing research demonstrates VR's potential in improving classroom management skills, self-efficacy, oral communication, and practical skill development. Studies have also focused on developing VR training systems for pre-service teachers, including project-based curricula, spherical view systems, and future advancements. However, a critical gap exists in the lack of user usability and behavioral testing, necessitating research on pre-service teachers' willingness to use and actual use behavior of VR training systems to inform system improvements. The UTAUT2 model, an extension of the UTAUT model, is chosen for its comprehensive approach to understanding technology acceptance, encompassing factors such as performance expectancy, effort expectancy, social influence, facilitating conditions, hedonic motivation, price value, habit, age, gender, and experience. Previous studies have applied UTAUT2 to various educational contexts, including MOOCs, e-learning systems, and Google Classroom, providing a foundation for this study's application in the VR training context.

This study employed a quantitative approach using a non-probability sampling method. Data was collected from 278 undergraduate and postgraduate students in the normal college of a university in China. These students had prior experience with the VR training system, which was implemented in their professional development classrooms in March 2023. The VR training system used was a desktop system, providing modules on teacher's moral practice, teaching practice, comprehensive education, self-development, and classroom emergency skills. An online questionnaire, using a seven-point Likert scale, was distributed via the Wenjuanxing platform from May to June 2023, resulting in 302 completed questionnaires. After removing 24 invalid responses, 278 valid responses were analyzed. The questionnaire included demographic information and items measuring the UTAUT2 variables (performance expectancy, effort expectancy, social influence, facilitating conditions, hedonic motivation, habit, behavioral intention, use behavior), along with added variables: perceived risk and self-efficacy. Data analysis involved PLS-SEM using SmartPLS4 software. Construct reliability and validity were assessed through Cronbach's alpha, composite reliability, factor loading, average variance extracted (AVE), and Fornell-Larcker criterion. Structural model evaluation assessed path coefficients, R-squared values, and statistical significance using bootstrapping (5000 resamplings).

The study found that all hypotheses were supported except for H6a (Habit's effect on Behavioral Intention) and H7 (Perceived Risk's effect on Behavioral Intention). Self-efficacy (SE), effort expectancy (EE), social influence (SI), performance expectancy (PE), facilitating conditions (FC), and hedonic motivation (HM) significantly and positively influenced pre-service teachers' behavioral intention (β values ranged from 0.281 for SE to 0.077 for HM, all p<0.05). Behavioral intention (BI), facilitating conditions (FC), and habit (HA) significantly and positively influenced use behavior (β values were 0.545 for BI, 0.182 for FC, and 0.171 for HA, all p<0.05). The R-squared values indicated a high level of explanatory power for behavioral intention (0.815) and use behavior (0.712). The final research model showed that self-efficacy, effort expectancy, social influence, performance expectancy, facilitating conditions, and hedonic motivation positively impacted behavioral intention, while behavioral intention, facilitating conditions, and habit positively impacted use behavior. Habit and perceived risk showed no significant relationship with behavioral intention.

The findings highlight the importance of self-efficacy as the strongest predictor of behavioral intention, emphasizing the role of internal motivation and confidence in adopting the VR training system. Effort expectancy, social influence, performance expectancy, facilitating conditions, and hedonic motivation also play significant roles. The positive impact of facilitating conditions underscores the importance of providing necessary resources and support. The strong influence of behavioral intention on use behavior suggests that fostering positive attitudes and intentions is crucial for promoting actual use. The lack of significant influence from habit and perceived risk on behavioral intention may be context-specific, given the mandatory use of the system in this educational setting. The results align with previous studies on technology acceptance in education, supporting the applicability of the refined UTAUT2 model in this context. However, the minimal impact of hedonic motivation suggests a need for balance between enjoyment and learning outcomes within VR-based education.

This study provides a comprehensive model for understanding pre-service teachers' adoption of VR training systems, extending the UTAUT2 framework and offering insights into both behavioral intention and use behavior. The high explanatory power of the model enhances its practical value. Findings highlight the importance of self-efficacy, facilitating conditions, and the strong link between intention and behavior. Future research should explore moderating effects of variables like gender, experience, culture, and interest, expand sample diversity, and incorporate qualitative methods to gain richer insights.

The study's limitations include its reliance on a single university sample, potentially limiting generalizability. The focus on quantitative methods may have overlooked nuances captured by qualitative approaches. The exclusion of several moderating variables from the UTAUT2 model also limits the breadth of understanding of the factors affecting adoption. Future research should address these limitations by expanding the sample size and including diverse contexts, incorporating qualitative data, and examining the moderating effects of additional variables.